548
Biochimica et Biophysica Acta, 497 (1977) 548--557
© Elsevier/North-Holland Biomedical Press
BBA 28221 PRODUCTION OF E N T E R O C H E L I N BY E S C t t E R I C H I A
C O L I 0111
HENRY J. ROGERS, C. SYNGE, B. KIMBER and P.M. BAYLEY National Institute for Medical Research, Mill Hill, London N W 7 1AA (U.K.)
(Received October 18th, 1976)
S ummar y The major neutral iron-transporting c o m p o u n d produced by E s c h e r i c h i a c o l i 0 1 1 1 / K 5 8 / H 2 has been isolated from iron-deficient cultures of the organism and co mp ar ed with the corresponding c o m p o u n d , enterochelin, p r o d u c e d by E. coli K12. The p r o d u c t contained serine and 2 , 3 - d i h y d r o x y b e n z o i c acid and f o r m e d a complex with Fe 3÷. Since the PMR spectra of the products from the two strains were identical, it was concluded that E. c o l i 0111 also secreted enterochelin u n de r iron-deficient conditions. Although it was n o t possible to establish the optical configuration of the serine residues in the molecule, the CD spectra o f the metal free and Fe 3÷, complexes were f o u n d to be of the same sign and magnitude. The spectra show that metal binding results in considerable conf or m a t i ona l changes in the enterochelin molecule. The biological properties of the two c o m p o u n d s appear to be identical as judged by their ability to abolish the bacteriostatic e f f e c t of serum on E. c o l i 0111.
Introduction A considerable a m o u n t of information is now available on the ability of iron c o m p o u n d s to enhance the growth of pathogenic bacteria bot h in serum in vitro and in animals [1]. In the case of E s c h e r i c h i a c o l i 0111, iron c o m p o u n d s can enhance the virulence 10 000-fold [2] and also abolish the bactericidal and bacteriostatic effects of serum [3]. F u r t h e r experiments suggested that highly virulent bacteria possess a mechanism for utilising transferrin-bound iron and t h at one f u n ction of a nt i body is to interfere with this process. This c o n c e p t was supported by comparative studies using E. c o l i 0111 and 0141; the latter is m o r e virulent than the f o r m e r and pr od uced iron-transporting c o m p o u n d s when growing in serum. The f or m er p r o d u c e d similar c o m p o u n d s in a synthetic medium but failed to multiply in serum. In this case it was concluded that antib o d y interfered either with the synthesis or release of the iron-transporting c o mp o u n d s . One of the c o m p o u n d s p r o d u c e d by E. c o l i 0141 was tentatively
549 identified as the m e t h y l ester of the linear trimer of 2 , 3 - d i h y d r o x y b e n z o y l serine [4]. Th e cyclic trimer of 2 , 3 - d i h y d r o x y b e n z o y l serine, enterochelin, has been isolated f r om iron-deficient cultures of E. coli K12, Aerobacter aerogenes [5] and Salmonella t y p h i m u r i u m LT2 [6]. This paper describes the isolation o f enterochelin f r om culture fluids of E. coli 0111, a strain causing gastroenteritis in infants. Materials Horse serum No. 3 was obtained in the frozen state from Wellcome Reagents Ltd. and stored at --20 ~C. Crystalline D-amino acid oxidase from hog kidney and L-amino acid oxidase t y p e I were obtained from Sigma. Chemicals were of analytical grade where available. E t h y l e n e diamine d i - o - h y d r o x y p h e n y l acetic acid (EDDA) [7] was estimated spectroscopically after reacting 10.0 ml of an approx. 1 mM solution in 0.01 M NaOH with 30.0 ml 0.25 mM ferrous a m m o n i u m sulphate using the extinction M coefficient E478 = 5.36 X 103. The technical p r o d u c t obtained from K and K Labs. Inc., which contained 65% EDDA was t herefore purified by the following m e t h o d . 14 g was stirred at 50°C for 30 min with 1 1 H20 containing 5.1 g b r o wn ferric a m m o n i u m citrate. The red solution obtained was clarified by filtration. After cooling to r o o m t em pe r at ure, solid citric acid was stirred in until th e pH fell to 1.8. The exchange reaction was allowed to proceed for 24 h at r o o m temperature. The pH was raised to 2.5 by careful addition of concentrated NH4OH. Crystallisation c o m m e n c e d immediately and after 24 h at 4 ° C, the p r o d u c t was filtered of f and washed with ice-cold water. On drying, 3.9 g of EDDA > 95% pure was obtained. Samples for the m e a s u r e m e n t of circular dichroism were dissolved in 50% ethanol. Th e gallium com pl e x of enterochelin was prepared by mixing 1.0 ml f o 2.0 mM enterochelin in ethanol with 4.0 ml o f 0.5 mM Ga(NO3)3 in 0.015 M HC104. Th e pH was raised to 7.35 by adding 1.0 ml of 0.1 M Tris. The final c o n c e n t r a t i o n of the com pl e x was 0.33 mM. E. coli 0 1 1 1 / K 5 8 (B4)/H2 has been used in all previous work [2--4]. E. coli K12 Hfr H thi- was obtained from R.S. B u x t o n o f this institute [8]. Methods Bacterial cultures. Logarithmic phase brot h cultures were used as the source o f inocula in all cases. Bacteria were collected by centrifugation and suspended in 10% (v/v) papain digest broth in saline. The bacterial population was estimated by n e p h e l o m e t r y with the aid of a standard curve. Production o f iron-transporting catechols. Bot h organisms were grown in 4 1 volumes o f medi um 56 f r o m which FeSO4 was o m i t t e d [9]. Glucose, 10 g per 1 was e m p l o y e d as carbon source and a trace element m i xt ure was added [10]. F or the growth of E. coli 0111, 2 #M ferric a m m o n i u m citrate was added whilst 5 pM ferric a m m o n i u m citrate plus 1 mg thiamine per 1 was added for the growth o f E. coli K12. Media were sterilised by m e m b r a n e filtration using 0.45 pM filters. EDDA dissolved in 5.0 ml 1 M NaOH was added to give a final c o n c e n t r a t i o n of 10 pM. The inocula were grown in 400 ml of the same
550 medium w i t h o u t EDDA, the t e m per a t ur e was 37°C and the pH maintained at 7.5 by automatic addition of 5 M NaOH. The aeration rate was initially 4 1 per min but when the bacterial population reached 6 • 108--7 • 108 per ml the culture was gassed with 2 1 oxygen per min. The bacterial population was estimated by n e p h e l o m e t r y and the c o n t e n t of neutral catechols was estimated by shaking 20 ml whole culture at pH 6.5 with 20 ml ethyl acetate and then shaking the ex tr a c t with 5.0 ml 1.0 mM ferric nitrilotriacetate, pH 8.0 [11] to form the red ferric complex. The extinction coefficient for the enterochelin M - 4.92 • 103 was used to estimate the catechol cont ent . iron co mp lex E488 At the end o f logarithmic growth (approx. 2 × 109 bacteria per ml) the pH of the whole culture was lowered to 6.5 by adding cone. HC1 and then immediately e xtr acted with 2 × 1200 ml ethyl acetate. Isolation o f neutral iron-binding compounds. The ethyl acetate extract was c on cen tr ated by r ot a r y evaporation at 37°C to approx. 100 ml and then shaken with 80 ml 2.0 mM ferric nitrilotriacetate (1.1 equivalents), pH 6.0. The aqueous phase was washed with 2 × 100 ml ethyl acetate after which the iron complex was d e c o m p o s e d by lowering the pH to 3.0 with 1 M HC1 and the p r o d u c t was extracted with 2 × 100 ml ethyl acetate which was then dried over anhydrous Na2SO4. The ethyl acetate solution was c o n c e n t r a t e d to dryness and the residue dissolved in 2.0 ml n-butanone. On careful addition of 3.0 ml benzene, 3.3 mg o f a brown precipitate was obtained (fraction I). Addition of a further 10 ml benzene p r o d u c e d a fine semi-crystalline material which was collected by centrifugation after 24 h at r oom t e m per a t ure and then washed with 5.0 ml 10% (v/v) n-butanone in benzene. After drying in air at 37°C, 38 mg of an offwhite p o w d e r was obtained (fraction II). Paper chromatography. Ascending c h r o m a t o g r a p h y was carried out on Whatman No. 1 paper using either 5% a m m o n i u m form at e plus 0.5% formic acid [5] or 6% acetic acid as solvents. The c o m p o u n d s were detected both by their fluorescence in ultraviolet light and the p r o d u c t i o n of Prussian blue after spraying with 1% K3Fe(CN)6 + 2% ferric a m m o n i u m citrate followed by washing in dilute HC1 [12]. A m i n o acid analysis. Approx. 0.5 pM of the c o m p o u n d was heated for 16 h at 100°C in 4 M HC1 in vacuo [13]. The hydrolysate was dried down in vacuo over conc. H2SO4 and NaOH pellets and then analysed on a Beckman model 120C amino acid analyser. Treatment with amino acid oxidases. Samples containing approx. 0.5 pmol were h y d r o ly s ed as described above and then diluted to 10.0 ml with water, 3 × 3.0-ml aliquots were dried in vacuo. One sample was incubated for 1 h at 37°C with 0.2 Sigma units of L-amino acid oxidase in 1.0 ml 0.05 M sodium phosphate, pH 7.2 [14]. A second sample was incubated for 1 h at 37°C in 1.0 ml 0.03 M sodium p y r o p h o s p h a t e , pH 8.3, containing 0.5 mg bovine serum albumin (fraction V), 0.1 mg FAD and 0.03 Sigma units of D-amino acid oxidase [15]. The third sample was incubated in the absence of D-amino acid oxidase. The reactions were terminated by adding 0.5 ml 1 M HC1, the samples were taken to dryness and then analysed for serine. Proton magnetic resonance spectra. The c o m p o u n d from E. coli 0111 (2.7 mg) and enterochelin from E. coli K12 (3.3 mg) were each dissolved in 0.5 ml D6 acetone. The samples were measured in 5-mm NMR tubes using a Varian -
-
551 XL-100-15 spectrometer in Fourier Transform mode; 8k data points were acquired resulting in 4096 o u t p u t data points in the phase-corrected absorption spectra. A sensitivity enhancement of 2.0 s was used. The spectra are referenced to tetramethyl silane in ppm (5). Circular dichroism measurements. CD spectra were recorded on a Jouan Dichrographe II (150 W xenon arc) with solutions at room temperature in quartz cells of path-length 1.0 cm to 1.0 mm. Concentrations were in the range 0.1--0.4 mM to keep sample absorbance less than 1.0 throughout. The experimental data were digitised at 0.5-nm intervals and processed by computer. Far ultraviolet data was smoothed by serial application of a 20 point, second-order polynomial; all data was plotted by Calcomp routines, in terms of molar circular dichroism Ae = ( e L - eR) (M -1 " cm -1) [16]. Growth curves o f E. coli 0111 in horse serum. In order to test the growthpromoting effects of the compounds a 1 mM solution in ethanol was suitably diluted in saline and 0.03 ml added to 3.0 ml horse serum under 5% CO2 + 10% 02 + 85% N2 at 37°C. Viable bacterial counts were made as previously described [4]. Results Production and isolation o f neutral catechols. Table I shows the bacterial population and neutral catechol contents obtained when E. coli 0111 and K12 were grown in medium 56 containing 10/aM EDDA at pH 7.5. It is probable that the high affinity of EDDA for Fe 3÷ (Kassoc = 1033"9) [7] limits the rate of supply of iron to the cell below that required for the induction of enterochelin synthesis. Satisfactory yields of enterochelin were obtained from E. coli K12 by adding 5 pM Fe 3÷. E. coli 0111 produced no enterochelin in the presence of 5 pM Fe 3÷ but did so when the concentration was lowered to 2 pM. Gassing with pure oxygen during the final 2 h of growth increased the yield of catechols but no compounds were produced at pH 6.8. In view of the sensitivity of enterochelin itself to both enzymic [17] and alkaline [5] hydrolysis, it was decided to employ a rapid extraction procedure which would leave compounds such as the acidic degradation products of enterochelin [5] in the culture fluid. This was achieved by extracting the whole culture at pH 6.5 with ethyl acetate. The TABLE I CATECHOL T i m e (h)
4.5 5.5 6.5 7.5
PRODUCTION Strain:
B Y E. C O L I G R O W I N G
IN M E D I U M 56 a
01 11 b
K12 c
Bacteria (m1-1 )
Catechols (rag/l)
Bacteria (m1-1 )
Catechols (mg/1)
5.4 6.6 • 1.1 • 1.5.
5.4 7.8 15.8 16.5
7.7 • 1.8 1.8 1.8.
7.7 9.1 17.4 21.8
108 108 109 109
a 10 pM EDDA added. b M e d i u m c o n t a i n e d 2 p M F e a+. e M e d i u m c o n t a i n e d 5 ~tM F e 3+.
108 109 109 109
552
catechols were separated from other neutral components present in the extract by conversion to the water-soluble iron complexes. Decomposition of these complexes by lowering the pH to 3 allowed the catechols to be re-extracted into ethyl acetate and after removal of the ethyl acetate the compounds were recovered as semi-crystalline solids from n-butanone by addition of benzene. Approx. 50% of the material present in the culture fluid was recovered by this method. Paper chromatography of the products. It was found that as the products from the two strains were purified, their RFvalues in the formate system [5] decreased, the pure products remaining at the origin. The two products had identical RFvalues in 6% acetic acid (0.29), n-butanol/acetic acid/water (4 : 1 : 5, v/v) (0.9) and benzene/acetic acid/water, (125 : 72 : 3, v/v) (0.9) [5]. Both compounds gave a weak, pale blue fluorescence under ultraviolet light and reacted strongly with the Fe3+-ferricyanide reagent which is specific for o- and p-dihydroxybenzene compounds [ 12 ]. Elemental analysis o f the product from E. coli 0111. Found: C, 53.76; H, 4.22; N, 6.35%. (A. Bernhardt, Elbach fi Engleskirchen, West Germany). C1oH905 requires: C, 53.80; H, 4.07; N, 6.28%. These figures are very similar to those obtained for enterochelin from E. coli K12 [5]. Reaction with ferric nitrilotriacetate. 3 ml of a solution containing 0.0637 mg per ml of the product from E. coli 0111 was titrated spectroscopically with 50-pl aliquots of 1.0 mM ferric nitrilotriacetate (Fig. 1). The end point indicates an equivalent weight of 631 whilst the extinction coefficient based on Fe 3+ is EMs = 4.92 • 103. Degradation products obtained by acid hydrolysis. After heating for 16 h in 4 M HC1 at l l 0 ° C in vacuo [13] amino acid analysis showed that serine was the only amino acid present in both products. Thus 0.522 pm, based on the reaction with Fe 3+ of the product, from E. coli 0111 gave 1.45 pmol serine or 2.79 M per M
0.5
v
w
0,4--
0.3
0.2 c
~
0.1
Biochimica et Biophysica Acta, 497 (1977) 548--557
© Elsevier/North-Holland Biomedical Press
BBA 28221 PRODUCTION OF E N T E R O C H E L I N BY E S C t t E R I C H I A
C O L I 0111
HENRY J. ROGERS, C. SYNGE, B. KIMBER and P.M. BAYLEY National Institute for Medical Research, Mill Hill, London N W 7 1AA (U.K.)
(Received October 18th, 1976)
S ummar y The major neutral iron-transporting c o m p o u n d produced by E s c h e r i c h i a c o l i 0 1 1 1 / K 5 8 / H 2 has been isolated from iron-deficient cultures of the organism and co mp ar ed with the corresponding c o m p o u n d , enterochelin, p r o d u c e d by E. coli K12. The p r o d u c t contained serine and 2 , 3 - d i h y d r o x y b e n z o i c acid and f o r m e d a complex with Fe 3÷. Since the PMR spectra of the products from the two strains were identical, it was concluded that E. c o l i 0111 also secreted enterochelin u n de r iron-deficient conditions. Although it was n o t possible to establish the optical configuration of the serine residues in the molecule, the CD spectra o f the metal free and Fe 3÷, complexes were f o u n d to be of the same sign and magnitude. The spectra show that metal binding results in considerable conf or m a t i ona l changes in the enterochelin molecule. The biological properties of the two c o m p o u n d s appear to be identical as judged by their ability to abolish the bacteriostatic e f f e c t of serum on E. c o l i 0111.
Introduction A considerable a m o u n t of information is now available on the ability of iron c o m p o u n d s to enhance the growth of pathogenic bacteria bot h in serum in vitro and in animals [1]. In the case of E s c h e r i c h i a c o l i 0111, iron c o m p o u n d s can enhance the virulence 10 000-fold [2] and also abolish the bactericidal and bacteriostatic effects of serum [3]. F u r t h e r experiments suggested that highly virulent bacteria possess a mechanism for utilising transferrin-bound iron and t h at one f u n ction of a nt i body is to interfere with this process. This c o n c e p t was supported by comparative studies using E. c o l i 0111 and 0141; the latter is m o r e virulent than the f o r m e r and pr od uced iron-transporting c o m p o u n d s when growing in serum. The f or m er p r o d u c e d similar c o m p o u n d s in a synthetic medium but failed to multiply in serum. In this case it was concluded that antib o d y interfered either with the synthesis or release of the iron-transporting c o mp o u n d s . One of the c o m p o u n d s p r o d u c e d by E. c o l i 0141 was tentatively
549 identified as the m e t h y l ester of the linear trimer of 2 , 3 - d i h y d r o x y b e n z o y l serine [4]. Th e cyclic trimer of 2 , 3 - d i h y d r o x y b e n z o y l serine, enterochelin, has been isolated f r om iron-deficient cultures of E. coli K12, Aerobacter aerogenes [5] and Salmonella t y p h i m u r i u m LT2 [6]. This paper describes the isolation o f enterochelin f r om culture fluids of E. coli 0111, a strain causing gastroenteritis in infants. Materials Horse serum No. 3 was obtained in the frozen state from Wellcome Reagents Ltd. and stored at --20 ~C. Crystalline D-amino acid oxidase from hog kidney and L-amino acid oxidase t y p e I were obtained from Sigma. Chemicals were of analytical grade where available. E t h y l e n e diamine d i - o - h y d r o x y p h e n y l acetic acid (EDDA) [7] was estimated spectroscopically after reacting 10.0 ml of an approx. 1 mM solution in 0.01 M NaOH with 30.0 ml 0.25 mM ferrous a m m o n i u m sulphate using the extinction M coefficient E478 = 5.36 X 103. The technical p r o d u c t obtained from K and K Labs. Inc., which contained 65% EDDA was t herefore purified by the following m e t h o d . 14 g was stirred at 50°C for 30 min with 1 1 H20 containing 5.1 g b r o wn ferric a m m o n i u m citrate. The red solution obtained was clarified by filtration. After cooling to r o o m t em pe r at ure, solid citric acid was stirred in until th e pH fell to 1.8. The exchange reaction was allowed to proceed for 24 h at r o o m temperature. The pH was raised to 2.5 by careful addition of concentrated NH4OH. Crystallisation c o m m e n c e d immediately and after 24 h at 4 ° C, the p r o d u c t was filtered of f and washed with ice-cold water. On drying, 3.9 g of EDDA > 95% pure was obtained. Samples for the m e a s u r e m e n t of circular dichroism were dissolved in 50% ethanol. Th e gallium com pl e x of enterochelin was prepared by mixing 1.0 ml f o 2.0 mM enterochelin in ethanol with 4.0 ml o f 0.5 mM Ga(NO3)3 in 0.015 M HC104. Th e pH was raised to 7.35 by adding 1.0 ml of 0.1 M Tris. The final c o n c e n t r a t i o n of the com pl e x was 0.33 mM. E. coli 0 1 1 1 / K 5 8 (B4)/H2 has been used in all previous work [2--4]. E. coli K12 Hfr H thi- was obtained from R.S. B u x t o n o f this institute [8]. Methods Bacterial cultures. Logarithmic phase brot h cultures were used as the source o f inocula in all cases. Bacteria were collected by centrifugation and suspended in 10% (v/v) papain digest broth in saline. The bacterial population was estimated by n e p h e l o m e t r y with the aid of a standard curve. Production o f iron-transporting catechols. Bot h organisms were grown in 4 1 volumes o f medi um 56 f r o m which FeSO4 was o m i t t e d [9]. Glucose, 10 g per 1 was e m p l o y e d as carbon source and a trace element m i xt ure was added [10]. F or the growth of E. coli 0111, 2 #M ferric a m m o n i u m citrate was added whilst 5 pM ferric a m m o n i u m citrate plus 1 mg thiamine per 1 was added for the growth o f E. coli K12. Media were sterilised by m e m b r a n e filtration using 0.45 pM filters. EDDA dissolved in 5.0 ml 1 M NaOH was added to give a final c o n c e n t r a t i o n of 10 pM. The inocula were grown in 400 ml of the same
550 medium w i t h o u t EDDA, the t e m per a t ur e was 37°C and the pH maintained at 7.5 by automatic addition of 5 M NaOH. The aeration rate was initially 4 1 per min but when the bacterial population reached 6 • 108--7 • 108 per ml the culture was gassed with 2 1 oxygen per min. The bacterial population was estimated by n e p h e l o m e t r y and the c o n t e n t of neutral catechols was estimated by shaking 20 ml whole culture at pH 6.5 with 20 ml ethyl acetate and then shaking the ex tr a c t with 5.0 ml 1.0 mM ferric nitrilotriacetate, pH 8.0 [11] to form the red ferric complex. The extinction coefficient for the enterochelin M - 4.92 • 103 was used to estimate the catechol cont ent . iron co mp lex E488 At the end o f logarithmic growth (approx. 2 × 109 bacteria per ml) the pH of the whole culture was lowered to 6.5 by adding cone. HC1 and then immediately e xtr acted with 2 × 1200 ml ethyl acetate. Isolation o f neutral iron-binding compounds. The ethyl acetate extract was c on cen tr ated by r ot a r y evaporation at 37°C to approx. 100 ml and then shaken with 80 ml 2.0 mM ferric nitrilotriacetate (1.1 equivalents), pH 6.0. The aqueous phase was washed with 2 × 100 ml ethyl acetate after which the iron complex was d e c o m p o s e d by lowering the pH to 3.0 with 1 M HC1 and the p r o d u c t was extracted with 2 × 100 ml ethyl acetate which was then dried over anhydrous Na2SO4. The ethyl acetate solution was c o n c e n t r a t e d to dryness and the residue dissolved in 2.0 ml n-butanone. On careful addition of 3.0 ml benzene, 3.3 mg o f a brown precipitate was obtained (fraction I). Addition of a further 10 ml benzene p r o d u c e d a fine semi-crystalline material which was collected by centrifugation after 24 h at r oom t e m per a t ure and then washed with 5.0 ml 10% (v/v) n-butanone in benzene. After drying in air at 37°C, 38 mg of an offwhite p o w d e r was obtained (fraction II). Paper chromatography. Ascending c h r o m a t o g r a p h y was carried out on Whatman No. 1 paper using either 5% a m m o n i u m form at e plus 0.5% formic acid [5] or 6% acetic acid as solvents. The c o m p o u n d s were detected both by their fluorescence in ultraviolet light and the p r o d u c t i o n of Prussian blue after spraying with 1% K3Fe(CN)6 + 2% ferric a m m o n i u m citrate followed by washing in dilute HC1 [12]. A m i n o acid analysis. Approx. 0.5 pM of the c o m p o u n d was heated for 16 h at 100°C in 4 M HC1 in vacuo [13]. The hydrolysate was dried down in vacuo over conc. H2SO4 and NaOH pellets and then analysed on a Beckman model 120C amino acid analyser. Treatment with amino acid oxidases. Samples containing approx. 0.5 pmol were h y d r o ly s ed as described above and then diluted to 10.0 ml with water, 3 × 3.0-ml aliquots were dried in vacuo. One sample was incubated for 1 h at 37°C with 0.2 Sigma units of L-amino acid oxidase in 1.0 ml 0.05 M sodium phosphate, pH 7.2 [14]. A second sample was incubated for 1 h at 37°C in 1.0 ml 0.03 M sodium p y r o p h o s p h a t e , pH 8.3, containing 0.5 mg bovine serum albumin (fraction V), 0.1 mg FAD and 0.03 Sigma units of D-amino acid oxidase [15]. The third sample was incubated in the absence of D-amino acid oxidase. The reactions were terminated by adding 0.5 ml 1 M HC1, the samples were taken to dryness and then analysed for serine. Proton magnetic resonance spectra. The c o m p o u n d from E. coli 0111 (2.7 mg) and enterochelin from E. coli K12 (3.3 mg) were each dissolved in 0.5 ml D6 acetone. The samples were measured in 5-mm NMR tubes using a Varian -
-
551 XL-100-15 spectrometer in Fourier Transform mode; 8k data points were acquired resulting in 4096 o u t p u t data points in the phase-corrected absorption spectra. A sensitivity enhancement of 2.0 s was used. The spectra are referenced to tetramethyl silane in ppm (5). Circular dichroism measurements. CD spectra were recorded on a Jouan Dichrographe II (150 W xenon arc) with solutions at room temperature in quartz cells of path-length 1.0 cm to 1.0 mm. Concentrations were in the range 0.1--0.4 mM to keep sample absorbance less than 1.0 throughout. The experimental data were digitised at 0.5-nm intervals and processed by computer. Far ultraviolet data was smoothed by serial application of a 20 point, second-order polynomial; all data was plotted by Calcomp routines, in terms of molar circular dichroism Ae = ( e L - eR) (M -1 " cm -1) [16]. Growth curves o f E. coli 0111 in horse serum. In order to test the growthpromoting effects of the compounds a 1 mM solution in ethanol was suitably diluted in saline and 0.03 ml added to 3.0 ml horse serum under 5% CO2 + 10% 02 + 85% N2 at 37°C. Viable bacterial counts were made as previously described [4]. Results Production and isolation o f neutral catechols. Table I shows the bacterial population and neutral catechol contents obtained when E. coli 0111 and K12 were grown in medium 56 containing 10/aM EDDA at pH 7.5. It is probable that the high affinity of EDDA for Fe 3÷ (Kassoc = 1033"9) [7] limits the rate of supply of iron to the cell below that required for the induction of enterochelin synthesis. Satisfactory yields of enterochelin were obtained from E. coli K12 by adding 5 pM Fe 3÷. E. coli 0111 produced no enterochelin in the presence of 5 pM Fe 3÷ but did so when the concentration was lowered to 2 pM. Gassing with pure oxygen during the final 2 h of growth increased the yield of catechols but no compounds were produced at pH 6.8. In view of the sensitivity of enterochelin itself to both enzymic [17] and alkaline [5] hydrolysis, it was decided to employ a rapid extraction procedure which would leave compounds such as the acidic degradation products of enterochelin [5] in the culture fluid. This was achieved by extracting the whole culture at pH 6.5 with ethyl acetate. The TABLE I CATECHOL T i m e (h)
4.5 5.5 6.5 7.5
PRODUCTION Strain:
B Y E. C O L I G R O W I N G
IN M E D I U M 56 a
01 11 b
K12 c
Bacteria (m1-1 )
Catechols (rag/l)
Bacteria (m1-1 )
Catechols (mg/1)
5.4 6.6 • 1.1 • 1.5.
5.4 7.8 15.8 16.5
7.7 • 1.8 1.8 1.8.
7.7 9.1 17.4 21.8
108 108 109 109
a 10 pM EDDA added. b M e d i u m c o n t a i n e d 2 p M F e a+. e M e d i u m c o n t a i n e d 5 ~tM F e 3+.
108 109 109 109
552
catechols were separated from other neutral components present in the extract by conversion to the water-soluble iron complexes. Decomposition of these complexes by lowering the pH to 3 allowed the catechols to be re-extracted into ethyl acetate and after removal of the ethyl acetate the compounds were recovered as semi-crystalline solids from n-butanone by addition of benzene. Approx. 50% of the material present in the culture fluid was recovered by this method. Paper chromatography of the products. It was found that as the products from the two strains were purified, their RFvalues in the formate system [5] decreased, the pure products remaining at the origin. The two products had identical RFvalues in 6% acetic acid (0.29), n-butanol/acetic acid/water (4 : 1 : 5, v/v) (0.9) and benzene/acetic acid/water, (125 : 72 : 3, v/v) (0.9) [5]. Both compounds gave a weak, pale blue fluorescence under ultraviolet light and reacted strongly with the Fe3+-ferricyanide reagent which is specific for o- and p-dihydroxybenzene compounds [ 12 ]. Elemental analysis o f the product from E. coli 0111. Found: C, 53.76; H, 4.22; N, 6.35%. (A. Bernhardt, Elbach fi Engleskirchen, West Germany). C1oH905 requires: C, 53.80; H, 4.07; N, 6.28%. These figures are very similar to those obtained for enterochelin from E. coli K12 [5]. Reaction with ferric nitrilotriacetate. 3 ml of a solution containing 0.0637 mg per ml of the product from E. coli 0111 was titrated spectroscopically with 50-pl aliquots of 1.0 mM ferric nitrilotriacetate (Fig. 1). The end point indicates an equivalent weight of 631 whilst the extinction coefficient based on Fe 3+ is EMs = 4.92 • 103. Degradation products obtained by acid hydrolysis. After heating for 16 h in 4 M HC1 at l l 0 ° C in vacuo [13] amino acid analysis showed that serine was the only amino acid present in both products. Thus 0.522 pm, based on the reaction with Fe 3+ of the product, from E. coli 0111 gave 1.45 pmol serine or 2.79 M per M
0.5
v
w
0,4--
0.3
0.2 c
~
0.1
