Both aerobically and photosynthetically grown wild-type Rhodobacter sphaeroides swarmed through soft nutrient agar. However, individual aerobically and photosynthetically grown tethered cells showed different responses to steps in concentrations of some attractants. Photosynthetically grown cells showed little response to a step-up in attractant, but large response to a step-down. Aerobically grown cells showed a large but opposite response to a step-up of chemoeffectors such as succinate and aspartate. The responses in che operon deletion mutants were also investigated and indicated that the aerobic response may depend on the protein products of che operon 1.
Rhodobacter sphaeroides, a purple non-sulfur bacterium, is metabolically versatile, growing in very diverse soil/water habitats. In the laboratory, it is grown under a range of defined conditions: photoautotrophically, photoheterotrophically, chemoheterotrophically and using anaerobic respiration with DMSO and TMAO as alternative electron acceptors. These different modes of growth require a complex sensing and expression system controlling the synthesis of numerous proteins and pathways for energy generation, transport and metabolism.
The motility of R. sphaeroides WS8 is probably one of the best understood outside the enteric bacteria [1,2]. It has one flagellar motor which rotates unidirectionally. The motor periodically stops, the cell is reorientated, the motor restarts and the cell swims off in a new direction . R. sphaeroides is chemotactic to a range of metabolites, i.e. sugars, amino acids and weak organic acids. Transport and induction for growth on the effector is required for some chemotactic responses [4–7].
Unlike the relatively uniform behaviour of Escherichia coli, the behaviour of R. sphaeroides cells in a population is diverse with cells showing a wide range of speeds and stopping frequencies [3,8–10]. Studies have shown that R. sphaeroides has an apparently variable flagellar rotation rate whilst actively swimming, can accelerate out of stops at variable rates and can change direction without measurable stopping [3,9]. There is also a stimulus-dependent speed change, chemokinesis, a sustained increase in the rate of flagellar rotation and a decrease in stopping .
R. sphaeroides has multiple copies of chemosensing genes arranged primarily in two operons. Operon 1 contains genes for CheY5, CheY1, CheY2, CheA1, CheW1, CheR1 and 3 transducers while operon 2 contains genes for CheY3, CheA2, CheW2, CheW3, CheR2 and CheB and one transducer. CheY4 is located elsewhere on the genome [12,13]. Deletion of operon 1 causes little loss of chemotaxis while deletion of operon 2 causes loss of chemotaxis and responses to light.
Responses to temporal changes in chemoeffector concentrations have been measured in tethered cells grown under photosynthetic conditions. No transient adaptive changes were seen to increases of effector concentration but when decreased the cell population stopped transiently and then adapted. This suggests that R. sphaeroides primarily senses a reduction in attractant concentration . In this study, the behaviour of R. sphaeroides grown under dark aerobic conditions was compared to that of photosynthetically grown cells.
Materials and methods
R. sphaeroides WS8N and the chemotaxis operon deletion mutants JPA403 (Δ operon 2) and JPA117 (Δ operon 1) [12,13] were grown to mid-exponential phase at 30°C either in the dark with shaking for aerobic chemoheterotrophic growth or in a sealed bottle in the light for photosynthetic growth .
Analysis of motile behaviour
Cells were harvested, washed and resuspended in 10 mM Na–HEPES pH 7.2 buffer containing 50 μg ml−1 of chloramphenicol which had been either aerated by shaking at 30°C or sparged with O2 free nitrogen for 40 min. For tethering the cells were concentrated 10×. The cells were starved for 45 min. For tethered cell experiments, cells were tethered with antiflagellar antibody in a flow chamber and behaviour analysed by motion analysis on the Bactracker using the software AROT7 (HTS Systems Ltd, Sheffield, UK) [8,14]. 10 mM Na–HEPES (pH 7.2) buffer containing known concentrations of chemoeffectors (Na salt) was flowed past the tethered cells and the response of the cells was measured. For free-swimming analysis, the cell suspension (resuspended to original volume) was drawn into a microslide (0.2 mm diameter, Camlab, UK) and tracked using the Bact software package.
Swarm plates consisted of Sistroms medium A with concentrations of Na–succinate or propionate from 0 to 1 mM solidified with 0.23% w/v Bacto agar (Difco) . The plates were inoculated in the centre with a drop of 5 μl of stationary culture and incubated at 30°C either in the dark or in the light in an anaerobic cabinet (Don Whitley Scientific Ltd, Shipley, UK) and repeated four times. For plug plates, plugs were made from Na–HEPES (pH 7.2) containing the chemoeffector solidified with 1.5% w/v agar. The cells, after harvesting, were resuspended in their original volume of 0.22% w/v Bacto agar in Na–HEPES. The responses of the cell populations to the plugs were recorded at 1 and 3 h.
Responses in swarm plates
Aerobically and photosynthetically grown R. sphaeroides responded to a broad range of chemoeffectors on swarm plates, however there were differences between the cells grown under the two growth conditions. Swarm plates containing concentrations from 100 nM to 1 mM of either Na–succinate or propionate were incubated aerobically or anaerobically for 72 h and the sizes of the tactic bands were observed (Fig. 1). Both aerobically and photosynthetically incubated cells produced larger swarms to propionate than succinate. Under both conditions it was visually observed that higher concentrations of attractant caused regularly spaced chemotactic bands to form within the colony. JPA403 (Δ operon 2) although motile did not show any swarming on swarm plates. No differences were apparent between JPA117 (Δ operon 1) and the wild-type behaviour.
Individual cell responses
Measurement of the tracks of aerobically grown free-swimming wild-type cells showed a mean population swimming speed of 13±0.3 μm s−1 with cells stopping on average every 1.4±0.1 s for a period of about half a second.
The behaviour of 50 tethered cells was analysed over 180 s and they stopped approximately once every 4 s (mean population stopping frequency 0.26±0.19 stops s−1) with a mean population stop duration of 0.7±0.5 s. The stop probability of chemoheterotrophically grown R. sphaeroides was 0.11. The parameters of motility for photosynthetically grown bacteria [11,15] did not significantly differ from those found for aerobic cells and thus cannot account for the differences on swarm plates.
Aerobically grown R. sphaeroides were tethered and Na–HEPES (pH 7.2) buffer with and without 1 mM attractant was flowed past the cells and compared to photoheterotrophically grown cells. For example, Fig. 2 shows the response of a population of 10 cells to the addition and removal of 1 mM sodium succinate. The stop probability increased from the prestimulus level of 0.1 to 0.5. Visually individual cells appeared to be stopping frequently rather than all the cells in the population showing a sustained stop.
Analysing the population data, on addition of succinate the population stopped twice as frequently and for longer (mean population stopping frequency increased from 0.16±0.1 to 0.24±0.1 stops s−1, the stop duration from 0.6±0.4 s to 1.4±2 s) with the overall percentage of stopped time increasing from 7% to 28%. The run lengths were reduced with the mean population run duration decreasing from 27.0±49.5 s to 4.5±4.6 s although the mean population rate of rotation only decreased from 5.5±1 Hz to 4.8±1 Hz. This is a very different response to that shown by photoheterotrophic cells which showed an apparent increase in the rate of rotation and smooth swimming, on the addition of 1 mM Na–succinate. In both cases the cells adapted to the level of chemoeffector.
Removal of attractant resulted in the cells returning to prestimulus behaviour after a transient increase in speed and a period of smooth rotation (stop probability 0.01) for 330 s (Fig. 2). An increase in the rate of rotation and decrease in stopping frequency to a step-down in concentration has not been observed previously. A number of other attractants gave an inverted response (compared to photoheterotrophic cells) on their addition (Table 1) but only succinate gave a transient speed increase on its removal.
|Attractants causing inverted responses||Attractants causing normal responses|
|All of the chemoeffectors, except malonate, were attractants.|
|Attractants causing inverted responses||Attractants causing normal responses|
|All of the chemoeffectors, except malonate, were attractants.|
Some attractants produced identical responses to those found for photoheterotrophically grown cells, defined as the ‘normal’ response (Table 1). Fig. 3 shows the response of a cell population to Na–propionate. The addition of 1 mM Na–propionate increased smooth rotation followed by a response on removal of the attractant with an increase of stopping probability from 0.1 to 0.86. The non-metabolisable analogue of succinate, Na–malonate, caused a small increase in stop probability on its addition but no adaptation.
Irrespective of the pattern of tethered cell responses all of the chemoeffectors used, with the exception of malonate, they were chemoattractants in swarm plates indicating that either pattern of response can lead to accumulation in real gradients.
Effect of concentration on the behavioural response
Photoheterotrophically grown cells show an increase in stopping frequency when high concentrations of attractant are added . The response of tethered aerobically grown cells to different concentrations of Na–succinate between 500 nM and 1 mM was investigated. These concentrations caused swarming in soft agar (sub-saturation to saturation of the sensory pathway). The addition of concentrations up to 500 μM did not result in any significant changes in rotation, however, the addition of 750 μM and above elicited an increase in the stop probability (Fig. 4). In plug plates, using concentrations of Na–succinate from 10 μM to 10 mM, responses were found to concentrations of 250 μM and above.
Responses of the chemotaxis operon deletion mutants JPA117 and JPA403
The behavioural responses of JPA117 (Δ operon 1) and JPA403 (Δ operon 2) were analysed using tethered cells (Fig. 5). The chemoattractants propionate and succinate were used at 500 nM and 1 mM concentrations. Aerobic and photosynthetic JPA117 showed identical responses to those of wild-type. Photosynthetically grown JPA403 showed inverted responses, a stop on addition to all of the attractants while aerobically grown cells showed much weaker but still inverted responses.
Both aerobically and photosynthetically grown wild-type R. sphaeroides swarmed through soft agar plates containing concentrations of a broad range of chemoeffectors, though there were some differences in swarm patterns between the two types of cell. When tethered, however, and subject to step increases and decreases in attractant the responses of the two cell types were found to be very different. Photosynthetically grown cells showed little response to a step-up (i.e. a positive stimulus) but stopped (i.e. tumble) on a step-down (i.e. a negative stimulus). This is very similar to E. coli, but the high rotational bias of the motor makes the negative response the most apparent behavioural change .
Aerobically grown wild-type cells however showed the reverse response to some chemoeffectors. They stopped and adapted when subjected to a step-up in attractant and returned to their prestimulus behaviour on its removal. Other attractants caused the same response as photosynthetically grown cells. These inverted responses have not been found in photosynthetically grown wild-type R. sphaeroides to any chemoeffector tested.
Although tethered cells showed these very different responses, both aerobically and anaerobically grown cells produced tactic swarms in soft agar around plugs containing all of the chemoeffectors.
JPA203, a mutant with a transposon interruption of operon 2 and JPA403, the operon 2 deletion mutant, both showed an inverted response to attractants in tethered experiments after photosynthetic growth, but did not show chemotaxis in swarm plates (, S. Porter, D. Shah and J.P. Armitage, submitted). The difference in swarming ability of aerobically grown cells and operon 2 deletion cells, which both gave inverted responses to step changes when tethered, suggests that responses to real gradients may be more complex than simple stimulus response microscopic assays would suggest.
Inverted chemotaxis responses have been reported in E. coli and Salmonella typhimurium. Some mutants in the motor protein FliM show inverted responses as do mutants with cheB (the methylesterase) deleted and some mutants in the receptor methyl accepting chemotaxis protein MCP, Tsr [16–19]. It has been suggested that the loss of cheB in E. coli results in over-methylation of MCPs resulting in an inverted response, however the inverted response in R. sphaeroides occurs in both the presence of CheB (wt and JPA203) and in a mutant lacking CheB (JPA403) suggesting over-methylation is not likely to be the cause.
The presence of two responses indicates that the sensory response in R. sphaeroides is complex. Models exist predicting that a direction changing response to negative stimuli could allow a cell population to move up a gradient but a model allowing accumulation when cells stop in response to increases in chemoeffector concentration is difficult. It is possible that this type of response could lead to trapping in regions of high chemoeffector concentration, but this is less efficient than chemotaxis and does not reflect the findings in this study as no difference was found in the strength of the chemotactic responses in plug and swarm plates.
In this study R. sphaeroides only responded to saturating stepwise increases in concentration in the tethered cell assay and the responses to these high concentrations may not reflect its ability to respond to shallow gradients. In swarm plates responses were seen at concentrations much lower than these very high levels. As bacteria would encounter natural gradients of the order of concentration change of 1 μM s−1 it is likely that R. sphaeroides responds to this concentration range under natural conditions and indicates that for bacteria with complex chemosensory pathways such as R. sphaeroides extrapolation for the behaviour of cells in swarm plates from large step increases in concentration should be interpreted with caution.
The authors would like to thank the NERC for their support of this research. H.L.P. is an NERC Advanced Research Fellow.