Induction of phenazine biosynthesis in cultures of Pseudomonas aeruginosa by L-N-(3-oxohexanoyl) homoserine lactone

ELSEVIER

FEMS Microbiology

Letters 140 (1996) 15-22

Induction of phenazine biosynthesis in cultures of Pseudomonas aeruginosa by L-N-( 3-oxohexanoyl) homoserine lactone Paul Stead Nutural

Products

*,

Brian A.M. Rudd, Helen Bradshaw, David Noble, Michael J. Dawson

Chemist?

Section, Glalo

Wellcome Medicines

Received

Research Centre, Gunnells

Wood Road, Stel~enage. Hem.

SGI 2NY, UK

1 February 1996: accepted 29 March 1996

Abstract A range of Pseudomonas spp. and other Gram-negative bacteria were screened for induction of antimicrobial activity in response to the autoregulatory factor L-N-(3-oxohexanoyl)homoserine lactone. In one of these, P. aeruginosa ATCC 10 145, the production of phenazine metabolites was shown to be inducible in a dose-dependent manner. The production of phenazine-I-carboxamide increased over SO-fold compared to control cultures when supplemented with 200 FE/ml of the autoregulator. In addition, the production of an unidentified polar antibacterial substance by this strain increased with

autoregulator concentration. Kewordsr

Phenazine;

Autoregulation;

N-(3-0xohexanoyl)homoserine

1. Introduction y-Butyrolactone autoregulators are widespread in nature and are increasingly implicated in the control of cell density-dependent phenomena in bacteria, including the production of secondary metabolites. A-factor and related butanolides have been identified in a number of Streptomyces species and have been shown to regulate the production of many classes of secondary metabolites, e.g. virginiamycin, anthracycline and aminoglycoside antibiotics [l-3]. N-(3-0xohexanoyl)homoserine lactone (HSL) was originally isolated from the marine bacterium Vibrio fischeri where it was shown to regulate bioluminescence in a

* Corresponding Federation

author.

of European

Microbiological

PI1 SO378-1097(96)00149-8

lactone:

cell density-dependent manner [4]. The same molecule was more recently demonstrated to regulate the production of a I-carbapenem antibiotic in the terrestrial plant pathogen Etwinia carotocoru [5,6]. Jones et al. [7] further showed that the expression of cell-wall-degrading exoenzymes was regulated by HSL in E. carotocora, and the related molecule N-(3-oxododecanoyl)homoserine lactone was shown by Pearson et al. [8] to regulate the expression of virulence genes in Pseudomonas aeruginosa PA0 1. Here we report an investigation of the effect of HSL on secondary metabolite production in a range of pseudomonads and other Gram-negative bacteria, using antimicrobial activity (inhibition zone bio-assay versus Saccharomyces cereuisiae, Escherichia coli and Bacillus subtilis) as the indicator phenotype.

Societies.

P.seudomonus wruginoso

2. Materials and methods Chemicals and reagents were purchased from Sigma-Aldrich and were of AR grade. Solvents were HPLC grade and were purchased from Rathbums Ltd., UK. Strains: S. cerelisiae NCYC232. E. coli DC2, B. subtilis ATCC 605 la and P. aeruginosa ATCC 10145 together with the other strains investigated were stored frozen in brain heart infusion broth containing 10% (v/v) glycerol. L-N-(3-Oxohexanoyl)homoserine lactone was kindly provided by Prof. B.W. Bycroft, University of Nottingham, UK. Media: Growth media were as follows (g/l in parentheses): SV2: glucose (I 5), glycerol (I 5). soy peptone (15), NaCl (3), CaCO, (l), pH 7. SM9: MOPS (20.9), L-proline (11.5), glycerol (23). NaCl (0.5), K,HPO, (0.52), EDTA (0.251, MgSO, 7H,O (0.49), CaCl, .2H,O (0.029) and 5 ml of a trace salts solution containing: ZnSO, . 7H20 (0.86). MnSO, . 4HZ0 (0.22), H 3B0, (0.062), CuSO, . 5H,O (0.125), Na?MoO, . 2HZ0 (0.048), CoC12 6Hz0 (0.048), FeSO, . 7H20 (1.8), KI (0.083) and 1 ml of 1 M H,SO,: pH 6.5. SM9 agar was prepared by adding agar (15 g/l) to the medium. SM84: L-glutamic acid Na salt 16.9 g, KH?PO, 0.4 1 g, distilled water 900 ml, with sucrose solution (30% w/v) 100 ml, 0.2 M MgSO, 7Hz0 10 ml and trace salts solution (5 ml) added after autoclaving. pH 7. YNBG: Difcobacto yeast nitrogen base (67), glucose (20), agar (17). Da1.i.~ Mingioli agar: K ?HPO, (7). KH,PO, (3), Na,C,HsO,. 2H,O (0.5), MgSO,. 7Hz0 (O.l), (NH,)>SO, (l), agar (20) with 50% w/v glucose (40 ml) added after autoclaving. Nutrient broth (NB): Lab-Lemco (Difco) (IO), peptone (lo), NaCl (5). Nutrient agar (NA) was NB solidified with agar (15 g/l).

SM84 and NB; as before). Liquid cultures were incubated at 28°C on an orbital shaker (250 rpm). Agar cultures were incubated in a stationary incubator at 28°C as 5 ml agar in 12-well Costar plates. To prepare larger batches of liquid culture for extraction, 50 ml volumes of production medium in 250 ml flasks were inoculated and grown as above. and pooled before extraction. 2.2. Inhibition :one bio-assays Cell suspensions of the test strains were inoculated from thawed cryotubes into nutrient broth (E. coli. 8. .subtilis) or nutrient broth containing 20 g/l glucose (S. cere~~isiae). Cultures were incubated at 37°C (S. cererisiae. E. coli) or 28°C (B. subtilis) on an orbital shaker (250 rpm) for 16 h. Molten agars were kept at 50°C in a water bath prior to inoculation. S. cerec~isiae culture (1% v/v) was inoculated into 250 ml of 1.79 w/v agar solution containing 10% (v/v> of 10 X YNBG. E. coli culture (0.29 v/v) was inoculated into Isosensitest agar (250 ml) and B. subtilis culture (0.2% v/v) into 260 ml Davis Mingioli agar. Seeded agar was immediately poured into level bio-assay plates (Nunc) and plates were allowed to set. Agar plugs were stamped out using a sterilised cork borer (d = 0.6 cm) and removed using a sterile toothpick. After addition of test solutions to wells. plates were placed at 4°C for 2 h to allow diffusion of sample into agar. then transferred to a static incubator (E. co/i and S. cerel,isiue 37°C; B. subtilis 28°C) overnight. Inhibition zones were measured using calipers. HSL was prepared as a sterile solution in methanol (I mg/ml) and added to media by aseptic transfer. The solution was dispensed into culture media prior to inoculation to give the required concentration.

2. I. Growth conditions

2.3. Time course measurements

The organisms used in this study were grown on three different test media in two forms: shaken liquid and agar. To prepare seed cultures, frozen cell suspensions were thawed and inoculated (5% v/v) into SV2 seed medium (10 ml in 50 ml shake flask) and incubated overnight at 28°C on an orbital shaker at 250 rpm. Inocula were transferred to production media (SM9,

All cultures were sampled initially at 1, 3 and 6 days. For the shaken liquid cultures, whole fermentation broth (0.5 ml) was mixed with methanol (0.5 ml), left to stand at ambient temperature for 1 h then centrifuged (13 000 rpm X 5 min) to pellet cells; the supematant was taken. For the plate bio-assays. 50 ~1 of the extract was added to wells. For analysis by HPLC. 20 ~1 was injected onto the column. For the

agar-based cultures, methanol (5 ml) was added to each culture and cultures left at ambient temperature for 1 h. Extract (2 ml) was pipetted off, centrifuged to remove solid material ( I3 000 rpm X 5 min) then supematant was collected.

Detection: phenazinephenazine-

2.4. HPLC

‘H NMR spectra were recorded using a Bruker AM500 instrument operating at 500.13 MHz and 298 K. Samples were dissolved in DMSO-d, and TMS was used as internal standard. Thermospray (TSP) mass spectra were recorded using a Finnigan Matt TSQ 700 instrument. Samples were flow injected using 50% (v/v) MeCN/H,O with 100 mM ammonium acetate as eluent. FAB mass spectra were recorded using a VG Autospec instrument via a direct insertion probe and using a glycerol/thioglycerol matrix. For both analyses the source temperature was 250°C and vapouriser temperature 85°C: the repeller voltage was 55 V.

HPLC analysis was performed using an LDC instrument (Constametric I and Constametric 3 pumps with gradient controller) equipped with a Hewlett Packard series 1050 UV-diode array detector monitoring at 210, 240 and 280 nm with wavelength scanning from 190 to 400 nm. Preparative HPLC was performed using a Gilson instrument (two Gilson 303 pumps with gradient controller, manometric module, mixer, Holochrome detector). The detection wavelength was 210 nm. 2.4. I. HPLC analysis of phenazine- 1-carboxamide and phenazine-I-carboxylic acid Column: Spherisorb 5 ,um C,, 15 cm X 0.46 cm. Eluent: 35% (v/v) acetonitrile plus 50 mM NH, H 2PO, in water with 1 ml H,PO, added per litre. Detection: UV at 240 nm. Flow rate: 2 ml/min. Phenazine- I -carboxamide eluted after 3.0 min and phenazine- I-carboxylic acid after 4.8 min. 2.4.2. Analysis of pyocynine Column: Kromasil 5 Km C, 15 cm X 0.46 cm. Eluent: 38% (v/v) acetonitrile plus 0.01 M NH,H,PO, in water with 3 ml H,PO, added per litre. After de-gassing 14.4 g sodium dodecyl sulfate was added to each litre of eluent. Detection: UV at 280 nm. Flow rate: 2 ml/min. Pyocyanine eluted after 4. I min. 2.4.3. Gmdient HPLC am&G of,fermenlation broth extracts Column: Spherisorb 5 pm C, 15 cm X 0.46 cm. Solvent A: 0.035 M NHJH,PO, in water with 0.35 ml H,PO, added per litre. Solvent B: 750 ml acetonitrile made up to I litre with 0.075 M NH, H, PO,, with 3 ml H,PO, added. Gradient program: t = 0 to f = 5 min, 100% A; t = 5 to r = 20 min, linear gradient 100% A to 100% B; t = 20 to t = 35 min, hold at 100% B; return to 100% A in 1.5 min.

UV at 210 nm. Under these conditions I -carboxamide eluted after 9.6 min and I -carboxylic acid after 10.4 min.

2.6. Isolntinn oj phenadnephenazine- 1-carboxylic acid

I -carboxamide

and

Methanol (1 litre) was added to whole fermented broth (1 litre. SM9) of P. aeruginosa ATCC 10145. After 1 h at ambient temperature cells were removed by centrifugation (2500 rpm X 30 min) and the supernatant was taken. Methanol was removed by rotary evaporation at 30°C. The aqueous solution was passed through a column (200 ml bed volume) containing Amberlite XAD-16 resin which had previously been washed with acetonitrile and equilibrated in water. The resin was washed with water (400 ml) and then sequentially eluted with 25% (v/v) acetonitrile in water (400 ml). 70% (v/v) acetonitrile in water (400 ml) and neat acetonitrile (200 ml). The 25% and 70% acetonitrile eluates were combined and evaporated to dryness at 40°C to give a dark green solid. This was taken up in the preparative HPLC eluent (15 ml) and chromatographed on Spherisorb 5pm C,, (25 cm X 2.5 cm) with 35% (v/v) acetonitrile. 50 mM NH,H.PO, in water + I ml/l H,PO, as the eluent. The -flow rate was 25 ml/min and the sample injection volume was 5 ml. Under these conditions phenazine- 1-carboxamide eluted after 14 min and phenazine- I -carboxylic acid after 18 min. The relevant fractions were pooled and

P. Stead et al. / FEMS Microbiology Letters 140 (I 9%) 15-22

18

acetonitrile was removed by rotary evaporation. Fractions were desalted by pumping the aqueous solutions back on to the HPLC column (after cleaning and equilibrating in water) and washing the column with water. Bound material was eluted with 90% (v/v) acetonitrile in water. The bulk of the acetonitrile was removed by rotary evaporation and the solutions were lyophilised to afford phenazineI-carboxamide (4 mg) as a yellow gum and phenazine- I-carboxylic acid (22 mg) as a yellow powder. 2.61. Phenazine-I-carboxamide ‘H NMR data (DMSO-d,): d 6.8-7.3 (broad, NH), 8.03-8.15 (m, 3H; H-3, H-7, H-81, 8.33 and 8.44 (both m, 2H; H-6 and H-9), 8.46 (dd, IH; H-4), 8.71 (dd, 1H; H-2), 9.76 (bs, 1H; NH). MS data: TSP +ve; M+H’+ 224. FAB + ve; M + H 1 + 224. FAB -ve; M-’ 223. UV,,, (nm, HPLC mobile phase); 205, 266, 37 1. 2.6.2. Phenazine-I -carboxylic acid ‘H NMR data (DMSO-d,): d 8.07-8.15 (m, 3H; H-3, H-7, H-8), 8.37 and 8.43 (both m, 2H; H-6 and

H-91, 8.54 and 8.56 (both dd, 2H; H-2 and H-4). MS data: TSP + ve; M + H 1 + 225. FAB + ve; M + H T + 225. FAB -ve; M-’ 224. UV,,, (nm, HPLC mobile phase); 208, 250, 375. 2.7. Isolation of pyocyanine Pyocyanine was isolated using a method modified from that of Von Saltza et al. [9]: a cell-free methanol extract of whole fermented broth (1 litre) of P. aeruginosa ATCC 10145 was obtained as described in the previous method. The pH of the extract was adjusted to 4 (cont. HCl) to stabilise pyocyanine and methanol was removed by rotary evaporation. The pH was raised to 7.5 (cont. ammonia) and the solution was extracted with dichloromethane (2 X 500 ml). Centrifugation (2000 rpm X 30 min) was used to separate emulsions. The pooled dichloromethane extracts were back-extracted with 0.05 N HCl (100 ml). The pH of the aqueous extract was re-adjusted to 7.5 and the solution was extracted with dichloromethane (3 X 25 ml). Combined extracts were back-extracted with 0.05 N HCl (7.5 ml>. The pH of the aqueous extract was adjusted to 4 and the

.

-0mgMHsL

-aDtruJtYlHsL

-53rrQhiHsL

-

1tTJ~

HSL

Fig. 1. The effect of HSL on bioassay zone size. Cultures were sampled (0.5 ml mixed with 0.5 ml methanol: cells removed by centrifugation) and extracts were inoculated into wells cut into agar that had been seeded with B. subtilis 6051a. Plates were incubated at 4°C for 2 h, then at 28°C for a further 16 h, and then the inhibition zone diameters were measured. A typical induction profile from four replicate experiments is shown here

P. Stead et al. / FEMS Microbiology

volume was reduced to ca. 2.5 ml by rotary evaporation. The pH was further adjusted to 8.0 and the solution was stored at 4°C overnight. Blue-black crystals were harvested by centrifugation (13 000 rpm X 10 min), washed with water (1 ml pH 8) and

Letter.7 140 (19Y61 IS-22

dried in vacua over P,O,

19

to yield pyocyanine

mALl

woo-

IOW-

Phenazinsl-carboxyiic acid

&x-

aa-

Phenezine-l-carboxamide

4w

200

i

\4 0

--v---v----

--

2

4

(11.5

mg). ‘H NMR data (DMSO-d,): d 3.87 (s, 3H; CH,), 5.92 (d, J = 7.8 Hz, 1H; H-4), 6.58 (d, J = 9.3 Hz, 1H; H-2), 7.49 (t, J = 7.8 Hz, IH: H-8). 7.56 (d,

_--_I

6

Fig. 2. Gradient HPLC analysis of P. aeruginosa

8

lb

ATCC 10145 culture extract.

12

mm

P. Stead et (II. / FEMS Microhiolog~ Letters 140

procyanine

PlEnaz&-l-carboxylic acid

Phemzbe-1-c

Fig. 3. Structures of phenazines

J = 8.7 Hz, 1H; H-6), 7.60 (dd, J = 9.3 and 7.8 Hz, 1H; H-3), 7.79 (td, J = 8.1 and 1.2 Hz, IH; H-7), 8.35 (dd, J = 7.8 and 1.2 Hz, 1H; H-9). MS data: TSP + ve; M + H 1 + 2 11. UV,,, (nm, HPLC mobile phase); 200, 280, 385.

3. Results and discussion A total of 18 Pseudomonas species and 27 miscellaneous Gram-negative organisms were tested for antimicrobial activity in three agar and three 3 shaken liquid media, with or without the addition of 20 lug/ml HSL. The majority showed antimicrobial

CI YMI 15-22

induced by HSL.

activity against at least one of the indicator strains in at least one of the six media, but only one strain, P. aeruginosa ATCC 10 145, showed evidence of increased antimicrobial activity with HSL addition, and this was in only one medium with only one bioassay microorganism. The strain showed no activity against E. coli or S. cereuisiae with or without added HSL, but significantly greater activity was given against B. subtilis when shaken liquid cultures in SM9 media had been supplemented with 20 pg/ml HSL compared to control cultures (see Fig. 1). The increase in antibacterial activity was also accompanied by an increase in pigmentation of the supplemented cultures relative to controls.

Fig. 4. The induction of phenazine metabolites in P. arruginosa ATCC 10145 by HSL. Extracts of HSL-induced cultures were prepared by mixing whole broth (100 gl) with HPLC mobile phase (400 ~1) and centrifuging to remove cells. 100 ~1 of the extracts were analysed by HPLC. The figure shows a typical induction profile from three replicate experiments.

P. Stead et 01./ FEMS Microbinlo~~ Letten 140 (19%) 15-22

In order to analyse this effect more closely, methanol extracts of HSL-supplemented and control cultures were prepared at various time points throughout the growth cycle and the extracts were analysed by gradient HPLC (Fig. 2). The analysis revealed that two major components of the extract were responsive to HSL and increased significantly as the concentration of added HSL was raised. These metabolites were isolated from a culture of ATCC 10145 and shown to correspond to phenazine- Icarboxamide and phenazine- 1-carboxylic acid (Fig. 3) by ‘H NMR and mass spectrometry. Neither metabolite, however, possessed significant activity against the test strain of B. subtilis under the conditions used. Since pyocyanine is a known phenazine metabolite of P. aeruginosa 191, a separate culture of ATCC 10145 was investigated for the presence of this metabolite, in an attempt to explain the increase in antibacterial activity and pigmentation which accompanied the addition of HSL to cultures. Pyocyanine was indeed shown to be present, but it too was inactive against B. subtitis either as the hydrochloride or as the free base. Neither did the levels of phenazines present in cultures appear to fully explain the increase in pigmentation observed. The magnitude of the induction of phenazine-lcarboxamide in SM9 shaken liquid culture is most striking (Fig. 4). The levels of this metabolite were very low throughout the growth cycle in cultures that had not been supplemented with exogenous HSL. However. when HSL was added to cultures at inoculation, titres of phenazine- I-carboxamide rose dramatically, increasing over 50-fold when 200 pg/ml HSL was added. The induction of phenazine- lcarboxylic acid and pyocyanine was somewhat smaller; titres approximately doubled at the peak of the dose-response curve. It is interesting to note that the dose-response curves are different for each phenazine. The titre of pyocyanine was maximised when the level of exogenously added HSL was ca. 20 pg/ml. The corresponding level for induction of phenazine-I-carboxylic acid was 50 ,cLg/ml and it is clear from Fig. 3 that the peak titre for phenazine-lcarboxamide had probably not been reached even when 200 @g/ml HSL was added. The general pattern of induction as illustrated in Fig. 4 was the same whether cultures were analysed at 24, 48 or 72 h, suggesting that HSL is indeed inducing signifi-

21

cantly increased levels of phenazine production overall, and not simply shifting a normal metabolite production profile to a point earlier in the growth cycle. The nature of the enhanced antibacterial effect of culture extracts vs. B. subtilis awaits full characterisation. The antibacterial metabolites produced by P. aeruginosu ATCC 10145 in SM9 shaken liquid were partially characterised by performing gradient HPLC with fraction collecting, then assaying each fraction against B. subtilis. The activity was located in two fractions; one eluted very near to the solvent front, indicating that the antibacterial substance was highly polar under the conditions of the HPLC analysis. The zone of inhibition caused by this substance was large and diffuse, and appeared to be inducible by HSL. The nature of the compound(s) responsible for this activity has yet to be elucidated. A much later running fraction from the gradient HPLC experiment elicited a small, but clear zone of inhibition in the B. subtilis inhibition zone assay. This second fraction was fully characterised and the antibacterial metabolites were shown to be rhamnolipids (data not shown). The clear zone of inhibition did not appear to increase markedly with HSL concentration. indicating that the rhamnolipids might not be subject to induction by HSL. It is clear from these results and observations that HSL is eliciting a broad based response in cultures of P. aeruginosu ATCC 10145. In addition to the metabolites identified in this work. other metabolites. e.g. pigments, appear to be subject to induction. The effect of HSL in this organism is characteristic of the pleiotropic effects of y-butyrolactone autoregulators in many other species so far studied. It is possible that HSL. or an analogue. functions as an endogenous autoregulator of phenazines and other secondary metabolites in this organism. since HSL has already been identified in P. arruginctsu [5]. and the structural analogue N-(3-oxododecanoyl)homoserine lactone has been shown to function as an autoregulator of virulence determinants in P. wruginmu PAOI 181. In addition. the biosynthesis of phenazine metabolites in Pseudmnonas uureoficciens has been shown to be regulated by p&R [IO], which shares considerable homology with IuxR. the regulatory gene of the lux operon which mediates bioluminescence in V. ,fisheri.

22

P. Stead et al. / FEMS Microbiology

In conclusion, HSL has been shown to induce the production of phenazines and other unidentified secondary metabolites in P. aeruginosu ATCC 10145 in a dose-dependent manner. Since 45 strains were tested for inducibility of antimicrobial activity with HSL, potentiation of antimicrobial production by HSL does not appear to be a very common phenomenon, at least under the conditions reported here.

Acknowledgements

The authors would like to thank Dr P.J. Sidebottom and Mr A. Blackaby for performing the NMR experiments, and Mr N. Taylor and Dr P.S. Marshall for mass spectrometry analysis.

References [l] Yamada, Y., Sugamura, K., Kondo, K., Yanagimoto, M. and Okada, H. (1987) The structure of inducing factors for virginiamycin production in Streptomyces cirginiae. J. Antibiot. 40, 496-504. [2] Grafe, U., Reinhardt, G., Schade, W., Eritt, I., Fleck, W.F. and Radics, L. (1983) Interspecific inducers of cytodifferentiation and anthracycline biosynthesis from Streptomyces bikinensis and Streptomyces cyneofuscatus. Biotechnol. Lett. 5, 591-596. [3] Khokhlov, A.S., Tovarova, I.I., Borisova, L.N., Pliner, S.A., Schevchenko, A., Kornitskaya, E.Ya., Ivkina, N.S. and Rapoport, LA. (1967) A-factor responsible for the biosynthe-

Letters 140 (lYY6) 15-22 sis of streptomycin by a mutant of Actinomyces .ttrepiomycini. Dokl. Acad. Nauk. SSSR 177, 232-235. [4] Eberhard, A., Burlingame, A.L., Ebcrhard, C., Kenyon, G.L.. Nealson, K.H. and Oppenheimer, H.J. (I 98 I) Structural identification of autoinducer of Photobucterium jischeri luciferase. Biochemistry 20, 2444-2449. [5] Bainton, N.J., Bycroft, B.W., Chhabra, S.R., Stead. P., Gledhill, L., Hill, P.J., Rees, C.E.D.. Winson, M.K., Salmond, G.P.C., Stewart, G.S.A.B. and Williams, P. (1992) A general role for the /UX autoinducer in bacterial cell signalling: control of antibiotic biosynthesis in Erwiniu. Gene 116. 87-91. 161Bainton, N.J., Stead. P., Chhabra, S.R., Bycroft, B.W., Salmond, G.P.C., Stewart, G.S.A.B. and Williams, P. (1992) N-(3.oxohexanoyl)homoserine lactone regulates carbapenem antibiotic biosynthesis in Erwinia carotorora. Biochem. J. 288, 997- 1004. [71 Jones, S., Yu, B., Bainton, N.J., Birdsall, M., Bycroft, B.W., Chhabra, S.R., Cox, A.J.R., Golby, P., Reeves, P.J., Stephens, S., Winson, M.K., Salmond, G.P.C., Stewart, G.S.A.B. and Williams, P. (1993) The lux autoinducer regulates the production of exoenzyme virulence determinants in Eminiu carotovora and Pseudomonas aeruginosa. EMBO J. I?. 2477-2482. Bl Pearson, J.P., Gray, K.M., Passador, L., Tucker, K.D., Eberhard, A., Iglewski, B.H. and Greenberg, E.P. (1994) Structure of the autoinducer required for the expression of Pseudomonas aeruginosa virulence genes. Proc. Natl. Acad. Sci. USA 91, 197-201. [91 Von Saltra, M.H., Last, J.A., Stapleton, P.G., Rathnum, M.L. and Neidleman, S.L. (1968) Cyanomycin, its identity with pyocyanine. J. Antibiot. 22, 49-54. [IO1Pierson, L.S., Keppenne, V.D. and Wood, D.W. (1994) Phenazine antibiotic biosynthesis in Pseudomonas aureofaciens 30-84 is regulated by PhzR in response to cell density. J. Bacterial. 176, 3966-3974.