Audio recording of this lecture

Previous or Next lecture

Reading for this lecture:

• JE Berleman, J Scott, T Chumley & JR Kirby 2008 Predataxis behavior in Myxocuccus xanthus. PNAS 105:17127-17132
• V Mariscal, A Herrero & E Flores 2007 Continuous periplasm in a filamentous heterocyst-forming cyanobacterium. Mol Microbiol 65:1139-1145

We generally think of Bacteria (and Archaea and protists, for that matter) as being unicellular - every bug for itself. But this is not a very sophisticated viewpoint. It's true that Bacteria don't develop into large interdependent clusters that walk around talking to one another on cell phones, but multicellularity is a spectrum, not just an "is" or "is not". Most Bacteria communicate by secreting small compounds called phermones; cell-to-cell communication is a kind of multicellularity, in the sense that the cells act and react as a population rather than individually. Many Bacteria form aggregates of various kinds during their life cycle in which cells differentiate. In some cases, there is no doubt that the bacterium is multicellular, for example the filamentous cyanobacteria in which some cell undergo terminal differentiation in nitrogen-fixing heterocysts, or the colonies of Streptomyces described above that undergo complex morphological development and programmed cell death. Cells in a group specializing into different forms, and especially if that differentiation is terminal (those cells will not contribute to future generations) is the hallmark of multicellularity.

One distinction between multicellular behavior in Bacteria (and Archaea and unicellular eukaryotes) in contrast to plants/animals/fungi is that the groups of cells are often composed of more than one species; for example, the photosynthetic mats. Here we have complex layers of different kinds of cells, working together to make a living; a tissue, of a sort, made up of cells of various kinds specialized for different functions. Not much different than the tissue of a plant of animal, except that the cells that make it up are of different species.

An aspect of multicellularity in most "single-celled" Bacteria is cell-to-cell communication, which in turn allows multicellular arrangements of cells to undergo coordinated development. One well-studied form of communication in Bacteria is quorum sensing, by which organisms measure their population density by measuring the concentration of secreted phermones. This is how Pseudomonas knows when to form a biofilm, or Vibrio knows when to emit light in the light organ of a squid. Neither of these energy-expensive processes are successful unless performed by a large local population of cells.

Predataxis behavior in Myxococcus xanthus

The epsilon-proteobacteria Myxobacteria use two (at least) forms of cell-to-cell communication during their life cycle. Myxobacteria are predators of other Bacteria; they lyse them, hydrolyze the macromolecules, and them absorb the nutrients. The first communication system they use is a quorum-sensing system (called the A-signal) based on a soluable signal that detects starvation in the population and induces cells to come together into colonies. The second is a contact-mediated signal (the C-signal) based on a surface protein, and guides the formation and development of the slug, fruiting body, and eventually the spores. In has more recently been shown also, and probably more importantly, to direct efficient group feeding during predation.

When Bacteria are starved for carbon, energy, or nitrogen, one of the major immediate results is that the concentration of amino acids drops, and spent tRNAs can't get recharged. This causes ribosomes to slow and stall during translation, and a ribosome-associated enzyme known as relA catayzes the synthesis of ppGpp (guanosine 5',3' bis diphosphate). ppGpp binds to RNA polymerase, turning off rRNA synthesis and activating the sigma35 stationary-phase specific transcription factor; this is called the "stringent response", and begins the cascade (directed by sigma35) that switches the cell to stationary phase. Once the stringent response has been activated, new rounds of DNA replication or cell division are not initiated. In many organisms, ppGpp can send organisms down other starvation-induced developmental pathwways; sporulation in Bacillus, swarmer-cell development in Bdellovibrio, or aggregation in Myxococcus.

In Myxobacteria, ppGpp also induces A-signal protease synthesis; this is an extracellular protease that releases amino acids and short peptides from the outside surface of the cell. This mixture is the A signal. Cells sense this with a receptor that activates a cascade that ultimately activates sigma-54. This process takes about 2 hours, and is dependent on the density of the local population, and thus serves as a slow-response measure of how many starved cells are in the vicinity. The combination of standard stringent response and A-signal serve to induce a series of genes of poorly understood function that cause the cells to swarm together by chemotaxis, and sets them up to start developing and responding to the C signal.

The C signal is mediated by a surface protein located at the ends of cells. There also is a receptor for this C protein, and the interaction of these proteins signals end-to-end contact of cells. One result of this C signal is the further induction of the gene encoded the C protein - and so C protein expression amplifies over time by positive reinforcement. The C signal also causes the activation of the FruA protein, whose original expression in inactive form is dependent on the A signal. Active FruA starts the developmental cascade that directs aggregation either for predation or for fruiting body formation, and also activates the genes that control the frequency of reversal in gliding motility. Early in the predation or aggregation processes, the C signal is weak, and cells reverse their direction of gliding frequently, causing cells to travel in waves as they move back and forth absorbing nutrients from their prey. This serves as a measure of cell density; the more cells there are around and the more contact, them stringer the C-signal. As they stay in end-to-end contact, however, and the stronger the C signal gets, this reduces the frequency of reversing. The result is that swarms of cells that start out moving in waves and clumping at traffic jams, become increasingly organized into streams that move uniformly in one direction; toward more prey or into a central gathering. If starved, these streams converge into a huge (in bacterial terms) organized aggregates called slugs, that then undergo differential development into fruiting bodies, containing basal cells, stalk cells, sporangium cells, and spores. Cells inside the fruiting bodies (sporangia) develop into spores, that are released into the wind for dispersal - these propogate the population, the remaing cells all die.

Fruiting bodies in the specie Myxococcus xanthus are pretty simple - basically individual sphaeroids on a short stalk. Other genera have much more elaborate fruting body structures, with many spoangia and branching talks.

BTW, gliding in Myxococcus is by a combination of two mechanisms: type IV pili-mediated twitching and the directed movement of adhesion sites down the cell body.

Some great movies of Myxococus:

• Gliding motility
• Adventurous swarming
• Colony swarming
• Another colony swarming
• Fruiting body development
• Swarming waves and fruiting body formation (warning - it's 207Mbytes!)

The paper for today:

JE Berleman, J Scott, T Chumley & JR Kirby 2008 Predataxis behavior in Myxocuccus xanthus. PNAS 105:17127-17132

Hypothesis: Rippling in swarms of Myxococcus xanthus is not just an intermediate phase in development toward fruiting body formation, but is also used to faciliate predation during vegetative growth.

Abstract:
Spatial organization of cells is important for both multicellular development and tactic responses to a changing envirnoment. We find that the social bacterium Myococcus xanthus utilizes a chemotaxis (Che)-like pathway to regulate multicellular rippling during predation of other microbial species. Tracking of GFP-labeled cells indicates directed movement of M. xanthus cells during the formation of rippling wave structures. Quantitative analysis of rippling indicated that ripple wavelength is adaptable and dependent on prey cell availability. Methylation of the receptor, FrzCD is required for this adaptation: a frzF methyltransferase mutant is unable to construct ripples, whereas a frzG methylesterase mutant forms numerous tightly-packed ripples. Both the frzF and frzG mutant strains are defective in directing cell movemenet through prey colonies. These data indicate that the transition to an organized multicellular state during predation in M. xanthus relies on the tactic behavior of individual cells, mediated by a Che-like signal transuction pathway.

In this paper, the authors are interested in chemotaxis and rippling wave formation as it's related to predation rather than fruiting body formation (which is much better studied).

Here is a better (color and larger scale) version of Fig 1A:

What they've done is plant a colony of E.coli (with a defective DAP-sythesis gene, so it can't grow in media without DAP) next to a pre-existing ex-colony of Myxococcus fruiting bodies. The spores have germinated, and swarmed over the E. coli colony. Notice the rippling waves of Myxococcus cells.

Here are a couple of movies of parts B and C of this figure that make this a lot clearer:

• Fig 1B movie
• Fig 1C movie

In these images, 1 cell in 50 is labeld with GFP - these are the dark ones. These serve as trces for individual cell, kind of like tracer bullets. The background grey shows you what the bulk of the cells are doing. In B, the swarming cells have not made the transition to rippling - they're still searching for food. In C, they are in contact with prey or prey guts, and are feeding using this rippling behavior.

In panels A and B, the authors are looking a direction of movement of individual cells. In the absence of prey (A), cells move around, but in random directions, but n the presence of prey (B), the cells move back-and-forth (more forth than back) in the direction of the swarm into the prey colony. In panel C, they're watching the waves - what they see are forward-moving and backwad-moving waves that create a kind of interference pattern.

In Fig 3, the authors are looking at how the wavelength of the ripples changes under different conditions. Panel A shows that the wavelenth increases increases over time during feeding, until they can't be measured when the prey is all moped up. In panel B, normal Myxococcus (black) form smaller ripples when given more prey, but mutants this the CHE system turnd on constitutively (frxG mutants - open circles) always form tiny ripples, whereas those with the CHE system turned off (frzF - dotted line) doesn't ripple at all. Neither mutant can adjust rippling based on prey density. Panel D shows that rippling wavelngth is not a function of predator Myxococcus cell density - this isn't controlled by cell-cell communication. Panel D measured the length of time required for rippling to start relative to the predator density - clearly there is some cell-cell communication, because rippling requires more time to get started with fewer predators.

In Fig 5, the authors are watching the methylation state of the CHE FrzCD receptor protein - this methylation state is what, in E. coli determines how long the runs are during motility.These are in sets of 3 - the first 3 are wild-type Myxococcus, the next 3 are the frzF mutant, the next 3 are the frzG mutant, and the last lane is a negative control. The first lane of each set is the initial state when the Myxococcuscells from a liquid culture are first presented (on the agar surface) with prey. In the absence of prey, the wild-type cells reduce methylation - no chemotaxis. In the presence of prey, they increase methylation, and so chemotaxis. Neither mutant responds to prey - the frzF mutant is constitutively unmethylated, the frzG mutant is constitutively methylated.

In the final figure, the authors show that motility is effective in the wild-type strain but not the mutants. They spot Myxococcus onto an existing stripe of E. coli, then measure swarming in the right directions (closed circles) or wrong directions (open circles). As you can see, wild-type Myxococcus swarms in the right directions (more right-to-left than top-to-bottom), and start rippling after about 24 hours. Niether of the mutants swarm as well, nor favor the right direction as well, and as before the frzG mutant doesn't ripple.

Continuous periplasm in a filamentous, heterocyst-forming cyanobacterium

V Mariscal, A Herrero & E Flores 2007 Continuous periplasm in a filamentous heterocyst-forming cyanobacterium. Mol Microbiol 65:1139-1145

Hypothesis: The periplasm of Anabaena is continuous along the filament, and material in this periplasm can diffuse from its source to nearby cells.

Summary (abstract):
The cynobacteria bear a Gram-negative type of cell wall that incudes a peptidoglycan layer and an outer membrane outside of the cytoplasmic membrane. In filamantous cyanobacteria, the outer membrane appears to be continuous along the filament of the cells. In the heterocyst-forming cyanobacteria, two cell types contriute specialized functions for growth: vegetative cells provide reduced carbon to heterocyts, which provide N2-derived fixed niotrogen to vegetative cells. The promoter of the patS gene, which is active specifically in developing proheterocycts and heterocyts of Anabaena sp. PCC7120, was used to direct the expression of altered versions of the gfp gene. An engineered green fluorescent protein (GFP) that was exported to the periplasm of proheterocycts through the twin-arginine transloaction system was observed also in the periphery of neighboring vegetative cells. However, if the GFP was anchored to the cytoplasmic membrane, it was observed in the periphery of the producing proheterocycts or heterocycts but not in adjacent vegetative cells. These results show that there is no cytoplasmic membrane continuity between heterocycts and vegetative cells and that the GFP protein can move along the filament in the periplasm, which is functionally continuous and so provides a conduit that can be used for chemical communication between cells.

Fig. 1. Cyanobcterial autofluorescence (red - from photopigments) anf GFP fluorescence (green) of filaments of the induced Anabaena strains. The filaments were taken from cultures that had been incubated in the absence of fixed nitrogen for 10hr (to start heterocyct formation) for 10hr (A left), 25-30 hr (A right,B and C). The arrowheads point to develping or mature heterocycsts, which show reduced autofluorescence depending on their degree of differentiation and produce the GFP. Panel C shows an overlay of the cyanobacterial autofluorescence and GFP fluorescence. The scale bar is 3um.

Strain CSVM17 produces GFP without a signal peptide, and so it is trapped in the cycoplasm of the producing heterocyst. The GFP in strain CSVM18 has a "twin-arginine" signal peptide for translocation across the cytoplasmic membrane, and a signal peptidase I site so that the protein is released into the periplasm. Notice that this GFP diffuses in the periplasm along the adjacent cells, and this increases as the time after induction increases. The GFP in strain CSVM19 also has the twin-arginine transloaction signal, but a signal peptidese II site (instad of SP I) - this causes the protein to be covalently linked to a membrane lipid during export, and the result is hat the GFP is attached to the cytoplasmic membrane. This GFP stays with the heterocyst, not diffusing along the membrane. In panel B, the incubation has been extended from 26 ro 30 hours - notice how much more distributed the GFP is.

In Fig. 2, the authos show that GFP expressd in the periplasm (closed circles) diffuses further and further along the filament as time goes by. If the GFP is trapped in the cytoplasm (open circles and squares) or cytoplasmic membane (triangles), it doesn't move.

At the early time points, there is more spread of GFP. This is part of the negotiation between these cells for who's going to develop into the heterocyct. When fixed nitrogen is taken away (time 0), all of the cells start producing PatS as if they're ready to develop into heterocysts. Once one cell makes the commitment, however, PatS expression in the other cells is repressed.

Fig 3. GFP fluorescence recovery after photobleaching. Blaching was accomplished by increasing the intensity of excitation irradiation (remember, this is UV fluorescnt microscopy) over a bleach point (A and C, at arrowheads) or larger area (B, D & E). Post-bleaching micrographs were taken every second (A-D) or 5 second intervals (E), and an image is shown from a time when fluorescence has recovered as discussed in the text. Samples were from 4-day N2-grown cultures (A-D) or from a 25hr incubation without fixed notrogen (E). Scale bar, 2um.

Panels A and B show that the GFP tethered to the cytoplasmic membrane is free to diffuse around in the membrane of the heterocyst, but cannot move to adjacent cells. Panels C-E show that GFP loose in the periplasm can diffuse around in the space arounf the heterocysts (C) or adjescent cells where the GFP has already moved to (D & E).

So, the conclusion is that the continuous periplasm of Anabaena serves as a conduit - a sort of circulatory system - for proteins (and presumably other things) between cells of a filament.

Questions for thought

• What's the advantage of a slow-reacting measurement like the A signal over an instantaneous signal?
  
  
• What might a cell of Myxococcus do if it becomes starved, but is alone so that it cannot receive either an A signal?
  
  
• How might gliding motility be different for individual cells and cells in an aggregate?
  
  
• Can you think of any way in which rippling might improve predation by Myxococcus? How would you test your hypothesis?
  
  
• What mechanisms can you imagine Myxococcus might use for lysing prey cells with at the same time self-destructing or killing other cells of Myxococcus?
  
  
• The main nutrients passed back and forth between Anabaena heterocysts and vegetative cells is probably glutamate (from veg cells to the heterocycts) and glutamine (from heterocycts to veg cells). These are small organic compounds and might be expected to diffuse out of the porins in the outer membrane and so leak quickly out into the environment. Do you think this periplasmic conduit might still somehow be used for these small molecules? If so, how might this work? How would you test this?
  
  
• Filaments of cyanobacteria don't grow indefinitely long - they are easily broken intio fragments of shorter length. What do you suppose happens in this cae to the outer membrane and periplasm? How might this be repaired?