Expression of cell polarity during Caulobacter differentiation

Expression of cell polarity during Caulobacter differentiation

seminars in D E V E L O P M E N T A L B I O L O G Y , Vol 6, 1995: pp 3-11 Expression of cell polarity during Caulobacter differentiation Craig Step...

823KB Sizes 0 Downloads 6 Views

seminars in D E V E L O P M E N T A L

B I O L O G Y , Vol 6, 1995: pp 3-11

Expression of cell polarity during Caulobacter differentiation Craig Stephens, Urs Jenal and Lucille Shapiro

The bacterium Caulobacter crescentus generates two distinct progeny cells, a motile swarmer cell and a sessile stalked cell, at every cell division. The dramatic morphological and physiological differences between the progeny are expressed in the predivisional cell prior to separation. We review recent work examining mechanisms responsible for differentiation of the incipient swarmer and stalked cell compartments. These include differential transcription of the newly replicated chromosomes, and targeting of proteins to specific poles. The biosynthesis of the polar flagellum is emphasized as a model for studying these processes. Hypotheses concerning the role of the cell poles in expression of asymmetry are discussed. Key words: Caulobacter / polarity / differentiation transcriptional regulation / protein localization


CELL DMSION EVENTSresulting in dissimilar progeny are integral to differentiation processes in many prokaryotes and single-called eukaryotes, and are essential for development of multicellular eukaryotes. The di_morphic bacterim Caulobacter crescentus has proven to be a useful model system for studying the generation of asymmetry during the cell cycle, as every division results in two morphologically and functionally distinct progeny cells (Figure 1). We will focus in this review on two general mechanisms underlying Caulobacter differentiation: asymmetric transcription from the newly replicated chromosomes of the predivisional cell, and differential localization of proteins to the poles of the predivisional cell. The Caulobacter cell cycle is typically depicted as beginning with the swarmer cell (Figure 1), as this is the cell type which can be specifically isolated for experiments requiring synchronous cultures. The swarmer cell is motile by virtue of a single polar flagellum. The chemosensory apparatus is also localized to the flagellated pole of the swarmer cell, 1 as are multiple pili. 2 The swarmer cell eventually differFrom the Department of Developmental Biology, Beckman Center for Molecular and Genetic Medicine, Stanford University, Stanford, CA 94305-5427, USA 9 Academic Press Ltd 1044-5 781/95/010003 + 0958. 00/(9

entiates into a sessile stalked cell, ejecting its flagellum and synthesizing a stalk at the same pole. The stalk is a thin extension of the cell wall and membrane which is tipped by an adhesive 'holdfast' used for anchorage to the substrate. Aside from motility, the outstanding functional distinction between the two cell types lies in their replicative abilities. Chromosomal replication does not occur in the swarmer cell. Initiation of replication only occurs upon, or coincident with, differentiation of the swarmer into a stalked cell (Figure 1), and in the stalked cell generated by cell division. The cell cycle proceeds with the stalked cell elongating into a predivisional cell while synthesizing a new flagellum at the pole opposite the stalk. Cell division ultimately results from progressive pinching of the cell wall at the division site. The biogenesis of the swarmer cell flagellum has been the most intensively studied developmental event in the Caulobacter life cycle. The synthesis of the flagellum presents us with two questions of fundamental interest in developmental biology. The first involves t i m i n g - - h o w is flagellar gene expression initiated in and limited to the predivisional stage of the cell cycle? The second ".revolves spatial distribu t i o n - how is the flagellum specifically placed at the pole opposite the stalk? Over the past decade, these questions have begun to yield to molecular genetic analysis. Over 40 gene products are necessary for flagellar assembly or function, s and the majority of these have now been cloned and sequenced. Epistasis experiments to examine the regulation of expression of flagellar genes were used to organize these genes into a three-tiered hierarchy, 4' 5 conceptually similar to the flagellar hierarchies of Escherichia coli and Salmonella typhimurium. ~ An important distinction between these systems is that in Caulobacter induction of the hierarchy is not under environmental control, but is instead cell cycle dependent. 'Class I' of the hierarchy is currently unoccupied, being reserved for the unidentified factor(s) that link the initiation of the hierarchy to cell cycle cues. The composition and temporal expression patterns of Classess II, III, and IV reflect the order of assembly of the gene products into the flagellar structure (Figure 1). Class II is composed primarily of proteins known or predicted to reside at

C. Stephens et al

1 DivisionUnit Predivisional S ~



Swarmer cell


" I

Flagellargene expression

:lass III


Class IV (flagellins)I

Figure 1. The Caulobactercell cycle. The biogenesis and loss of polar appendages (flagellum, pili and stalk) described in the text are shown. The nucleoid is condensed and non-replicating in the swarmer cell, then decondenses and initiaties replication upon differentiation into a stalked cell. The period of chromosomal replication and flagellar gene expression are indicated beneath the cell cycle. The full length of the cell cycle is referred to as one 'division unit'.

the inner membrane; two important exceptions to this rule are transcription factors required for expression of Class III and IV, discussed below. Class III encodes structural components of the basal body and hook, and Class IV encodes the flagellin subunits making up the flagellar filament.

(RNA polymerase holoenzyme containing o~4) requires an additional 'enhancer-binding factor' to activate gene expression. 1~ These factors generally function while bound to DNA roughly 100 bp upstream of the o~4 promoter, and require phosphorylation to stimulate transription, o~4-activators from different species are closely related, forming a distinct subclass of response regulators within the superfamily of bacterial 'two-component' regulatory systems,is The enteric NtrC protein, which controls nitrogen assimilation, might be considered the prototype member of this family. The Caulobacter flbD gene product is closely related to NtrC and other 054 activators. 14 flbD is a Class II flagellar gene, and thus necessary for Class III and IV transcription, and purified FlbD protein has been shown to interact directly with a conserved DNA sequence ('ftr') upstream of certain Class III and IV promoters, is A third protein, IHF (Integration Host Factor), also appears to be necessary for the activation of at least some FlbD-dependent promoters. I H F is a DNA bending protein which binds between the promoter and the enhancer site, and is thought to facilitate interactions between Eo"54 and FIbD by bending and looping out the intervening sequences, a6' 1~ as has

Asymmetric transcription

Polarized expression of flageUar structural genes Though several Caulobacter Class II flagellar genes share a unique promoter sequence which does not appear to resemble other known bacterial promoters, no RNA polymerase species or transcription factors have been identified which transcribe these genes. 7"9 Progress is being made, however, in dissecting the regulation of Class III and IV genes. Promoters for Class III/IV genes are dependent on RNA polymerase containing the 0`54 subunit, a~ encoded by the rpoN gene. The Class II rpoNgene is transcribed just prior to the expression of the Class III genes. 11 0`54 is found in many bacterial species, and is utilized for a wide range of transcriptional regulatory functions. Eo"54 4

Expression of Caulobacter cell polarity been shown with o~4-dependent promoters in other species, la The fir-containing promoters activated by FlbD have the interesting property of being expressed preferentially in the swarmer pole of the predivisional cell (Figure 2).1~' 19 This has been demonstrated by pulselabelling predivisonal cells with S~S-methionine, isolating swarmer and stalk cell populations after division, and examining the distribution of radiolabelled product in each cell type. Of course, assembled flagellar proteins are expected to partition to the swarmer cell. However, transcriptional fusions to reporter genes (lacZ, neo or lux) have demonstrated that neither the coding sequence of the flagellar gene nor the leader sequence of the flagellar mRNA is necessary for partitioning of the reporter protein, m Thus, the promoter itself is preferentially active in one cell pole. How might this occur? One possibility is that differences in transcription factor (0~4, FlbD, IHF) concentrations in the incipient swarmer and stalk poles are responsible for pole-

P flgK

specific expression. However, WoN and flbD are maximally expressed in the stalk and predivisional cell prior to any pinching, when the cytoplasm appears continuous from pole to pole. It is thus likely that these proteins are distributed throughout the predivisional cell; this has in fact been shown directly for FIbD. 15 The concentration of IHF peaks in the predivisional cell, 17 but its intracellular distribution is not known: An alternative possibility is that the activity of one or more of these factors is spatially regulated. The kinase which phosphorylates FIbD to activate it has not been isolated, though it has been shown that phosphorylation of FlbD is cell cycle regulated. 15 When these authors generated a phosphorylationindependent mutant form of FIbD, it was found that a FlbD-dependent promoter (flgK) was now expressed equivalently in both cell poles. It was concluded that pole-specific phosphoryation of FlbD controls differential transcription of Class III and IV flagellar genes in the predivisional compartments. It should be noted that normal temporal expression of flgK was


9 .








Figure 2. Temporal and spatial regulation Of gene expression during the cell cycle. Two examples of polarized promoter activities are shown. (These examples are described in the text.) ' P ~ i is an Eo~a/FIbD dependent promoter which transcribes the Class IV flgK gene, encoding the 25 kDa flagellin.15 'P~,~r' refers specifically to 'P,~o,g', the strong, cell cycle regulated bane promoter which ooverlaps the Caulobacterchromosomal origin of replication. 21 P~,r is maximally active in the predivisional cell and is biased to the swarmer pole. Ph~,~ is maximally on in the stalked cell, but is induced again in late predivisional cell, where it is biased to the stalked pole. The height of the bars indicates promoter activity at each stage point relative to other stages for the same promoter; though P/tgK and P,~ng are of comparable strength, the bars are not intended'for direct comparisons between promoters.

C. Stephens et al maintained even in the presence of the phosphoryladon-independent FIbD, demonstrating that activation of FlbD alone is insufficient to explain cell cycle regulation of transcription in vivo. Nevertheless, identifying the FlbD kinase and factors controlling its activity and distribution is clearly crucial to understanding pole-specific transcription of flagellar genes.

Asymmetric expression of non-flagellar genes Are there non-FlbD-dependent promoters which exhibit asymmetric expression patterns during Caulobacterdevelopment? Several have been reported which are preferentially active in the stalked pole rather than the swarmer pole. These include the promoters for gyrB, which encodes the gyrase B subunit; orfl, an open reading frame of unknown function upstream of gyrB; and heinE, which encodes a heme biosynthetic enzyme. 2~ The heinE promoter region is particularly fascinating in that it overlaps the Caulobacter chromosomal origin of replication. 21 In fact, there are two promoters upstream of heinE, one of which is relatively weak but constitutively active, and another which is strong ('Pst~ong') and cell cyclle regulated. The activity of P~t~ongand origin function have thus far been inseparable by mutation, leading to the suggestion that transcription from Pst~o,g may be involved in some aspect of initiation of replication. Pstrong is maximally active at the transition from swarmer to stalked cell, coincident with initiation of replication, and is induced again around the time of cell division, when the progeny stalked cell reinitiates replication (Figure 2). Interestingly, as cell division approaches, Pstrong activity becomes increasingly biased to the incipient stalked pole, where chromosomal replication will occur. Clearly it is of great interest to identify the factors involved in activation of Pstrong-

The basis for polarized transcription Timing of promoter activity may be very relevant to the issue of spatial localization. Class II flagellar promoters, which generally show little polar bias, are maximally active at around 0.6 division units, when there is no visible pinching of the predivisional cell (see Figure 1). The swarmer-pole biased Class III and IV promoters, in contrast, are maximally active from 0.7 to 0.9 division units, when pinching is readily

apparent in the predivisional cell. is active through the whole predivisional phase, but becomes increasingly biased to the stalked-pole late in the cell cycle. In order to achieve spatial localization of the cytoplasmic transcriptional reporter proteins (e.g. [~ralactosidase) used in these experiments, some physical barrier must be present which precludes diffusion between compartments of the predivisional cell. To date, no structure resembling a septum has been observed in sectioned predivisional cells viewed by electron microscopy, but the partitioning of cytosolic proteins strongly suggests that a physical boundary arises at the division site while pinching is occurring. What is the molecular basis for differential transcription in the incipient swarmer and stalked cell compartments? At least two factors are known which may contribute. First, the polar regions of each compartment are distinct. A stalk resides at one pole. At the other, by the time Class III genes are expressed, the early components of the flagellar basal body complex have been or are being assembled. Some structural or biochemical signal may emanate from these poles to affect events in the respective compartments. For example, the kinase responsible for phosphorylation of FlbD might require binding to an early flagellar intermediate at the swarmer pole for activation. Polar signaling mechanisms will be discussed in more detail in a later section. Secondly, the nucleoid of the swarmer cell is significantly more compact than that of the stalked cell, based on density gradient centrifugation experiments. 2~'24 (The nucleoid is a supramolecular complex consisting of the chromosome and associated p~-oteins, including DNA and RNA polymerases, nascent RNAs, and polysomes.) The transition from a 'fast-sedimenting' (swarmer cell) nucleoid to a 'slow-sedimenting' nucleoid occurs abruptly at the swarmer-to-stalked cell transition. The precise temporal and regulatory relationship of the nucleoid transition to morphological changes and initiation of DNA replication are not known. The newly replicated chromosome in the swarmer pole of the predivisional cell appears to be converted back to the fast-sedimenting form prior to cell separation. It is possible that polar identity has a regulatory role in signaling nucleoid condensation/ decondensafion; it is known that the nucleoids in both cell types are membrane bound. 2a' 25 The distinct physical states of the two nucleoids of the predivisional cell could account for differential transcription in the two compartments. Alternatively,

Expression 0fCaulobacter cell polarity chromosome condensation and differential transcription may be independent consequences of the compartmental differentiation signal.

Protein localization

Differential transcription of late flagellar genes is presumed to take place after a physical boundary has been formed between the two compartments of the predivisional cell, preventing the newly synthesized mRNA and proteins from diffusing to the opposite pole. At this point in the cell cycle, the identity of the swarmer pole has already been established by assembly of early component of the flagellar structure, the Class II gene products. To understand how the identity of the swarmer pole is initially established, we must understand how the Class II proteins are targeted and assembled at the pole opposite the stalk. Furthermore, flagellar components are not the only proteins which must be localized to build asymmetrically positioned cell suface structures in Caulobact~ including pili, which are formed and retracted at the swarmer pole coordinately with the appearance and shedding of the flagellum,~' 26 and of course, the stalk. Additional non-structural proteins are known to be, or are likely to be, localized to the cell poles, including proteins involved in chemotaxis, 1 pilus synthesis27 and stalk synthesis.28

has been inferred from genetic data that these proteins interact with a highly conserved domain (HCD) located at the carboxy-terminus of the chemoreceptor. 33 Observation of McpA (the chemoreceptor which has been studied most thoroughly) during the Caulobacter cell cycle by immunoelectron microscopy revealed that wild-type McpA is present only at the flagellated pole of the predivisional cell and remains at that pole in the progeny swarmer cell, then disappears when the swarmer cell differentiates into a stalked cell (Figure 3). 1 While amino-terminal sequences target the protein to the membrane, polar segregation of the newly synthesized chemoreceptor McpA depends on sequences in the carboxy-terminal domain. 29 Deletion of the carboxy-terminus, including the HCD, yields a protein that is still associated with the membrane, but has completely lost its spcificity for the swarmer pole. Surprisingly, Maddock and Shapiro s~ have found that in E. coli, MCPs also localize to the poles of the cell, even though multiple flagella in this organism are distributed over the entire surface of the cell. Targeting of MCPs to the E. coli cell pole requires the presence of CheA and CheW, and it was proposed that subcellular localization of MCPs occurs in an MCPCheA-CheW ternary complex, CheA and CheW most likely interacting with the HCD of the chemoreceptor, s~ When an E. coli chemoreceptor gene is expressed in Caulobacteg, the E. coil protein is also

Targeting of chemotaxis proteins to the swarmer pole One of the best understood examples of proteins localized to a specific site in the bacterial cell are the chemoreceptors.1,29, s0 Bacterial chemoreceptors, referred to as methylaccepting chemotaxis proteins, or 'MCPs', are integral membrane proteins with two membrane spanning domains. ~l A periplasmic domain interacts with ligands (attractants or repellents) causing a conformational change that mediates the transmission of a chemotactic signal to the flagellar motor. This is accomplished through the interaction of two cytoplasmic components of the signal transduction pathway (CheA and CheW) with the cytoplasmic domain of the MCP. CheA and CheW are part of a well-characterized phospho-relay signal transduction pathway that links the chemoreceptor to the switch apparatus of the flagellar motor, sl The signal to the flagella facilitates net movement of tl4e bacterium towards or away from high concentrations of the ligand. CheA and CheW were shown in vitro to form a long-lived ternary complex with MCPss2 and it

"1 Figure 3. Localization of the chemoreceptor McpA at different stages in the Caulobactercell cycle?Swarmer cell, stalked cell and predivisional cell are shown schematically and the black dots at the swarmer pole represent the location of the chemoreceptor molecules.

C. Stephen~ et al

targeted to the pole of the swarmer cell but is not turned over during the cell cycle. 1'29 This suggests that the Caulobacter CheA and CheW proteins are able to interact with the E. coli MCP and direct the heterologous protein complex to the cell pole. The targeting of Caulobacter MCPs is likely to function similarly to that proposed for polar localization of the E. coli chemoreceptors. The lack of cell cycle degradation of a heterologous E. coli chemoreceptor in Caulobacter suggests that a signal unique to the Caulobacter MCP results in localization to the incipient swarmer pole. The Caulobacter McpA sequence has a short carboxy-terminal amino acid sequence that is not present in the E. coli chemoreceptors. 1 Deletion of 14 amino acids at the carboxy-terminus of Caulobacter McpA results in a protein that, like the E. coli MCPs, is not degraded during the swarmer-to-stalked cell transition, and as a result cells show a bipolar distribution of McpA. 29 Thus, temporally and spatially controlled proteolysis contributes to the asymmetric positioning of the McpA chemoreceptor during the cell cycle. However, it is clear that proteolysis is only involved in the asymmetric distribution of McpA and not in polar targeting, since in the absence of proteolytic signals, McpA is still targeted to the cell poles. Two models arise from these data that could explain the mechanism of polar tergeting of the chemoreceptor in Caulobacter. McpA could either be randomly inserted into the cytoplasmic membrane and then be moved to the cell pole, or it could initially be targeted to the pole of the cell, where it is inserted into the membrane by a pole-specific export apparatus. In both cases the aggregation with CheA and CheW might trap chemoreceptor complexes at the pole by restricting lateral diffusion. Alternatively, distinctive polar periplasmic and membrane subcompartments which restrict protein movement have been proposed for E. coli, and are equally feasible in Caulobacter (refs 34, 35; see below).

have to be targeted to the swarmer pole. These proteins may serve as polar nucleation sites for the flagellar subunits expressed later in the cell cycle. The polar site for the initiation of flagellar assembly could be selected by the insertion of a single membrane protein, which is followed by tlae assembly of the rest of the flagellar proteins at that site. Alternatively, several flagellar proteins could be targeted independently to the cell pole. The enteric flagellar M ring protein (FliF), an integral membrane protein, is the first known structural component to be laid down at site of flagellar biogenesis, and its insertion is independent of other known flagellar genes, a6 The Caulobacter M ring protein (FliF) is thus a likely candidate for the first structural component targeted to the cell pole. The cell cycle pattern of synthesis and degradation of the Caulobacter FIiF protein and the McpA chemoreceptor are very similar (U.Jenal, L. Shapiro, unpublished results). Although polar localization of the McpA protein is not dependent on flagellum assembly, 1 the targeting of McpA and early flagellar components, such as FliF, to the swarmer pole could be accomplished by similar mechanisms. Not all Caulobacter proteins involved in flagellar function are targeted to the flagellated pole. The FIiL protein is encoded by a Class II flagellar gene (t/L) and has been shown to be required for function but not assembly of the flagellum, a7 FliL resides in the inner membrane but is not polarly localized; it is distributed throughout the membrane of all cell types. Furthermore, unlike other flagellar and chemotaxis proteins, FIiL is not turned over during the loss of the flagellum, as its concentration remains constant throughout the cell cycle, s7 Although a more general role for FliL can not be excluded, FliL function is most likely flagellar specific. Thus, proteins involved in motility do not necesarily have to colocalize with the flagellar motor.

Localization of non-flagellar proteins

Prote/n/oca//zat/on and flage//ar assemb/y The flagellum is assembled by adding proteins to the nascent structure in acell-distal order, beginning with the transmembrane rotor, continuing with the hook, and ending with the distal filament. Although the localization of the early structural components of the flagellum is not as well understood as the positioning of chemotaxis proteins, some parallels can be drawn. Flagellar proteins expressed early in the cell cycle (Class II) are involved in the initial assembly steps and

Synthesis of pilin subunits is initiated in the stalked cell well before the pili structures are assembled and become visible on the cell surfaceY At the time of assembly, pilin protein is thus drawn from a cytoplasmic pool that is distributed equally between swarmer and stalked cell poles, ss Localized assembly of the pilin subunits at the swarmer pole is therefore not dependent on compartmentalization of the subunits, but rather must occur by segregation of the assembly machinery to this site of the cell.

Expression of Caulobacter cell polarity previous division site to direct future protein targeting, as is the case in BUD3 and BUD4 localization to the bud scar of the previous budding site in yeast. 4s This organizing center may be transient, only being required to endure until early flagellar proteins can recognize it and initiate assembly, thereby generating a swarmer pole. Unidentified cell cycle signals could then activate further maturation of the swarmer pole to a stalked pole, after cell division has occurred. It is possible that a r e m n a n t of the basal body, such as the M ring (composed of the FIiF protien), is retained after the ejection of the flagellum to serve as a marker for stalk positioning. T h o u g h there is no physical evidence for a polar organizing center in Caulobacter cells, there is evidence from enteric bacteria that distinct structures and compartments are generated in the polar region during cell division, and appear to be stably maintained in the progeny cells. In E. coli, adhesion zones

There is also evidence for proteins being specifically targeted to the stalked c o m p a r t m e n t of the predivisional cell. Pools of the two heat-shock inducible proteins Lon and DnaK synthesized in the predivisional cell are segregated predominantly to the stalked cell progeny, though a third heat shock protein, GroEL, distributes equally to both the swarmer and the stalked cell. ~9 Little is known about stalk biosynthesis to date, 4~ but the eventual identification of factors specifically involved in stalk structure and assembly will almost certainly yield more proteins localized to this pole.

Concluding remarks m the role o f the cell p o l e s in a s y m m e t r y

Caulobacter cells undergo numerous morphological and physiological changes over the course of the cell cycle, both internally and externally. If, however, there is a single key to understanding the generation and maintenance of asymmetry in Caulobact~ it may lie in the cell poles. It is instructive to follow the fate of the poles through the swarmer and stalked cell cycles (Figure 4). When Caulobacter divides, two new poles are formed opposite of each of the original poles marked by the stalk and the flagellum, respectively. In both the swarmer and the stalked cell progeny, this new pole is eventually destined to become a swarmer pole, though the assumption of this identity is delayed in the swarmer progeny until after the current swarmer pole has converted into a stalked pole. A swarmer pole always differentiates into a stalked pole, which represents the terminally differentiated form of the cell pole. Thus, Caulobacter shows a clear succession pattern in pole development: the new pole generated in each cell division becomes a swarmer pole, which in turn differentiates into a stalked pole (Figure 4). The transition of a new pole into a swarmer pole is characterized by the appearance of morphological markers including the flagellum, holdfast and pili, as well as protein complexes involved in chemotaxis. The transition from swarmer to stalked pole is accompanied by the removal of the flagellum, pili, and the chemoreceptors, a n d by the initiation of stalk biogenesis. Targeting of flagellar proteins specifically to the new pole of the predivisional cell requires an identification mechanism. It has been proposed that 'a marker is laid down at the site of cell division which could serve as an 'organizational center' at the new poles.4L 42 Thus, cells would retain a m e m o r y of the

NP SWP 0 "~II~ 0 - - ~


Figure 4. Pole development during the Caulobaaercell cycle. The new cell poles (NP) formed at each cell division site are indicated by I-7. New poles become swarmer poles (SWP; 0) later in the swarmer (top) and stalked cell cycle (bottom). In the next swarmer cell cycle the swarmer poles turn into stalked poles (STP; A). The succession of pole development is indicated below. Note that the progeny swarmer cell resulting from the first division shown is flipped relative to the preceding predivisional cell. 9

C. Stephem a al

m u t a n t strains form filamentous cells, indicating failure o f an early step in division at a relatively high frequency. Such feedback between polar m o r p h o g e netic events and subsequent cell cycle regulated process (gene expression and cell division) suggests that polar morphogenesis may'be an integral compon e n t o f the Caulobacter cell cycle 'clock'. It is clear that we are just beginning to u n d e r s t a n d how Caulobacter cellular d e v e l o p m e n t depends on dynamic interactions between temporally and spatially regulated gene expression, and the positioning and assembly o f the gene products.

between the inner m e m b r a n e and the o u t e r membrane-murein complex are observed adjacent to invaginating or fully f o r m e d septa, a4 These 'periseptal annuli' are proposed to define domains that restrict essential elements o f the cell division machinery to this location, s4'44 Photobleaching experiments with fluorescently-labeled periplasmic proteins provide support for such a domain structure, as fluorescence recovery is uniformly low at the poles, indicating that this region is biochemically sequestered from the r e m a i n d e r o f the periplasm, s5 A t t a c h m e n t sites are also present at the leading edge of septum formation during the cell division process, and an a p p a r e n t r e m n a n t o f these, attachment sites remains at the newly f o r m e d poles after cell division as a form o f birth scar. 45 By analogy, a polar c o m p a r t m e n t may exist in Caulobacter cells within the periplasm and i n n e r m e m b r a n e at the new pole (the site o f the previous cell division), and this c o m p a r t m e n t may retain a polar m a r k e r which presents a docking structure for protein targeting. In addition to serving as a nucleation site for assembly o f polar structures, the Caulobacter cell poles may actively generate signals for developmental events occurring in the cytoplasm. A prime example o f such signalling is the feedback between assembly of the flagellum and transcriptional activation o f flagellar genes. Consider that an inactivating mutation in any Class II gene prevents the expression o f Class III genes, and at the same time causes over expression of Class II promoters. 4 T h e known Class II genes, with the exception o f rpoN and flbD, do not a p p e a r to be transcription factors or DNA bindig proteins. It seems unlikely, therefore, that the presence o f the Class II gene products directly modulates gene expression. A m o r e reasonable scenario is that the successful assembly o f the i n n e r m e m b r a n e portion o f the basal body somehow activates Class III promoters a n d represses Class II promoters. An assembly 'checkpoint' has been shown to regulate flagellar gene expression in Salmonello~ in which an anti-sigma factor is expelled from the cell when an e x p o r t - c o m p e t e n t stage has b e e n achieved in the basal body, thereby allowing expression o f h o o k and filament genes. 46 We do not know how the assembly status o f the flagellum is 'sensed' in Caulobaaer O n e possibility with respect to activating Class III/IV genes is that the kinase which phosphorylates FIbD associates with the nascent basal body, and only becomes catalytically active u p o n completion o f a specific structure. Interestingly, it seems that successful completion o f early flagellar assembly in Caulobacter is linked to cell division as well. Many Class II

Acknowledgements We think the members of the Shapiro laboratory past and present for helpful discussions. This work was supported by National Institutes of Health grants GM32506 (to L.S.) and GM14179 (to C.S.), American Cancer Society grant NP938B (to L.S.) and a grant from the Swiss National Science Foundation (to u.J.).

References 1. AlleyMRK,MaddockJ, Shapiro L (1992) Polar localization of a bacterial chemoreceptor. Genes Dev 6:825-836 2. SchmidtJ (1966) Observation on the absorption of Caulobacter bacteriophage containing ribonucleic acid. J Gen Microbiol 45:347-353 3. Ely B, EIyT (19889) Use of pulsed field gel electrophoresis and transposon mutagenesis to estimate the minimal number of genes required for motility in Caulobaaer crescents. Genetics 123:649-654 4. Xu H, Dingwall a, Shapiro L (1989) Negative transcriptional regulation in the Caulobaaerflagellar regulatory heirarchy. Proc Nail Acad Sci USA 86:6656-6660 5. Newton A, Ohm N, Ramakrishnan G, Mullin D, Raymond G (1989) Genetic switching in the flagellar gene hierarchy requires both positive and negative regulation of transcription. Proc Nat Acad Sci USA 86:6651-6655 6. Jones C, Aizawa SI (1991) Genetic control of the bacterial flagellar regulon. Curr Opin Genet Dev 1:319-323 7. DingwallA, Zhuang W, Quon K, Shapiro L (1992) Expression of an early gene in the flagellar regulatory hierarchy is sensitive to an interuption in DNA replication. J Bacteriol 174:1760-17768 8. Van WayS, Newton A, Mullin A, Mullin D (1993) Identification of the promoter and negative regulatory element, fir4, that is needed for cell cycle timing of fl/F operon expression in Caulobaaer r J Bacteriol 175:367-376 9. Stephens C, Shapiro L (1993) An unusual promoter controls cell-cycle regulation and dependence on DNA replication of the CaulobaaerfliLM early flagellar operon. Mol Microbiol 9:1169-1179 10. Mullin D, Minnich S, Chen L, Newton A (1987) A set of positively regulated flagellar gene promoters in Caulobaaer cresctnt~ with sequence homology to the nifgene promoters of K/ebs/d/apneumonia~J Mol Biol 195:939-943 10

Expression of Caulobacter cell polarity 11. Brun Y, Shapiro L (1992) A temporally controlled o factor is required for polar morphogenesis and normal cell division in Caulobacter. Genes Dev 6:2395-2408 12. Kustu S, Santero E, Keener J, Popham D, Weiss D (1989) Expression ofo ~4 (ntrA)-dependent genes is probably united by a common mechanism. Microbiol Rev 53:367-376 13. Volz K (19993) Structural conservation in the CheY superfamily. Biochemistry 32:11741-11753 14. Ramakrishnan G, Newton A (1990) FIbD of Caulobactercrescentus is a homologue of NtrC (NRI) and activates 054 dependent flagellar gene promoters. Proc Natl Acad Sci USA 87:2369-2373 9 15. Wingrove J, Mangan E, Gober J (1993) Spatial and temporal phosphorylation of a transcriptional activator regulates polespecific gene expression in Caulobact~ Genes Dev 7:1979-1992 16. GoberJ, Shapiro L (1990) Integration host factor is required for the activation of developmentally regulated genes in Caulobact~ Genes Dev 4:1494-1504 17. Gober J, Shapiro L (1992) A developmentally regulated Caulobacter flagellar promoter is activated by 3' enhancer and IHF binding elements. Mol Biol Cell 3:913-926 18. Hoover T, Santero E, Porter S, Kustu S (1989) The integration host factor (IHF) stimulates interaction for nitrogen fixation operons. Cell 63:11-22 19. GoberJ, Champer R, Reuter S, Shapiro L (1990) Expression of positional information during cell differentiation in Caulobacte~. Cell 64:381-391 20. Rizzo M, Shapiro L, GoberJ (1993) Asymmetric expression of the gyrase B gene from the replication-competent chromosome in the Caulobacter crescentus predivision~il cell. J Bacteriol 175:6970-6981 21. Marczynski G, Lentine K, Shapiro L (1994) A developmentally regulated transcript overlaps the Caulobacter origin of replication. Submitted 22. Evinger M, Agabian G (1979) Caulobacter crescentus nucleoid: analysis of sedimentation behavior and protein composition during the cell cycle. Proc Nail Acad Sci USA 76:175-178 23. Soboda U, Dow C, Vitkovic L (1982) Nucleoids of Caulobacter crescent~ CB15. J Gen Microbiol 128:279-289 24. GoberJ, Shapiro L (1991) Temporal and spatial regulation of developmentally expressed genes in Caulbacter. BioEssays 13:277-283 25. Evinger M, Agabian N (1977) Envelope-associated n u c l e o i d from Caulobacter crescentus stalked and swarmer cells. J Bacteriol 132:294-301 26. Sommer J, Newton A (1988) Sequential regulation of developmental events during polar morphogenesis in Caulobacter crescentur, assembly of pili on swarmer cells requires cell separation. J Bacteriol 170:409-415 27. Stair J, Agabian N (1982) Caulobacter crescentus pili: analysis of production during development. Dev Biol 89:23%247 28. SchmidtJ, Stanier R (1966) The development of cellular stalks in bacteria. J Cell Biol 28:423-436 29. Alley MR_K, MaddockJ, Shapiro L (1993) Requirement of the carboxyl terminus of the bacterial chemoreceptor for its targeted proteolysis. Science 259:1754-1757

30. Maddock J, Shapiro L (1993) Polar location of the chemoreceptor complex in the Escherichia coli cell. Science 259:1717-1723 31. Parkinson J 91993) Signal transduction schemes of bacteria. Cell 73:857-871 32. GegnerJ, Graham D, Roth A, Dahlquist F (1992) Assembly of an MCP receptor, CheW, and kinase CheA complex in the bacterial chemotaxis signal transduction pathway. Cell 70:975-982 33. Liu J, Parkinson S (1991) Genetic evidence for interaction between the CheW and Tsr proteins during chemoreceptor signaling by Eacherichia cols J Bacteriol 173:4941-4951 34. MacAlisterr T, MacDonald B, Rothfield L (1983) The periseptal annulus: an organeile associated with cell division in gramnegative bacteria. Proc Nati Acad Sci USA 80:1372-1376 35. Foley M, BrassJ, Birmingham J, Cook W, Garland P, Higgins C, Rothfield L (19989) Compartmentalization of the periplasm at the cell division sites in Escherichia coli as shown by fluorescence photobleaching experiments. Mol Microbiol 3:1329-1336 36. Kubori T, Shimamoto N, Yamagnchi S, Namba K, Alzawa S-I (1992) Morphological pathway of flagellar assembly in SodmoneUa Oephimurium. J Mol Biol 226:433-446 37. Jenal U, White J, Shapiro L (1994) Caulobacter flagellar function, but not assembly, requires FiiL: a non-polarly localized membrane protein present in all cell types. J Mol Biol 243:227-244 38. SmitJ (1987) Localizing the subunit pool for the temporally regulated polar pili of Caulobacter crescentus. J Cell Biol 105:1821-1828 39. Reuter S, shapiro L (1987) Asymmetric segregation of Caulobacter heat shock proteins upon cell division. J Mol Biol 194:653-662 40. Brun Y, Marczynski G, Shapiro L (1994) The control of timing and spatial organization during Caulobactercell differentiation. Annu Rev Biochem 63:419-450 41. Bender R, Agabian N, Shapiro L (1978) Cell differentiation in Caulobacter crescentus, in The Molecular Genetics of Development (Leighton T, ed). Academic Press, New York 42. Huguenel E, Newton A (1982) Localization of surface structures during prokaryotic differentiation: role of cell division in Caulobacter crescentus. Differentiatior~ 21:71-78 43. Chant J, Herskowitz I (1991) Genetic control of bud site selection in yeast by a set of gene products that constitute a morphogenic pathway. Cell 65:1203-1212 44. Cook W, MacAlister T, Rothfield L (1986) Compartmentalization of the periplasmic space at division sites in gram-negative bacteria. J Bacteriol 168:1430-1438 45. Macalister T, Cook W, Weigand R, Rothfield L (1987) Membrane-murein attachment at the leading edge of the division septum: a second membrane-murein structure associated with morphogenesis of the gram-negative bacterial division septum. J Bacteriol 169:3945-3951 46. Hughes K, Gillen K, Semon M, Karlinsey J (1993) Sensing structural intermediates in bacterial flagellar assembly by export of a negative regulator. Science 262:1277-1280