JOURNAL OF BACTERIOLOGY, Nov. 2000, p. 6042–6048 0021-9193/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.
Vol. 182, No. 21
Energy Taxis Is the Dominant Behavior in Azospirillum brasilense GLADYS ALEXANDRE,1† SUZANNE E. GREER,1
AND
IGOR B. ZHULIN2*
Department of Microbiology and Molecular Genetics, School of Medicine, Loma Linda University, Loma Linda, California 92350,1 and School of Biology, Georgia Institute of Technology, Atlanta, Georgia 30332-02302 Received 5 June 2000/Accepted 6 August 2000
Energy taxis encompasses aerotaxis, phototaxis, redox taxis, taxis to alternative electron acceptors, and chemotaxis to oxidizable substrates. The signal for this type of behavior is originated within the electron transport system. Energy taxis was demonstrated, as a part of an overall behavior, in several microbial species, but it did not appear as the dominant determinant in any of them. In this study, we show that most behavioral responses proceed through this mechanism in the alpha-proteobacterium Azospirillum brasilense. First, chemotaxis to most chemoeffectors typical of the azospirilla habitat was found to be metabolism dependent and required a functional electron transport system. Second, other energy-related responses, such as aerotaxis, redox taxis, and taxis to alternative electron acceptors, were found in A. brasilense. Finally, a mutant lacking a cytochrome c oxidase of the cbb3 type was affected in chemotaxis, redox taxis, and aerotaxis. Altogether, the results indicate that behavioral responses to most stimuli in A. brasilense are triggered by changes in the electron transport system.
as carbon and energy sources (20). A. brasilense utilizes fructose via the Entner-Doudoroff pathway and possesses a complete tricarboxylic acid (TCA) cycle; it lacks a catabolic Embden-Meyerhof-Parnas pathway and a hexose monophosphate pathway (14, 26) and grows poorly on amino acids as sole carbon and energy sources (19). Roots of grasses and cereals exude significant amounts of organic acids (mainly those of the TCA cycle), sugars, and amino acids that are major organic compounds in the rhizosphere (11, 23, 29). Motility and chemotaxis are thought to be important factors for efficient plant colonization by the bacteria (47). Aerotaxis is a major behavioral response in A. brasilense, which guides the bacteria to a preferred low oxygen concentration (4 M), which appears to be the optimal oxygen concentration for energy generation and nitrogen fixation (49). The aerotactic response of azospirilla is so strong that it often masks responses to other stimuli, especially in capillary assays (6). A. brasilense is the first bacterium in which redox taxis has been demonstrated (18). Positive chemotaxis to certain nutrients has been demonstrated in A. brasilense, but no repellents were found (5, 32, 48, 51). In the present study, we show that energy taxis is the dominant behavior in A. brasilense, with most of known chemoeffectors being processed via this mechanism, and that changes in the electron transport system govern most behavioral responses. (A preliminary report of these studies has been presented elsewhere [G. Alexandre, S. E. Greer, and I. B. Zhulin, Abstr. 100th Gen. Meet. Am. Soc. Microbiol., abstr. I-56, p. 393. 2000].)
The ability of motile bacteria to detect and respond tactically to physical parameters that are involved in energy generation, such as light, oxygen, and oxidizable substrates, has long been known (10). Energy taxis encompasses aerotaxis, phototaxis, redox taxis, taxis to alternative electron acceptors, and chemotaxis to oxidizable substrates (3, 40, 41). Through energy taxis, bacteria seek the most favorable environment supporting optimal cellular energy levels (13, 27, 38, 40, 41). In Escherichia coli, chemotaxis to most chemical stimuli does not require their metabolism (1). Typically, attractants and repellents are sensed directly by interaction with the periplasmic sensing domain of membrane-spanning chemoreceptors (34). In E. coli, only taxis to proline (9), glycerol (50), and succinate (8) appears to be dependent on the oxidation of the substrate. Metabolism-dependent responses to several sugars were described for the ␣-proteobacterium Rhodobacter sphaeroides (4, 22) and proposed for some attractants for another ␣-proteobacterium, Sinorhizobium meliloti (4, 15). Chemotaxis of Spirochaeta aurantia to xylose may also be metabolism dependent, since addition of xylose caused changes in membrane potential (16, 17). Two mechanisms for sensing in the metabolism-dependent chemotaxis have been proposed. Bacteria are able to respond to changes in the electron transport system during oxidation of a substrate; i.e., signaling occurs via the energy taxis mechanism, as was shown for glycerol chemotaxis in E. coli (50). Alternatively, bacteria may detect intracellular metabolic intermediates, as was proposed for chemotaxis mediated by the cytoplasmic McpA receptor in R. sphaeroides (4, 45). The ␣-proteobacterium Azospirillum brasilense is a diazotrophic soil organism, able to colonize the rhizosphere of many agronomically important cereals and grasses. This microaerophilic bacterium has an oxidative type of metabolism and can use nitrate as an alternative electron acceptor. A. brasilense is a metabolically versatile bacterium and grows optimally when fructose or organic acids, such as malate or succinate, are used
MATERIALS AND METHODS Bacterial strains and growth conditions. A. brasilense Sp7 (ATCC 29145), wild type for chemotaxis, and its cytN mutant FAJ851, lacking the cbb3 terminal oxidase (25), were used in this study. The cells were grown at 30°C in a minimal medium (18) supplemented with a carbon source of choice to a final concentration of 10 mM. For aerobic growth, the cells were incubated at 200 rpm on a rotary shaker. For anaerobic growth, cells were incubated in an anaerobic hood, and an alternative electron acceptor (either sodium nitrate or dimethyl sulfoxide [DMSO]) was added to a final concentration of 10 mM. The FAJ851 strain was grown in medium supplemented with 25 g of kanamycin per ml. Swarm plates. Chemotactic responses in a spatial gradient were observed on swarm plates containing chemotaxis buffer (K2HPO4, 10 mM; KH2PO4, 10 mM; EDTA, 0.1 mM [pH 7.0]), 0.3% (wt/vol) agar (Difco Laboratories, Detroit, Mich.), and the carbon source to be tested as an attractant at a final concentration of 0.01, 0.1, 1, or 10 mM. Plates were inoculated with 10 l of an overnight culture and incubated at 30°C. Chemotaxis to alternative terminal electron ac-
* Corresponding author. Mailing address: School of Biology, Georgia Institute of Technology, Atlanta, GA 30332-0230. Phone: (404) 894-3700. Fax: (404) 894-0519. E-mail:
[email protected]. † Present address: School of Biology, Georgia Institute of Technology, Atlanta, GA 30332-0230. 6042
ENERGY TAXIS IN A. BRASILENSE
VOL. 182, 2000 ceptors (nitrate, nitrite, and DMSO) was assessed on swarm plates containing chemotaxis buffer, 10 mM malate as a carbon source, and a terminal electron acceptor at a final concentration of 5, 10, 50, or 100 mM. The swarm plates were inoculated with cells that were grown anaerobically with the terminal electron acceptor to be tested. Plates were incubated anaerobically at 30°C. Chemical-in-plug assay for chemotaxis. The chemical-in-plug assay was used essentially as described by Tso and Adler (43) to assess negative taxis to substituted quinones. Cells grown in a minimal medium (18) supplemented with malate (10 mM) were washed twice with chemotaxis buffer and resuspended in the same buffer containing malate (20 mM). The cell suspension was mixed with an equal volume of 0.6% agar in chemotaxis buffer. Plugs of 2% agar in chemotaxis buffer contained malate (10 mM), the substituted quinone to be tested, and 2 mM potassium ferricyanide to maintain the quinones in their oxidized forms. The plates were incubated at 30°C. Miniplug assay for chemotaxis. The semiquantitative method of Grishanin et al. (18) was used with minor modifications to measure taxis in a spatial gradient of organic acids, carbohydrates, amino acids, aromatics, and substituted quinones. A small plug of 1.5% agarose in chemotaxis buffer containing the chemical to be tested as a chemoeffector was placed in a microchamber constructed as described by Grishanin et al. (18); 100 l of cell suspension in chemotaxis buffer was introduced in the microchamber and covered with a coverslip. The formation of a chemotactic band away from (repellent effect) or near (attractant effect) the plug was observed and recorded using a videomicroscope. Different concentrations of chemoeffectors were introduced into the plug in order to determine the threshold for attractants and repellents. Chemoeffectors were tested in a wide concentration range, typically from 1 M to 10 mM. Plugs containing substituted quinones were supplemented with 2 mM ferricyanide to maintain the quinones in an oxidized state. Ferricyanide alone did not affect bacterial behavior. Temporal gradient assay for chemotaxis. Optimal conditions for the use of the temporal assay (33) for A. brasilense were as follows. Cells were grown in minimal medium with 10 mM carbon source to be tested as a chemoeffector to an optical density at 600 nm (OD600) of 0.3 to 0.4, washed three times, and resuspended in chemotaxis buffer. Cells were grown to a low OD600 (⬍0.4) in order to avoid the intracellular accumulation of poly--hydroxybutyrate (PHB), whose presence drastically affected the behavior (see Results). A 9-l drop of a diluted bacterial suspension (107 cells/ml) in chemotaxis buffer was placed on a microscope slide. The compound to be tested (1 l) was added to the suspension, and changes in swimming behavior were recorded with a videomicroscope. Chemoeffectors were tested in a wide concentration range, typically from 1 M to 10 mM. In some experiments (see Results), temporal assays were performed under fully aerobic conditions, where a cell suspension was equilibrated with 21% oxygen (49). To measure the response in a temporal gradient of an alternative electron acceptor, a cell suspension was equilibrated with 100% nitrogen (49). Reversal frequency of free-swimming cells was determined by computerized motion analysis with a VP110 video processor (Motion Analysis Corp., Santa Rosa, Calif.) and a program developed by using EXPERTVISION software, as previously described (49). Video records were digitized at 10 frames/s. Aerotaxis assay. A spatial gradient assay for aerotaxis was performed in an optically flat microcapillary (inner dimensions, 0.1 by 2 by 50 mm; Vitro Dynamics Inc., Rockaway, N.J.). The formation of a band of bacteria near the air-liquid interface was observed, and the distance between the meniscus and the band was measured as described previously (49). The cell suspension contained 1 mM carbon source used as an electron donor. Measurement of membrane potential. Membrane potential was measured using a tetraphenylphosphonium ion (TPP⫹)-selective electrode (World Precision Instruments, Inc., Sarasota, Fla.), as described previously (49, 50). The TPP⫹-selective electrode and a semimicro calomel reference electrode (Orion, Boston, Mass.) were connected to an ion meter (Corning, Wilmington, Del.) and a MacLab MKIII datum recording system (Analog Digital Instruments, Milford, Mass.). The data collected were analyzed and stored in a Macintosh computer using Chart version 3.3 software (Analog Digital Instruments). All measurements were performed in a 10-ml closed vessel at 30°C and pH 7.5. The vessel was ventilated with oxygen gas (21%), and chemicals were added to the cell suspension at a final concentration of 1 mM. The cells were permeabilized to TPP⫹ by treatment with EDTA. Membrane potential was calculated by using the Nernst equation after correction for nonspecific TPP⫹ binding and estimation of the cell volume. Measurement of respiration rates. Respiration rates of bacterial suspensions were measured with a Clark-type electrode and an oxygen monitor (Yellow Springs Instruments, Yellow Springs, Ohio) upon addition of chemicals to be tested at a final concentration of 1 mM. All measurements were performed in a 1-ml closed vessel at 30°C and pH 7.5. The output of the oxygen monitor was collected in a channel of the MacLab datum recording system.
RESULTS Correlation between the efficiency of a chemical as a growth substrate and an attractant. Major organic compounds that are present in root exudates of grasses and cereals (11, 23, 29) were tested as chemoeffectors. Three methods, swarm plates, a
6043
TABLE 1. Correlation between the efficiency of a chemical as a growth substrate and as a chemoeffectora
Chemical
Doubling time for growth (min)b
Malate Succinate Fructose Galactose Glutamate Aspartate Proline Glycerol Alanine
120 130 150 295 440 475 495 770 820
Chemotaxis Threshold in a miniplug assay (M)
Mean response time in a temporal assay (s) ⫾ SDc
1 1 1 7.5 10 10 10 102 103
78 ⫾ 9 62 ⫾ 9 48 ⫾ 8 25 ⫾ 7 22 ⫾ 5 21 ⫾ 4 14 ⫾ 4 18 ⫾ 4 10 ⫾ 3
a Results obtained by using the swarm plate assay directly correlated with those obtained by using miniplug and temporal assays. b Cells were grown in minimal medium (see Materials and Methods) containing 10 mM chemical to be tested as the sole carbon and energy source. c All chemicals were tested at 10 mM.
miniplug assay, and a temporal gradient assay, were used to test the efficiency of a chemical as a chemoeffector for A. brasilense. Strong, moderate, and weak attractants were defined based on the results obtained by all three methods (selected results are presented in Table 1). Organic acids of the TCA cycle (malate, succinate, oxaloacetate, and fumarate), pyruvate, and fructose were strong attractants, whereas oxalate, galactose, ribose, arabinose, aspartate, glutamate, asparagine, and proline were moderate attractants. Alanine, phenylalanine, formate, and glycerol were weak attractants. Glucose, xylose, glycine, histidine, isoleucine, threonine, valine, arginine, lysine, and methionine were not chemoeffectors. These results are in agreement with the lists of attractants previously reported by us (48, 51) and by others (5, 32). In spatial gradient assays, growing the cells on the substrate to be tested as a chemoeffector was not necessary except for galactose taxis: cells showed taxis to galactose only if the sugar was present in the growth medium. In the temporal assay, a chemotactic response of A. brasilense cells toward moderate and weak attractants was observed only when cells were grown on the corresponding substrates, suggesting that metabolism was required to elicit the response. Strong attractants elicited longer responses in a temporal assay when cells were grown on corresponding substrates. The response to all chemoattractants was significantly stronger if the cells were starved before being chemotactically stimulated. To evaluate the relationship between chemotaxis and metabolic efficiency, we measured the doubling times of cells grown on selected chemoeffectors as a sole carbon and energy source (Table 1). A direct correlation between the efficiency of a chemical as a growth substrate and as a chemoeffector was observed. TCA cycle intermediates, such as malate and succinate, and fructose were the best growth substrates for A. brasilense. Galactose, aspartate, glutamate, and proline were relatively poor growth substrates, whereas alanine and glycerol were very poor growth substrates for A. brasilense (Table 1). To investigate further the role of metabolism of a substrate in chemotaxis, we estimated temporal changes in the cell energy level by measuring respiration and membrane potential upon addition of a chemoeffector (Table 2). There was a direct correlation between the efficiency of a substrate as a chemoeffector and the increase in the intracellular energy level upon substrate addition. The strong attractants malate, succinate, and fructose triggered higher increases in oxygen uptake and
6044
ALEXANDRE ET AL.
J. BACTERIOL.
TABLE 2. Temporal changes in respiration and membrane potential of A. brasilense Sp7 upon addition of attractants Mean increase ⫾ SD in: Chemical
Respiration (10⫺11 mol of O2/min/cell)a
Membrane potential (mV)b
Malate Succinate Fructose Galactose Glutamate Alanine
24 ⫾ 4 22 ⫾ 6 18 ⫾ 3 4⫾1 0.2 ⫾ 0.1 0.5 ⫾ 0.2
16 ⫾ 7 23 ⫾ 13 15 ⫾ 2 3⫾1 Nondetectable Nondetectable
a The level of endogenous respiration was (6 ⫾ 2) ⫻ 10⫺11 mol of O2/min/ cell. b The prestimulus level of membrane potential was 119 ⫾ 6 mV.
membrane potential than the moderate attractants, galactose and glutamate, whereas the weak attractant alanine had no detectable effect (Table 2). Chemicals that do not support growth are not attractants. Chemotaxis to some aromatic compounds that are present in plant root exudates, such as catechol, benzoic acid, para-hydroxybenzoic acid, and caffeic acid, was tested. A. brasilense Sp7 could not grow on any of the aromatic compounds as the sole carbon and energy source. None of the aromatic compounds elicited a chemotactic response on swarm plates and in miniplug assays in a wide concentration range. Chemotaxis was then tested in a temporal assay with cells grown on malate, galactose, or glutamate in the presence of the aromatic compound to be tested. No changes in swimming behavior were detected upon addition of any of the aromatic compounds that were tested in a wide concentration range. Maleate and sorbose, nonmetabolizable analogues of malate and fructose, respectively, did not support cell growth, did not cause any increase in respiration or membrane potential, and were not chemoeffectors for A. brasilense. Glucose and most amino acids do not support growth of A. brasilense Sp7 (19, 37) and were not chemoeffectors for this strain when tested in a wide concentration range. Nonmetabolizable analogs and inhibitors of metabolism prevent chemotaxis to specific chemoeffectors. If the metabolism of a chemical is required to elicit a chemotactic response, then inhibiting the metabolism will inhibit chemotaxis. To verify this hypothesis, we first tested chemotaxis toward malate or fructose after incubating the cells for 30 min with 10 mM maleate or sorbose, respectively (Table 3). No change in the swimming pattern was observed in cells incubated with maleate or sorbose. Cells incubated with maleate did not respond chemotactically to malate, whereas cells showed normal chemotaxis to fructose or glutamate. Similarly, sorbose prevented chemotaxis to fructose but did not affect chemotaxis to malate or glutamate. Sodium fluoride inhibits growth of A. brasilense on fructose but not on malate (26) and was used to test the effect of a metabolic inhibitor on chemotaxis. No chemotactic response to sodium fluoride was observed in a temporal assay. Cells incubated with 5 mM sodium fluoride for 45 min showed no change in swimming pattern. However, the cells could no longer respond chemotactically to fructose, whereas chemotaxis to malate and glutamate was unaffected (Table 3). During prolonged growth in a rich medium, A. brasilense cells accumulate PHB as an energy and/or carbon storage material (35–37). PHB oxidation involves a specific NADH-dependent dehydrogenase (35), which competes with the TCA cycle intermediates for the electron transport system (36). The presence of intracellular PHB in the cells (growth above
OD600 ⫽ 0.5) prevented chemotaxis to any chemoattractant (Table 3), while the cells remained fully motile and able to respond to repellents. Altering the electron flow through the electron transport system affects chemotaxis. We tested the effects of the respiratory inhibitors myxothiazol and 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO), a quinone analogue, on chemotaxis. Myxothiazol is a specific inhibitor of the ubiquinol reduction by the Rieske iron-sulfur protein, which is specific for the bc1 complex that donates electron directly to cytochrome c oxidase (42). HQNO is a noncompetitive inhibitor of ubiquinone (42). Incubating the cells in chemotaxis buffer with 2.5 M either of the two respiratory inhibitors completely abolished chemotaxis to all chemoeffectors (Table 3). Addition of the respiratory inhibitors to the cells caused a high reversal frequency bias (Fig. 1), indicating that the inhibitors acted as repellents. Similarly, substituted quinones that are competitive inhibitors of ubiquinone were found to be repellents. In a temporal assay, addition of quinones caused a distinct repellent response (high reversal frequency bias) (Table 4). The repellent effect of myxothiazol lasted for approximately 10 min when cells were incubated with 2.5 M inhibitor, allowing testing for the effect of bypassing electron donors. Addition of ascorbate and N,N,N⬘,N⬘-tetramethyl-p-phenylenediamine (TMPD) (500 and 250 M, respectively), which donate electrons directly to cytochrome c, bypassing the bc1 complex (inhibition site for myxothiazol), had an attractant effect: cells became smooth for 27 ⫾ 5 s and then returned to the high reversal frequency bias caused by the presence of myxothiazol (Fig. 1B). In a control experiment, ascorbate and TMPD did not cause a behavioral response when added to cells incubated in the absence of myxothiazol (data not shown). To further test the hypothesis that the functional electron transport system is required for chemotaxis, the behavior of a cytN mutant of A. brasilense (FAJ851) lacking the cytochrome cbb3 terminal oxidase was analyzed. The motility pattern of the mutant was identical to that of the wild type. The mutant consistently formed an aerotactic band at a lower oxygen concentration than the wild type (Fig. 2A). The distance between the band and the meniscus in the mutant suspension was 142% ⫾ 18% of that in the wild-type suspension (six independent experiments). The effect was observed in the presence of any of the following substrates (at 10 mM): malate, succinate, fructose, glutamate, proline, or glycerol. No difference was observed in the respiration rates and doubling times of the mutant and the wild type grown on any of these substrates. On swarm plates, chemotaxis of FAJ581 to organic acids, carbo-
TABLE 3. Effects of nonmetabolizable analogs, an endogenous substrate, and metabolic and respiratory inhibitors on chemotaxis of A. brasilense Chemical
Malate Fructose Galactose Aspartate Glutamate Proline
Chemotaxisa in the presence of: Buffer PHB
⫹ ⫹ ⫹ ⫹ ⫹ ⫹
⫺ ⫺ ⫺ ⫺ ⫺ ⫺
b
Sorbose Maleate NaF Myxothiazol HQNO
⫹ ⫺ NDc ND ⫹ ND
⫺ ⫹ ND ND ⫹ ND
⫹ ⫺ ND ND ⫹ ND
⫺ ⫺ ⫺ ⫺ ⫺ ⫺
⫺ ⫺ ⫺ ⫺ ⫺ ⫺
a Assessed in both temporal and miniplugs assays. ⫹, presence of the response in both assays; ⫺, absence of the response in both assays. b Cells were grown to the late exponential growth stage (OD600 ⫽ 0.8) under high aeration in order to achieve high levels of the intracellular PHB (35, 36). c ND, not determined.
ENERGY TAXIS IN A. BRASILENSE
VOL. 182, 2000
FIG. 1. Repellent response of A. brasilense Sp7 to myxothiazol in a temporal gradient assay. (A) Examples of paths of individual cells traced for 5 s within 20 s upon addition of a buffer or myxothiazol (2.5 M). (B) Reversal frequency of free-swimming cells upon addition of myxothiazol followed by addition of an artificial electron donor (ascorbate ⫹ TMPD). The dashed line indicates a prestimulus level of reversal frequency. Reversal frequency was determined by computerized motion analysis.
hydrates, and amino acids was similar to that of the wild type. However, in the more sensitive temporal assay, the cytN mutant showed a significant reduction in the response time to all chemoeffectors tested compared to the wild type (Fig. 2B). The cytochrome cbb3 oxidase supports microaerobic growth of A. brasilense (25). In a chamber ventilated with 21% oxygen, no difference was observed in the response time of the wild-type and cytN mutant cells. The response to 10 mM fructose was 32 ⫾ 8 s in the wild type and 24 ⫾ 11 s in the mutant. The response to 10 mM malate was 28 ⫾ 4 s in the wild type and 24 ⫾ 4 s in the mutant.
TABLE 4. Effects of exogenous substituted quinones on the behavior of A. brasilense wild type (Sp7) and its cytN mutant (FAJ851) Threshold for a repellent response (M)
Quinonea
Reduction potential Chemical(mV)b in-plug assayc Sp7
1,4-BQ 2,6-Dimethyl-1,4-BQ Tetramethyl-1,4-BQ 2-Hydroxy-1,4-NQ a
⫹99 ⫺80 ⫺240 ⫺415
102 103 104 NR
c
Temporal assay
FAJ851 Sp7 FAJ851 Sp7 FAJ851
104 105 NRd NR
BQ, benzoquinone; NQ, naphthoquinone. Data from Wardman (46). Performed as described by Tso and Adler (43). d NR, no response. b
Miniplug assay
0.1 1 5 10
10 102 103 NR
0.1 1 2.5 10
10 102 103 NR
6045
FIG. 2. The cytN mutant is affected in both aerotaxis and chemotaxis. (A) Aerotactic bands formed by the wild type (Sp7) and the cytN mutant (FAJ851) in a glass capillary. Cells were suspended in chemotaxis buffer supplemented with 10 mM malate. Magnification, ⫻200. (B) Chemotaxis of wild type (Sp7) and the cytN mutant (FAJ851) in a temporal gradient assay. The cells were challenged with 10 mM attractant, and the mean smooth-swimming response times were determined.
Electron acceptor and redox taxis. Under anaerobic conditions, A. brasilense Sp7 and the cytN mutant FAJ851 both showed taxis to alternative electron acceptors, nitrate and DMSO, with malate as an electron donor but did not show any response to nitrite. In temporal assay, there was no difference in the response time of the wild type and mutant to 1 mM nitrate (48 ⫾ 8 and 46 ⫾ 5 s, respectively) or 1 mM DMSO (36 ⫾ 10 and 40 ⫾ 7 s, respectively). This was expected, since electron transport to alternative electron acceptors does not require the cytochrome cbb3 oxidase. Substituted quinones were repellents for both the wild type and the cytN mutant. In chemical-in-plug, miniplug, and temporal assays, there was a direct correlation between the repellent effect and the reduction potential of the quinone (Table 4). The repellent effect of substituted quinones was significantly less pronounced in the cytN mutant than in the wild type (Table 4). DISCUSSION Prior to this investigation, energy taxis was observed in several microbial species but did not appear to be the primary determinant of behavior in any of them (40). The results obtained in the present study provide experimental evidence that energy taxis is a dominant behavior in the ␣-proteobacterium A. brasilense; i.e., changes in the electron transport system trigger most tactic responses in this microorganism that are relevant to its ecology.
6046
ALEXANDRE ET AL.
Chemotaxis in A. brasilense is metabolism dependent. In contrast to the best-studied systems, such as E. coli, Salmonella enterica serovar Typhimurium and Bacillus subtilis, chemotaxis in A. brasilense is dependent on the metabolism of attractants. The following lines of experimental evidence support this notion. (i) Only metabolizable substrates are attractants for A. brasilense. Nonmetabolizable analogues of strong attractants do not attract the bacteria. Moreover, they inhibit chemotaxis to (and only to) their metabolizable analogs. Similarly, inhibitors of the metabolism of a substrate specifically inhibit chemotaxis toward this substrate. The presence of an intracellular metabolizable substrate, PHB, which competes with metabolizable attractants for the electron transport system (35, 36), prevents chemotaxis to any extracellular attractant. (ii) There is a direct correlation between the efficiency of a chemical as a growth substrate and as a chemoeffector: the most efficient growth substrates are the stronger attractants for A. brasilense. Such a correlation is absent in E. coli. (iii) There is a direct correlation between temporal changes in cells behavior and temporal changes in intracellular energy levels. The better growth substrates caused a larger increase in respiration and membrane potential and elicited a longer smooth swimming response. Figure 3 summarizes the interrelationship between behavior and metabolism in A. brasilense. Major attractants (and best substrates) are fructose and the TCA cycle intermediates that donate reducing equivalents directly to the electron transport system. This fact indicates that the signal for metabolism-dependent chemotaxis in A. brasilense can be originated within the electron transport system; i.e., behavioral responses to major attractants may be the part of energy taxis. Chemotaxis requires a functional electron transport system. Several lines of experimental evidence led us to conclude that in metabolism-dependent chemotaxis, azospirilla respond to changes in the electron transport system rather than to accumulation of some intracellular metabolites. First, inhibition of the electron transport causes a repellent response. Only chemicals that interact directly with the electron transport system were found to be repellents for A. brasilense; inhibitors of early catabolic events were not repellents although they prevented chemotaxis to oxidizable substrates (Fig. 3). In contrast, repellents in E. coli are not always harmful chemicals and do not specifically interact with the electron transport system (43). Competitive inhibitors of the electron transport, such as substituted quinones, and noncompetitive inhibitors, such as myxothiazol, both caused a repellent response in A. brasilense. In the presence of myxothiazol, supplying reducing equivalents beyond the inhibition site (ascorbate plus TMPD [Fig. 3]) caused the smooth-swimming response. When myxothiazol was not present and the electron transport system was saturated with the reducing equivalents coming from malate or fructose oxidation, no response to ascorbate plus TMPD was observed. This points toward the electron transport system as the origin of the tactic signal. The same conclusion can be drawn from the observed correlation between the magnitude of the chemotactic response and electron transport-related parameters (oxygen uptake and membrane potential). If attractants are sensed independently of the electron transport system (metabolic intermediates, direct interaction between a ligand and a receptor, etc.), then a mutation in the electron transport system should not affect chemotaxis. This is not the case for A. brasilense behavior: a mutation in a particular branch of the electron transport system (cytochrome c oxidase of the bb3 type) significantly diminished chemotaxis to all major attractants. This was observed under microaerobic conditions that are typical of the temporal gradient assay (49). The cytochrome
J. BACTERIOL.
FIG. 3. Scheme showing energy metabolism in A. brasilense and the sites of action for attractants and repellents. Attractants are shown in bold; action of metabolic and respiratory inhibitors is shown by dashed arrows.
cbb3 branch is expected to be active under these conditions (25). Under fully aerobic conditions (active ventilation with 21% oxygen), where the cbb3 oxidase is not functional, chemotactic responses in the mutant and the wild type were identical. Similarly, the cytN mutant was significantly less sensitive to the repellent effect of substituted quinones. This was expected since we have previously demonstrated that the loss of a cytochrome c containing branch allows azospirilla to resist the respiratory inhibitory effect of substituted quinones (2). This set of experiments provided evidence that the signal for chemotaxis toward major attractants and repellents is originated within the functional electron transport system. The electron transport system is required for other types of behavior. A. brasilense have other behavioral responses that are part of the overall energy taxis: aerotaxis, electron acceptor taxis, and redox taxis. The bacteria are attracted to a preferred oxygen concentration that supports a maximal cellular energy level (49). In a spatial oxygen gradient, cells accumulate at the
VOL. 182, 2000
preferred oxygen concentration, which is determined by the oxygen affinity of major terminal oxidases (49). In the cytN mutant, the electron transport pathway and the overall oxygen affinity of terminal oxidases are modified due to the lack of the cbb3 terminal oxidase (Fig. 3). That is why the preferred oxygen concentration of the cytN mutant is different from the wild type (Fig. 2A). In a spatial redox gradient, both A. brasilense and E. coli form a sharply defined band at their preferred redox potential (7, 18). This behavior is known as redox taxis. In the present study, we showed that permeable redox molecules, such as quinone analogues, are repellents for A. brasilense as they are for E. coli. As in E. coli (7), the magnitude of the behavioral response in A. brasilense was dependent on the reduction potential of the quinone. In E. coli, quinones interrupt electron flow by diverting electrons from the electron transport system. This elicits a behavioral response. The same mechanism is likely to take place in A. brasilense. Taxis to alternative electron acceptors was previously demonstrated in E. coli (39) and R. sphaeroides (12) and is known to require the functional electron transport system. In this work, we have found an expected tactic response of A. brasilense to nitrate and DMSO. Both compounds are used by azospirilla as alternative electron acceptors. The response occurs only under strictly anaerobic conditions. Nitrite, which cannot be utilized by the A. brasilense Sp7 strain as an alternative electron acceptor (28), did not elicit a behavioral response. Energy taxis as a dominant behavior. Altogether our results conclusively establish that the functional electron transport system is required to mediate all known behavioral responses in A. brasilense, including chemotaxis to most attractants and all repellents. The present study does not address the question of which of the parameters, redox state of components of the electron transport system or an ion motive force, is the signal for chemotaxis in A. brasilense. Chemotaxis signal transducers most likely can respond to changes in both parameters (41). Several transducer-like proteins are present in A. brasilense Sp7 (S. E. Greer, G. Alexandre, and I. B. Zhulin, unpublished results) as well as all major chemotaxis proteins, including the CheB methylesterase and the CheR methyltransferase (D. Hauwerts, S. K. Das, G. Alexandre, J. Vanderleyden, and I. B. Zhulin, unpublished results). Therefore, the signal transduction mechanism in azospirilla is expected to be similar to that in other chemotactic bacteria. It is possible that in A. brasilense some chemoattractants signal through metabolismindependent ligand-receptor interactions in addition to or instead of the mechanism of energy taxis. In the case of galactose taxis, signaling may involve SbpA (44), which is related to periplasmic sugar-binding proteins ChvE from Agrobacterium tumefaciens and MglB from E. coli. However, based on the results of this investigation, we expect at least some transducers in A. brasilense to be sensors of energy-related parameters. So far, only three types of energy taxis-related chemoreceptors have been identified. The aerotaxis transducer Aer has a flavin adenine dinucleotide cofactor associated with a PAS domain that senses redox changes in the electron transport system in E. coli (8, 31). An Aer homolog was also found to mediate aerotaxis in Pseudomonas putida (30). Another E. coli transducer, Tsr, also senses energy changes and has been suggested as a putative proton motive force sensor (31). A myoglobin-like, heme-based signal transducer mediates aerotaxis in the archaeon Halobacterium salinarum and in the gram-positive bacterium B. subtilis (21). Aromatic compounds, such as benzoic acid, hydroxybenzoic acid, and catechol, were previously described by Lopez-deVictoria and Lovell (24) as strong attractants for A. brasilense
ENERGY TAXIS IN A. BRASILENSE
6047
Sp7, with a threshold for a chemotactic response being in a picomolar range when tested in a capillary assay. However, our results contradict this finding. We found that benzoic acid, hydroxybenzoic acid, and catechol do not support growth of A. brasilense Sp7 and do not elicit a chemotactic response. The discrepancy may come from the different methods used to assess chemotaxis. The major disadvantage of using a capillary assay for azospirilla is that the aerotactic response, triggered by the use of the endogenous substrate PHB, often masks a chemotactic response and leads to ambiguous results. As previously suggested by Barak et al. (6), chemotaxis in A. brasilense should be examined by methods in which the aerotactic response is avoided. This excludes the capillary and other related methods and renders the temporal assay a necessity. We have used three independent methods, including the temporal gradient assay, and found no evidence that aromatics are attractants. Chemotaxis in Azospirillum spp. was previously described as strain specific (32). These results can be explained by the finding that behavior of azospirilla is metabolism dependent. Various species and strains of Azospirillum respond to different spectra of attractants that are correlated with chemicals typical of the rhizosphere of the host plant, the best attractants being also the best growth substrates. Therefore, A. brasilense behavior is likely to reflect the environment in which a cell finds itself (soil and plant rhizosphere) and its general physiological traits. In contrast to E. coli, bacteria of the genus Azospirillum are diazotrophic and therefore not dependent on exogenous sources of nitrogen (37). A. brasilense appear to be solely dependent on energy taxis and preferentially seek organic acids and sugars as carbon and energy sources. Such a behavior would allow the cells to reach and maintain themselves in an environment that supports optimal energy levels, the plant rhizosphere. ACKNOWLEDGMENTS We thank J. Vanderleyden for the gift of the FAJ851 strain and B. L. Taylor for access to his laboratory facilities. This study was supported in part by grant 96-35305-3795 from the U.S. Department of Agriculture and by funds from the National Medical Technology Testbed (to I.B.Z.). REFERENCES 1. Adler, J. 1969. Chemoreceptors in bacteria. Science 166:1588–1597. 2. Alexandre, G., R. Bally, B. L. Taylor, and I. B. Zhulin. 1999. Loss of cytochrome c oxidase and acquisition of resistance to extracellular quinones in laccase-positive Azospirillum lipoferum. J. Bacteriol. 181:6730–6738. 3. Armitage, J. P. 1997. Behavioural responses of bacteria to light and oxygen. Arch. Microbiol. 168:249–261. 4. Armitage, J. P., and R. Schmitt. 1997. Bacterial chemotaxis: Rhodobacter sphaeroides and Sinorhizobium meliloti—variations on a theme? Microbiology 143:3671–3682. 5. Barak, R., I. Nur, and Y. Okon. 1983. Detection of chemotaxis in Azospirillum brasilense. J. Appl. Bacteriol. 53:399–403. 6. Barak, R., I. Nur, Y. Okon, and Y. Henis. 1982. Aerotactic response of Azospirillum brasilense. J. Bacteriol. 152:643–649. 7. Bespalov, V. A., I. B. Zhulin, and B. L. Taylor. 1996. Behavioral responses of Escherichia coli to changes in redox potential. Proc. Natl. Acad. Sci. USA 93: 10084–10089. 8. Bibikov, S. I., R. Biran, K. E. Rudd, and S. J. Parkinson. 1997. A signal transducer for aerotaxis in Escherichia coli. J. Bacteriol. 179:4075–4079. 9. Clancy, M., K. A. Madill, and J. M. Wood. 1981. Genetic and biochemical requirements for chemotaxis to L-proline in Escherichia coli. J. Bacteriol. 146:902–906. 10. Clayton, R. K. 1958. On the interplay of environmental factors affecting taxis and motility in Rhodospirillum rubrum. Arch. Microbiol. 29:189–212. 11. Fan, T. W., A. N. Lane, J. Pedler, D. Crowley, and R. M. Higashi. 1997. Comprehensive analysis of organic ligands in whole root exudates using nuclear magnetic resonance and gas chromatography-mass spectrometry. Anal. Biochem. 251:57–68. 12. Gauden, D. E., and J. P. Armitage. 1995. Electron-transport dependent taxis
6048
ALEXANDRE ET AL.
in Rhodobacter sphaeroides. J. Bacteriol. 177:5853–5859. 13. Glagolev, A. N. 1980. Reception of the energy level in bacterial taxis. J. Theor. Biol. 82:171–185. 14. Goebel, E. M., and N. R. Krieg. 1984. Fructose catabolism in Azospirillum brasilense and Azospirillum lipoferum. J. Bacteriol. 159:86–92. 15. Go ¨tz, R., N. Limmer, K. Ober, and R. Schmitt. 1982. Motility and chemotaxis in two strains of Rhizobium with complex flagella. J. Gen. Microbiol. 128: 789–798. 16. Goulbourne, E. A., Jr., and E. P. Greenberg. 1981. Chemotaxis of Spirochaeta aurantia: involvement of membrane potential in chemosensory signal transduction. J. Bacteriol. 148:837–844. 17. Goulbourne, E. A., Jr., and E. P. Greenberg. 1983. A voltage clamp inhibits chemotaxis of Spirochaeta aurantia. J. Bacteriol. 153:916–920. 18. Grishanin, R. N., I. I. Chalmina, and I. B. Zhulin. 1991. Behaviour of Azospirillum brasilense in a spatial gradient of oxygen and in a “redox” gradient of artificial electron acceptor. J. Gen. Microbiol. 137:2781–2785. 19. Hartmann, A., H. Fu, and R. H. Burris. 1988. Influence of amino acids on nitrogen fixation ability and growth of Azospirillum spp. Appl. Environ. Microbiol. 54:87–93. 20. Hartmann, A., and W. Zimmer. 1994. Physiology of Azospirillum. p. 15–39. In Y. Okon (ed.), Azospirillum/plant associations. CRC Press, Boca Raton, Fla. 21. Hou, S., R. W. Larsen, D. Boudko, C. W. Riley, E. Karatan, M. Zimmer, G. W. Ordal, and M. Alam. 2000. Myoglobin-like aerotaxis transducers in Archaea and Bacteria. Nature 403:540–544. 22. Jeziore-Sassoon, Y., P. A. Hamblin, C. A. Bootle-Wilbraham, P. S. Poole, and J. P. Armitage. 1998. Metabolism is required for chemotaxis to sugars in Rhodobacter sphaeroides. Microbiology 144:229–239. 23. Kloss, M., K.-H. Iwannek, I. Fendrik, and E.-G. Niemann. 1984. Organic acids in the root exudates of Kallar grass (Doplachna fusca (Linn.) Beauv.) Environ. Exp. Bot. 24:179–188. 24. Lopez-de-Victoria, G., and C. R. Lovell. 1993. Chemotaxis of Azospirillum species to aromatic compounds. Appl. Environ. Microbiol. 59:2951–2955. 25. Marchal, K., J. Sun, V. Keijers, H. Haaker, and J. Vanderleyden. 1998. A cytochrome cbb3 (cytochrome c) terminal oxidase in Azospirillum brasilense Sp7 supports microaerobic growth. J. Bacteriol. 180:5689–5696. 26. Martinez-Drets, G., M. Del Gallo, C. Burpee, and R. H. Burris. 1984. Catabolism of carbohydrates and organic acids and expression of nitrogenase by azospirilla. J. Bacteriol. 159:80–85. 27. Miller, J. B., and D. E. Koshland, Jr. 1977. Sensory electrophysiology of bacteria: relationship of the membrane potential to motility and chemotaxis in Bacillus subtilis. J. Mol. Biol. 111:183–201. 28. Neyra, C. A., J. Dobereiner, R. Lalande, and R. Knowles. 1977. Denitrification by N2-fixing Spirillum lipoferum. Can. J. Microbiol. 23:300–305. 29. Neyra, C. A., and R. G. Hageman. 1976. Relationships between carbon dioxide, malate and nitrate accumulation and reduction in corn (Zea mays L.) seedlings. Plant Physiol. 58:726–730. 30. Nichols, N. N., and C. S. Harwood. 2000. An aerotaxis transducer gene from Pseudomonas putida. FEMS Microbiol. Lett. 182:177–183. 31. Rebbapragada, A., M. S. Johnson, G. P. Harding, A. J. Zuccarelli, H. M. Fletcher, I. B. Zhulin, and B. L. Taylor. 1997. The Aer protein and the serine chemoreceptor Tsr independently sense intracellular energy levels and transduce oxygen, redox, and energy signals for Escherichia coli behavior. Proc. Natl. Acad. Sci. USA 94:10541–10546. 32. Reinhold, B., T. Hurek, and I. Fendrik. 1985. Strain-specific chemotaxis of
J. BACTERIOL. Azospirillum spp. J. Bacteriol. 162:190–195. 33. Spudich, J. L., and D. E. Koshland, Jr. 1975. Quantitation of the sensory response in bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 72:710–713. 34. Stock, J. B., and M. G. Surette. 1996. Chemotaxis, p. 1103–1129. In F. C. Neidhart, R. Curtiss III, J. I. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella, 2nd ed., vol. 1. ASM Press, Washington, D.C. 35. Tal, S., P. Smirnoff, and Y. Okon. 1990. Purification and characterization of D(⫺)--hydroxybutyrate dehydrogenase from Azospirillum brasilense Cd. J. Gen. Microbiol. 136:645–649. 36. Tal, S., P. Smirnoff, and Y. Okon. 1990. The regulation of poly--hydroxybutyrate metabolism in Azospirillum brasilense during balanced growth and starvation. J. Gen. Microbiol. 136:1191–1196. 37. Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978. A taxonomic study of the Spirillum lipoferum group, with descriptions of a new genus, Azospirillum gen. nov. and two species, Azospirillum lipoferum (Beijerinck) comb. nov. and Azospirillum brasilense sp. nov. Can. J. Microbiol. 24:967–980. 38. Taylor, B. L. 1983. Role of proton motive force in sensory transduction in bacteria. Annu. Rev. Microbiol. 37:551–573. 39. Taylor, B. L., J. B. Miller, H. M. Warrick, and D. E. Koshland, Jr. 1979. Electron acceptor taxis and blue light effect on bacterial chemotaxis. J. Bacteriol. 140:567–573. 40. Taylor, B. L., and I. B. Zhulin. 1998. In search of higher energy: metabolism dependent behaviour in bacteria. Mol. Microbiol. 28:683–690. 41. Taylor, B. L., I. B. Zhulin, and M. S. Johnson. 1999. Aerotaxis and other energy-sensing behavior in bacteria. Annu. Rev. Microbiol. 53:103–128. 42. Trumpower, B. L. 1990. Cytochrome bc1 complexes in microorganisms. Microbiol. Rev. 54:101–129. 43. Tso, W.-W., and J. Adler. 1974. Negative chemotaxis in Escherichia coli. J. Bacteriol. 118:560–576. 44. Van Bastelaere, E., M. Lambrech, H. Vermeiren, A. Van Dommelen, V. Keijers, P. Proost, and J. Vanderleyden. 1999. Characterization of a sugarbinding protein from Azospirillum brasilense mediating chemotaxis to and uptake of sugars. Mol. Microbiol. 32:703–714. 45. Ward, M. J., D. M. Harrison, M. J. Ebner, and J. P. Armitage. 1995. Identification of a methyl-accepting chemotaxis protein in Rhodobacter sphaeroides. Mol. Microbiol. 18:115–121. 46. Wardman, P. 1989. Reduction potentials of one-electron couples involving free radicals in aqueous solution. J. Phys. Chem. Ref. Data 18:1637–1755. 47. Zhulin, I. B., and J. P. Armitage. 1992. The role of taxis in the ecology of Azospirillum. Symbiosis 13:199–206. 48. Zhulin, I. B., and J. P. Armitage. 1993. Motility, chemokinesis, and methylation-independent chemotaxis in Azospirillum brasilense. J. Bacteriol. 175: 952–958. 49. Zhulin, I. B., V. A. Bespalov, M. S. Johnson, and B. L. Taylor. 1996. Oxygen taxis and proton motive force in Azospirillum brasilense. J. Bacteriol. 178: 5199–5204. 50. Zhulin, I. B., E. H. Rowsell, M. S. Johnson, and B. L. Taylor. 1997. Glycerol elicits energy taxis of Escherichia coli and Salmonella typhimurium. J. Bacteriol. 179:3196–3201. 51. Zhulin, I. B., S. E. Tretyakova, and V. V. Ignatov. 1988. Chemotaxis of Azospirillum brasilense towards compounds typical of plant root exudates. Folia Microbiol. 33:277–280.
