Some Properties of Pyoverdine, the Water-Soluble Fluorescent Pigment of the Pseudomonads R

1958] PROPERTIES OF PYOVERDINE 241

metabolites. The Chemical Society, Special Publication Roussos, G. G. AND VINING, L. C. 1956 Isolation and prop- No. 5, London, England. erties of pure actinomycins. J. Chem. Soc., 2469-2474. KOFFLER, H., EmERSON, R. L., PERLMAN, D., AND BURRIS, SAUNDERS, A. AND SYLVESTER, J. C. 1947 Synthetic media R. H. 1945 Chemical changes in submerged penicillin for the production of streptomycin. Abstracts of Papers, fermentations. J. Bacteriol., 60, 517-548. 112th meeting, Am. Chem. Soc., 9A-1OA. O'BRIEN, E., WAGMAN, G. H., AND PERLMAN, D. 1952 Syn- SCHMIDT-KASTNER, G. 1956 Actinomycin E und Actino- thetic media for growth and streptomycin production by mycin F, zwei neue biosynthetische Actinomycin gemische. Streptomyces griseus. Bacteriol. Proc., p. 25. Naturwissenschaften, 43, 131-132. PUGH, L., KATZ, E., AND WAKSMAN, S. A. 1956 Antibiotic WAKSMAN, S. A. AND WOODRUFF, H. B. 1940 Bacteriostatic and cytostatic properties of the actinomycins. J. Bac- and bactericidal substances produced by a soil actino- teriol., 72, 660-665. myces. Proc. Soc. Exptl. Biol. Med., 45, 609-614.

Some Properties of Pyoverdine, the Water-soluble Fluorescent Pigment of the Pseudomonads R. P. ELLIOTT Food and Drug Administration, U. S. Department of Health, Education, and Welfare, Federal Office Building, San Francisco, California Received for publication December 3, 1957 The fluorescent water-soluble pigment complex pro- were probably determined on mixtures, and therefore duced by pseudomonads has been variously called not valid. Giral (1936), Chodat (1951), and Naves "bacterial fluorescein" or "fluorescin." These terms (1955) all claim to have separated blue and yellow frac- are unfortunate for they are synonymous with re- tions by chromatographic means. Giral (1936) con- sorcinolphthalein and resorcinolphthalin, respectively sidered pyoverdine to be a lyochrome or a pterine. (Merck Index, 1952). The bacterial pigment is not re- Birkhofer and Birkhofer (1948) reported that it was lated to either of these compounds. However, the term composed in part of riboflavin. "pyoverdine," suggested by Turfreijer (1941), spe- Giral (1936) found pyoverdine to be soluble in water, cifically identifies the bacterial pigment complex and formic acid, aqueous alcohol, aqueous acetone, and 90 will be used here. per cent pyridine. Turfitt (1937) found it was also Pyoverdine has interested bacteriologists for years, soluble in phenol and acetic acid. Giral (1936), Turfitt but few intensive studies have been made on it. Some (1937), and Turfreijer (1941) have presented absorp- investigators have studied media for maximum pig- tion spectra of pyoverdine in various reagents. Meader ment production. This work was adequately reviewed et al. (1925) described the indicator characteristics of by Seleen and Stark (1943) except for the important pyoverdine, and Giral (1936) described the effect of a contribution of Sullivan (1905), who was the first to large number of reagents on its appearance. He stated describe a simple synthetic medium composed of that air favored its production, but that light had the asparagine, magnesium sulfate, and potassium hydrogen opposite effect. phosphate. King et al. (1948), Baghdiantz (1952), and The object of the present study was to obtain enough Totterand Moseley (1953) studiedthe stimulation of pig- information about the properties of pyoverdine to form ment production by added minerals. King et al. (1954) a basis for determining it in frozen whole egg. The described simple solid media. resultant method is described in the second paper of No one has yet determined the exact chemical nature this series (Elliott, 1958). of pyoverdine or its fractions, nor has anyone yet succeeded in crystallizing it. Various authors have used EXPERIMENTAL activated carbon to remove it from aqueous solution. Equipment and materials. The Pseudomonas ovalis When thus adsorbed it can be eluted with aqueous and Pseudomonas fluorescens cultures used in the ex- acetone or alcohol (Giral, 1936; Turfitt, 1937; Tur- periments had been isolated earlier from fluorescent freijer, 1941). Turfreijer succeeded in separating four eggs (Elliott, 1954) and identified by the method of fractions based on solubilities of phosphotungstic acid Haynes (1953). These cultures have been deposited at precipitates, and described many properties of these the Northern Utilization Research Branch, Agricul- fractions. Empirical formulae suggested by Turfitt tural Research Service, U. S. Department of Agricul- (1937), Turfreijer et al. (1938), and Bonde et al. (1957) ture, Peoria, Illinois, and have been numbered B1595 242 R. P. ELLIOTT [VOL. 6 and B1613, respectively. Asparagine broth for the organism was there any evidence of antibacterial production of pyoverdine consisted of asparagine, 0.1 action. per cent; MgSO4.7H20, 0.05 per cent; and K2HPO4, Absorption spectra. The absorption spectrum offered 0.05 per cent in distilled water. The filtrates of fluores- a possible means of identification and measurement of cent cultures used in several experiments were prepared the pigment. Five mg of crude pyoverdine, prepared by inoculating asparagine broth with one of the above as described above, when dissolved in 25 ml of water organisms, incubating it in the dark for 1 to 3 weeks presented the absorption curve shown in figure 1. at room temperature, and filtering it through a Seitz Similar spectra were obtained on the fluorescent fil- pad or a Mandler filter, previously washed free of trates of cultures (figures 2-5) and on this same pyo- soluble fluorescent materials. verdine preparation dissolved in 0.5 per cent AlCl3. Absorption spectra were determined on a Becknman' The curves obtained agreed well with those presented model DU spectrophotometer, and pH measurements by Giral (1936), Turfitt (1937), and Turfreijer (1941). on a Beckman' model H2 pH meter. Fluorescence emis- Fluorescence emission spectra and the effect of pH sion spectra were obtained using the equipment de- change. Fluorescence emission spectra were determined scribed by French (1955). All emission spectra were at three pH values on a fluorescent culture filtrate of corrected for the instrument response to a standard Pseudomonas ovalis and on pure riboflavin solutions. light source. Fluorescence measurements were made on Absorption spectra then were run on the same solutions. a Coleman2 model 12B photofluorometer. Visual fluores- Results are plotted in figure 2. The pH shift in absorp- cence of materials was determined in subdued lighting tion maxima in pyoverdine solutions as described by or in the dark using a Vogelite3 model 6W101 hand Turfitt (1937) is evident here. Also, when the pH is fixture previously described (Elliott, 1954). A Vogel either raised or lowered, the emission maximum shifts blue to green fluorescence color comparison chart was toward the red. Total emission is lessened by such pH used for recording small changes in black-light fluores- changes, the greater decrease occurring in alkali. The cent colors (Benson and Vogel, 1955). Reagents and same pH variations cause nearly complete quenching glassware were tested frequently for the presence of of fluorescence emission of riboflavin, though little foreign fluorescent materials. In general, chemically change in its absorption spectrum. pure reagents were nonfluorescent, whereas technical That pyoverdine is a sensitive pH indicator was grades had to be purified. Darco G604 was the activated shown by adjusting the pH values of an asparagine carbon used for adsorption of pyoverdine. broth culture of P. ovalis using HC1 and NaOH. The Solvents. As a means of separating pyoverdine from colors noted are shown in table 1. All color changes were frozen whole egg magma, a pyoverdine solvent im- reversible. miscible in water would be ideal. A large number of Riboflavin content. If riboflavin were a free component such solvents was tried, but no suitable one was found of pyoverdine, as described by Birkhofer and Birk- since pyoverdine has a strong affinity for the aqueous hofer (1948), the identification and measurement of phase in any extraction system. Water and aqueous the pigment could be accomplished easily by established alcohol were found most practical for the present work. methods. Riboflavin was found to be absent when a Attempts at purification. A crude preparation of P. ovalis culture was analyzed by the official method pyoverdine was obtained by adsorption on activated (AOAC, 1955). Furthermore, the absorption spectrum carbon from a culture filtrate of P. ovalis, followed by elution with aqueous alcohol. This material dried to a brown amorphous residue but did not crystallize. On heating, it began to soften at 230 C, but even at 252 C it did not melt completely. Although this indicates .8001 that pyoverdine is a mixture, it will be considered a / i single entity in further discussion here. Unsuccessful .6001 attempts were made to fractionate this material by paper and column chromatography. In z3 .400 Antibiotic properties. Brief experiments using Micro- a coccus pyogenes var. aureus and Escherichia coli with -J the agar-cup-plate method showed a concentrated a. .200 aqueous solution of crude pyoverdine to diffuse rapidly 0 ./ into the medium surrounding the cup. With neither 250 300 350 400 450 Soo 1 Beckman Instruments, Inc., Fullerton, California. 2 Coleman Instruments, Inc., Maywood, Illinois. WAVELENGTH my 3 Ultra-Violet Products, Inc., San Gabriel, California. Figure 1. Absorption spectrum of crude pyoverdine, 0.2 mg 4Atlas Powder Company, New York, New York. per ml in water. 1958] PROPERTIES OF PYOVERDINE 243 of pyoverdine is not that of riboflavin, and its pH indi- Pyoverdineformation and oxidation-reduction potential. cator characteristics are different. It was concluded When a culture of the facultative organism Pseudo- that riboflavin is not a free component of pyoverdine. monas ovalis is growni in a colorless asparagine broth, Naves (1955) came to this same conclusion. It is of then examined under black light, pyoverdine formation interest to note, however, that species of Pseudomonas begins with a light blue color throughout the medium, that produce riboflavin are known (for example, then the top few millimeters become yellow-green. Ganguly, 1955; Landenburger, 1952.) It is possible Finally the yellow-green color diffuses throughout the that Birkhofer and Birkhofer isolated one of these. medium. This more rapid production of pyoverdine at Separation of pyoverdine from riboflavin is not easy. the surface must be related to the higher oxygen tension Beer's law. Measurements made on a P. ovalis fluores- there. It is not related to pH, as the following data show: cent filtrate at 412 m,u at various stages of dilution show-ed optical density to be proportional to concen- Portion Color under black light pH tration. Similarly, the fluorescence of a culture filtrate Top Green BG 60 7.5 when diluted was proportional to concentration, using Bottom Blue BG 80.8 7.5 the B1 and PC2 filters of the photofluorometer. An Giral described greater production of pyoverdine extract of fluorescent eggs, prepared in accordance with during aeration. An experiment was conducted which the quantitative method described in the second paper confirmed his findings. Results, presented in figure 3, of this series, was also diluted quantitatively. Fluores- show the rapid production of pyoverdine in the culture eince of this material was likewise proportional to aerated during growth with a continuous bubbling of oncentration. Thus pyoverdine obeys Beer's law in sterile air. On the other hand, the culture held aerobic he concentrations expected in egg spoilage. without this flow of air produced much less, and the FLUORESCENT CULTURE FILTRATE RIBOFLAVIN

.2 pH 8.1 6

ASSORPT ION MISSION ~ 0 . I S \E 4 Z 2Zz

0:D a c: i- pH 11.9 pH 1 1.8 01 10- ABSORPTION EMISSION ABSORPTION EMISSION j 4 __.4 . 1 4 .11

2w 2 W z uJ z U)i U) a _ .2 pH 1.1 0 pH 1.0 0 0. ABSORPTION EMISSION .64 ABSORPTION EM ISSION 64- IL 0 .4 .11 4 2 2

r_ * * z r . 350 400 500 600 350 400 500 600 WAVELENGT H my WAVELENGTH mp

Figure 2. Absorption and fluorescence emission spectra of a filtrate of a fluorescent culture of Pseudomonas ovalis and of pure riboflavin. 244 R. P. ELLIOTT [VOL. 6 one in which anaerobic conditions were maintained by has shown that pyoverdine is readily decomposed by a layer of oil on the surface produced almost none of the ultraviolet light, but less readily so by visible light. In pigment. acid solution it appears to have greater susceptibility This study shows that air is necessary in the produc- to long wave ultraviolet light; but in neutral solution tion of pyoverdine. Subsequent studies showed a cessa- it is more susceptible to the short ultraviolet bands. tion of pyoverdine production in spoiling liquid egg by It is readily decomposed by laboratory daylight that the time the oxidation-reduction potential had fallen has passed through window glass. Fluorescence of one to -0.3 v. This observation is illustrated in table 2; solution dropped 30 per cent on 3 hr exposure to however, these data do not indicate the exact voltage laboratory daylight out of direct sunlight. at which production ceases. Pyoverdine will not decompose on several hours' Effect of light. A preliminary test showed that culture exposure to light from incandescent or fluorescent filtrates changed in fluorescent color from green (BG lamps that emit little or nothing in the ultraviolet 50.8) to blue (BG 80.8) when exposed to direct sun- range. It is worthy of note, however, that pyoverdine light and to bluish green (BG 60.8) when exposed to a solutions will break down even in the dark if stored black-light lamp. Figure 4 shows the drastic effect of several weeks. direct sunlight in breaking down the pigment on ex- Effect of concentration on color. There is a correla- posure in silica cells. tion between pyoverdine concentration and its fluores- Further experimentation with known wave bands cent color. That the yellow-green color may indicate

.800O .400 I / ~~~~~BEFORE .6001 \AERATED .300 ~~~~~~EXPOSURE Ii AFTER

.400k _ 1 I- .1 z EL .200~ uJ \. ._,_._ ! AEROBIC 100 J0 11 o ANAEROBIC 250 300 350 400 450 500 550 WAVELENTH my

250 300 350 400 450 500 Figure 4. Effect of 132 min of sunlight passed through silica WAVELENGTH mP glass on the absorption spectrum of the sterile filtrate of a Pseudomonas ovalis culture.

92-j .300 / P. vaIls 8a. I . 7Fx D- .200 AEROB u !, 6 m *- -VIABLE BACTERIA "1 in- - COLO R 59 z ..\/ \ /. 0 .100~ 4> V /ANAEROBIC~ ANAEROBIC WITH DEXTROSE ., ,_. ._ _ _ _ 0 ., _--_I__ 3 _ -' . \ uLI o 0 i P. fluorescens 2 3 4 DAYS AT 200C i Figure S. Comparative production of pyoverdine by Pseudo- 250 300 350 400 450 500 monas in ovalis B1595 asparagine broth cultures (1) continu- WAVELENGTH mnp ously aerated during growth, (2) aerobic but not aerated, (3) anaerobic by reason of a 5-mm layer of sterile mineral oil on Figure 5. Absorption spectra of filtrates of identically its surface, and (4) anaerobic with added 0.5 per cent glucose. treated cultures of the "blue" Pseudomonas fluorescens B1613 Only the 2-day absorption spectra are presented. and the "yellow-green" Pseudomonas ovalis B1595. 1958] PROPERTIES OF PYOVERDINE 245 high concentration and the blue color a low concentra- ovalis. A batch of fresh eggs was immersed in a cold- tion was shown in the following series of experiments. water suspension of the organism. Then the eggs were A culture of Pseudomonas ovalis in asparagine broth divided, and half of them were immersed immediately fluoresced a brilliant yellow-green (BG 50.8). When in melted paraffin. All were held at 15 C. All resulting diluted 1 to 50 with asparagine broth of the same pH fluorescent eggs in the paraffin-coated lot were blue (8.1), it fluoresced blue (BG 80.8). when examined with the black-light candler; all those Pseudomonas fluorescens B1613, which had been in the uncoated lot were yellow-green. isolated from a blue fluorescent egg, produced blue fluorescence consistently when inoculated into shell ACKNOWLEDGMENTS eggs. On the other hand, Pseudomonas ovalis B1595, Dr. C. Stacy French of the Department of Plant isolated from a yellow-green fluorescent egg, produced Biology, Carnegie Institution of Washington, Stanford, yellow-green fluorescence when inoculated into shell California, kindly consented to allow the use of his eggs. These eggs were broken out and stored as separate equipment for fluorescence spectra determinations, and lots in the freezer. Subsequent analysis by the quanti- he assisted in the calculation of the curves. Dr. T. Y. tative method described in a second study (Elliott, Kingma Boltjes of the University of Amsterdam fur- 1958) showed that the egg pulp prepared from the nished a translated summary of the thesis of Turfreijer light blue fluorescent eggs contained one third as (1941). much pyoverdine as that from the yellow-green fluores- Personnel of the Food and Drug Administration at cent eggs. Absorption spectra run on identically treated San Francisco assisted as follows: Herman J. Meuron culture filtrates of these two organisms showed a made the determinations of riboflavin. Doris H. Tilden similar difference in pyoverdine concentration (figure 5). assisted in the literature search and in attempts at Anaerobic cultures of Pseudomonas ovalis in as- chromatographic separation of pyoverdine. Anthoniy paragine broth produced only small quantities of W. Daly translated some of the foreign literature. pyoverdine and fluoresced blue (figure 3). Likewise, artificial conditions which created a low oxygen tension SUMMARY resulted in blue fluorescence in eggs inoculated with P. The term "pyoverdine" is recalled from the work of Turfreijer (1941) as the most suitable for the water- TABLE 1 soluble fluorescent pigment of the pseudomonads. A Effect of alteration of pH on the color of an asparagine crude preparation of pyoverdine would not crystallize, broth culture of Pseudomonas ovalis and had no precise melting point. Its absorption and pH Color by Daylight Color by Black Light fluorescence emission spectra are presented. Riboflavin is not a free component of pyoverdine. Pyoverdine is a 3.1 Colorless Bluish white reversible pH indicator. It is decomposed by short and 6.6 Colorless Bright blue (BG 80.6) by long wave ultraviolet light but ilot by visible light. 7.3 Very light yellow Greenish blue (BG 70.8) It is more rapidly and in greater quantity 7.9* Light yellow green Blue green (BG 50.8) produced 9.0 Light yellow green Blue green (BG 50.8) in aerated media than in media with lesser oxygen 12.2 Yellow green Green (BG 30.8) tension. Its production by pseudomonads ceases in spoiling egg by the time the oxidation-reduction po- * Unadjusted. tential falls to -0.3 v. Fluorescence of pyoverdine is proportional to concentration. It has no antibiotic TABLE 2 activity against Micrococcus pyogenes var. aureus or Spoilage of liquid whole egg at 22 to 26 C after inoculation against Escherichia coli. with Pseudomonas ovalis Some color variations in pyoverdine may be due to Viable Bacteria A Time Eh* Pyover-dinet Odor test: per g variation in concentration. strain of Pseudomonas fluorescens produced less pyoverdine in eggs and cul- hr v X 106 tures than did a strain of Pseudomonas ovalis. Cultures 0 +0.274 9.3 Normal 0.247 of the former fluoresced blue consistently, whereas 19 +0.044 10.6 Normal 147 43 -0.341 54.0 Decomposed 2,300 those of the latter fluoresced yellow-green. 67 -0.326 59.3 Decomposed 5,000 REFERENCES 92 -0.306 55.9 Decomposed 10,300 Association of Official Agricultural Chemists. 1955 Official * Beckman pH meter with platinum and calomel electrodes methods of analysis, Ed. 8, par. 38.33, 38.37-38.40. in continuous contact with the spoiling egg. Includes correc- BAGHDIANTZ, A. 1952 Role of zinc in appearance of com- tion of +0.244 v for use of calomel electrode. ponent of the pigment of Pseudomonas fluorescens (Flugge- t Expressed as gg of riboflavin per 100 g egg. Average of 2 Migula). Arch. sci. (Geneva), 5, 47-48. Brit. Abstr., determinations. AIII, 1935 (1952). t Three persons. BENSON, R. C. AND VOGEL, M. J. 1955 Principles of identi- 246 R. P. ELLIOTT [VOL. 6

fication and measurement of vulvar fluorescence. J. monas). Can. J. Research, 26C, 514-519. Brit. Abstr. Clin. Endocrinol. and Metabolism, 15, 784-800. AIII, 1208 (1949). BIRKHOFER, L. AND BIRKHOFER, A. 1948 Riboflavin, a com- KING, E. O., WARD, M. K., AND RANEY, D. E. 1954 Two ponent of "bacterial fluorescein." Z. Naturforsch. 3b, simple media for the demonstration of pyocyanin and 136. Brit. Abstr., AIII, 779 (1949). fluorescin. J. Lab. Clin. Med. 44, 301-307. Biol. Abstr., BONDE, G. J., JENSEN, C. E., AND THAMSEN, J. 1957 A 29, 14036 (1955). water-soluble fluorescing bacterial pigment which de- LANDENBERGER, R. 1952 The inhibitor and the growthi- polymerizes hyaluronic acid. Acta Pharmacol. Toxicol., promoter in Pseudomonas fluorescens. Z. Naturforsch., 13, 184-93. Chem. Abst. 51, 8199i. 7b, 630-632. Chem. Abstr., 47, 4954b (1953). CHODAT, F. 1951 Procede chromatographique rapide pour MEADER, P. D., ROBINSON, G. H., AND LEONARD, V. 1925 l'analyse des de Pseudomonas fluorescens Pyorubrin, a red water soluble pigment characteristic of pigments Flugge- Bacillus pyocyaneus. Am. sci. J. Hyg., 5, 682-708. Migula. Arch. (Geneva), 4, 188-192. Merck Index of Chemicals and Drugs. 1952 Ed. 6, pp. 437- ELLIOTT, R. P. 1954 Spoilage of shell eggs by pseudomonads. 438. Merck & Co., Inc., Rahway, New Jersey. Appl. Microbiol., 2, 158-164. NAVES, R. G. 1955 Contribution A l'6tude de la pigmentation ELLIOTT, R. P. 1958 Determination of pyoverdine, the de Pseudomonas fluorescens. Doctorate Thesis, U. of fluorescent pigment of pseudomonads, in frozen whole Geneva. 77 pp. egg. Appl. Microbiol., 6, 247-251. SELEEN, W. A. AND STARK, C. N. 1943 Some characteristics FRENCH, C. S. 1955 Fluorescence spectrophotometry of of green-fluorescent pigment producing bacteria. J. photosynthetic pigments. The luminescence of biological Bacteriol., 46, 491-500. systems, pp. 51-74. Edited by Frank H. Johnson. Ameri- SULLIVAN, M. X. 1905 Synthetic culture media and the can Association for the Advancement of Science, Wash- biochemistry of bacterial pigments. J. Med. Research, ington, D. C. 14, 109-160. GANGULY, S. 1955 A new riboflavin-producing bacterium. TOTTER, J. R. AND MOSELEY, F. T. 1953 Influence of con- Its identification and characterization. Ann. Biochem. centration of iron on the production of fluorescin by and Exptl. Med. (Calcutta) 15, (4) 215-218. Biol. Ab- Pseudomonas aeruginosa. J. Bacteriol., 65, 45-47. stracts, 30, 31806 (1956). TURFITT, G. E. 1937 XXV. Bacteriological and biochemical in GIRAL, F. 1936 Sobre los liocromos caracteristicos del relationships the pyocyaneus-fluorescens group. II. glUpo Investigations on the green fluorescent pigment. Bio- de bacterias fluorescentes. Anales soc. espafi. fis. y chem. J., 31, 212-218. quim., 34, 667-693. TURFREIJER, A. 1941 Pyoverdinen, de groen fluoresceerende HAYNES, W. C. 1953 Key to the species of genus Pseudo- kleurstoffen van Pseudomonas fluorescens. Doctorate monas. Mimeo. 7 pp. Presented at Symposium on Thesis. U. of Amsterdam, 85 pp. Brit. Abstr., 16, 16578 Taxonomy, Aug. 13, 1953. Fifty-third general meeting (1942). of the Society of American Bacteriologists, San Francisco. TURFREIJER, A., WIBAUT, J. P., AND KINGMA BOLTJES, T. Y. KING, J. V., CAMPBELL, J. J. R., AND EAGLES, B. A. 1948 1938. The green fluorescent pigment of Pseudomnonas Mineral requirements for fluorescin production (by Pseudo- fluorescens. Rec. trav. chim., 57, 1397-1404.