Bacteria utilize quorum sensing to regulate the expression of cell density-dependant phenotypes such as biofilm formation and virulence. Zoospores of the marine alga Ulva intestinalis exploit the acyl-homoserine lactone (AHL) quorum sensing system to identify bacterial biofilms for preferential settlement. Here, we demonstrate that AHLs act as strong chemoattractants for Ulva zoospores. Chemoattraction does not involve a chemotactic orientation towards the AHL source. Instead, it occurs through a chemokinesis in which zoospore swimming speed is rapidly decreased in the presence of AHLs. The chemoresponse to AHLs was dependant on the nature of the acyl side chain, with N-(3-oxododecanoyl)-homoserine lactone (3O-C12-HSL) being the most effective signal molecule. Mean zoospore swimming speed decreased more rapidly over wild-type biofilms of the marine bacteria Vibrio anguillarum relative to biofilms of the vanM mutant, in which AHL synthesis is disrupted. These data implicate a role for AHL-mediated chemokinesis in the location and preferential settlement of Ulva zoospores on marine bacterial assemblages. Exposure to AHLs did not inhibit the negative phototaxis of Ulva zoospores, indicating that chemoattraction to bacterial biofilms does not preclude the response to a light stimulus in substrate location.
Eukaryotes have evolved signalling mechanisms to detect and respond to the presence of bacteria. In plants, detection of specific bacterial signals allows the pre-emptive induction of host defence systems or the identification of specific symbionts (Dangl & Jones 2001; Riely et al. 2004). The recent discovery that algae and higher plants can respond directly to bacterial quorum sensing signals greatly expands the potential role for chemical signalling in plant–microbial interactions (Joint et al. 2002; Mathesius et al. 2003). Characterizing the range of plant responses to quorum sensing signals will allow us to identify novel mechanisms through which plants and bacteria interact.
Quorum sensing enables bacterial populations to communicate intercellularly using specific diffusible chemical signals to modulate gene expression in accordance with changing population density (Bassler 2002). The best-defined quorum sensing system is the acyl-homoserine lactone (AHL) system present in a diverse range of Gram-negative bacteria, although many other signalling molecules have also been identified (Taga & Bassler 2003). At a low cell density, AHLs are produced at constant low level. As cell density increases, AHL concentration reaches a critical threshold, enabling regulatory circuits to modulate the expression of genes involved in many diverse functions such as biofilm formation, virulence and bioluminescence (Parsek & Greenberg 2000; Lazdunski, Ventre & Sturgis 2004). Specificity in AHL quorum sensing systems is conferred by variations in the length of the acyl side chain and its respective substitutions, allowing individual species to distinguish their signals from those of others.
It is now apparent that quorum sensing signals are not exclusive to bacterial communication but can be exploited by plants to detect and locate their prokaryotic neighbours (Joint et al. 2002). Given the importance of quorum sensing signalling in regulating pathogenic and symbiotic bacterial phenotypes (Winzer & Williams 2001; Gonzalez & Marketon 2003), it is not surprising that eukaryotes have evolved mechanisms to detect quorum sensing signalling. For example, the proteome of the legume Medicago truncatula demonstrates specific responses to different AHLs from symbiotic or pathogenic bacteria (Mathesius et al. 2003). In addition, plants may influence their neighbouring bacterial communities by disrupting quorum sensing signalling. The marine alga Delisea pulchra disrupts AHL signalling by producing halogenated furanones that bind the AHL-responsive LuxR transcription factor from Vibrio fischeri and promote its proteolytic degradation (Givskov et al. 1996; Manefield et al. 2002). Medicago truncatula and the unicellular green alga Chlamydomonas reinhardtii produce AHL mimics, which may stimulate or inhibit the quorum sensing circuits of surrounding bacteria (Mathesius et al. 2003; Teplitski et al. 2004).
The marine alga Ulva intestinalis, formerly Enteromorpha intestinalis (Hayden et al. 2003), exploits quorum sensing signalling molecules in order to locate bacterial biofilms. The asexual stage of Ulva's reproductive life cycle involves the massive release of motile quadriflagellate zoospores, which enable dispersal and location of suitable substrates for settlement. Settlement results in irreversible attachment of zoospores to a substrate and is influenced by many different factors (Callow et al. 1997). Negative phototaxis and gravitaxis enable swimming zoospores to orientate their swimming towards a solid substrate. Once a substrate is located, its surface chemistry, wettability and topography are important in determining settlement (Callow & Callow 2000). Microbial biofilms present on all submerged surfaces in the marine environment will determine many of these characteristics. Ulva zoospores were shown to settle preferentially onto bacterial assemblages from natural sea water relative to a control surface (Joint et al. 2000). Subsequent studies using the marine bacterium Vibrio anguillarum identified a requirement for AHLs in this preferential settlement (Joint et al. 2002). Settlement of Ulva zoospores is positively correlated with the density of wild-type V. anguillarum biofilms, but strains unable to produce AHLs through deletion of AHL synthases or expression of an AHL lactonase (aiiA from Bacillus sp.) do not promote settlement (Tait et al. 2005). The response is specific to the AHL signal molecule, as preferential settlement occurs on agarose films impregnated with synthetic AHLs and on Escherichia coli biofilms expressing AHL synthases from V. anguillarum (Joint et al. 2002).
It is evident that AHLs are important in the settlement process of Ulva zoospores, although the mechanisms through which they act are unclear. In this paper, we present data to demonstrate that bacterial signalling molecules influence the settlement process by modulating zoospore swimming to allow location of bacterial biofilms. The study explores the nature of this response and demonstrates that attraction to AHL-producing biofilms occurs via chemokinesis rather than chemotaxis.
MATERIALS AND METHODS
Fertile thalli of U. intestinalis were collected from Wembury, Devon, England (50°18′N: 40°2′W) and stored at 15 °C. Zoospores were released by immersion in filtered sterile sea water as described previously (Callow et al. 1997). In addition, zoospores were further separated from any contaminating gametes by phototaxis. Zoospores are negatively phototactic and collect on the side of a glass Petri dish furthest from a light source, whereas gametes are positively phototactic and collect nearest the light source. Zoospores were collected and used immediately.
Bacterial strains and growth conditions
Wild-type V. anguillarum strain NB10 and the vanM mutant have been described previously (Milton et al. 2001). For biofilm formation, overnight cultures were grown in tryptic soy broth (TSB), harvested by centrifugation and washed in sterile sea water. The resulting cell suspension was used to inoculate sterile culture vessels containing microscope coverslips immersed in sea water. After incubation for 48 h at 20 °C, the coverslips were thoroughly rinsed in sterile sea water to remove unattached cells. Biofilm coverage was determined using the method described by Tait et al. (2005) to ensure that an equal cell density was present for both strains.
Synthesis of AHLs
The AHLs used in this study were synthesized, purified and characterized as described previously (Chhabra et al. 1993; Camara, Daykin & Chhabra 1998). For convenience, AHL names are abbreviated; with 3H-C6-HSL, 3O-C10-HSL and 3O-C12-HSL representing N-(3-hydroxyhexanoyl)-homoserine lactone, N-(3-oxodecanoyl)-homoserine lactone and N-(3-oxododecanoyl)-homoserine lactone, respectively.
Capillary chemotaxis assay
Glass capillaries (length 7.5 mm, inner diameter 0.5 mm, Harvard Apparatus, Kent, UK) were filled with test solutions of AHLs in sea water. Capillaries were arranged horizontally around a glass cavity slide containing 300 µL of zoospore suspension (2.4 × 106 zoospores mL−1). Only the tip of each capillary was in contact with the zoospore suspension. Capillaries containing sea water were used as control. Assays were performed in the dark at 20 °C for 15 min. Capillaries were immediately removed from the zoospore suspension, the exterior blotted very carefully and the contents expelled into 5 µL of 1% Lugol's iodine. Cell numbers were analysed with a Reichert-Jung Polyvar microscope (Vienna, Austria) and an Optronics Magna Fire SP camera (Goleta, CA, USA) with Image Pro-Plus image analysis software (Media Cybernetics, Silver Spring, MD, USA). Macroalgal zoospores may settle in capillary tubes, leading to an underestimation of cell number when the contents are expelled and counted. However, we assessed the number of zoospores settling over the course of the assay, using flattened capillary tubes so that internal cell numbers could be counted, and settlement within flattened capillary tubes was found to be negligible.
Video motion analysis of zoospore swimming
Cells were imaged using a Nikon Diaphot microscope (Tokyo, Japan) and a JVC TK-1085E colour video camera (Wayne, NJ, USA). Zoospores were imaged using dim red light (700 nm long pass filter) to minimize phototaxis (Callow & Callow 2000). Zoospore suspensions were mixed with an equal volume of AHL test solution and added to a glass cavity slide. Zoospores are highly thigmotactic and were imaged swimming at the glass surface in all cases, unless otherwise stated. Swimming patterns were recorded for 1 s (25 frames per second) at appropriate time intervals using Ulead Video Studio v4.0 video capture software (Ulead Systems, Kaarst, Germany). The direction and speed of swimming zoospores were analysed using Image Pro-Plus v5.1 motion tracking software (Media Cybernetics). To minimize errors due to pixel drift, swimming speed for each individual zoospore was defined as the cumulative distance moved over a one second period, and zoospore position was measured every 200 ms. Unless otherwise stated, the mean swimming speed for a population of zoospores was determined from a minimum of 200 tracks, with each track representing the distance travelled by an individual zoospore in one second. For all swimming speed determinations, error bars denote one standard error. Swimming direction was defined as the vector angle between the initial position and the final position of the zoospore after 1 s. Zoospores swimming very slowly (i.e. less than 20 µm s−1) were not included in the analysis of mean swimming direction. Using a specific settlement assay based on the accumulation of the membrane dye FM 1-43 (Thomson et al., unpublished data), we confirmed that non-motile zoospores had actually settled following AHL exposure
To monitor zoospore behaviour around a dried spot of AHL, either 3O-C10-HSL or 3O-C12-HSL (10–20 nmol) was spotted onto the surface of a small glass dish and allowed to dry. One millilitre of zoospore suspension was added to the dish, and video microscopy was used to record swimming patterns as detailed above. Phototactic experiments were conducted in square cell culture dish (100 × 100 mm, 25-well plate, Bibby Sterilin, Staffs, UK) using a Schott KL200 cold light source (Mainz, Germany) with fibre optic guides. Zoospore swimming was imaged prior to illumination and again at 30 s after the onset of illumination.
The mean direction of swimming spores was determined using Oriana circular statistics package (Kovach Computing, Anglesey, UK). The Rayleigh test was applied to measure whether the sampled population is distributed randomly around a circle. In addition, the V-test, a modified form of the Rayleigh test, was applied to identify mean orientation towards a specific angle. One-way analysis of variance (anova) followed by Tukey's honest significant difference (HSD) test were performed, where appropriate, using Minitab software (Minitab, Warwicks, UK).
AHLs are chemoattractants for Ulva zoospores
We have previously demonstrated that production of AHL quorum sensing signalling molecules is essential for Ulva zoospores for increased settlement on bacterial biofilms (Joint et al. 2002; Tait et al. 2005). The addition of synthetic AHLs to the surrounding sea water prevented preferential settlement on AHL-producing biofilms, suggesting that zoospores may employ chemotaxis towards a diffuse AHL concentration gradient. To investigate the ability of AHLs to act as chemoattractants, we examined the distribution of Ulva zoospores around a point source of AHL. AHLs with longer acyl side chains (C10–C14) are poorly soluble in aqueous solution (Yates et al. 2002), and a dried spot of AHL will dissolve slowly to produce a diffusion gradient around the point source. The addition of a zoospore suspension to a point source of 3O-C12-HSL resulted in a massive accumulation of zoospores around the AHL source within 10 min, suggesting that there is a strong attraction to this signalling molecule (Fig. 1). The area of accumulated zoospores extended much further than the point source itself, indicating that zoospores are responding to a concentration gradient of dissolved attractant. Similar results were obtained with 3O-C10-HSL, although no accumulation was observed around the point source of 3H-C6-HSL (data not shown). However, the increased solubility of 3H-C6-HSL relative to the longer chain AHLs may restrict its ability to act as a chemoattractant in this assay, as it dissolves rapidly from the surface.
Zoospores do not exhibit chemotaxis towards AHLs in the capillary assay
The massive accumulation of zoospores around a point source of AHL suggests that there is a chemotactic response to the bacterial signalling molecules. We initially tested this hypothesis using the capillary tube assay for chemotaxis. This method has been employed previously in the identification of chemotactic responses in microalgae and macroalgal zoospores (Sjoblad & Frederikse 1981; Amsler & Neushul 1989). Chemotaxis was assessed by counting the number of zoospores that have swum into glass capillary tubes containing the test attractant relative to the sea-water control. Surprisingly, we found that a range of 3O-C12-HSL concentrations had no significant effect on zoospore accumulation within the test capillary (Fig. 2). An alternative AHL, 3H-C6-HSL, also failed to elicit a chemotactic response, suggesting that zoospores may not use chemotaxis in their localization of AHLs.
The term chemotaxis is commonly used to refer to chemoattractant-dependant changes in cell motility. However, it is often used erroneously, as true chemotaxis is the directional orientation of a cell following exposure to a chemoattractant (Amsler & Iken 2001). In contrast, chemical-mediated changes in swimming speed and turning frequency are chemokinetic responses. It should be noted that the capillary assay may not identify certain chemokinetic responses. For example, a chemokinesis resulting in decreased swimming speed could lead to an accumulation of zoospores within the test capillary, but may also serve to reduce cell numbers entering the test capillary.
AHLs induce a rapid decrease in zoospore swimming speed
To examine in detail the nature of zoospore swimming following exposure to AHL and to distinguish between chemotactic and chemokinetic responses, we employed video motion analysis. Zoospores were monitored swimming at the surface of a glass slide after the addition of different concentrations of 3O-C12-HSL (Fig. 3a). Control zoospores began to settle during the course of the assay, as demonstrated by the marked decrease in mean swimming speed after 25 min. However, there is no appreciable decrease in the swimming speed of control zoospores after 4 min, which is in strong contrast with zoospores swimming in the presence of 125 or 25 µm 3O-C12-HSL. In these treatments, zoospore swimming speed after 4 min was reduced to 27 and 47%, respectively, of the control. A similar response was not elicited by 5 µm 3O-C12-HSL, suggesting that there is a threshold concentration for the decrease in swimming speed.
In order to examine the specificity of the AHL-mediated swimming response, we investigated the efficacy of a range of AHL molecules with differing acyl side chains in the video motion analysis assay. The 3O-C12-HSL signalling molecule was clearly the most effective in reducing swimming speed, eliciting an 82% decrease after 10 min relative to the control (P < 0.001, one-way anova, Tukey's HSD, Fig. 3b). Zoospore swimming speed was also significantly reduced following exposure to 3O-C10-HSL (P < 0.05), although this was only a 39% reduction relative to the control. Zoospores swimming in the presence of other AHLs, notably the 3-oxo group or other longer chain AHLs, exhibited reduced swimming speeds although these were not found to be statistically significant from the control.
Zoospores exhibit chemokinesis to a diffuse AHL concentration gradient
The dramatic and rapid AHL-mediated decrease in zoospore swimming speed suggests that chemokinesis, rather than chemotaxis, may be responsible for the accumulation of zoospores around the AHL point source. Thus, zoospores may not orientate themselves towards the source of AHL, but accumulate in areas containing high AHL concentrations due to a rapid decrease in their swimming speed. To test this hypothesis, we used video motion analysis to examine the swimming of zoospores around point sources of 3O-C10-HSL and 3O-C12-HSL. After 8 min, swimming speed was measured at the glass surface at increments up to 15 mm away from the point source. There is a clear correlation between zoospore swimming speed and distance to the source of 3O-C12-HSL, with zoospores closer to the AHL source swimming much more slowly (Fig. 4a). To establish whether contact with the glass surface influenced the chemoresponse, the microscope's focal plane was altered to capture the swimming paths of zoospores that were not in contact with the surface. These spores exhibited no significant reduction in swimming speed, indicating that proximity to the surface is necessary for chemokinesis.
The mean swimming speed was also greatly reduced in zoospores adjacent to the source of 3O-C10-HSL (Fig. 4b). 3O-C10-HSL was less effective at 5 and 10 mm distances, which is perhaps a reflection of the lower efficacy of this AHL relative to 3O-C12-HSL, as determined previously. Nevertheless, it is apparent that zoospores are responding to a diffuse gradient originating from either AHL source. It is conceivable that the observed reduction in mean swimming speed may be due to an AHL-induced increase in settlement rather than a direct effect on swimming speed. To investigate further the nature of the chemokinesis, we compared frequency distribution histograms of individual swimming speeds adjacent to and 15 mm away from an AHL source (Fig. 4c). In the presence of 3O-C10-HSL, there was a large increase in the proportion of zoospores that either had settled or were swimming very slowly (0–20 µm s−1), suggesting that AHLs may promote settlement. However, the proportion of zoospores swimming at higher speeds (greater than 100 µm s−1) was also dramatically reduced, indicating a direct effect on swimming speed itself. These data suggest that the chemokinetic response may serve not only to localize zoospores around an AHL source but also to promote their settlement.
AHLs do not alter the orientation of swimming zoospores
The dramatic chemokinetic response following exposure to either 3O-C10-HSL or 3O-C12-HSL explains the accumulation of zoospores around an AHL source. However, the presence of an additional directional response to an AHL stimulus could not be discounted. We therefore examined the orientation of zoospore swimming tracks around a point source of either 3O-C10-HSL or 3O-C12-HSL. The circular frequency histograms indicate no apparent orientation of the spores towards either AHL source (Fig. 5a). Data were assessed statistically using the Rayleigh test for uniformity. There was no significant orientation of the swimming tracks at any of the distances examined. In addition, the V-test indicated there was no mean orientation towards 270°, where the AHL source was positioned (data not shown). In combination with the previous data, this evidence suggests that zoospores accumulate around an AHL source due to a dramatic reduction in swimming speed and not because of a chemotactic response resulting in a change of swimming direction.
AHLs do not inhibit phototaxis in swimming zoospores
The settlement of Ulva zoospores may be influenced by many different factors. Light has a strong influence on zoospore orientation, as the zoospores are negatively phototactic. In order to understand the contribution of AHL-mediated chemokinesis in the complex settlement process, we compared the phototactic responses of zoospores with the chemokinetic response and examined their possible interaction. The directional swimming tracks of zoospores in the dark are distributed randomly, with no significant orientation (Fig. 5b). Following exposure to a directional light source, there is an immediate switch in the orientation of swimming paths. After 30 s, the vast majority of cells are swimming directly away from the light, as indicated by the highly significant Rayleigh test (P < 0.001) and V-test (90°, P < 0.001). The apparent orientation of zoospores swimming in a light gradient is in clear contrast with those in an AHL gradient.
The integration of photosensory and chemosensory signalling pathways has been studied in the unicellular green alga Chlamydomonas (Ermilova, Zalutskaya & Gromov 1997; Govorunova & Sineshchekov 2003). Chemoattractants were shown to inhibit not only phototaxis, but also the generation of the membrane photocurrent, which is the earliest detectable event in the light-dependant signalling pathway. Thus, there is a close integration of photosensory and chemosensory responses at the very onset of the signalling process. We examined phototactic responses of Ulva zoospores at time intervals following exposure to 3O-C12-HSL. Throughout the course of the experiment, the phototactic response of swimming zoospores was very pronounced, with significant orientation of swimming paths in all cases (Table 1). Thus, there is no apparent inhibition of phototaxis in zoospores that continue to swim. However, it must be considered that the rapid decrease in swimming speed due to addition of 3O-C12-HSL results in an increasing percentage of non-swimming cells. Therefore, while phototaxis is not inhibited directly by AHLs, the chemokinetic effect of this attractant will result in a decrease in the number of cells responding to a light stimulus.
Table 1. Analysis of zoospore phototaxis following AHL addition
Time (min) Directional light
Zoospores were imaged at time intervals after the addition of 60 µm 3O-C12-HSL, firstly in the absence of light, and then after 30 s adaptation to directional light positioned at 270° (light intensity 50 µmol m−2 s−1). Orientation towards 90° indicates negative phototaxis and is assayed by the V-test.
Rayleigh test (P)
V-test, 90° (P)
Zoospore swimming speed is decreased around V. anguillarum biofilms producing AHLs
Our previous work has indicated that zoospores preferentially settle on bacterial biofilms producing AHL quorum sensing signalling molecules (Joint et al. 2002; Tait et al. 2005). We therefore examined whether the AHL-mediated chemokinesis could contribute to the increased settlement on bacterial biofilms. Zoospores were imaged swimming over a biofilm of wild-type Vibrio anguillarum, which produces three AHL signalling molecules (C6-HSL, 3H-C6-HSL and 3O-C10-HSL), and over a biofilm of the vanM mutant which is unable to synthesize any AHL (Milton et al. 2001). Mean zoospore swimming speed decreased more rapidly over the wild-type biofilm, in comparison with the vanM mutant (Fig. 6a). As mentioned previously, the decrease in mean swimming speed may be due in part to an increased proportion of settled zoospores in the presence of AHLs. Therefore, in order to resolve whether biofilm-derived AHLs acted to directly decrease zoospore swimming speed, we ignored the contribution of settling zoospores and compared mean swimming speeds using data for zoospores swimming faster than 20 µm s−1. At all time points examined, mean swimming speed was reduced for zoospores swimming over a wild-type biofilm relative to a vanM mutant (Fig. 6b). The analysis was repeated with four independent replicate biofilms of each strain. After 10 min, both mean swimming speed and mean swimming speed excluding settling zoospores were significantly reduced over the wild-type biofilms relative to the vanM mutant (P < 0.001 and P < 0.034, respectively, one-way anova), whereas the percentage of settling zoospores (swimming speed < 20 µm s−1) was significantly increased (P < 0.001, one-way anova). Thus, Ulva zoospores swimming over AHL-producing biofilms exhibit both a decrease in swimming speed and an increase in the apparent rate of settlement.
The initial discovery of quorum sensing regulated bioluminescence in Vibrio fischeri has led to the identification of multiple quorum sensing regulatory circuits in diverse bacterial species (Fuqua, Parsek & Greenberg 2001). This has, in turn, changed our perception of bacteria from selfish individuals to coordinated altruistic populations. Dissection and characterization of the responses of Ulva zoospores to quorum sensing signals will allow us to ascertain whether similar responses are widespread among other eukaryotes and perhaps re-evaluate our perception of plant–microbial interactions.
Our previous studies have identified a requirement for AHLs in the settlement of Ulva zoospores on bacterial biofilms. This present study aimed to determine the mechanism through which AHLs promote zoospore settlement and to identify their significance to the ecological success of Ulva. Our results demonstrate that exposure to specific AHLs rapidly decreases the swimming speed of Ulva zoospores, but swimming paths are not orientated towards a source of AHL. The effect on swimming speed becomes more pronounced with proximity to a point source of AHL, resulting in a massive accumulation of zoospores around the point source. The decreased swimming speed may therefore contribute to increased settlement on bacterial biofilms by both concentrating zoospore numbers and increasing the likelihood that settlement will occur. In accordance with this, zoospore swimming speed decreased more rapidly over an AHL-producing biofilm of wild-type Vibrio anguillarum relative to the AHL-deficient vanM mutant.
The nature of the Ulva zoospore chemoresponse is intriguing and appears to be a chemokinetic rather than a chemotactic response to the AHL signal. Both responses would result in accumulation, and therefore it is likely that the adopted mechanism reflects both the nature of the plant–microbial interaction and the environment in which it occurs. Chemosensory mechanisms related to nutrient availability, antifouling compounds and reproduction have been investigated in other macroalgal propagules. Chemotaxis towards a nitrogen source was observed in zoospores of the kelps Macrocystis pyrifera and Pterygophora californica (Amsler & Neushul 1989). Video motion analysis indicated a directional orientation towards the nitrate source, confirming that the accumulation of zoospores observed in a capillary assay was due to ‘true’ chemotaxis. In contrast, zoospores of the brown alga Ectocarpus siliculosus did not exhibit a change in swimming direction in response to a broad nutrient source (Amsler et al. 1999). Motile zoospores of the brown alga Hincksia irregularis exhibited erratic swimming patterns on exposure to natural antifouling agents present in echinoderm extracts (Iken et al. 2003). The chemoresponses relating to gamete fusion in the brown algae have been extensively studied. Female gametes secrete sexual pheromones attracting massive numbers of male gametes. The nature of the swimming response is complicated and species-specific (Maier 1995). Male Ectocarpus gametes exhibit a response defined as chemothigmoklinokinesis (Geller & Muller 1981). On initial exposure to the pheromone ectocarpene, male gametes exhibit a strong thigmotactic response, resulting in reduced swimming speed in constant contact with the surface. Increasing ectocarpene concentrations result in increasing turning frequency, locking the male gamete into the vicinity of the female gamete through a trajectory of ever-decreasing circles. In contrast, male gametes of the kelp Laminaria digitata directionally orientate themselves towards female gametes, thus exhibiting a true chemotactic response (Maier & Muller 1990). The lamoxirene pheromone of Laminaria also induces a thigmotactic response, maintaining contact between the male gametes and the surface.
The chemokinetic response of Ulva zoospores is similar in some respects to the swimming patterns of male Ectocarpus gametes. On exposure to AHL, zoospores in contact with a surface exhibit a greatly reduced swimming rate, ensuring that the swimming cells stay in the proximity of the source of chemoattractant. However, Ulva zoospores did not circle a point source of AHL, as observed for male Ectocarpus gametes around a source of ectocarpene. This reflects the requirements of each propagule in that the male Ectocarpus gamete must locate precisely the female gamete in order to be successful, whereas the localization of the Ulva zoospore is influenced by many other factors. Thus, in response to AHLs, Ulva zoospores will remain in the vicinity of a suitable bacterial biofilm but will continue to respond to other settlement cues such as surface topography and chemistry.
The small size of macroalgal propagules and the turbulent nature of the intertidal zone imply that spore motility will not play a major role in the settlement process until the spore enters the undisturbed boundary layer surrounding a solid substrate (Amsler, Reed & Neushul 1992; Fletcher & Callow 1992). However, on entry to the boundary layer, thigmotactic, phototactic and chemotactic responses of macroalgal propagules are all important in determining substrate location (Amsler et al. 1992; Fletcher & Callow 1992). It was proposed that in male gametes of Ectocarpus, chemoattraction is a very short-distance orientation mechanism of only a few millimetres, with longer distance orientation governed by phototactic responses (Maier 1995). A similar mechanism is apparent in Ulva zoospores where the strongly negative phototaxis (or gravitaxis) is used to promote contact with the solid substrate followed by short-range chemoattraction to bacterial biofilms. This is a reflection of the nature of the stimuli, as light gradients will be orientated in a constant direction throughout the boundary layer, but chemical gradients for poorly soluble compounds, such as AHLs or mating pheromones, are likely to exist over a very small distance (Amsler & Iken 2001). Thus, light rapidly orientates swimming for all zoospores away from the light source, whereas AHLs decrease swimming speed for zoospores in contact with a surface. Without the requirement for surface contact, the chemokinetic response could be counterproductive and could prevent localization of a solid substrate. The concentrations of 3O-C12-HSL in biofilms of Pseudomonas aeruginosa have been measured in excess of 600 µm, although much lower nanomolar concentrations were detected in the corresponding aqueous phase (Charlton et al. 2000). This suggests that, at least for 3O-C12-HSL, there will be a very steep gradient of decreasing AHL concentration in the immediate vicinity of a biofilm. The observed threshold for the Ulva chemoresponse (25 µm 3O-C12-HSL) implies that it will only occur at the biofilm surface or in the immediate vicinity of a biofilm and not at the much lower concentrations found in the aqueous phase. This is analogous to the specific inhibition of Candida albicans filamentation by 200 µm 3O-C12-HSL (Hogan, Vik & Kolter 2004), suggesting that certain interspecies interactions may only occur within the context of a biofilm.
A survey of natural biofilms demonstrated that different bacterial strains were stimulatory or inhibitory, therefore indicating specificity in the settlement process (Patel et al. 2003). There is also specificity in the chemoresponse, as 3O-C12-HSL was clearly the most effective AHL in reducing swimming speed. Previous studies demonstrated preferential settlement of zoospores on AHLs with longer acyl side chains, which may in part be attributable to their greater stability in sea water (Tait et al. 2005). However, settlement was not increased on 3O-C12-HSL relative to other longer chain AHLs with 3-oxo or 3-hydroxy substitutions, suggesting that AHLs may influence aspects of the settlement process other than the swimming response of zoospores. Nevertheless, the marked specificity of the chemoresponse may be instrumental in our understanding of the interactions between Ulva zoospores and bacterial biofilms. Biofilms represent a beneficial and nutritious substrate for settlement, although there is evidence for a much more specific bacterial interaction. The thalli of many green algae do not develop normally in axenic culture (Fries 1975). Matsuo et al. (2003) identified specific bacteria from the Cytophaga-Flavobacterium-Bacteriodes (CFB) group that are required for the correct morphological development of Ulva and Monostroma. A specific chemical inducer called thallusin, which restores morphological development in axenic culture, was recently isolated from these strains (Matsuo et al. 2005). The strong chemoresponse to 3O-C12-HSL indicates specific targeting of bacteria producing this quorum sensing signal and therefore represents a putative mechanism for locating bacterial symbionts. However, AHL dependant signalling mechanisms have not been identified in members of the CFB group (Gram et al. 2002; Manefield & Turner 2002). Nevertheless, the strong chemokinetic response may be responsible for locating bacterial assemblages that are associated with CFB strains.
It is becoming increasingly apparent that bacterial quorum sensing signals can elicit specific responses in eukaryotes relating to the nature of the microbial interaction. The next challenge in this field is to identify further the range of these interactions and to identify the cellular mechanisms through which they occur. The chemokinetic response of Ulva zoospores resembles the chemotactic and phototactic responses of Chlamydomonas, which are mediated by calcium-dependant modulation of flagellar beating patterns (Harz & Hegemann 1991). We are currently investigating the presence of similar AHL-specific signal transduction pathways in Ulva zoospores. Elucidating this pathway may enable us to devise methods of inhibiting the settlement of this important fouling organism. Furthermore, it will enable the identification of novel microbial interactions through the characterization of similar signal transduction pathways in other eukaryotes.
We are indebted to Paul Williams and Miguel Camara (University of Nottingham, UK) for the supply of AHLs and to Debra Milton (Umea University, Sweden) for the gift of V. anguillarum NB10 and the vanM mutant. We also thank James Callow and Maureen Callow (University of Birmingham, UK) for their advice and guidance. This work was supported by a Leverhulme Trust research grant (GLW, number F/0982/A).