the of chemistry vol. 266, no. 36, issue of 25, pp. 24268 ... · the journal of biological...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 266, No. 36, Issue of December 25, pp. 24268-24275,1991 Printed in U.S.A.

Chitin Utilization by Marine Bacteria CHEMOTAXIS TO CHITIN OLIGOSACCHARIDES BY VIBRIO FURNZSSZP

(Received for publication, October 18, 1990)

Bonnie L. BasslerS, Patrick J. Gibbons, Charles Yu, and Saul Roseman From the McCollum-Pratt Institute and the Department of Biology, The Johns Hopkins University, Baltimore, Maryland 21218

The adhesion/deadhesion apparatus of the marine bacterium Vibrio furnissii (Yu, C., Lee, A., Bassler, B. L., and Roseman, S. (1991) J. Biol. Chem. 266,24260- 24267) probably catalyzes the first step in colonizing chitin. Evidence is presented here for a second step, chemotaxis to chitin hydrolysis products. V. furnissii swarm8 toward chitin oligomers (GlcNAc),, n = 1-6, a t initial concentrations as low as 10 MM. A modified capillary assay was used for quantitation; the cells exhibit low level constitutive taxis to GlcNAc but not to the oligosaccharides. A mutant defective in the GlcNAc receptor (IINag of the phosphotransferase system) showed inducible taxis to the oligosaccharides. Two (or more) independently inducible receptors with overlapping specificities recognize (GlcNAc),, n = 2- 4. (GlcNAc)a and (GlcNA& were inactive in the capillary assay; expression of this receptor(s) apparently require special induction conditions. The (GlcNAc),, n = 1-4, chemoreceptors of V. furnissii may be the most potent reported for bacteria.

L-Amino acids were weak, constitutive attractants; glutamine, not known to be an attractant in other bacteria, was the most effective amino acid. The most potent receptor in Escherichia coli, Tar (aspartate), is not expressed in V. furnissii.

The chemotactic responses were greatly affected by growth and induction conditions and the presence of nutrients in the assay media. Taxis to GlcNAc and GlcNAc oligomers was optimally induced by growth in lactate medium containing 0.6 mM sugar, while growth on the sugar per se resulted in poor taxis. Chemotaxis to the sugars increased 2- to %fold when the cells were starved. Nutrients in the assay medium, especially compounds that feed into or are part of the Krebs cycle, were potent inhibitors of taxis to the sugars and Gln. With the exception of isocitrate, inhibition of taxis correlated with the rate of oxidation of these compounds. The results suggest a link between catabolism and taxis in this organism, i.e. interactions or “cross- talk’’ between systems that are regulated by protein phosphorylation (Stock, J. A., Ninfa, A. J., and Stock, A. M. (1989) Microbiol. Rev. 53, 450-490).

In the accompanying papers (1, Z), we note that chitin is

* This work was supported by Contract N0001485-K-0072 from the Office of Naval Research. This paper is Contribution No. 1477 from the McCollum-Pratt Institute of The Johns Hopkins University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Supported by National Institutes of Health Training Grant 5T32 GM07321 and by W. R. Grace and Co. Present address: The Agouron Institute, 505 Coast Blvd. South, La Jolla, CA 92037.

an important component of both the carbon and nitrogen cycles in the sea, and that marine bacteria, such as Vibrios, play critical roles in converting this highly insoluble polysac- charide to a biologically useful form. in Fig. 1 (l), we propose a scheme for chitin degradation by organisms such as Vibrio furnissii, and evidence was presented for one step, adhesion. The adhesion phenomenon is, in fact, more complex than anticipated, since it is part of an adhesion/deadhesion apparatus which appears to act as a signal transduction mechanism for continuously monitoring the nutrient status of the environment. We report here that V. furnissii expresses another signal transduction system for analyzing the environment; that is, it recognizes and migrates toward low concentrations of chitin oligosaccharides.

Most reports on bacterial chemotaxis have utilized Esche- richia coli and Salmonella typhirnurium (3), while only a few papers have been published on marine prokaryotic systems. In one such study, Mitchell et al. (4) found that a motile marine Pseudomonad, which preys upon (and degrades) the fungus Pythium debaryanum, is attracted to cellulose and cellulose oligosaccharides released into the medium by the fungus. In addition, other marine Pseudomonads exhibited chemotaxis to D-glUCOSe, and to D- and L-amino acids. Chet and Mitchell (5) also reported that V. alginolyticus, which preys upon the alga Skeletonema costatum, exhibited chemotaxis toward the algal extracellular products acrylate and glycolate ( 5 ) . Sar et al. (6) found that both polar and lateral flagella were involved in chemotaxis by V. parahaemolyticus. McCarter et al. (7) had already demonstrated that the polar flagellum was responsible for movement of V. parahaemolyticus in a liquid environment while lateral flagella allowed migration over solid surfaces. Mutants in either the polar or lateral flagella recognized the same chemoattractants, which suggested a common regulatory mechanism in both modes of chemotaxis (6). Finally, Alam and Glagolev (8) showed that V. harueyi responded positively to several amino acids and divalent metals, while hydrophobic amino acids, uncouplers, indole, and acetate were repellents. These researchers pro- posed a specific ApH+-receptor, a “protometer” that governs nonspecific repellent taxis, aerotaxis, and phototaxis in V. harueyi.

We report here that V. furnissii expresses chemotactic systems capable of recognizing GlcNAc and GlcNAc oligosaccharides at initial concentrations as low as 10 pM. Low constitutive taxis occurs to the monomer, which increases mark- edly after induction, and induction is required for chemotaxis to the oligosaccharides. Definitive evidence is presented for the existence of at least three distinct chemoreceptors, one for GlcNAc, and two that recognize a limited but overlapping group of small oligosaccharides ((GlcNAc)e to (GlcNAch). NO chemotaxis occurs toward GlcNH2 and its oligomers (derived from chitosan). Additionally, the cells show weak taxis to most of the amino acids, but, surprisingly, none toward as-

24268

Chemotaxis to Chitin Oligosaccharides

partate. Thus, V. furnissii may lack the well characterized tar (aspartate) receptor found in E. coli (3). PTS sugars,’ includ- ing GlcNAc, are chemoattractants for E. coli and S. typhimurium (9, lo), and V. furnissii taxis to PTS sugars is described in an upcoming report.’ However, some studies

24269

by Hazen et al. (20).3 This assay was adapted for use with V. furnissii, grown in the “lactate-labeling medium” (1) containing 32P,. Midex- ponential phase cells were harvested by centrifugation at 4,000 x g at room temperature in a Sorvall SS34 rotor, washed three times with “taxis buffer” (Hepes-buffered 50% ASW containing 0.1% NH&l and 0.002% K,HPO,), and resuspended to the original volume in a 15-ml Corex tube (Asdo from 0.2 to 0.3). Specific activities were determined for each experiment and, in general, were about 0.04 cpm/ cell. The specific activity did not change during the time of the experiment. Putative chemotactic inducers were included in the labeling medium for the entire growth period. Unless otherwise specified, the final inducer concentration was 0.6 mM.

using GlcNAc are presented here as a necessary part of characterization of the oligosaccharide chemoreceptors.

Evidence is also presented for an interaction between the V. furnissii chemotaxis systems and the Krebs cycle or closely related catabolic events. While it has been suggested that such interactions, designated “cross-talk,” may occur (12-15), to our knowledge this is the first evidence of a regulatory link between chemotaxis and the processes of intermediary metab- olism.

Preliminary communications have been presented (16, 17).

EXPERIMENTAL PROCEDURES

Materials-The following substances* were purchased from Sigma: GlcNAc, (G~cNAc)~,-(G~cNAc)~, the GlcNAc-glycosides PN- GlcNAc, PN-(GlcNAch, PN-a-GlcNAc, PN-@-thio-GlcNAc, phenyl @GlcNAc, MUF-@-GlcNAc, methyl a- and p-GlcNAc, and p-aminophenyl fl-thio-GlcNAc, L-amino acids, and dipeptides. The chitin oliaosaccharides (GlcNAch. (GlcNAch. and (GlcNAch were obtained from Seikagaku America lnc: (St. Petersburg, FL). Supplies used for preparation of bacterial media were of the highest purity available and came from Difco Laboratories. High purity chemical reagents were purchased from J. T. Baker. Hepes was obtained from Research Organics Inc. (Cleveland, OH), and radioisotopes from Du Pont-New England Nuclear. Chitosan oligosaccharides were prepared as described (18).

Growth o/Cells-The isolation, characterization, maintenance, and growth of V. jurnissii and of the IIN” mutant, AP801, are described in the accompanying report (2). E. co& AW405, a kind gift of Dr. J. Adler (Universitv of Wisconsin. Madison WI), was stored at room temperature in medium containing 1% tryptone, 0.5% NaCl, and 1.5% agar, and was grown overnightin 1% t&tone, 0.5% NaCl broth at 37 “C with rotarv shaking (200 rpm). Bacteria from the overnight culture were inoculated at a 1:26 dilution into minimal medium containing (in g/liter): glycerol, 5 (autoclaved separately); K,HPO,, 4.8; (NH&SO,, 2; MgSO,.7 H20, 0.25; Fe2(S0&, 0.0005; and casa- mino acids, 0.0025. The culture was grown overnight in minimal medium and inoculated again into minimal medium containing 10 &i/ml i4C-amino acids or [‘Clglucose and grown to midexponential phase for each experiment.

Assays for Chemotaris of V. furntisii-Two assays were used to study chemotaxis. The qualitative, swarm plate method has been described (16). Briefly, the method consists of inoculating bacteria in the center of a soft agar plate (0.3%) containing a putative chemoattractant. Oligosaccharides were used at concentrations ranging from 10 pM to 5 mM. The plates were allowed to stand for 24-48 h at room temperature, and swarm rings were assessed visually.

The quantitative capillary assay method of Adler (19) was modified

i All sugars are of the D-configuration and glycosides are pyrano- sides unless otherwise specified. The abbreviations used are: ASW, artificial sea water; Hepes, 4-(2-hydroxyethyl)-l-piperazineethane- sulfonic acid; GlcNAc or Nag, N-aCet~l-D-dUCOSamine: the 8.1+4- linked oligosaccharides derived from chitin are designated (GlcNAc)?, N,N’-diacetvlchitobiose: (GlcNAch. N.N’.N”-triacetvlchitotriose: (G~cNAc)~, N,N’,N”,g “-tetraacetylchitotetraose; (GlcNAch, N; N’ N” N”’ N”” N”‘N:“’ ,+,

-pentaacetylchitopentaose; (GlcNAc),, N,N’,N”, -hexaacetylchitohexaose; methyl a-GlcNAc, methyl

a-N:acetylglucosaminopyranoside, methyl @-GlcNAc, methyl P-N- acetyl-glucosaminopyranoside; MUF, 4-methylumhelliferone; MUF- fl-GlcNAc, 4-methylumbelliferyl N-acetylglucosaminopyranoside; phenyl &GlcNAc, phenyl @-N-acetylglucosaminopyranoside; PN- GlcNAc, p-nitrophenyl @-N-acetylglucosaminopyranoside; PN- (G~cNAc)~, p-nitrophenyl p-n-N,N’-diacetylchitobioside; PTS, phos- phoenolpyruvate:glycose phosphotransferase system. The PN abbre- viation designating p-nitrophenyl is also used with a-GlcNAc and with /3-thio-GlcNAc to designate the corresponding glycosides of N- acetylglucosamine.

* C. Yu, B. L. Bassler, J. A. Stock, and S. Roseman, submitted for publication.

Aliquots of the washed cell suspension (150 ~1) were placed in test tubes (6 x 50 mm). A 5-111 capillary (sealed at the top), containing the specified chemoattractant in taxis buffer, was placed vertically in the tube so that the open end of the capillary was immersed to about half the depth of the cell suspension. The capillary was held in position by an adhesive (Fun-Tak, Beecham Products, Dayton, OH) at the mouth of the test tube. After incubation at 28 “C for 1 h, each capillary was removed, the outside was washed with deionized water and wiped dry, and the contents were expelled directly into vials containing 4 ml of Hvdrofluor liauid scintillation fluid (National Diagnosti& Manville, -NJ). Radioactivity was quantitated in a Pack- ard liquid scintillation spectrometer, giving the number of cells in each capillary. More than 98% of the radioactivity contained in the capillaries was transferred to the scintillation vials by this method, as determined by measuring the radioactivity remaining in the capillary tubes after they were crushed. All experiments were conducted in triplicate, i.e. using three separate capillaries per data point.

Competition assays were performed using GlcNAc oligosaccharides or glycoside analogues in pairwise combinations to determine the number of independent oligosaccharide chemoreceptors expressed by V. furnissii. In these assays, one of the attractants was present only in the capillary tube, while the putative competitor was included both in the capillary tube and in the bacterial suspension at a concentration sufficient to saturate the chemoreceptor (1 or 10 mM). Theoretically, if a single chemoreceptor recognizes both compounds, chemotaxis is blocked by the saturating level of the compound in the cell suspension. However, if independent receptors exist, chemotaxis to the first compound occurs in the presence of the competing compound. Inhi- bition of chemotaxis was performed exactly as described for the competition assay except that the compounds added to the bacterial suspension and the capillaries (in general, intermediary metabolites) were not themselves chemoattractants or repellents.

E. coli A W405 Chemotaxis, Capillary Assay-E. coli AW405 was grown overnight as described above and inoculated (120) into minimal medium containing 10 &i/ml [i4C]glucose or “C-amino acids and grown to midexponential phase as determined by absorbance at 590 nm. The cells w&e centrifuged at 5,000 x g at rodm temperature, washed three times with E. coli CHE buffer (19) (10 mM notassium phosphate, pH 7.0, 10 PM EDTA) and resuspended to an*Aasa from 0.05 to 0.10. The capillary assay was performed exactly as described for V. furnissii, except that the cells and attractants were in CHE buffer. The specific activity of the cells remained constant throughout the experiment.

Oxygen Consumption-V. furnissii was grown overnight in peptone-yeast extract medium, inoculated at a I:25 dilution into lactate, 50% ASW medium, and grown (25 “C, 200 rpm, 500 ml in a 2-liter flask) to midexponential phase. The cells were harvested by centrifugation (7000 x g, 10 min), washed three times with taxis buffer, and resuspended. Oxygen consumption was monitored at room temperature using a Yellow Springs Instruments Model 4004 Clark-type oxygen electrode. The cell suspension (7 ml containing 2 X 10’ cells/ ml) was poured into the chamber, vigorously mixed, and, after the baseline rate of oxygen consumption was established, the test compound, in taxis buffer (70 ~1 of a 100 mM solution), was injected. Thus, each compound was examined at a final concentration of 0.99 mM. The rate of oxygen consumption was quantitated using oxygen- saturated water as the standard.4

Chemoreceptor Methylation-Analysis of methylation of chemoreceptors was essentially by the method of Stock et al. (21), after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (22). Results are presented in detail in an upcoming report.*

3 The method described here is based on a procedure developed by Dr. N. Williams, to whom the authors are most grateful.

’ We are grateful to Todd Pihl and Dr. Robert Maier for their help in the oxygen consumption studies.

24270 Chemotaxis to Chitin Oligosaccharides

RESULTS

Swarm Plates-Our results with swarm plates have been published (16) and can be summarized as follows. V. furnissii cells were grown on lactate, but do not exhibit chemotaxis toward this compound. In contrast, GlcNAc and the chitin oligosaccharides proved to be chemoattractants for V. furnissii at concentrations as low as 10 p ~ . Swarm plates were also used to test V. furnissii chemotaxis to other compounds that support growth of the cells, but no swarming was detected toward glycerol and Krebs cycle intermediates (pyruvate, succinate, citrate, fumarate, a-ketoglutarate, malate, and isocitrate) at concentrations ranging from 1 to 5 mM.

Test of Modified Capillary Assay Method-E. coli AW405, obtained from Dr. J. Adler, is considered a wild type chemotactic strain. This organism was therefore employed to determine the validity of the modified capillary assay. Recent results (3) obtained with E. coli AW405 by the standard capillary assay method are summarized in Table I and are compared with those obtained in the present experiments using the modified assay. The data obtained with the modified and with the standard Adler capillary assays are comparable. We therefore conclude that the modified assay is a valid and simple method for measuring chemotaxis.

Since V. furnissii is uncharacterized with respect to its chemotactic behavior, it appeared necessary to repeat the earlier work of Adler (19) to determine optimum conditions for quantitation of chemotaxis in the capillary assay. The parameters investigated included time of incubation, cell density, pH, ionic strength, and the presence of heavy metals, NH3, or phosphate (17). All experiments were performed at 28 "C. As a result of our studies, taxis buffer (Hepes-buffered 50% ASW containing NH&1 and K2HP04, pH 7.4) was used as the medium for experiments with V. furnissii in the capillary assay. The assays were performed for 1 h at densities between 2 X lo7 and 2 X 10' cells per ml; these conditions were observed to be within the linear response range for time and cell density. Similar to E. coli, chemotaxis was partially inhibited by divalent cations. However, we decided that the behavior of the cells in the more physiologically relevant environment (50% ASW) was of greater importance than the increased signal observed in buffer containing no divalent cations. It should be emphasized that only the magnitude of the response (about 40% decrease) not the concentration dependence (shape of the curve) was affected by the selected conditions.

Chemotaxis to GLcNAc and GlcNAc Oligosaccharides- V. furnissii exhibited low level, constitutive taxis to GlcNAc, and induction by growth of the cells in the presence of this sugar greatly increased the effect. In the case of the oligosaccharides, no taxis was observed using cells grown on lactate alone,

TABLE I Comparison of chemotaxis of E. coli A W405 in the standard and in

the modified capillary assays The magnitudes of the peak responses of E. coli AW405 to aspar-

tate, serine, and GlcNAc reported for the standard capillary assay (3) are compared with the values obtained in the modified assay. The cells were grown, harvested, and washed as reported (3, 9). The modified assays were performed for 1 h at 30 "C.

Peak values (number of cells) Attractant Standard Modified

technique technique X IO-?

Aspartate 310 280 Serine 310 220 GlcNAc 60 60

whereas induction elicited large responses to the oligomers. The optimal induction conditions were growth of the cells in the lactate, 50% ASW medium containing 0.6 mM concentrations of the inducers (data not shown). When the capillaries contained increasing concentrations of the potential attractant, such as (G~cNAc)~, a typical "bell-shaped" curve was observed (Fig. I).

The concentration-dependent response of V. furnissii was investigated toward each of the chitin oligomers, (GlcNAc),, n = 1-6. Ideally, when chemotaxis is measured over a suffi- ciently broad range of concentrations, a bell-shaped curve is obtained (giving both the threshold and peak responses) by plotting the number of cells in the capillaries against the log of concentration of chemoattractant in the capillary. Such results were obtained for chemotaxis to GlcNAc, (GlcNAc)z, and (G~CNAC)~ over a 20,000-fold concentration range (10 p~ to 200 mM). In each case, the threshold and peak responses were 0.1 mM and 10 mM, respectively. Because of their insol- ubility, it was not possible to test (G~CNAC)~, (GICNAC)~, and (GlcNAc)G at the highest concentrations (100 and 200 mM). However, as shown above, the peak response to (G~CNAC)~ was 1.0 mM and the threshold at 0.1 mM. No taxis to the penta- and hexasaccharide was observed under the above conditions, and these oligosaccharides are considered separately below.

Evidence for the existence of at least three independently inducible chemoreceptors responsible for recognition of GlcNAc to (G~CNAC)~ was obtained as follows. V. furnissii was grown overnight in peptone-yeast extract and inoculated into lactate medium without inducer or with (GlcNAc),, n = 1-6, at 0.6 mM concentration each. In every case, the culture was assayed for chemotaxis to the entire series ( n = 1-6) of compounds, using concentrations ranging from 0.01 to 10 mM. No taxis to (GlcNAc)5 and (GlcNAch was observed in these experiments.

The quantitative data are presented in Table I1 and can be explained in terms of three independently induced oligosaccharide chemoreceptors. The receptor for the monosaccharide is the Enzyme IINag, analogous to that found in E. coli. Additionally, two other receptors are expressed, one which recognizes small oligomers and is induced by the disaccharide, and a second which recognizes oligosaccharides of larger size (induced by the tetrasaccharide).

200 7

unmduced z

0 1 2 3 4 Log (GIcNAc), (pM)

FIG. 1. Effect of induction and concentration of (G1cNAc)s on chemotaxis. V. furnissii was grown in lactate, 50% ASW medium (uninduced) (A), or lactate, 50% ASW medium plus 0.6 mM (GICNAC)~ (induced) (0). Chemotaxis experiments were performed with the capillary method for 1 h at 28 "C and at a density of 1.5 X loa cells/ml.

Chemotaxis to Chitin Oligosaccharides 24271

TABLE I1 Chemotaxis of V. jurnissii to chitin oligosaccharides

The peak chemotactic response of V. fumissii to each of the chitin oligosaccharides ((GlcNAc),, n = 1-6) obtained in the modified capillary assay are presented in the table. V. jurnissii was grown in lactate-labeling medium in the absence of inducer or in the presence of a 0.6 mM concentration of each of the chitin oligosaccharides (GlcNAc),, n = 1-6. In each case, the cultures were grown to midexponential phase, harvested, and washed as described (see “Experi- mental Procedures”), and chemotaxis to the entire series of compounds (GlcNAc),, n = 1-6 was assayed. All experiments were performed for 1 h at 28 “ C . All values are corrected for the controls, cadlaries lacking attractants.

Attractant“ Inducer (GlcNAc),, n = 1-6

None 1 2 3 4 5 6

cel[s /capiky (X IO-a) GlcNAc 58 378 224 47 112 35 53

(G~cNAc)~ 4 4 246 65” 82 0 30 (GIcNAc)~ 0 0 146 148” 152” 5” 4” (GlcNAc), 0 16 31 50 112 3 5 (GIcNAc)~ 6 0 0 10 8 7 8 (G1cNAc)a 4 23 0 8 0 4 2 ’ Each compound was assayed over the concentration range shown

in Fig. 1. The number of cells in the capillaries at the peaks of the curves are presented. Peak responses were obtained at 10 mM concentrations excepting those marked with an “a,” where the maximal responses were at 1 mM concentrations.

At t roc ton t

FK. 2. Chemotaxis of V. furnissii mutant APSO1 following (G1cNAc)a induction. V. furnissii AP801 was grown to midexponential phase in lactate-labeling medium containing 0.6 mM (GlcNAc),. The cells were harvested, and chemotaxis to Gln, GlcNAc, (GlcNAc),, and (GlcNAc), was assayed at 10 mM concentrations by the capillary method.

The accompanying paper (2) describes periplasmic hydrolases that cleave the oligosaccharides to lower oligomers and GlcNAc. Thus, it was conceivable that the chemotaxis results were obtained via a single receptor, lINag, which responded to the “true” attractant, GlcNAc, formed in the periplasmic space. This idea was tested with a V. furnissii mutant, AP801, which is defective in IINRg but which expresses the oligosaccharide hydrolases (2). The mutant was “induced” with GlcNAc, but did not respond to the monosaccharide in the capillary assay at concentrations as high as 200 mM. However, growth in the presence of an inducing concentration of (G1cNAc)a resulted in a chemotactic response to the di- and trisaccharide (similar to wild type V. furnissii), but unlike wild type V. furnissii, the induced cells did not respond to the monosaccharide (Fig. 2). Thus, the data in Table I1 cannot be explained in terms of a single (IINag) chemoreceptor.

Chemotaxis toward (G1cNAc)s and (G1cNAc)s was not observed in the capillary assay. However, V. furnissii swarmed avidly to these compounds in the swarm plate assay. We believe growth on these higher oligosaccharides induces enzymes and/or porins and possibly additional chemoreceptor(s) which are not induced by (GlcNAc),, n = 1-4 (see the accompanying report (2)). To illustrate this, Fig. 3 shows growth curves for V. furnissii on the penta- and hexasaccharide as

Time of Growth (h)

FIG. 3. Effect of inoculum on growth of V. furnissii on (G1cNAc)a and (G1cNAc)e. V. furnissii was grown at 25 “C with shaking (200 rpm) in buffered 50% ASW medium with 2.5 mM (GICNAC)~ (A) or (GlcNAc)e ( B ) as the sole source of carbon. The media were inoculated with V. furnissii previously grown in lactate, 50% ASW liquid medium containing 0.6 mM (G1cNAc)z (0) or with V. jurnissii harvested from the edges of swarms on plates containing 2 mM (GlcNAcIs (A) or (G~CNAC)~ ( B ) (A).

the sole sources of carbon. The cells were inoculated into these media after prior growth either in liquid lactate, 50% ASW medium containing 0.6 mM (G1cNAc)z or growth on swarm plates containing only (G1cNAc)a or (GlcNAc)s as the carbon source. In both cases (penta- and hexasaccharide), the cells grew more rapidly when the inocula were taken from the respective swarm plates; these cultures also grew to 2-3.5 higher cell densities compared to cultures inoculated with (GlcNAc)s-induced cells. Our routine growth conditions (lactate, 50% ASW medium containing 0.6 mM oligosaccharide as inducer) may be insufficient to induce the porins and chemotactic apparatus necessary for a response to the higher oligomers (GlcNAc)s and (GlcNAc)6. Chemotaxis of V. furnissii to the larger chitin oligosaccharides is now under inves- tigation.

The monosaccharide receptor, IINag, does not recognize any higher oligomer, and, likewise, GlcNAc is not recognized by the oligosaccharide receptors. However, the absolute number and specificity of the oligosaccharide chemoreceptors have not yet been resolved. We attempted to determine the specificity of each receptor through competition assays in which combinations of the oligosaccharides were studied in pairs. However, because more than a single oligosaccharide is recognized following induction by each oligomer, potent competition occurred in every case. Thus, the competition assays gave no useful information as to the true number and specificity of the oligosaccharide receptors except that the receptor specificities overlap.

Synthetic glycosides of GlcNAc were tested both as attractants and as competitors ( ix . inhibitors) of chemotaxis to (GlcNAc),. PN-GlcNAc and PN-(GlcNAc)Z were the only analogues that had any activity as chemoattractants, but the cell suspensions turned yellow during these experiments, im- plying that the sugars formed by hydrolysis were the actual attractants. Inactive analogues included: PN-a-GlcNAc, PN- p-thio-GlcNAc, phenyl-6-GlcNAc, MUF-0-GlcNAc, methyl a-, and p-GlcNAc, and p-aminophenyl-@-thio-GlcNAc. How- ever, when the same compounds were tested as inhibitors of chemotaxis using cells similarly induced (0.6 mM (GlcNAc):! in lactate medium), approximately 50% inhibition was observed with one derivative, methyl-@-GlcNAc.

The cells were also tested for taxis toward GlcNH2 and the corresponding di- and trisaccharides after growth in lactate medium containing the amino sugars as inducers. No taxis was observed toward these compounds.

A comparison of the quantitative responses of E. coli and V. furnissii shows that the chitin oligosaccharides (in V. furnissii) are among the most potent chemoattractants ever reported in bacteria. Asp is the strongest attractant for E.


coli, and, in our capillary assay, the peak response of E. coli to Asp is 82% of that obtained with V. furnissii to GlcNAc. Among the sugars, GlcNAc is considered a strong attractant in E. coli, but the peak response of V. furnissii to this sugar is 6-fold greater than that of E. coli. Similarly, the response of V. furnissii to (GlcNAc)2 is about the same magnitude as that of E. coli to Asp. These large signal to noise ratios in V. funissii may make it the organism of choice for studying chemotaxis, particularly with regard to the link between the PTS and the Che systems.

Chemotaxis Following Growth on (GlcNAc),-As noted above, the maximal chemotactic response to the chitin oligosaccharides occurred when the cells were grown in lactate medium containing 0.6 mM inducer. Chemotaxis by cells grown on (GlcNAc), as the sole carbon source was also tested. Two such compounds, GlcNAc and (GlcNAc)4, were chosen for study. As a preliminary experiment for these studies, the rates of growth of V. furnissii on the (GlcNAc), oligosaccharides were first determined. V. furnissii grew with the same generation time, approximately 1.5 h, on (GlcNAc),, n = 1, 2, 3, and 4 as on lactate. Growth on (GlcNAc)s and (GlcNAc)6 as the sole carbon sources required special conditions, as explained above.

V. furnissii was grown to midexponential phase in lactate medium containing 0.6 mM GlcNAc or in medium with 55 mM GlcNAc substituted for the lactate, and capillary assays were conducted with 1 and 10 mM GlcNAc as the attractant (Fig. 4A). No taxis to 1.0 mM and an estimated &fold decrease in taxis to 10 mM GlcNAc occurred when the cells were grown exclusively on GlcNAc as the carbon source. Similar results were obtained using (GlcNAc)4 as the sole carbon source, compared with cells grown on lactate containing 0.6 mM (GlcNAc), (Fig. 4B). These potentially important results are considered under "Discussion."

Chemotaxis to PTS Sugars and Amino Acids". furnissii exhibited positive chemotaxis to other PTS sugars, as well as to GlcNAc. These responses are described in an upcoming report.'

The chemotactic behavior of V. furnissii toward the amino acids is of particular interest. Aspartate and serine are the most potent chemoattractants known for E. coli and S. typhimurium, and the corresponding chemoreceptors, tar (aspartate) and tsr (serine), are among the best characterized receptors in these organisms. The naturally occurring amino acids were analyzed as chemoattractants for V. furnissii by the capillary assay, over the concentration range 0.01 to 10 mM. All of the amino acids except the acidic amino acids, aspartate and glutamate, were chemoattractants, with peak responses

FIG. 4. Effect of growth on GlcNAc and (GlcNAc)r on chemotaxis. A, V. jurnissii was grown either in lactate medium containing 0.6 mM GlcNAc for induction (open bar) or in medium containing GlcNAc at 55 mM as the only carbon source (stippled bar). Panel B, V. jurnissii was grown in lactate medium containing 0.6 mM (GlcNAc)4 (open bar) or in medium with 5 mM (GlcNAc)4 as the sole carbon source (stippled bar). At midexponential phase, the cultures were harvested, washed, and tested for taxis to 1 and 10 mM GlcNAc ( A ) or (GlcNAc)4 ( B ) , respectively, in the modified capillary assay.

300r I n

FIG. 5. Chemotaxis to amino acids. A single culture of V. furnissii was grown in lactate-labeling medium, harvested, washed, and tested for chemotaxis to all 20 amino acids at both 10 mM and 1 mM in the modified capillary assay. In all cases, a larger response was found at 10 mM than at 1 mM; only the 10 mM results are shown.

at 10 mM for 18 amino acids (Fig. 5 ) . The presence of any or all of the amino acids during growth did not change the results. Therefore, we conclude that induction is not necessary for V. furnissii chemotaxis to these compounds. Glutamine was the most potent amino acid as an attractant, and, insofar as we are aware, this has not been reported to be an attractant in other bacteria.

Dipeptides are attractants (3) for E. coli and S. typhimurium, and a number of dipeptides, most of which are potent attractants for E. coli (Gly-Leu, Leu-Gly, Leu-Phe, Met-Phe, Pro-Val, Gly-Ser, Gly-Gly, and Gly-Thr), were therefore tested with V. furnissii, but the responses to the dipeptides were never greater than to either of the single amino acids contained in the dipeptide.

The unexpected negative results with aspartate were con- sistently obtained under a variety of conditions, and thus it appears that V. furnissii lacks the tar receptor found in E. coli and 5'. typhimurium, and, further, that serine is not the most potent chemoattractant in V. furnissii. One of the important features of the membrane receptors such as tsr and tar is that they are reversibly methylated by a methyltrans- ferase and a methylesterase (3,23, 24). The methylation step is critical in the adaptation phase of the chemotactic response in enteric bacteria. Experiments were performed with V. furnissii to determine whether this organism also generates methyl-esterified membrane proteins. Essentially the same results were obtained with V. furnissii as with E. coli (21) and are presented in detail in an upcoming report.'

Effects of Lactate and Glycerol on Chemotaxis-The effects of growth conditions were among the parameters studied to determine optimal conditions for chemotaxis. A number of carbon sources were found either to reduce chemotaxis when they were used for growth of the bacteria or to inhibit chemotaxis when they were included in the chemotaxis assay medium. As shown above (Fig. 4), growth of V. furnissii on the (GlcNAc), compounds reduced the chemotactic response.

Neither lactate nor glycerol is a significant a t t ra~tan t .~ When the cells were grown on 0.5% lactate or glycerol as the primary carbon source with 0.6 mM GlcNAc for induction, taxis to GlcNAc was roughly 3.5-fold greater with the lactate-

V. jurnissii can grow on D- or L-lactate, or a racemic mixture of the two compounds. These substances do not act as chemoattractants on swarm plates, and, in many assays by the capillary method, lactate gave no significant values above the control (capillaries containing only the buffered salts solutions).


grown cells. In these, and the inhibition experiments reported below, motility was determined by phase contrast microscopy. No decrease was detected in glycerol-grown cells, SO it appears that the reduction in chemotaxis is not the result of an inability of the bacteria to swim or to generate flagella.

Not only was the choice of carbon source used for cell growth of critical importance, but, most unexpectedly, the chemotaxis of responsive V. furnissii cells was completely inhibited when lactate was present in the chemotaxis assay medium. Some results are shown in Fig. 6. Addition of 1 rnM lactate to the washed cell suspensions and the capillaries (so that there was no lactate gradient) resulted in complete inhibition of taxis to GlcNAc and to (GlcNAc),. Glycerol was a less effective inhibitor. These surprising results led to a series of experiments on the effects of catabolites as inhibitors of chemotaxis.

Inhibition of Chemotaxis by Krebs Cycle Intermediates and Other Metabolites-An array of Krebs cycle intermediates and compounds which feed into the cycle were tested as inhibitors. Each compound was also tested separately as an attractant or repellant. The inhibition experiments were performed under conditions where the cells showed strong taxis toward three compounds, the inducible GlcNAc (PTS) system, the inducible (GlcNAc), system, which may involve a periplasmic binding protein, and the constitutive glutamine system.

In each experiment, the washed .cell suspensions were di- vided into aliquots and assayed as follows. (a) Control (un- inhibited) chemotactic values were obtained for GlcNAc, (GlcNAc)2, and glutamine in the usual manner. ( 6 ) Each putative inhibitor was assayed at 1 mM concentration in the capillary to determine whether it behaved as a chemoattractant per se. ( c ) Since the inhibitors might act as repellents, the capillary assays were conducted with binary mixtures containing one of the attractants (GlcNAc, (GlcNAc),, or glutamine) and one of the test compounds. The possible repellent activity of the test compounds was determined by comparing these results with those obtained in the controls. (d) The experimental mixtures were prepared as follows. The capillaries (containing one of the attractants) and the cell suspensions were supplemented with 1 mM concentration

A. GlcNAc

- 3001 1

I O 0 . . ..: ., :.. ,. . . . . .. . ..

0 0.1 0.5 I 5 Inhibitor (mM)

FIG. 6. Inhibition of chemotaxis of V. furnissii by lactate and glycerol. V. furnissii was grown in 50% ASW medium containing 55 mM lactate as the carbon source and induced for chemotaxis with 0.6 mM GlcNAc (Panel A) or (GlcNAc), (Panel B ) . The cells were assayed for taxis to the same compound used for induction. In the inhibition experiments, the indicated concentrations of lactate (open bars) or glycerol (stippled burs) were added to both the capillary and the assay medium.

each of lactate, pyruvate, succinate, citrate, fumarate, glycerate, a-ketoglutarate, malate, acetate, propionate, or isocitrate. All stock solutions of the test compounds were in taxis buffer, final pH 7.4. Since the cell suspensions and the capillary tubes contained 1 mM concentration of the putative inhibitors, there was no gradient of the test compounds in the assay system.

None of the compounds listed above acted either as a chemoattractant or as a repellent, but some were potent inhibitors of taxis to the known attractants. A wide range of inhibitory effects was observed. The results are compared in Table 111.

Regardless of the chemoattractant (GlcNAc, (GlcNAc),, or glutamine), lactate, pyruvate, succinate, and fumarate gave close to complete inhibition. Malate, glycerate, acetate, and propionate gave partial inhibition (propionate did not inhibit taxis to Gln). Finally, citrate, a-ketoglutarate, and isocitrate either had no effect or stimulated taxis. Thus, the results indicate that this inhibition is a general phenomenon, i.e. the inhibition is not restricted to a single attractant, and inhibitory effects were obtained with a number of catabolites.

Insofar as we are aware, there have been no previous reports linking the Krebs cycle and associated metabolic reactions with regulation of chemotaxis. Our results suggest this important possibility.

Why are some precursors and intermediates of the Krebs cycle potent inhibitors of chemotaxis, whereas others are not? Conceivably, the effectiveness of inhibition is correlated with the ability of the cells to catabolize the test compounds, and this idea was investigated as follows.

Each substance was added to a suspension of the cells, and oxygen consumption was monitored. The results (Table 111) showed that, except for isocitrate, the rate of oxygen consumption correlated with the effectiveness of chemotaxis inhibition for each intermediate. Compounds which were potent (80-100%) inhibitors gave oxygen consumption rates of greater than 35 nmol per min per 1 X 10’ cells. Moderate inhibitors (20-60%) were utilized more slowly, with oxygen consumption rates of 10 to 30, and noninhibitory compounds gave oxygen consumption rates of less than 10. The notable exception to this correlation is isocitrate, which did not inhibit chemotaxis but was avidly utilized by V. furnissii (oxygen consumption = 55 nmol per min per 1 X 10’ cells). This important, apparent anomaly is considered under “Discus- sion.”

Effect of Starvation on Chemotaxis-Marine bacteria are frequently subjected to periods of nutrient deprivation (11,

TABLE I11 Effects of Catabolites on chemotaxis

Chemotaxis results are presented as percent of cells in capillaries in the presence of the inhibitors relative to the controls (no inhibitors). The catabolites were used at 1 mM concentrations (see teY+) __

Inhibitor Chemotaxis to

0 2

GlcNAc (GlcNAcL Glutamine consumption

None Lactate Pyruvate Succinate Fumarate Malate Glycerate Acetate Propionate Citrate a-Ketoglutarate Isocitrate

100 100 0 0 0 0

20 0 20 0 50 40 50 60 48 67 59 76

100 100 100 110 100 130

100 0 0 0 0

40 70 63 98

200 130 100

99 49 44 36 31 11 29 14 4 6

55


25, 26). The effect of starvation on chemotaxis by V. furnissii was therefore determined.

Following induction by (GlcNAc)z, the culture was washed with taxis buffer, and chemotaxis to 10 mM GlcNAc assayed immediately or after shaking for 2 h at 25 "C. The results (Fig. 7) show that starvation for 2 h resulted in about a 3-fold increase in the magnitude of chemotaxis toward GlcNAc. In a more extensive experiment, (GlcNAc)z-induced V. furnissii was tested for chemotaxis to (GlcNAc)2 immediately after harvesting, or after a 2-h, 4-h, or overnight (21-h) starvation. The level of chemotaxis increased after 2 and 4 h of starvation, but decreased to the initial (induced) chemotactic level after overnight starvation. Thus, at the minimum, the induced level of chemotaxis is stable for at least 21 h of starvation, and increases in chemotaxis are evident at earlier times.

When uninduced cells were starved, they exhibited no increased taxis to GlcNAc and the oligosaccharides over the controls, i.e. starvation does not substitute for induction of the chemotactic receptors under growing conditions.

DISCUSSION

In a preliminary communication (16) we reported that V. furnissii swarmed to chitin oligosaccharides, (GlcNAc),, n = 1-6, at initial oligosaccharide concentrations as low as 10 MM in the agar. Swarming may even occur at lower concentrations, but would not have been visible because of insufficient carbon to support significant cell growth.

The present report quantitates these results by the use of a modified capillary assay, establishes the requirements for induction of the receptors, shows that more than one receptor exists, and, finally, presents the first evidence for a direct connection between catabolism and chemotaxis.

Enzyme IINng of the bacterial phosphoeno1pyruvate:glycose phosphotransferase system is the chemoreceptor for GlcNAc in V. furnissii as it is in E. coli and S. typhimurium. However, IINag is not the receptor for the oligosaccharides, and the receptors for the latter do not recognize GlcNAc.

At least two separately inducible oligosaccharide chemoreceptors were detected in V. furnissii. Optimum induction was obtained by growth of the cells in lactate medium in the presence of 0.6 mM concentrations of the inducer. Except for IINag, the chemoreceptors have overlapping specificities, and an accurate definition of each awaits cloning of the respective genes. We propose that the oligosaccharide receptors are periplasmic solute binding proteins, which would also serve the important function of salvaging the lower oligomers during degradation (2).

Unlike the results obtained on the swarm plates, the cells

Hours of Starvation - 400 'z 300 2 200

loo

0

- - 0)

GlcNAc (GlcNAc),

FIG. 7. Effect of starvation on V. furnissii chemotaxis. V. furnissii was grown in lactate-labeling medium containing 0.6 mM (GlcNAc)n. After harvesting and washing, chemotaxis to 10 mM GlcNAc was analyzed in the modified capillary assay. The experiment was performed immediately or after an additional 2 h during which the cells were shaken at 25 "C in taxis buffer (no carbon source). The same experiment was performed using (G1cNAc)p as the attractant and where the starvation period was extended for 21 h.

Attractant

did not show taxis in the capillary assay to (GlcNAc)6 or (G1cNAc)G. However, since growth of V. furnissii on these compounds requires special conditions, it is likely that proteins (e.g. porins) must be induced for these oligosaccharides to be recognized and utilized. These conditions exist on the swarm plates.

The cells showed low to no taxis to the sugars in the capillary assays after growth on GlcNAc or the oligosaccharides as the sole sources of carbon, which appears to be entirely contrary to the results obtained on the swarm plates. The apparent contradiction may be explained because the cells consume the sugars as they swarm and are thus continuously exposed for many hours to low concentrations of inducers in the agar, rather than to the 5-55 mM concentrations used in liquid culture.

Once induced for the oligosaccharide chemoreceptors, V. furnissii retains its ability to respond to the chitin oligomers following starvation for at least 21 h. In fact, during the early stages of starvation, the cells show a 2-fold greater chemotactic response than the nonstarved cells (Fig. 7).

Perhaps the most exciting and unexpected results reported here were obtained by studying the effects of catabolites and growth media on the chemotactic behavior of V. furnissii. In these experiments, there were no obvious changes in cell motility as determined microscopically. The catabolite effect was explored using three different chemotaxis systems, the inducible GlcNAc, inducible (GlcNAc)z, and the constitutive Gln systems (Table 111). While the degree of inhibition varied with the catabolite tested, the results were independent of the chemoattractant used. Lactate, pyruvate, succinate, and fumarate gave close to complete inhibition of taxis to all three compounds. Other catabolites gave partial inhibition, had no effect, or even stimulated taxis. The inhibitory effects correlated with the ability of the cells to oxidize the compounds. With one exception, substances that were oxidized rapidly were the most effective inhibitors. The exception, isocitrate, was oxidized at a very rapid rate, but did not inhibit chemotaxis. The latter result must be considered in light of the idea of cross-talk (12-15), since isocitrate lyase is phosphorylated at a His residue, as are proteins of the phosphotransferase system, some of the pivotal Che proteins, and some Krebs cycle enzymes.

Insofar as we know, there has been no evidence linking the Krebs cycle and chemotaxis, other than that presented in this report. From the pattern of responses, it is obvious that these interactions must be complex. Since most of the catabolites are inhibitory, we speculate that the chemotactic response functions normally when the Krebs cycle is operating at a low rate, but is inhibited when the cycle functions at a high rate.

The other data discussed above are consistent with this general notion. That is, when the medium contains sufficient carbon (and nitrogen) to fulfill the metabolic requirements of the cell, then the chemotaxis apparatus is either inhibited or is not induced. As in the adhesion/deadhesion apparatus (l), and in the chitin oligosaccharide-degradative pathways (21, V. furnissii once again exhibits an unusual and surprising degree of sophistication in its ability to monitor and respond to the nutritional status of its environment.

REFERENCES 1. Yu, C., Lee, A., Bassler, B. L., and Roseman, S. (1991) J. Biol.

Chem. 266,24260-24267 2. Bassler, B. L., Yu, C., Lee, Y. C., and Roseman, S. (1991) J. Biol.

Chem. 266,24276-24286 3. Macnab, R. M. (1987) in Escherichia coli and Salmonella typhi-

murium: Cellular and Molecular Biology (Nerdhardt, F. C., Ingraham, J. L., Low, K. B., Magasanik, B., Schaechter, M., and Umbarger, H. E., eds) Vol. 1, pp. 732-759, American


Society for Microbiology, Washington, D. C.

76,110

239

Bacterid. 172,334-341

4. Mitchell, R., Chet, I., and Asketh, P. (1976) Goo. Rep. Announce.

5. Chet, I., and Mitchell, R. (1976) Annu. Reu. Microbiol. 30, 221-

6. Sar, N., McCarter, L., Simon, M., and Silverman, M. (1990) J.

7. McCarter, L., Hilmen, M., and Silverman, M. (1988) Cell 5 4 ,

8. Alam, M., and Glagolev, A. N. (1982) FEBS Lett. 143, 144-146 9. Adler, J., and Epstein, W. (1974) Proc. Natl. Acad. Sci. U. S. A.

10. Lengler, J., Auburger, A,"., and Percher, A. (1981) Mol. & Gen.

11. Smith, F. G. W. (ed) (1974) Handbook of Marine Science, Vol. I,

12. Nixon, B. T., Ronson, C. W., and Ausubel, F. M. (1986) Proc.

13. Ronson, C. W., Nixon, B. T., and Ausubel, F. M. (1987) Cell 49,

14. Bourriet, R. B., Hess, J. F., Borkovich, K. A., Pakula, A. A., and

345-351

71, 2895-2899

Genet. 183,163-170

CRC Press, Boca Raton, FL

Natl. Acad. Sci. U. S. A. 83, 7850-7854

579-581

Simon, M. I. (1989) J. Bid. Chem. 264, 7085-7088

15. Stock, J. A., Ninfa, A. J., and Stock, A. M. (1989) Microbiol. Reu.

16. Bassler, B. L., Gibbons, P. J., and Roseman, S. (1989) Biochem. Biophys. Res. Commun. 161,1172-1176

17. Bassler, B. L., and Roseman, S. (1990) Chemotaxis and Chitin Degradation in the Marine Bacterium Vibrio furnissii, pp. 65- 94. Ph.D. thesis, The Johns Hopkins University

18. Horowitz, S. T., Roseman, S., and Blumenthal, H. (1957) J. Am. Chem. SOC. 79,5046-5049

19. Adler, J. (1973) J. Gen. Microbiol. 74, 77-91 20. Hazen, T. C., Dimock, R. V., Jr., Esch, G. W., Mansfield, A., and

21. Stock, J. B., Clarke, S., and Koshland, D. E., Jr. (1984) Methods

22. Laemmli, U. K. (1970) Nature 227, 680-685 23. Kort, E. N., Goy, M. F., Larson, S. H., and Adler, J. (1975) Proc.

24. Springer, M. S., Goy, M. F., and Adler, J. (1979) Nature 280,

25. Goldberg, E. D. (ed) (1974) The Sea, Vol. 5, Wiley-Interscience,

26. Parson, T., Takahashi, M., and Hargrave, B. (1977) Biological

53,450-490

Raker, M. L. (1984) Curr. Microbiol. 10, 13-18

Enzymol. 106,310-321

Natl. Acad. Sci. U. S. A. 72,3939-3943

279-284

New York

Oceanographic Processes, 2nd ed., Pergamon Press, Oxford

the of chemistry vol. 266, no. 36, issue of 25, pp. 24268 ... · the journal of biological...

Documents