Behavioral responses in bacteria.

ANNUAL REVIEWS

Further

Quick links to online content Annu. Rev. Physiol. Copyright

© 1992

1992.54:683-714

by Annual Reviews Inc. All rights reserved

Annu. Rev. Physiol. 1992.54:683-714. Downloaded from www.annualreviews.org Access provided by University of New England - Australia on 01/22/15. For personal use only.

BEHAVIORAL RESPONSES IN BACTERIA Judith P. Armitage Microbiology Unit, Department of Biochemistry, University of Oxford, Oxford OXI

3QU, England KEY WORDS:

bacterial motility, bacterial chemotaxis, chemotaxis, aerotaxis, phototaxis

INTRODUCTION The majority of bacterial families contain members that move actively about their environment, either swimming by means of flagella or gliding along the substratum. Investigations into bacterial movement and its control by environ mental stimuli suggest that the mechanisms probably developed very early in evolution; groups as diverse as gram-positive and negative-species share the same basic mechanisms as archaebacteria and spirochaetes. The synthesis and operation of flagella, and probably the gliding apparatus, is highly energy demanding, therefore if a bacterium is motile, moving almost certainly provides a survival advantage for that species. In several species, flagellar synthesis is inhibited by growth on rich media, which suggests that flagella synthesis, although energetically expensive, becomes worthwhile in limiting growth conditions. In this review I try to bring together data on motility and behavior in a range of bacteria and emphasize the common features and variations imposed by evolution on specific species. The majority of experimental work has been carried out on enteric bacteria, and this will necessarily provide the backbone of the review, but it should be borne in mind that the best understood and most studied system does not necessarily make it the most representative when applied to bacterial species in vivo. 683

0066-4279/92/0315-0683$02.00

684

ARMITAGE


MECHANISMS OF BACTERIAL MOVEMENT Bacteria are in general too small to sense gradients across their length, therefore they make tempo ral comparisons of the different sensory stimuli relevant to that species while randomly moving through that environment. Bacteria swim in a world of negligible Reynolds number, where the major force is not inertia, but viscosit y, which adds constraints to their movement and the mechanisms of gradient sensing. Despite this, they can swim at speeds of over 100 }-tm sec-1 (over 50 body lengths per second). Even at these speeds, their small size and three-dimensional random swimming pattern means that the distances covered and the gradients sensed are relatively small.

Flagella Bacterial flagella, with some variation, (see below), apparently have the same basic design (reviewed in 58,79). The flagellar filament is usually made up of monomers of a single protein, flagellin, polymerized to form a semi-rigid helix. The movement of ions down the transmembrane electrochemical gra dient through a membrane-spanning motor complex results in the rotation of the filament. The cell body counter-rotates in response, and the cell swims

forward (Figure l a). FLAGELLAR FILAMENT Flagellin proteins from different species can vary in size from about 25 to 60 kd (58). There is a high degree of homology between the C-terminal and N-terminal regions of the different species, sometimes allowing co-polymerization, with the differences that lie in the core of the protein ( 1 49) providing the antigenic character. The flagellin subunits form a cylinder of 11 protofilaments with a 60 A

central hole. The flagellin subunits in each protofilament can be in one of two

conformations (147), one of which produces a shorter protofilament. The normal functional enteric bacterial flagellum has nine p rotofilam ents with subunits in the long conformation and two protofilaments in the short con formation, forcing the filament to adopt the classical corkscrew shape with a 1 }-tm pitch (80). A reversal in the direction of flagellar rotation appears to change the number of prot ofilament s in each conformation, which results in changes of handedness and helical pitch of the total filament and in cell tumbling ( 146). VARIATIONS IN FLAGELLAR STRUCTURE

Several species have more than one flagellin (5,6, 104, 143, 144). In Caulobacter crescentus and Rhizobium meliloti the different, but related, flagellins are expressed and incorporated


BACTERIAL BEHAVIOR

685

sequentially (108, 150) and all are required for normal swimming. Two related flagellins comprise the complex flagella found in many rhizobial species. These complex flagella have the same basic structure as simple flagella, but the pattern of subunit interaction results in a stiffer filament that can only rotate in one direction (147). The stiffer filament may make it possible for rhizobial species to swim through more viscous media than species with simple flagella. The best swimmers in viscous environments are the spirochaetes (39, 43). These flexible, spirally shaped bacteria have their polar flagella internalized between the protoplasmic cylinder and the outer cell wall. The five or so identified flagellins appear closely related to other flagellins, but have addi tional unrelated proteins providing a filament sheath that may aid rotation of the filament in its membrane-bound space (104). Various species have slight ly different mechanisms of swimming (39), but generally it is thought that the rotation of the periplasmic filaments between the sheath and cell cylinder causes the cell cylinder to rotate in one direction while the cell sheath rotates in the other (16). If the filaments at the two poles rotate in the same direction, then the cell corkscrews through viscous environments, whereas if they rotate in opposite directions, the cells flex, equivalent to tumbling (40). Some species, e.g. Leptospira and Treponema phagdenis have short, non-overlap ping periplasmic filaments with a wavelength different from that of the cell cylinder. Rotation in the direction of the cell body helix produces a spiral shape imposed on the cell structure, whereas rotation in the opposite direction produces a hook shape (23). Therefore cells have two mechanisms of forward thrust: rolling of the cell to allow movement through viscous media and gyration at the front of the cell that generates a backward moving spiral wave to allow movement through less viscous media. The archaebacterium Halobacterium halobium swims using polar tufts of flagella. Its glycosylated flagellins allow the filaments to slip against each other when switching rotational direction which, unlike the enteric bacteria, allows the complete bundle to change rotational direction without the fila ments separating (6). The hook joins the flagellar filament to the basal body. It is normally curved and has a similar helical arrangement of protein subunits ( 148), which probably allows the two structures to interact at the junction (59) with the aid of two hook-associated proteins (50). The curved hook may allow flagella to come together as a bundle while the straight hook of the monoflagellate Rhodobacter sphaeroides may simply couple the filament to the motor (1 l9a). The basal body drives the rotation of the helical flagellum. It is comprised of a rod connecting the filament to the motor and a series of rings (Figure 1 a). Only part of the total motor apparatus

THE HOOK AND BASAL BODY

ARMITAGE


686

has been physically isolated while the switch complex and the motor proteins appear to remain in the membrane (3). The Sand M rings may in fact be part of a single gene product (79, 126). Additional structures identified by rapid freeze electron microscopy may be the intracellular protein complex responsi ble for controlling flagellar rotation (64). Extra periplasmic rings identified in C. crescentus and R. sphaeroides ( 1 19a, 1 27) may be involved in the shedding of flagella during certain phases of the bacterial life cycle or in targeting the flagellum to a specific site in the cytoplasmic membrane. A large protein plate associated with the basal body

a

/

HAP2 (FliD) Class 3a

b fliA c-AMPICAP

filament (FliC)

Class 1

� fliE fliF fliL

/liC motA I2r

11rl/A _J

Class 3b

Class 2

hook (FlgE)

exterior

if rod --------:;;0-1 0;:-

(FlgB-FlgC �

FlgF-FlgG)

n f=.g;:�h

)() Mot complex

cytoplasm

(MotA-MotB)

switch complex

(FIiG-FIiM-FliN)

export apparatus

20nm

1---1 Figure 1

(0) A diagrammatic representation of the E. coli bacterial flagella in the cell mem

brane. Gene names in brackets. (b) Genetic control pathway for flagellar synthesis showing the catabolite control of expression of the different operons. flhD codes for a sigma factor controlling expression of class 2 and 3a operons, fliA codes for a sigma factor controlling class 3b operon

expression. The fliD operon probably codes for a repressor of class 3 operous (from 79 with permission).

BACTERIAL BEHAVIOR

687

has been identified in several species of bacteria, especially those with polar flagella, e.g. Wolinella, but a role for this plate has not been identified (71).


Possible Mechanisms of Rotation The protein products of the motA and motB genes, forming a 8-12 subunit circlet around the motor, are involved in transforming the electrochemical ion gradient into mechanical rotation (63). MotA is a proton channel with MotB possibly controlling the interaction of MotA with the basal body (18, 18a, 151). The mechanisms involved in turning either the electrical or chemical component of the ion gradient into mechanical rotation are not understood; however , there are many hypotheses (17, 62, 93, 157). Control of Flagellar Synthesis In E. coli and S. typhimurium. flagella synthesis is under the control of at least 40 genes, and in C. crescentus it involves at least 48 genes. This makes flagella synthesis one of the most complex systems identified in bacteria, only exceeded by the ribosomal system (for reviews see 29, 77, 79). Perhaps the c

if11llicated in o flagellum-specific

flhe flhD

=

export prac e 55

I777iAl If!i!J

Flgc

FigS

[rod]

[rod] fliM[SWitCh] fliP

[flgJ fliJL fI G ] IUZEl fliD fliR ( l g A o flhB fliE

FlgG

FlgF [ r d)

-

(rod)

..!il..

-

Fig! (Pring)

.i..-

=�

::B:

{lgD

-

rivet

filA

FlgK (HAP1)

FlgL (HAP3J

FHO (HAP2)

illii

[switch]

==

F lgH

;fr(� :0:

[f�tJ

� � � ""

H88 Figure 1

complex

FlgE [hook] fliK

�0

1!t

membrane """"" ./ - pe ptldog Iycan laye r

--,

cytoplasmic membrane

(c) Order of assembly of flagellar components. The proteins/genes required for each

step of assembly are shown above the arrows. For details see text and References 57 and 79 (from

79 with permission).


688

ARMITAGE

need for such complexity is due to the numerous components of the structure, its interactions with the membrane, and the complications of assembling an organelle outside the cell. The analysis of non-motile mutants and promoter fusions with indicator protein genes has contributed to a model of sequential gene expression and protein assembly of functional flagellum. Figure 1 b shows the regulation pathway of about 14 operons clustered in four regions under the catabolite control of cAMP. Several genes, jliA, jlhDlC, may be specific sigma factors involved in control of gene expression. Other gene products feed back on their own promoters, such as FliD that caps the flagellar filament and also acts as a repressor of the filam�nt genes. Shearing of filaments reduces the intracellular concentration of FliD and consequently derepresses the genes for filament synthesis. In B. subtilis the situation is more complex with very large operons (averaging 26 kb) under the control of sigma 28 promoters. However, the use of a limited number of sigma factors for transcriptional regulation cannot provide the temporal control of expression that is needed for flagellar assem bly. In comparison in C. crescentus there is evidence that a complex interac tion of a number of trans-acting factors with sigma 54-dependent promoters allows for temporal control (95, 116, 153). Figure lc shows the order of assembly identified by electron microscopy of partial structures and by immuno-electron microscopy (switch and motor com ponents are not isolated by these procedures) and an analysis of conditional mutants (51, 57, 135). The order is closely related to the order of expression of the genes. The membrane components are exported using the classical signal-se quence mechanism, while the extracellular proteins lack these sequences and are exported as monomers down the hollow center of the basal body and filament for assembly at the distal end of the structure (52, 57, 79). Gliding is used by a wide range of species from cyanobacteria to the myxobacteria, but the mechanisms involved are poorly understood. While Myxococcus glides as either single cells or more generally as groups of cells at speeds of 2-4 JLmlmin at solid/liquid or liquid/air interfaces, no motility organelle has been identified, and no cellular changes equivalent to those seen in gliding eukaryotes cells have been seen. The electrochemical proton gra dient may be involved (37). Although the cells often glide along slime trails, slime is not essential. For a review of possible gliding mechanisms see (30, 158, 159). GLIDING

UNSTIMULATED SWIMMING Flagellate bacteria swim randomly about their environment, swimming smoothly in one direction for a period of time and then changing direction. The mechanism of direction changing differs slightly across species.


BACTERIAL BEHAVIOR

689

The result of switching patterns is always the same, however, and results in a three-dimensional pattern of swimming. Gliding bacteria have also been shown to randomly switch their direction of gliding, but in two rather than three dimensions. In E. coli and S. typhimurium the six to eight individual flagella rotate independently, spending about 95% time rotating counterclock wise and 5% clockwise (81, 78). When the majority of flagella filaments rotate counterclockwise, they are forced together into a bundle that drives the cell forwards, but the switch of some of the filaments to clockwise causes a change in wavelength and handedness in those filaments, thus forcing the flagella bundle apart and causing the cell to tumble (82). When the majority of filaments return to counterclockwise rotation the bundle reforms and the cell swims smoothly again, usually in a new direction. Each flagellum has its own switching frequency, and therefore the overall tumble frequency is dependent on the number of flagella at any one time that switch to clockwise rotation (55, 81). Organisms with polar flagella change direction simply by reversing the direction of flagellar rotation. In all these cases Brownian motion will almost certainly ensure that the reversal results in a new swimming direction. R. sphaeroides and Rhizobium meliloti use transient stops in flagellar rotation to change the swimming direction (12, 41). These can occur while the driving force for rotation remains maximal, which suggests that this organism can either disengage the motor from the driving force or put on a brake. GRADIENT SENSING AND RESPONSES As bacteria are too small to sense a gradient across their length, they sample their environment with time. A cell swims farther in the direction of increased attractant concentration; however, it will still intermittently change directions, with the length of swimming in the new direction depending on whether the attractant concentration continues to increase or decrease. In this way the random walk is biased up an attractant gradient (Figure 2a). Bacteria simultaneously sense a wide range of stimuli; the range and the response depending on the niche of that species and the current growth conditions. The responses must be integrated to maintain or guide the bacteri um to its optimum environment (Figure 2b). To allow for future responses to a change in either a current or a new stimulus, bacteria must be able to adapt (115). Therefore bacteria have a memory of past concentrations, with cells returning to unstimulated swimming patterns after a time in a stable environment. Receptor-Dependent Chemotaxis Probably the best understood behavioral system in biology involves the dedicated sensory receptors in bacteria (reviewed in 2, 78, 133, 138). Many


b

PTS

transport

osmolarity RECEPTORS

Ught

'--L

attractants/ repellents

/ integration

t

Figure 2

temperature MOTOR

Overview of bacterial behavior (a) Representation of a biased three-dimensional

random walk and (b) the range of stimuli that must be integrated.


BACTERIAL BEHAVIOR

691

years ago Adler and co-workers showed that some chemoeffectors caused a chemotactic response without being transported, thus indirectly implicating surface receptors (1), Competition experiments have identified four specific receptors, each responding to a range of chemicals. In E. coli and S. typhi murium these are Tsr, which mediates response to serine, temperature, and pH; Tar, mediating responses to asparate and ribose; Trg, maltose and galactose; and Tap, dipeptides (78). There are several hundred of each receptor per cell; a change of occupancy of only about four causes a response, thus making it a very sensitive system (20). The limited range of chemoeffec tors sensed through these receptors, which are also known as methyl accepting chemotaxis proteins (MCPs),probably reflects both a compromise over membrane space and the nature of chemical gradients in natural environ ments, i.e. a gradient of one amino acid indicating a general nutrient gradient. The response to sugar chemoeffectors sensed through MCPs is indirect, occurring via the periplasmic-binding proteins of the sugar transport system. Therefore Tar, for example, can respond to either the direct binding of aspartate or the binding of the large ribose periplasmic-binding protein, complexed with ribose. Tar and the other MCPs act as both receptors and transducers. The genes for these membrane-spanning receptors/transducers have been cloned and sequenced and recently subjected to site-directed mutagenesis and chemical cross-linking. Therefore we know a great deal about their structure and the possible mechanisms involved in both signal transduction and the adaptation pathway (Figure 3a) (48). Sequence analysis has shown that the transducers are structurally related, which suggests a common evolutionary origin (21,69). Not only are the four 60-kd enteric transducers related to each other, but antibodies raised to the E. coli Trg protein cross-react with mem brane preparations from a wide range of species,from S. aurentia through B. subtilis to H. halobium, and at least part of the sequence is common to a gene involved in aggregation in M. xanthus (4,21,91). The cytoplasmic sequence is highly conserved, particularly in the regions involved in signaling and in adaptation, whereas the periplasmic receptor regions are, as expected, less well conserved (Figure 3a). Each transducer has two membrane-spanning hydrophobic sequences with regions of structural homology. The transducers exist as dimers with the membrane-spanning regions coming into close con tact in the center of the membrane (92). Chimeric fusion proteins comprised of the periplasmic regions of one transducer and the membrane spanning and cytoplasmic region of a different MCP confirm that the periplasmic region specifies the range of responses (68). The different chemoeffectors interact with independent periplasmic sites, although mutant studies suggest that these are located in similar struc tural regions of the molecule (94, 105). Modeling studies suggest that these ligand-binding sites may be at the top of a four helix bundle structure because

692

ARMITAGE

ligand binding domains a


pertplasm

signalling domalns

cytoplasm Rl

methylation domains COOH

switch

adaptation

b

cw'"

"H::'=:O,�� :��, � � U

....

decrease -d attractant

..

....

...... ccw

�;fe'i.!se

H3CO

OCH3

H3cO

adaptation [no signal) Figure 3

Structure and function of MCPs. (a) Structure of an MCP monomer showing different

functional domains and (b) structural changes in MCP caused by ligand binding and subsequent methylation and adaptation. See text for detail, (adapted from 22 and 48).


BACTERIAL BEHAVIOR

693

specific response mutants all have single amino acid substitutions in these regions (48). The binding of the ligand to the receptor region of the transducer appears to cause a conformational change in the membrane-spanning segments of the protein. Cysteine substitutions within the hydrophobic segments of the protein have been used to show protein-protein interaction, both within the mono mers, by NEM reactivity, and between the dimers, by disulfide cross-linking. The binding of a chemoeffector ligand causes the stabilization of these interactions and alters the rate of disulfide bridge formation (34, 48). The transmembrane-spanning regions may roll against each other without dis sociating and thus transmit binding information to the cytoplasmic side of the membrane, or binding may change the vertical position of the transducer in the membrane. The periplasmic side of the protein is resistant to proteolysis, which suggests that part of protein may interact with the membrane. The cytoplasmic side of the transducers is highly conserved and shows strong homology to the sensor sequences identified in two-component sensing

attractant

c

repellent

CheW

EI

�

r

r

EI-P pyruvate

I

) ::::l8/::: Cthr Che�_ � + I CheA

PEP

t t

'-

T EI

.? pcp-I?

t

8

adaptation

inner membrane

CheA

HPr-P

Figure 3 (c) Integration of signaling and adaptation pathways from the Mep receptors and the PTS transport system in enteric bacteria. See text for details. (Adapted from 10, with permission)


694

ARMITAGE

systems (see below). The homologous domain is about 320 amino acids, and analysis of a large number of mutants has identified specific functional regions (Figure 3a). Two regions, designated Kl and Rl, are the sites of methylation during adaptation (Kl has three or four sites, and Rl has a single methylation site) (140, 99). Other regions are involved in sending a signal to the flagellar motor for clockwise or counterclockwise rotation. A subset of Tsr mutants showed that signaling for clockwise flagellar rotation is near the Rl region and for counterclockwise is near Kl, with a switch region between them (8) . Information about transducer binding is transmitted to these rota tional switch regions through the region between TM2 and Kl. How does the' information from the signaling region of the transducers reach the flagellar motors to bring about clockwise or counterclockwise rotation? By local stimulation of long cells of cell division mutants, Berg et al showed that the signal traveled a limited distance without amplification (55). In addition, bacteria that were voltage-clamped behaved normally (86). This suggests that the signal takes the form of a small diffusible chemical rather than a membrane potential change, although membrane activity may be involved in some larger polar flagellate species (42, 67). Four cytoplasmic gene products have been identified as essential for signal transduction, CheA, CheW, CheY, and CheZ (Figure 3c). The breakthrough in understanding cytoplasmic signal transduction came with the isolation of the individual proteins in vitro (reviewed in 22, 90, 106, 132). Purified CheA can be autophosphorylated in vitro and can transfer its phosphate to either CheY or to CheB (49). CheA exists in two forms CheAL (76 kd) and CheAs (66 kd), coded for by two overlapping genes with common C-termini, but different translational start sites ( 119) . CheAL has both autophosphorylation and phosphotransferase ability, but the role of CheAs is unknown. CheA can be isolated as a complex with CheW from lysed cells (36). This complex shows increased autophosphorylation levels and an enhanced affinity for CheY . The phosphorylation activity is also increased if transducers are present in the reaction mixture . When isolated from mutants locked in smooth swimming mode, i.e. counterclockwise, the level of CheA phosphorylation was less than 10% of wild-type levels, whereas clockwise-locked mutants had higher than normal phosphorylation levels (48). Therefore it appears that the transducers and CheW are responsible for the autophosphorylation state and thus the signaling state of the cytoplasmic proteins. Eisenbach, using cell envelope preparations with locked motors, showed that phosphorylated CheY is essential for clockwise motor switching (1 14). CheY is the phosphorylation substrate for the CheA-P/CheW complex. CheZ dephosphorylates CheY-P. Thus low level phosphorylation of CheY and its dephosphorylation by CheZ probably controls the background switching frequency seen in unstimulated cells (70). CheY is a small 129-amino acid protein that is homologous to a range of bacterial regulatory proteins (130, 131). It has now been crystal-


BACTERIAL BEHAVIOR

695

lized, and detailed structural analysis is underway. Phosphorylation occurs at Asp-57 and probably results in the active conformation. Cells lacking CheZ tumble constantly as a result of the high CheY-P levels, but can still be stimulated to smooth swimming by the addition of chemoattractants. The response is slow because it depends on dephosphorylation of CheA, which results in lower CheY P levels via autodephosphorylation rather than CheZ activity. Because the signaling pathway has been reconstructed from mutational studies and in vitro biochemistry, it is the sequence of events resulting from the removal of an attractant from a transducer, i.e. the induction of tumbling by clockwise flagellar rotation, that has been examined in most detail. The current hypothesis is that the information that less ligand is now bound to the transducer is transmitted to the transducer's cytoplasmic side by changes in the interaction of the transmembrane a helices of the transducer dimers. This alters the conformation of the cytoplasmic domain of the transducer signal to rotate. clockwise. As a result the CheW-CheA complex is released and accelerates autophosphorylation of CheAL. Phosphorylated CheA then phos phorylates CheY, which is able to interact with the switch complex of the motor, resulting in clockwise rotation. CheY-P is subsequently dephos phorylated by CheZ. Adaptation

Methionine auxotrophs are unable to respond to chemical stimuli that are sensed through the membrane-spanning transducers (15). Early experiments showed that the different transducers were methylated in response to their specific chemoeffectors in the absense of protein synthesis, which resulted in a series of apparently different molecular weight proteins (6 1 ) . These appar ent changes in molecular weight are caused by methylation of four specific glutamate residues in Kl and Rl. The extent of periplasmic chemoeffector binding is reflected in the number of cytoplasmic sites methylated. Initially, binding of a chemoeffector produces a signaling conformation in the cytoplasmic domain that is eventually reset to the non-signaling state by the methylation of these specific residues. If the chemoeffector is removed, a signaling state of the transducer is again induced, and demethylation of these residues returns the protein to its previous state. If signaling and (de)methyla tion happen on different time scales, then sensing and adaptation are possible (Figure 3b). Methylesterification results from a specific methyl transferase transferring methyl groups from s-adenosyl methionine (156). If the predicted alpha helical structure is assumed for the cytoplasmic domains of the transducers, then the amino acids surrounding each glutamate are similar. Alteration of any of these amino acids changes the rate of methylation (8). At least half the


696

ARMITAGE

specific glutamates are coded for by glutamines, which are modified by an esterase acting as an amidase (60). The activity of the methyl transferase CheR remains constant, whereas the methyl esterase CheB activity varies in response to CheA activity and chemotactic stimulation (129). Phosphoryla tion of CheB by CheA-P, if the concentration of a chemoeffector falls (128), increases its activity (76). The esterase demethylates the transducer gluta mates, previously methylated by the transferase in response to the earlier increase in the chemoeffector'concentration. Sequence analysis suggests that the alpha-helical regions of the cytoplasmic part of the MCP could change their interactions in the methylated and demethylated forms, and this change in conformation may regulate CheW-CheA activity (36). Other Transducer Activities E.

coli and S. typhimurium swim toward their growth temperature if placed in a temperature gradient (84, 53). An increase in temperature causes smooth swimming and a decrease results in tumbling. The response can be blocked by a saturating concentration of serine or aspartate, which suggests that the effect may involve the transducers Tsr and Tar (83). Recent work has shown that, in addition, Trg can act as a warm receptor and Tap as a cold receptor (96). The temperature change probably causes a conformational change in the transduc er, which is equivalent to the binding of an attractant to the periplasmic side of the membrane. Saturation of the attractant-binding sites and subsequent adaptation to the saturated receptors blocks any subsequent change and inhibits the temperature response. Many bacterial species are repelled by a range of chemicals. In enteric bacteria these are classically heavy metal ions, indole, leucine, organic acids and glycerol (102, 145). The mechanisms involved in repellent responses are still not well understood, but for at least the last two cases, they appear to involve changes in the sensory transducers (102). Ten percent glycerol (this concentration acts as a repellent) locks the periplasmic domain of the transducers into the compact conformation. Similarly weak acids appear to work by altering the intracellular pH which, in tum, alters the conformation of the cytoplasmic side of the Tsr and Tap transducers. In all these cases MCP-minus strains show no response (66). R. sphaeroides, an organism known to lack MCPs, does not respond to glycerol (1. Armitage, unpublished data) and shows a positive response to organic acids. Early competition data suggested that there were specific repellent recep tors, but until recently none had been identified. All repellents operate at high (millimolar) concentrations suggesting weak, low affinity binding. These are probably the MCPs, although whether there is any specificity is unclear (31, 154). Early suggestions that changes in membrane fluidity may be involved appear incorrect (31). Eisenbach and colleagues showed that the E. coli

BACTERIAL BEHAVIOR

697

flagella not only rotated clockwise and counterclockwise, but could pause for fractions of a second (32). The frequency of these pauses increased on addition of a repellent, even in transducer-minus mutants (3 1). The signifi cance of these results is not yet certain.


Other Transducers Obviously each species of bacteria will have evolved transducers that respond to their most useful chemoeffectors, e.g. Agrobacterium tumejaciens seems to respond to a low concentration of the wound exudate acetosyringone via specific transducers (75), and Ps. putida responds to benzoate (a repellent for E. coli) (47). In these cases the basic structure of the transducer shows homology with those from enteric bacteria. The archaebacterium H. halobium has methylatable transducers that cross react with E. coli anti-Trg antibody (4), but methylate in response to either chemoeffectors or photostimulation. The response is somewhat different from that of enteric bacteria and perhaps more closely related to that of B. subtilis ( 10 1 , 142) because the H. halobium 90- 135-kd proteins show a transient change in the rate of demethylation whether the stimulus given is positive or negative, the kinetics of the turnover determining the type of response. It is still uncertain whether the methylation system is involved in adaptation or just in excitation ( 1 4 1 ). The addition of ethionine to reduce the cytoplasmic concentration of s-adenosyl methionine increases the response time, which shows a role for methylation in signaling. A role in adaptation is suggested only by demethylation rates that are similar to measured adaptation rates. Similarly in B. subtilis. the total number of methyl groups on the transducer proteins does not change, but the turnover rate does ( 1 4 1 ). Therefore the enzyme homologous to CheB operates as an excitatory enzyme, possibly still under the control of a "CheA" protein, which is also responsible for phosphorylation of "CheY" (97). The current hypothesis for signaling in B . subtilis i s that, o n attractant binding, CheW/CheA autophosphorylates and phosphorylates CheB. Activated CheB transfers methyl groups from the transducers to an unidentified intermediate, and this intermediate interacts with the switch to cause smooth swimming while the transducer is remethylat ed. Adaptation is probably the result of the reduced rate of CheA auto phosphorylation, which reduces the level of CheB-P. The addition of a repellent would similarly result in both CheB and CheY being phosphory lated, as well as an increase in methylation of the intermediate. The gliding bacterium M. xanthus genes, termedjrz. show strong sequence homology to cheW. A. Y. B. and R as well as to the transducer Tar ( 1 58, 1 59). The role of these is unknown because the slow two-dimensional movement of the cells makes the sensing of a diffusing gradient impossible. The product of thejrzCD gene, equivalent to Tar, has the same molecular weight as a protein found to be methylated in vitro (9 1 ). The time course of methylation is long,

698

ARMITAGE

however, and the protein product is soluble rather than membrane spanning. Despite these differences, the presence of the equivalent set of proteins in M. xanthus, and the altered behavior of mutants in those genes suggest a similar role in this species. The slow, two-dimensional gliding of the organism may


alter its time course and sensing mechanism. For example, a normal gliding

M. xanthus cell will change direction about every 7-8 min, whereas a smooth mutant only changes direction every 30--40 min, and tumbling mutants change every 4-5 min. Relationship to Two Component Sensing Systems

About 20 metabolic pathways have been identified that use a common mech anism for inducing genes in response to the presence of specific metabolites

(7, 22, 1 32). These two-component systems depend on specific sensor pro teins in the membrane or cytoplasm, which when activated by a metabolic signal undergo histidine phosphorylation. This allows the protein to interact with a second regulatory protein and transcription of the genes of that specific metabolic pathway follows. There are conserved sequences at the C-terminus of the first sensor proteins and at the N-terminus of the second regulatory . proteins. The common mechanism of action involves phosphorylation of a protein kinase at a histidine in one protein and transfer of this to an aspartate on the second protein. The receptor-dependent chemotaxis pathway has been identified as a member of this environmental sensing and control system

family; however, it lacks the transcriptional activation of the other pathways. (Figure 4). CheA belongs to the histidine kinase receptor group, whereas CheY and CheB are regulator type proteins (132). The similarities between the different two-component systems allows for cross-talk between them, which suggests a common evolutionary origin for environmental sensing and response. Thus the chemotaxis system possibly evolved from one originally involved in tran scriptio nal activation . This hypothesis is supported by data from R. sphaeroides, where a two-component system used for transcriptional control may also be used for chemotactic signaling (see below). Small Molecules and Signaling

Recent data suggest that in addition to CheY P several intermediate metabo lites may play a role in switching. In H. halobium, switching has been restored to smooth swimming mutants by the addition of a few molecules per cell of fumarate (89). Switching occurs in E. coli mutants lacking components of the sensory transduction pathway if acetyl adenylate is present, and this compound is required for normal switching in wild-type strains ( 1 52). How these two compounds interact with the motor or with CheY is unknown, as is their -

BACTERIAL BEHAVIOR

699

general role in tactic signaling. Acetyladenylate is not an intermediate in all bacterial species and therefore cannot have a universal role in switching.

RECEPTOR-INDEPENDENT SIGNALING Most bacterial species also respond to a wide range of stimuli through


metaboliC pathways, using at least partial metabolic processing to generate a signal that must be integrated with the signals from the dedicated sensory transducers. These stimuli include many of the most basic requirements for bacterial growth, i.e. oxygen and other electron acceptors, e.g. fumarate (24), light in the case of photosynthetic bacteria, most sugars, and a range of electron donors, e.g. proline. The responses of enteric bacteria to some carboxylic acids also appears to depend on their metabolism. It is possible that these responses, although weak in the laboratory, may be more important in the nutrient-limited natural environment (reviewed in ( 1 0, 137».

Chemotaxis to Phosphotransferase Sugars Although E.

coli responds to the sugars galactose and maltose through MCPs,

these organisms still show a positive response to a wide range of carbohy drates in the absence of MCP's, or under conditions of s-adenosyl methionine or CheB/R deletion (74, 139). The chemotactic response is dependent on specific phosphoenolpyruvate-dependent phosphotransferase sugar transport (PTS)

(73). Transport of each carbohydrate requires specific membrane

spanning enzymes II(EII) and EIII, which may be a part of a single protein with Ell, as in n-acetyl glucosamine transport, or a separate soluble protein, as in glucose transport ( 112). During transport the carbohydrate is phosphory lated via a non-specific enzyme EI, which transfers phosphate from phos phoenolpyruvate to a histidine protein (HPr), and thence to the ElI/III mem brane complex. All PTS chemotaxis mutants analyzed show a correlation between transport and chemotaxis ( 1 07). A maximum response is seen only when the specific ElI/III is fully induced and the kinetics of transport and chemotaxis correlate. Sugars, which can be transported in the phosphorylated form, give no chemotactic response, whereas non-metabolizable analogues, which can be transported and subsequently phosphorylated, do give a response, even though they act as an energy drain on the cell (73). Hybrid proteins, with the

EIII of one complex (n-acetyl glucosamine) replaced by the soluble EIII of the glucose PTS system (74), respond to and transport n-acetyl glucosamine, but not glucose. Sorbose PTS ElI/III from a naturally non-motile species,

siella pneumoniae,

when transferred to E.

coli,

Kleb

allows both transport and

chemotaxis to sorbose (121). Mutants have also been isolated that lack the EI kinase activity, but retain some ElI/III activity. They allow slow transport


700

ARMITAGE

of the sugar in the absence of phosphorylation, but no chemotaxis occurs. Therefore, all the evidence suggests that chemotaxis to PTS carbohydrates depends on the change in phosphorylation of HPr in response to phosphoryla tion-dependent transport of the carbohydrates. What is the signal from the transport and phosphorylation system to the flagella? Mutants that retain normal chemotaxis to MCP-dependent chemoat tractants, but have lost the response to PTS sugars, are mutants in either EI or HPr. Fructose, which is transported through a similar system, but uses a simpler FPr rather than HPr protein for phosphorylation, elicits no chemotac tic response. Although FPr can substitute for HPr in transport, it cannot substitute in chemotaxis. HPr mutants with a reduced phosphorylation abi lity show a correlation between the level of HPr expression and the extent of chemotactic signaling (44). These data suggest that there may be an intracellular protein (phosphoryl-chemotaxis protein, PCP) that responds to the phosphorylation state of the HPr intermediate and signals this to either CheA or CheY (Figure 4) (74). Possibly the balance of HPr/HPr-P alters the activity of the CheA-CheY path, which results in more dephos phorylated CheY and increased smooth swimming in response to the in creased transport of a PTS sugar. Dephosphorylated HPr could cause re duced levels of CheY -P by either inhibiting CheA autophosphorylation or in creasing CheY dephosphorylation. Adaptation, an essential part of any chem otactic system, could result from different timescales for the restoration of equilibrium within the PTS transport system and the resetting of the Che pathway. The MCP and PTS pathways are not the only systems involved in sugar chemosensing. R. sphaeroides responds to a wide range of metabolite sugars, but has only one FPr-dependent PTS system and no MCP system. (W. A. leziorska & 1. P. Armitage, unpublished). R. sphaeroides responds well to fructose, which suggests an alternative signaling pathway. Chemoresponse in Rhodobacter sphaeroides In addition to PTS carbohydrates, responses to several other chemoeffectors require at least limited metabolism. As yet little is known about the nature of this signal or the mechanisms involved in integrating these signals with the Che pathway. Receptor-independent chemotaxis is best understood in the purple non sulfur bacterium R. sphaeroides. This species lacks the membrane-span ning MCP and also the PTS-dependent chemotaxis systems. ( 1 3, 120). In vestigation at the DNA and antibody level has failed to identify any com ponents of the intracellular Che signaling pathway (CheA, CheB, CheY); however, this species shows good chemoattractant responses, which suggests an alternative mechanism for controlling flagellar switching. Several as-


BACTERIAL BEHAVIOR

70 1

pects of swimming and chemotactic behavior in this species are different from those identified in other species. which may be indicative of the alternative signaling pathway. R. sphaeroides only rotates its single flagellum in a clockwise direction, and direction changing is brought about by stopping and starting flagellar rotation ( 12). Computerized motion analysis shows great variation in swim ming speeds and stopping frequencies between cells, and when chemotacti cally stimulated, only a percentage of the population responds, rather than the all-or-none response seen in other species ( 1 3). The lack of an all-or-none response, the variability in swimming behavior, and unidirectional flagellar rotation may all reflect the lack of the Che pathway, since in other species the equilibrium of the CheA-P/Che Y fine-tunes the switching frequency, and the dedicated Che sensory pathway ensures a complete response to chemoeffec tors. PTS sugars produce a much weaker response in enteric bacteria com pared with the chemoeffectors sensed through the MCPs, possibly because the signal transduction process feeds into the Che pathway through a metabolic change rather than through a dedicated component of the Che pathway. The major attractants are organic acids, sugars and polyols, nitrogenous compounds such as glutamate or ammonia, and cations such as K+ and Rb + . All the chemoeffectors identified in R. sphaeroides are metabolites and non-metabolizable, but transportable compounds cause no response, e.g. 2-aminoisobutyrate. Similarly, if a chemoeffector is transported but metabo lism is blocked, e.g. glutamate by methyl sulfoxamine, then the chemoattrac tant response is also lost, which suggests that not only transport but at least limited metabolism is required (109). For several years the mechanisms involved in the chemotactic responses of R. sphaeroides remained confusing because the major response to the addition of chemoattractants was a sustained increase in swimming speed, a response incompatible with an attractant response to a gradient. Recently it has become apparent that R. sphaeroides can show two independent behavioral responses to chemoeffectors: a sustained increase in swimming speed (chemokinesis) with an increase in limiting metabolite, and a change in stopping frequency (chemotaxis) in response to a chemoeffector gradient. CHEMOKINESIS A step-up in the concentration of a currently limiting metabolite causes a sustained increase in the mean swimming speed of the population, the speed increasing by up to 20% depending on the extent of starvation for that metabolite ( 1 1 1). Although the speed increase results from an increased rate of flagellar rotation, it is independent of the proton motive force (pmt) or changes in electron transport and is sustained for as long as the metabolite is present at the increased concentration ( 1 10). Chemokinesis correlates with measurable cell swelling in response to chemoeffector uptake


702

ARMITAGE

(J. Armitage, unpublished data). It is only seen in compounds that cause an increase in cell size, including non-metabolites such as betaine, but not in chemoeffectors such as sugars, or non-transported osmolites such as sucrose. Cell swelling may cause a physical change in the interaction of the Mot proteins with the flagellar motor. Alternatively there may be a sensory pathway from stretch receptors in the outer membrane to the flagellar motor (87). A chemokinetic response results in the popUlation of bacteria spreading within the environment in response to an increase in metabolite concentration. The bacteria, although now generally swimming faster, can still sense and respond to a gradient of chemoeffector. R. sphaeroides also shows a chemotactic response for which transport and limited metabolism are essential (54). Support for this comes from the observation that all identified chemotactic mutants have metabolic defects. The phenotype of one such mutant is consistent with a defect in the regulator of the two-component sensing systems involved in both transcrip " tional regulation and chemotactic signaling (Figure 4). (See also discussion of these sensor-regulator systems above). The Che pathway of enteric bacteria seems to have evolved from these two-component systems that are involved in transcription of genes for inducible metabolic pathways, and the Che-minus species, such as R. sphaeroides, may have evolved to utilize the same transcription regulators in chemotaxis that are needed to induce the metabo lism of its chemoattractants. CHEMOTAXIS

Electron-Transport-Dependent Taxis Most motile bacterial species respond to electron acceptors or compounds that perturb the rate of electron transport. The aerotactic response, described in 1883, by Engelmann, showed that different bacterial species accumulate at different distances from an air bubble on a microscope slide (33). It is probable that for most free-living species the responses to electron acceptors may be among the most important, i.e. maintaining bacteria in environments where the pmf can be sustained. Pmf-driven motility has a lower threshold than other pmf-dependent systems, e. g. ATP synthesis or transport, which reflects the importance of motility for bacteria under environmental stress (65). Even though electron transport taxis was recognized long ago, it is one of the least understood responses. Any change in electron transport also affects the driving force for flagellar rotation, and the separation of a direct effect on the motor from the real chemotactic response has proved difficult. However, in all cases the system has been shown to be independent of the Che and MCP systems (10).

+ attractant

a

t repellent

P18


--

H+ H+ [ Oo! , nv,OMSOetc J

b

2 component gene expression ATP change in

environmental metabolite

metabolic pathway regulator -----I.� operons (DNA binding)

� sensor

+-

Che dependent chemotaxis chemoeffectors -----I.�

alternative sigma factor

ATP MGPs

..

-� -"---..,�

GheA

...

_ _ _

�

./ adaptation CheB CheY

", motor

Possible environmental sensing in R. sphaeroides

-iATP

metabolic chen'loeffector

sensor

alternative sigma factor

Figure 4

metabolic pathway operons regulator

motor

(a) Possible mechanisms for integration of signals through MCP, PTS, metabolic, and

electron transport pathways (from 91). (b) Comparison between two-component sensing systems and the chemotactic pathways in enteric bacteria and the possible sensing system in R .

sphaeroides.

AEROTAXIS No aerotaxis mutants have been isolated with nonnal respira tory electron transport, and results from a wide range of species suggest that active electron transport is essential for a sensory response to oxygen; simply binding of oxygen to a tenninal cytochrome does not generate a signal. Aerotaxis can therefore be considered receptor-independent, metabolism dependent behavior. Bacteria are able to induce different tenninal reductase


704

ARMITAGE

enzymes under different growth conditions, which result in complex branched pathways with the branch point and site of the terminal cytochrome dependent on the redox potential of the terminal acceptor. E. coli, when grown anaerobically on nitrate, responds positively to the electron acceptor nitrate, but when grown aerobically, or if nitrate reductase is inhibited by the presence of oxygen, lose the response to nitrate ( 1 1 8, 1 37) . Active electron transport to the terminal acceptor is therefore necessary for a response. Inhibition of the electron transport also causes the loss of response unless the block is bypassed by a more electropositive electron donor. In addition, a change in the pmf may also be necessary, since R. sphaeroides shows an aerotactic response only when electron transport causes a change in pmf ( 1 1) . The sharing of components of electron transport between different pathways allows the integration of different sensory signals, e. g. a pulse of oxygen (which would be an attractant under aerobic conditions) causes a transient change in photosynthetic electron transport and thus a repellent response in photosynthetically growing cells. In vivo this may serve to stop photosynthetically growing bacteria from entering aerobic environments. A change in redox potential of an electon transport component rather than the change in pmf could be the signal, as the two are difficult to separate. A role for the redox state on an electron transport component is supported by the finding that Azospirillum brasilense accumulates at a specific position in an artificially generated redox gradient (43a, I. Zhulin, personal communica tion). Many obligate anaerobes are repelled by oxygen, and even aerobes are repelled by hyperbaric oxygen (117). The repellent response has different concentration characteristics from the attractant response, which suggests a different receptor system for repellent action. The receptor that may respond to a toxic byproduct of oxidation, e.g. superoxide or hydrogen peroxide, has not been identified. EUBACTERIAL PHOTOTAXIS

As with aerotaxis, phototaxis in eubacteria is complicated by photosynthetic electron transport generating the pmf and also controlling cellular metabolism. The most obvious photoresponse shown by photosynthetic bacteria is the photophobic response, where bacteria reverse their swimming direction when swimming over a light/dark boundary (for review see 45). No response is obvious if cells swim over the reverse darkJlight boundary, which suggests that only a step-down in light intensity can generate a change in switching behavior, although a transient increase in swimming speed may be seen after a step-up in intensity (J . P. Armitage, unpublished observation). In theory this response to a fall in light intensity can operate in the absence of adaptation; bacteria swimming normally under constant light reverse into


BACTERIAL BEHAVIOR

705

higher light when they swim over a light boundary. A large number of photosynthetic species do have polar flagella and do change direction by reversal, making this a feasible mechanism. Species such as R. sphaeroides also accumulate in the light but stop to change direction. Increased stopping on crossing a light/dark boundary would trap bacteria in the dark, not the light. Therefore there must be some memory and subsequent adaptation in the system to allow for accumulation. If a reversing species such as R. rubrum is subjected to a step-down in light and maintained at the new low level, reversals continue for about 30 sec, which shows a memory and adaptation similar to chemotaxis systems (25). The time course of the altered switching frequency is longer than would be expected for photosynthetic electron transport to re-equilibrate, therefore motor switching is probably controlled by a cellular metabolite. Phototaxis has been used frequently to isolate photosynthetic mutants and shows -the close correlation between photosynthetic and the photophobic response. Clayton showed that the photophobic and photosynthetic action spectra were very closely aligned (25, 26). Reaction center mutants with normal pigment complements have lost phototaxis, but not other tactic re sponses (9), and inhibition of photosynthetic electron transport with specific inhibitors causes the loss of the photoresponse, all suggesting a close link between photosynthetic activity and the behavioral response (46) . The tran sient change in photosynthetic activity caused by a 1 % decrease in light intensity results in a response in R. rubrum (25). Responses are additive, and if bacteria are subjected to two sub-threshold reductions in light intensity, there is still a photophobic response, providing that the gap between stimuli is less than 3 sec and more than 250 msec (27). The strength and duration of the response is also dependent on the size of the stimulus and its duration, e.g. a 1.5 sec step-down close to the response threshold will cause a response similar to a larger 0 . 5 sec step-down. The additive behavior of the response and the time course suggest that the signal that causes the response is not the pmf change directly interacting on the flagellar motor, but rather a metabolic signal responding to the photosynthetic change. Halobacterial Phototaxis Much is known about the processes involved in the positive and negative photoresponses shown by halobacteria because they are dependent on specific retinal-containing photoreceptors. These are dedicated sensory proteins that have no effect on metabolic activity and no effect on the membrane poten tial of the cell ( 1 00). Thus they can be studied without the complications of pmf changes or alterations in metabolic activity (reviewed in 10 1, 1 23, 1 25).

ARMITAGE

706


Itransducerl

(SR_U480) slow

fast

I

�

r�

methylation system

�

I - � -7 �-7 motor

1/

Itransducerl

SR-D350

Figure 5

Photosensory pathways in H. halobium. Photocycles from the different sensory

rhodopsin send sensory signals to the motor. The photocycle of the constitutive SRII produces a repellent signal, while the inducible SRI produces a repellent signal when stimulated through a fast photocycle by light at 373nm but an attractant signal from the slow cycle produced by 587nm. See text for detail. Adapted from 89, ! O I .

When grown aerobically using respiratory electron transport, H. halobium has a blue-light receptor SRII that absorbs light at 480 nm, undergoes a simple photocycle, and generates a repellent response (88, 136). This constitutive pigment serves to keep H. halobium out of strong damaging UV and blue light . If oxygen and nutrient levels are low, however, H. halobium induces the pigment bacteriorhodopsin, which acts as a light-driven proton pump to maintain the pmf across the cytoplasmic membrane. Under these conditions a second photopigment, SRI, is induced, which is responsible for keeping the cells both in regions of orange light, and also out of blue light ( 124). The pigment goes through a complex cycle , it absorbs light at 587 nm and generates an intermediate that can absorb the repellent wavelength of light (373 nm). If there is no repellent light, then the pigment shows a slow (750 ms) thermal decay back to 587 nm, and this generates a positive response, whereas if light at 373 nm is absorbed by the intermediate, there is a fast decay (80 ms), and a repellent signal is generated (Figure 5). In both cases the signal is generated by the retinal undergoing an all cis to all trans iso merization ( 1 55). As mentioned earlier, H. halobium uses MCPs for both chemoreception and photoreception, and there is most likely a common sensory pathway ( 134). However, unlike the chemotactic stimuli, photosti muli cause no measurable change in the methylation pattern of the MCPs, although release of radiolabeled methyl groups from the MCPs can be meas ured. This suggests that during phototaxis , as with B. subtilis chemotaxis , the turnover rate of methyl groups changes, but the absolute level does not ( 122)-

B ACTERIAL BEHAYIOR

707

Allowing for different time courses for chemoreception and photoreception

(4), it is possible that a specific MCP is involved in photoresponses. How MCP-dependent signaling interacts with fumarate, which is shown to be released from the cell membrane after a flash of blue-light (89), and can cause switching is currently unknown.


OTHER TAXES

Bacteria can show responses to a range of diverse stimuli

that have been much less studied than the responses described above. These

(87 , 1 13), the viscosity of the environment (28), and the earth' s magnetic field (19, 85).

involve responses to changes in the osmotic state of the environment

SUMMARY As has been stated, bacteria are able to sense a wide range of environmental stimuli through a variety of receptors and to integrate the different signals to produce a balanced response that maintains them or directs them to an optimum environment for growth. In addition, these simple, neuron-less organisms can adapt to the current concentration or strength of stimuli, i . e . they have a memory o f the past. Although different species show responses to different chemicals or stimuli, depending on their niche, a consistent pattern is starting to emerge that links environmental sensing and transcriptional control to the chemosensing system, either directly, as in R. sphaeroides and the PTS system, or indirectly, as in the MCP-dependent system. This suggests a common evolutionary pathway from transcriptional activators to dedicated sensory systems . Currently the majority of detailed investigations into bacte rial behavior have been carried out on single stimuli under laboratory con ditions using well-fed cells. Only limited analysis , using a range of rhizo sphere and pathogenic species , has been carried out on the role of behavioral responses in the wild. While laboratory studies are needed to provide the backbone for eventual in vivo investigations, we should remember the re sponses of whole cells to changes in their environment under laboratory conditions are essentially artificial compared to the natural environment of most species. Once the basic system is understood, it will be possible to investigate the role of these responses in vivo, under competitive, growth limiting conditions with multiple gradients. ACKNOWLEDGMENTS

The work described on R. sphaeroides was supported by grants from the Wellcome Trust, the SERC and the Leverhulme Foundation.

708

ARMITAGE


Literature Cited 1 . Adler, J. 1 969. Chemoreceptors in bac teria. Science 166: 1588-97 2. Adler, J. 1987. How motile bacteria are attracted and repelled by chemicals: an approach to neurobiology. BioI. Chem. Hoppe-Seyler 368 : 1 63-78 3. Aizawa, S.-I. , Dean, G. E . , Jones, C. J., Macnab, R . M., Yamaguchi, S . 1985 . Purification and characterization of the flagellar hook-basal body complex of Salmonella typhimurium J. Bacterial. 1 6 1 :836-49 4. Alam, M . , Lebert, M . Oesterhelt, D . , Hazelbauer, G . L. 1989. Methyl accepting taxis proteins in HaLobacteri um halobium. EMBO J. 8:631-39 5. Alam, M . , Oesterhelt, D. 1984. Morphology, function and isolation of halobacterial flagella. J. MoL. BioL. 1 76:459-75 6. Alam, M . , Oesterhelt, D. 1987 . Purification, reconstruction and poly morphic transition of halobacterial flagella. J. MoL. BioL. 194:495-99 7 . Albright, L. M . , Huala, E . , Ausubel, P. M . 1989. Prokaryotic signal transduc tion mediated by sensor and regulator protein pairs. Annu. Rev. Genet. 23: 3 1 1-36 8. Ames, P . , Chen, J . , Wolff, C . , Parkin son, J. S . 1988. Structure-function stud ies of bacterial chemosensors. CoLd

Spring Harbor Symp .

Quant. BioL.

53:59-65 9. Armitage , J. P . , Evans, M. C. W. 1 98 1 . The reaction centre in the phototactic and chemotactic responses of Rhodo

pseudomonas sphaeroides. FEMS Mic robiol. Lett. I I :89-92

1 0 . Armitage, J. P . , Havelka, W. A . , Sock ett, R. E. 1 990. Methylation-independ ent taxis in bacteria. See Ref. l l a, pp. 177-97 I I . Armitage , J. P . , Ingram, C . , Evans, M . C. W . 1985. Role of proton motive force in phototactic and aerotactic responses in

Rhodopseudomonas sphaeroides. BacterioL. 1 6 1 :967-72

J.

l l a. Armitage, J. P . , Lackie , J . M . , eds. 1 990. Biology of the Chemotactic Re

sponse. Soc. Gen. Microbiol. Symp. 46th. Cambridge: Cambridge Univ.

Press 1 2 . Armitage , J. P . , Macnab, R. M. 1987. Unidirectional intermittent rotation of the flagellum of Rhodobacter sphaer oides. J. BacterioL. 169:5 14-18 13. Armitage, J. P., Poole, P . S . , Brown, S. 1990. Sensory signalling in Rhodobacter sphaeroides. In MolecuLar Biology of

Membrane

Bound

Complexes

in

Phototrophic Bacteria, ed. G . Drews.

pp. 463-72. New York: Plenum. 14. Deleted in proof 1 5 . Armstrong, J . B . 1972. An S-adenosyl methionine requirement for chemotaxis in Escherichia coli. Can. J. Microbial. 1 8 : 1 695-1701 1 6 . Berg, H. C., Bromley, D. B . , Charon, N. W. 1978. Leptospiral motility. Symp. Soc. Gen. Microbiol. 28:285-94 1 7 . Berg, H. C . , Khan, S. 1983. A model for the flagellar rotary motor. See Ref. 133a, pp. 486-97 1 8 . Blair, D. P . , Berg, H. C. 1990. The MotA protein of E. coli is a proton conducting component of the flagellar motor. Cell 60:439-49 ! 8a. Blair, D . P . , Kim, D . Y . , Berg, H. C . 1 99 1 . Mutant MotB proteins i n Escherich ia coli. J. Bacteriol. 173:4049-55 19. Blakemore , R. P. 1982. Magnetotactic bacteria. Annu. Rev. Microbial. 36:2 1 738 20. Block, S. M . , Segall, J. E . , Berg, H. C . 1982. Impulse responses i n bacterial chemotaxis. Cell 3 1 :2 1 5-26 2 1 . Bollinger, J . , Park, C . , Harayama, S . , Hazelbauer, G . L . 1984. Structure o f the Trg protein: homologies with and differ ences from other sensory transducers of

Escherichia coli. Proc. Natl. Acad. Sci. USA 8 1 :3287-9 1

22. Bourret, R. B . , Borkovich , K. A . , Si mon, M. I. 1 99 1 . Signal transduction pathways involving protein phosphor ylation in prokaryotes. Annu. Rev. Biochem. 60:401-44 23. Charon, N. W . , Goldstein, S. P. , Curci, K, Limberger, R . J . 199 1 . The bent end morphology of Treponema phagede nis is associated with short, left-handed, periplasmic flagella. J. Bacteriol. 1 73: 4820-26 24. Clancy, M . , Madill, K. A . , Wood, J . M . 1 98 1 . Genetic and biochemical re quirements for chemotaxis to L-proline in Escherichia coli. J. Bacteriol. 146: 902-6 25. Clayton, R. K. 1953. Studies in the phototaxis of Rhodospirillum rubrum . II. The relation between phototaxis and photosynthesis. Arch. Microbiol. 19: 1 25-40 26. Clayton, R . K. 1953 . Studies in the phototaxis of Rhodospirillum rubrum. I. Action spectrum, growth in green light and Weber-law adherence. Arch Micro bioi. 1 9 : 1 07-24 27. Clayton, R. K. 1953. Studies in the phototaxis of Rhodospirillum rubrum. III. Quantitative relationship between

BACTERIAL BEHAVIOR

28.

29.


30.

31.

32.

33.

34.

35.

stimulus and response. Arch. Microbial. 19: 1 4 1-65 Daniels, M. J . , Longland, J. M . 1984. Chemotactic behavior of spiroplasms. Curro Microbial. 10: 19 1-94 Dingwall, A . , Shapiro, L . , Ely, B . 1990. Flagellar biosynthesis in Caulobacter. See Ref. l l a, pp. 155-76 Dworkin, M . , Keller, K. H . , Weisberg, D. 1983. Experimental observations consistent with a surface tension model of gliding motility of Myxococcus xan thus. J. Bacterial. 155: 1367-7 1 Eisenbach, M . , Constantinou, C . , Alo ni, H . , Shinitzky, M. 1990. Repellents for Escherichia coli operate neither by changing membrane fluidity nor by be ing sensed by periplasmic receptors dur ing chemotaxis. J. Bacterial. 172:52 1 824 Eisenbach, M . , Wolf, A . , Welch, M . , Caplan, S . R . , Lapidus, I. R . , e t al. 1990. Pausing, switching and speed fluctuations of the bacterial flagellar motor and their relation to motility and chemotaxis. J. Mol. Bioi. 2 1 1 :55 1-63 Engelmann, T. W. 1883. Bakterium photometricum. Ein Beitrag zur ver gleichenden Physiologie des Licht-und Farbensinnes. Pflugers Arch. Gesamte Physiol. Menschen Tiere 42: 183-86 Falke, J. J . , Dernburg, A. F. , Sternberg, D . A . , Zalkin, N . , Milligan, D. L . , Koshland, D . E. Jr. 1988. Structure o f a bacterial sensory receptor. J. Bioi. Chem. 263: 14850-58 Falke, J. J . , Koshland, D. E. Jr. 1987. Global flexibility in a sensory receptor: A site-directed cross-linking approach.

Science 237 : 1 596-1600

36. Gegner, J. A. , Dahlquist, F. W. 199 1 . Signal transduction i n bacteria: CheW forms a reversible complex with the pro tein kinase CheA. Proc. Natl. Acad. Sci. USA 88:750-54 37. Glagolev, T. N . , Glagolev, A. N . , Gusev, M . V . , Nikitina, K . A . 1980. Proton motive force supports gliding in cyanobacteria. FEBS Lett. 1 17 :49-53 38. Goldman, D. J . , Ordal, G. W. 1984. In vitro methylation and demethylation of methyl-accepting chemotaxis proteins in Bacillus subtilis. Biochemistry 23: 2600-6 39. Goldstein, S . F. , Charon, N. W. 1988. Motility of the spirochete Leptospira. Cell Mati!. Cytoskeleton 9: 101-10 40. Goldstein, S. F. , Charon, N. W . 1990. Multiple-exposure photographic analysis of a motile spirochete. Proc. Natl. Acad. Sci. USA 87:4895-99 4 1 . Gotz, R . , Schmitt, R. 1987. Rhizobium meliloti swims by unidirectional in-

709

termittent rotation of right-handed flagellar helices. J. Bacterial. 169: 3 1 46-50 42. Goulbourne, E. A. Jr. , Greenberg, E. P . 1983. A voltage clamp inhibits chemo taxis of Spirochaeta aurentia . 1. Bac terial. 1 53:9 16-20 43. Greenberg, E. P . , Canale-Parola, E. 1977. Relationship between cell coiling and motility of spirochetes in viscous environments. J. Bacterio!. 1 3 1 :960-69 43a. Grishanin, R. N . , Chacmina, I. I . , Zhulin, 1. B . 1991. Behaviour of Azos pirillium brasilense in a spatial gradient of an artificial electron acceptor. 1. Gen. Microbial. In press 44. Griibl, G . , Vogler, A. P . , Lengeler, J. W. 1990. Involvement o f the histidine protein (HPr) of the phosphotransferase system in chemotactic signaling of Es cherichia coli K- 12. Bacterial. 172: 587 1-76 45. Hader, D.-P. 1987. Photosensory be havior in prokaryotes. Microbial. Rev. 5 1 : 1-21 46. Harayama, S. 1977. Phototaxis and membrane potential in the photosynthet ic bacterium Rhodospirillum rubrum. J. Bacterial. 1 3 1 :34-4 1 47. Harwood, C. S. 1989. A methyl accepting protein is involved in benzoate taxis in Pseudomonas putida . J. Bacteri al. 1 7 1 :4603-8 48. Hazelbauer, G. L . , Yaghmai, R . , Bur rows, G. G . , Baumgartner, J. W. , Dut ton, D. P . , Morgan , D. G. 1990. Transducers: transmembrane receptor proteins involved in bacterial chemotax is. See Ref. I l a, pp. 107-34 49. Hess, J. F . , Bourret, R. B . , Oosawa, K . , Matsumura, P., Simon, M . I. 1988. Protein phosphorylation and bacterial chemotaxis. Cold Spring Harbor Symp . Quant. BioI. 53:41-48 50. Homma, M . , DeRosier, D. J . , Macnab, R. M. 1990. Flagellar hook and hook associated proteins of Salmonella typhi murium and their relationship to other axial components of the flagellum. J. Mol. Bioi. 2 13:8 19-32 5 1 . Homma, M . , lino, T. 1985. Location of hook-associated proteins in flagellar structures of Salmonella typhimurium . J. Bacterial. 162: 183-89 52. Homma, M . , Kutsukake, K . , Hasebe, M . , lino, T . , Macnab, R. M. 1 990. FJgB , F1gC, FlgF and F1gG . A family of structurally related proteins in the flagel lar basal body of Salmonella typhimur ium. J. Mol. Bioi. 2 1 1:465-77 53. Imae, Y . 1985. Molecular mechanism of thermosensing in bacteria. In Sensing and Response in Microorganisms, ed.

710

54.

55.


56.

57.

ARMITAGE M. Eisenbach, M. Balaban, pp. 73-8 1 . Amsterdam: Elsevier Ingham, C. J . , Armitage, J. P. 1987. Involvement of transport in Rhodobacter sphaeroides. J. Bacteriol. 1 69:580 1-7 Ishihara, A., Segall, J. E . , Block, S . M . , Berg, H . C. 1983. Coordination of flagella on filamentous cells of Es cherichia coli. J. Bacteriol. 155:228-37 Jin, S . , Roitsch, T. , Christie, P. J . , Nes ter, E. W. 1 990. The regulatory VirG protein specifically binds to a cis-acting regulatory sequence involved in trans criptional activation of Agrobacterium tumefaciens virulence genes. J. Bacteri ol. 172:531-37 Jones, C. J . , Macnab, R. M. 1990. Flagellar assembly in Salmonella typhi murium: Analysis with temperature sensitive mutants . J. Bacteriol. 172:

1 3 27-39 5 8 . Joys, T. M. 1988. The flagellar filament protein. Can . J. Microbiol. 34:452-58 59. Kagawa, H . , Morishita, H . , Enomoto, M. 1 98 1 . Reconstruction in vitro of

flagellar filaments onto hook structures attached to bacterial cells. J. Mol. Biol.

15 3:465-70 60. Kehry, M. R . , Bond, M. W . , Hunkapil ler, M. W . , Dahlquist, F. W. 1983 .

61.

62.

63.

64.

65.

66.

Enzymatic deamidation of methyl accepting chemotaxis proteins in Es cherichia coli catalyzed by the cheB gene product. Proc. Natl. Acad. Sci. USA 80:3599-3603 Kehry, M. R. , Dahlquist, F. W. 1982. The methyl-accepting chemotaxis pro teins of Escherichia coli. Identification of the multiple methylation sites on methyl-accepting chemotaxis protein 1 . J. Bioi. Chem. 257: 1 0378-86 Khan, S . , Dapice, M . , Humayun, I. 1 990. Energy transduction in the bacte rial flagellar motor. Effects of load and pH. Biophys. J. 57:779-96 Khan, S . , Dapice, M . , Reese, T. S . 1988. Effects of mot gene expression on the structure of the flagellar motor. J. Mol. Bioi. 202:575-84 Khan, S . , Khan, I. H . , Reese, T. S . 1 99 1 . New structural features o f the flagellar base in Salmonella typhimur ium revealed by rapid-freeze electron microscopy. J. Bacterial. 173:2388-96 Khan, S . , Macnab, R. M. 1 980. The steady state counter-clockwise/clock wise ratio of the bacterial flagellar motors is regulated by proton motive force. J. Mol. Bioi. 138:563-97 Kihara, M. , Macnab, R. M. 1 98 1 . Cytoplasmic pH mediates pH taxis and weak-acid repellent taxis of bacteria. J. Bacteriol. 145: 1 209-2 1

67. Kreig, N. R . , Tomeity, 1. P . , Wells, J . S . 1967. Inhibition o f flagellar coordina

tion in Spirillum volutans. J. Bacterial.

94:143 1-36 68. Krikos, A . , Conley, M. P . , Boyd, A . , Berg, H . C . 1985 . Chimeric che

mosensory transducers of Escherichia coli. Proc. Natl. Acad. Sci. USA

82: 1326-30 69. Krikos, A . , Mutoh, N . , Boyd, A . , Si mon, M. 1983 . Sensory transducers of

E. coli are composed of discrete structu ral and functional domains. Cell 33:6 1 5-

22 70. Kuo, S. C . , Koshland, D. E. Jr. 1 987.

Roles of chef and cheZ gene products in controlling flagellar rotation in bacterial chemotaxis in Escherichia coli. J. Bac teriol. 169: 1 307- 1 4 7 1 . Kupper, J . , Wildhaber, I., Gao, Z . , Baeuerlein, E. 1989. Basal-body associated disks are additional structural elements of the flagellar apparatus iso lated from Wolinella succinogens. J. Bacteriol. 1 7 1 :2803-1 0 72. Laszlo, D . J., Fandrich, B . L . , Sivaram, A . , Chance, B . , Taylor, B. L. 1 984. Cytochrome 0 as a terminal oxidase and receptor for aerotaxis in Salmonella typhimurium. J. Bacteriol. 1 59:663-67 73. Lengeler, J . , Auburgcr, A-M . , Mayer, R . , Pecher, A. 1 98 1 . The phosphoenol pyruvate-dependent carbohydrate:phos photransferase system enzyme II as che moreceptors in chemotaxis of Es cherichia coli K-12. Mol. Gen. Genet. 1 83: 1 63-70 74. Lengeler, J. W . , Volger, A. P. 1989.

Molecular mechanisms of bacterial che motaxis towards PTS-carbohydrates. FEMS Microbial. Rev. 63:8 1-92 75 . Loake , G. J . , Ashby, A. M . , Shaw, C . H . 1988. Attraction o f Agrobacterium tumefaciens C58Cl towards sugars in volves a highly sensitive chemotaxis system. J. Gen. Microbial. 1 34: 1 42732 76. Lupas, A . , Stock, J. 1989. Phosphoryla

tion of an N-terminal regulatory domain activatcs the CheB methylesterase in bacterial chemotaxis. J. Biol. Chem.

264: 17337-42 77. Macnab, R. M . 1987. Flagella. See Ref. 96a, pp. 70-83 78. Macnab, R. M. 1987. Motility and chemotaxis. See Ref. 96a, pp. 73259 79. Macnab, R. M. 1990. Genetics, struc

ture, and assembly of the bacterial flagellum. See Ref. 1 1 a, pp. 77-107 80. Macnab, R. M . , DeRosier, D. J. 1988. Bacterial flagellar structure and func tion. Can. J. Microbial. 34:442-5 1


BACTERIAL BEHAVIOR

711

8 1 . Macnab, R. M . , Han, D. P. 1983 . Asynchronous switching of flagellar motors on a single cell. Cell 32: 1 09-1 7 8 2 . Macnab, R. M . , Ornston, M. K . 1977. Normal-to-curly flagellar transitions and their role in bacterial tumbling. Stabil ization of an alternative quaternary structure by mechanical force. 1. Mol. Bioi. 1 1 2: 1-30 8 3 . Maeda, K . , Imae, Y. 1979. Thermosen sory transduction in Escherichia coli: in hibition of the thermoresponse by L serine. Proc. Natl. Acad. Sci. USA 76:

1987. Additive and independent re sponses in a single receptor; aspartate and maltose stimuli on the Tar protein. Cell 50: 1 7 1-80 95. Mullin, D. A . , Newton, A. 1 989. Ntr like promoters and upstream regulatory sequence ftr are required for transcrip tion of a developmentally regulated Caulobacter crescentus flagellar gene. J. Bacterial. 1 7 1 :3 2 1 8-27 96. Nara, T . , Lee, L . , Imae, Y. 199 1 . Therrnosensing ability of Trg and Tap chemoreceptors in Escherichia coli. J.

84. Maeda, K . , Imae, Y . , Shioi, J.-I. , Oosawa, F. 1976. Effect of temperature on motility and chemotaxis of Es cherichia coli. J. Bacterial. 1 27 : 1 0391 046 85. Mann, S . , Sparks, N . H. C . , B oard , R . G. 1990. Magnetotactic bacteria: Micro biology, biomineralization, paleomag netism and biotechnology. Adv. Microb. Physiol. 3 1 : 125-81 86. Margolin, Y . , Eisenbach, M. 1 984. Voltage clamp effects on bacterial che motaxis. J. Bacterial. 1 59:605-10 87. Martinac , B . , Buechner, M., Delcour, A. H . , Adler, J . , Kung, C. 1987. Pres sure-sensitive ion channel in Escherichia coli. Proc. Natl. Acad. Sci. USA 84: 2297-2301 88. Marwan, W . , Oesterhelt, D. 1 987. Sig nal formation in the halobacterial photophobic response mediated by a fourth retinal protein (P480). 1. Mol. BioI. 1 95:333-42 89. Marwan, W . , Schafer, W . , Oesterheit, D. 1990. Signal transduction in Halo bacterium depends on fumarate . EMBO J. 9:355-62 90. Matsumura, P . , Roman , S . , Volz, K . , McNally, D . 1 990. Signalling com plexes in bacterial chemotaxis. See Ref. 1 1 a, pp. 1 35-54 9 1 . McBride, M. J . , Weinberg, R. A . , Zus man, D. R. 1 989. "Frizzy" aggregation genes of the gliding bacterium Myxococ cus xanthus show sequence similarities to the chemotaxis genes of enteric bac teria. Proc. Natl. Acad. Sci. USA 86: 424-28 92. Milligan, D. L . , Koshland, D. E. Jr. 1 988. Site-directed cross-linking: Es tablishing the dimeric structure of the aspartate receptor of bacterial chemotax is. J. BioI. Chem. 263:6268-75 93. Mitchell, P. 1984. Bacterial flagellar motors and osmoelectric molecular rota tion by an axially transmembrane well and turnstile mechanism. FEBS Lett. 176:287-94 94. Mowbray, S . L . , Koshland, D. E. Jr.

96a. Neidhardt, F. C . , Ingraham, J . L . , Low, K . B . , Magasanik, B . , Schaech ter, M . , Umbarger, H. E. , eds. 1987.

91-95

Bacteriol. 173: 1 1 20-24

Escherichia coli and Salmonella typhi murium: Cellular and Molecular Biolo gy. Washington, DC: Am. Soc. Micro bioI . 97. Nettleton, D . O . , Ordal, G. W. 1989.

Functional homology of chemotactic methylesterases from Bacillus subtilis and Escherichia coli. ' 1. Bacterial.

1 7 1 : 1 20--2 3

98. Ninfa, A. J . , Ninfa, E. G., Lupas, A. N . , Stock, A . , Magasanik, B . , Stock, J. 1988. Crosstalk between bacterial che motaxis signal transduction proteins and regulators of transcription of the Ntr reg ulon: evidence that nitrogen assimilation and chemotaxis are controlled by a com mon phosphorotransfer system. Proc. Natl. Acad. Sci. USA 85:5492-96 99. Nowlin, D. M . , Bollinger, J . , Hazel bauer, G . L 1988. Site-directed muta tions altering methyl-accepting residues of a sensory transducer protein. Proteins 3 : 1 02-12 10 0 . Oesterhelt, D . , Marwan, W . 1987. Change in membrane potential is not a component of the photophobic transduc tion chain in Halobacterium halobium. J. Bacterial. 169:35 1 5-20 1 0 1 . Oesterhelt, D . , Marwan, W . 1990. Sig nal transduction in Halobacterium halo bium. See Ref. 1 1 a, pp. 2 1 9-40 1 02 . Oosawa, F. , Imae, Y. 1983. Glycerol and ethylene glycol: members of a new class of repellents of Escherichia coli chemotaxis. J. Bacteriol. 154: 1 04-12 103. Ordal, G. W . , Villani, D. P. 1980. Ac tion of uncouplers of oxidative phosphorylation as chemotactic repel lents of Bacillus subtilis. J. Gen. Micro bioi. 1 1 8:471-78 104. Parales, J. Jr. , Greenberg , E. P. 1 99 1 . N-terrninal amino acid sequences and amino acid compositions of the Spir ochaeta aurantia flagellar filament polypeptides. J. Bacterial. 1 73 : 1 35759


712

ARMITAGE

1 05 . Park, C . , Hazelbauer, G. L 1986. Mutations specifically affecting ligand interaction of the Trg chemosensory transducer. 1. Bacterial. 167 : 1 0 1-9 106. Parkinson, J. S. 1988. Protein phos phorylation in bacterial chemotaxis. Cell 5 3: 1-2 1 07 . Pecher, A . , Renner, t , Lengeler, J . 1983. The phosphoenolpyruvate-de pendent carbohydrate: phosphotrans ferase system enzymes II, a new class of chemosensors in bacterial chemotaxis. See Ref. 133a, pp. 5 1 7-3 1 108. Pleier, E . , Schmitt, R. 1989. Identifica tion and sequence analysis of two related fJagellin genes in Rhizobium meliloti. 1. Bacteriol. 1 7 1 ; 1 467-75 109. Poole, P. S . , Armitage, J. P. 1 989. Role of metabolism in the chemotactic re sponse of Rhodobacter sphaeroides to ammonia. 1. Bacteriol . 1 7 1 :2900-2 1 10. Poole, P. S . , Brown, S . , Armitage, J. P. 1990. Swimming changes and chemotactic responses in Rhodobacter sphaeroides .do not involve changes in the steady state membrane potential or respiratory electron transport. Arch. Microbiol. 153:614-18 1 1 1 . Poole, P. S., Williams, R. L, Armi tage, J. P. 1 990. Chemotactic responses of Rhodobacter sphaeroides in the ab sence of apparent adaptation. Arch. Mi crobiol. 153:368-72 1 1 2 . Postma, P. W . , Lengeler, 1. W. 1985. carbohydrate Phosphoenolpyruvate: phosphotransferase system of bacteria. Microbiol. Rev. 49:232-69 1 1 3 . Qi, Y . , Adler, J. 1989. Salt taxis in Escherichia coli bacteria and its lack in mutants . Proc. Natl. Acad. Sci. USA 86:8358-62 1 1 4. Ravid, S . , Matsumura, P. , Eisenbach, M. 1986. Restoration of flagellar clock wise rotation in bacterial envelopes by insertion of the chemotaxis protein CheY. Proc. Natl. Acad. Sci. USA 83:71 57-61 1 I 5 . , Schnitzer, M. J . , Block, S. M . , Berg, H. C , Purcell, E. M. 1 990. Strategies for chemotaxis. See Ref. l 1 a, pp. 1 5-34 1 1 6 . Shapiro, E . , Shapiro, B . , Shapiro, L. 1989. The molecular genetics of differ entiation . Genetics 123:427-29 1 17 . Shioi, J . , Dang; C. Y . , Taylor, B . L. 1 987. Oxygen as attractant and repellent in bacterial chemotaxis. 1. Bacteriol. 169:3 1 1 8-23 1 1 8 . Shioi, J . , Tribhuwan, R. C . , Berg, S . T . , Taylor, B . L. 1988. Signal transduc tion in chemotaxis to oxygen in Es cherichia coli and Salmonella typhimur ium. 1. Bacteriol. 170:5507-1 1

1 1 9. Smith, R. A . , Parkinson, J. S. 1980. Overlapping genes at the cheA locus of

Escherichia coli. Proc. Nat!. Acad. Sci. USA 77:5370-74 1 1 9a. Sockett, R. E . , Armitage, J. P. 199 1 .

Isolation, characterization, and com plementation of a paralyzed flagellar mutant of Rhodobacter sphaeroides WS8. 1. Bacteriol . 173:2786--90 1 20. Sockett, R. E. , Armitage, J. P . , Evans, M. C. W. 1987. Methylation-inde pendent and methylation-dependent che motaxis in Rhodobacter sphaeroides and

Rhodospirillum rubrum . 1. Bacteriol.

1 69:5808-1 4 1 2 1 . Sprenger, G . A . , Lengeler, J . 1984. L sorbose metabolism in Klebsiella pneumoniae and Sor+ derivatives of Es cherichia coli K- 1 2 and chemotaxis towards sorbose. 1. Bacteriol. 157:3945 1 22. Spudich, E. N . , Takahashi, T. , Spu dich, 1 . L. 1989. Sensory rhodopsin I and II modulate a methylation! demethylation system in Halobacterium

halobium. Proc. Natl. Acad. Sci. USA

86:7746--50 123. Spudich, J. L. 1985 . Bacterial sensory rhodopsin (SR), a dual attractant and re pellent phototaxis receptor. In Sensing and response in microorganisms, ed. M . Eisenbach, M . Balaban, pp. 1 1 9-27. Amsterdam: Elsevier Science 124. Spudich, J. L . , Bogomolni, R. A. 1984. Mechanism of colour discrimination by a bacterial sensory rhodopsin. Nature 3 1 2 :509- 1 3 1 2 5 . Spudich, J. L . , Bogomolni, R . A . 1 988. Sensory rhodopsins of halobacteria. Annu. Rev. Biophys. Chem. 1 7 : 1 93215 126. Stallmeyer, M . J . B . , Aizawa, S . -I . , Macnab, R . M . , DeRosier, D . J. 1989. Image reconstruction of the flagellar basal body of Salmonella typhimurium . 1. Mol. Bioi. 205 : 5 1 9-28 127. Stallmeyer, M . 1. B . , Hahnenberger, K . M . , Sosinsky, G. E . , Shapiro, L . , De Rosier, D. J. 1989. Image reconstruc tion of the flagellar basal body of

Caulobacter crescentus. 1. Mol. BioI.

205 : 5 l!- 1 8 1 2 8 . Stewart, R . C , Roth, A . F . , Dahlquist, F. W. 1990. Mutations that affect con trol of the methylesterase activity of CheB, a component of the chemotaxis adaptation system in Escherichia coli. 1. Bacterial. 172:3388-99 129. Stewart, R. C , Russell, C. B . , Roth, A . F. , Dahlquist, F. W. 1988. Interaction of CheB with chemotaxis signal transduction components in Escherichia

BACTERIAL BEHAVIOR coli: Modulation of the methylesterase

activity and effects on cell swimming behavior. Cold Spring Harbor Symp.

Quam. Bioi. 53:27-40


130. Stock, A . M . , Mottonen, J. M . , Stock, J. B . , Schutt, C. E. 1989. Three

dimensional structure of CheY , the re sponse regulator of bacterial chemotax is. Nature 337:745-49 1 3 1 . Stock, J . , Koshland, D. E. Jr . 1985. Homologies between the Salmonella typhimurium CheY protein and proteins involved in the regulation of chemotax is, membrane protein synthesis and spor ulation. Proc. Natl. Acad. Sci. USA

82:7989-93

1 32 . Stock, J. B . , Ninfa, A . J . , Stock, A. M . 1989. Protein phosphorylation and regu lation of adaptive responses in bacteria.

Microbiol. Rev. 53:450-90 1 33 . Stock, J . B . , Stock, A . M . , Mottonen, J. M . 1990. Signal transduction in bac teria. Nature 344:395-400

1 33a. Sund, H . , Veeger, C . , eds.

1983.

Mobility and Recognition in Cell Biolo gy. Berlin: de Gruyter

134. Sundberg, S. A . , Alam, M . , Lebert,

M . , Spudich, J. L . , Oesterhelt, D . , Hazelbauer, G. L. 1 990. Characteriza tion of Halobacterium halobium mutants defective in taxis. J. Bacteriol. 172:

2328-35 1 35 . Suzuki, T . ,

Komeda, Y. 198 1 . In flagellar structures in Es cherichia coli mutants. J. Bacteriol.

complete

145: 1036-1041

713

G . W . 1 990. Methyl group turnover on methyl-I -accepting chemotaxis proteins during chemotaxis by Bacillus subtilis.

J. Bioi. Chem. 265 : 1 928-32 142. Thoelke, M. S . , Kirby , J. R . , Ordal, G . W. 1989. Novel methyl transfer during chemotaxis in Bacillus subtilis. Bio

chemistry 28:5585-89

143. Thomashow , L. S . , Rittenberg , S. C . 1985. Waveform analysis and structure

of flagellar and basal body complexes from Bdellovibrio bacteriovorus 109J.

J. Bacteriol. 163: 1038-1046 144. Thornton, S. A . , Logan, S. M . , Trust, T. J . , Guerry, P. 1 990. Polynucleotide sequence relationships among flagellin genes of Campylobacter jejuni and

Campylobacter coli.

Infect.

Immun.

58:2686--89

145. Toews, M. L . , Goy, M. F . , Springer, M. S . , Adler, J. 1979. Attractants and

repellents control demethylation of methylated chemotaxis proteins in Es

cherichia coli. Proc. Natl. Acad. Sci. USA 76:5544--48 146. Trachtenberg, S . , DeRosier, D. J. 199 1 .

A molecular switch: subunit rotations in volved in the right-to-left handed transi tions of Salmonella typhimurium flagel lar filaments. J. Mol. Bioi. 220:67-77 147. Trachtenberg, S . , DeRosier, D. J . , Mac nab, R. M. 1987. Three-dimensional structure of the complex flagellar fila ment of Rhizobium lupini and its relation to the structure of the plain filament. J.

Mol. Bioi. 195:603-20

1 36 . Takahashi, T . , Yan, B . , Mazur, P . ,

148. Wagenknecht,

Biochemistry 29:8467-74 1 3 7 . Taylor, B. L. 1983. Role of proton mo

87 149. Wei, L.

1 38 .

Bioi. 1 86:79 1-803 150. Weissborn, A . , Steinman, H. M . , Sha

Derguini , F. , Nakanishi, K . , Spudich, J . L. 1990. Color regulation i n the archaebacterial phototaxis receptor pho borhodopsin (sensory rhodopsin II).

1 39 .

140.

141 .

tive force in sensory transduction in bac teria. Annu. Rev. Microbiol. 37:55 1-73 Taylor, B . L . , Johnson, M. S . , Smith, J . M . 1988. Signaling pathways i n bacte rial chemotaxis. Bot. Acta 1 0 1 : 1 0 1-4 Taylor, B . L . , Lengeler, J. W. 1990. Transductive coupling by methylated transducing proteins and permeases of the phosphotransferase system in bacte rial chemotaxis. In Membrane Transport and Information Storage, ed. R . C. Aloia, C. C . Curtain, L. M . Gordon, pp. 69-90. New York: Liss Terwilliger, T. C . , Bogonez, E . , Wang, E. A . , Koshland, D. E. Jr. 1983. Sites of methyl esterification in the aspartate receptor involved in bacterial chemotax is. J. Bioi. Chem. 258:9608-1 1 Thoelke, M . S . , Casper, J . M . , Ordal,

T . , DeRosier, D . J . , Aizawa, 5 . -1. Macnab, R . M . 1 982. Flagellar hook structures of Caulobacter and Salmonella and their relationship to filament structure . J. Mol. Bioi. 162:69-

N . , Joys, T. M . 1985. Covalent structure of three phase-l flagellar fila ment proteins of Salmonella . J. Mol.

piro, L. proteins flagellar filament

1982. Characterization of the of the Caulobacter crescentus filament. Peptide analysis and organization. J. Bioi. Chem.

257:2066--2074 1 5 1 . Wilson, M. L . , Macnab, R . M . 1990.

Co-overproduction and localization of the Escherichia coli motility proteins MotA and MotB. J. Bacterial. 172: 3932-39

152. Wolfe, A. J . , Conley, M. P . , Berg, H. C. 1988. Acetyladenylate plays a role in controlling the direction of flagellar rota tion. Proc. Natl. Acad. Sci. USA 85: 67 1 1 - 1 5

714

ARMITAGE

153. Xu, H . , Dingwall, A . , Shapiro, L. 1989. Negative transcriptional regula tion in the Caulobacter flagellar hierar chy. Proc. Natl. Acad. Sci. USA 86: 665fr60 154. Yamamoto, K . , Macnab, R . M . , Imae, Y. 1990. Repellent response functions

of the Trg and Tap chemoreceptors of


Escherichia coli. J. Bacteriol. 1 72:38388

155. Yan, B . , Takahashi, T . , Johnson, R . , Derguini, F., Nakanishi, K . , Spudich, J . L . 1990. All-transl13-cis isomerization of retinal is required for phototaxis signaling by sensory rhodopsins in Halo bacterium halobium. Biophys. J. 57:

807-14 156. Yonekawa, H., Hayashi, H . , Parkinson, J. S. 1983. Requirement of the cheB

function for sensory adaptation in Es cherichia coli. J. Bacteriol. 156: 1228-

35 157. Yoshida, S . , Sugiyama, S . , Hojo , Y . , Tokuda, H . , Imae, Y . 1990. In

tracellular Na + kinetically interferes with the rotation of the Na + -driven flagellar motors of Vibrio alginolyticus. J. Bioi. Chern. 265:20346-50 158. Zusman, D. R . , McBride, M. J . , McCleary, W. R . , O'Connor, K. A . 1 990. Control of directed motility in Myxococcus xanthus. See Ref. l l a, pp.

1 99-2 1 9 159. Zusman, D. R . , O'Connor, K. A. 1 99 1 . Development in Myxococcus xanthus in

volves aggregation and sporulation as well as non-aggregation and non sporulation. Dev. Bioi. In press

Behavioral responses to chemical cues by bacteria.

Behavioral and electroantennogram responses of periplanone analogs.

Behavioral inhibition of circadian responses to light.

Behavioral avoidance of pathogenic bacteria by Caenorhabditis elegans.

Behavioral Responses in Animal Model of Congenital Muscular Dystrophy 1D.

Behavioral responses to encounter of fishing boats in wandering albatrosses.

Increased behavioral responses to ethanol in Lmo3 knockout mice.

Behavioral responses elicited by intraperitoneal neurotensin in guinea pigs.

The MCPP challenge test in schizophrenia: hormonal and behavioral responses.

Social information embedded in vocalizations induces neurogenomic and behavioral responses.

Autonomic, Behavioral, and Subjective Pain Responses in Alzheimer's Disease.

Physiological and behavioral responses to starvation in the golden hamster.

Endorphins do not affect behavioral stress responses in mice.

Shock-prod burying test in rats: autonomic and behavioral responses.

Mismatched anti-predator behavioral responses in predator-naïve larval anurans.

Behavioral responses to intracerebroventricularly administered neurohypophyseal peptides in mice.

Subjective and behavioral responses to intravenous fentanyl in healthy volunteers.

Gender Difference in Bacteria Endotoxin-Induced Inflammatory and Anorexic Responses.

Neurophysiological and Behavioral Responses of Mandarin Lexical Tone Processing.

Behavioral responses to noxious stimuli shape the perception of pain.

Behavioral and physiological responses of dairy goats to food thwarting.

Behavioral responses of Atlantic cod to sea temperature changes.

Behavioral responses associated with a human-mediated predator shelter.

Modifications by lithium of behavioral responses to methamphetamine and tetrabenazine.