Role of the entC gene in enterobactin and menaquinone biosynthesis in Escherichia coli

ARCHIVES

OF BIOCHEMISTRY

Vol. 276, No. 2, February

AND

BIOPHYSICS

1, pp. 331-335,199O

Role of the entC Gene in Enterobactin Biosynthesis in Escherichia co/i Annette Institut

Kaiser

and Eckhard

fiir Pharmazeutische

Leistner’

Biologic, Rheinische Friedrich-

Received June 1,1989, and in revised form September

Wilhelms-Universitat

Isochorismic acid is a metabolite that may be formed from chorismic acid either by an addition-elimination reaction (1) or by a sigmatropic shift (2). The enzyme catalyzing this reaction has been termed isochorismic synthetase (3, 4). Isochorismic acid is the precursor of the salicylic acid moiety in mycobactins (5), the 2,3dihydroxybenzoic acid units in enterobactin (enterochelin), an iron ionophore (6), and o-succinylbenzoic acid should be addressed.

000%9861/90 $3.00 Copyright Q 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

Bonn, Bonn, Federal Republic of Germany

11,1989

TnlO mutants of Escherichia coli MC4100 were screened for their inability to grow under iron deficiency and for their inability to grow under anaerobiosis in the presence of fumarate as an electron acceptor. A strain so obtained (E. coli PBBl) lacked the ability to convert chorismic acid to isochorismic acid. This shows that the gene (ento encoding isochorismate synthase was mutated. E. coli PBB 1 did not produce any detectable amounts of menaquinones (vitamin K,) or enterobactin. When supplemented with isochorismic acid this strain produced menaquinones, indicating that isochorismic acid is involved not only in enterobactin but also in menaquinone biosynthesis. The entC gene was isolated and was shown to be part of the enterobactin gene cluster: It was located on a DNA fragment (9 kb in length) which also carried the entA gene. The DNA fragment was identified by restriction site mapping and was compared to a previously published map of the enterobactin gene cluster. The entC gene on this fragment responds not only to conditions (iron deficiency) that stimulate enterobactin biosynthesis but also to anaerobiosis which results in increased isochorismic acid formation and increased menaquinone biosynthesis. We conclude that isochorismic acid, isochorismic synthase, and the gene (en%) encoding this enzyme are involved in catalytic events at a metabolic branch point from which both enterobactin and menaquinones originate. ‘c 1980 Academic Press, Inc.

’ To whom correspondence

and Menaquinone

(OSB) (7-9) which in turn is an intermediate in menaquinone (vitamin K,) (10-12) (Fig. l), anthraquinone (13), and pierardine (14) biosynthesis. An enzyme system from Escherichia coli that catalyzes the formation of OSB’ was first described by Meganathan (15). His experiments were based on the findings of Young (16) and Young et al. (17) who assumed that E. coli AN191 (entC) and E. coli AN154 (entC) are blocked in the conversion of chorismic to isochorismic acid and produced normal levels of menaquinones. This seemed to indicate that chorismic acid is the precursor of OSB. It was later shown, however, that the levels of menaquinone present in these mutants are not normal and that the mutants seemed to be “leaky.” In addition, experiments on the OSB synthase showed that isochorismic rather than chorismic acid was the true precursor of this reaction series (7, 8). Thus, isochorismic acid is an important metabolite from which both enterobactin and menaquinones originate (Fig. 1). Enterobactin is an ionophore which is essential for iron uptake during aerobic growth, whereas menaquinones are involved in electron transport processes (18, 19). Thus, isochorismate synthase has a dual function. The enzyme is encoded by the entC gene which is part of the enterobactin gene cluster in E. coli (20-28). This cluster maps at 13.8 min on the E. coli linkage map (29). We describe in this paper experiments which corroborate our previous finding (7, 8) that isochorismate synthase and the entC gene not only play a role in enterobactin biosynthesis but also are functionally related to the men genes encoding enzymes which catalyze the formation of menaquinones. These genes are located at 49 and 88 min on the E. coli linkage map (29). We also show that the enterobactin gene cluster responds to conditions (anaerobiosis) which stimulate menaquinone biosynthesis. ’ Abbreviations used: OSB, o-succinylbenzoic acid; MK8, menaquinone with 8 isoprene units; DMKB, demethylmenaquinone with 8 isoprene units; TSB, tryptic soy broth; men, gene encoding a protein of the menaquinone pathway; ent, gene encoding a protein of the enterobactin pathway; TBE, Tris-borate-EDTA buffer. 331

332

KAISER

AND

LEISTNER

cool-l HOOC

(..+oo;;‘

HOOC Y

OSB

C$gismic

FIG. 1. Branched pathway leading to enterobactin

AND

METHODS

Bacteria and phuges. E. coli MC4100, phage XNK370, and cosmid pHC79 were kindly provided by Dr. A. Bock, Institut fur Mikrobiologie Ludwig Maximilians-Universitiit Miinchen, FRG. Cosmids of pHC79 carrying genomic DNA of E. coli MC4100 were provided by W. Leinfelder from the same institute. E. coli strains AN92 and AN154 were from Dr. I. G. Young, Canberra, Australia. Genotypes of the strains used in this study are listed in Table I. Media for determination of isochorismate synthase activity and growth curves. Bacteria were grown in tryptic soy broth (30 g liter-‘, TSB medium, Difco). A preculture (150 ml TSB medium) was grown for 12 h at 37°C and 150 rpm on a rotary shaker and used to inoculate 1 liter medium which was also kept at 37°C for 12 h, however, on a reciprocal shaker at 70 strokes per minute. When grown anaerobically, bacteria were incubated in TSB medium (30 g liter-‘) with Na,fumarate (6.4 g liter-‘) and glycerol (3.683 g liter-‘) at pH 7.0. PBBP and PBB3 were kept in the presence of either 30 pg ml-’ (PBBP) or 100 pg ml i (PBB3) ampicillin. Solid media contained 16 g liter-’ agar and were kept in a Becton-Dickinson Gaspak incubator under anaerobiosis. Colonies were used to inoculate a preculture (100 ml medium) with a 2-cm paraffin layer on top. After 24 h at 37°C this culture was used to inoculate the main culture (1 liter) with a lo-cm paraffin layer which was also incubated at 37°C for 24 h.

TABLE I E. coli Strains

and Phage XNK370 Used in This Study

Strain

Genotype

Reference

MC4100

F ara D 139, (arg F-lac) U 169, pts F 25, deo Cl, rel A 1 fEbB 5301, rpsL 150, X proA2, argE3, pheA1, tyrA4, trp-401, aroB Same as strain AN92, but entC401 Same as strain AN92, but entA403 Same as strain AN92, but entC401 Same as MC4100 but entC Same as AN154, but recA, tet’ with cosmid pHC79 carrying genomic DNA (20-30 kb) including the entC gene Same as PBBP, but pHC79 shortened to 9 kb carrying entA and entC b221cI857 cl1 71: :TnlO O(UGA)261

(30)

AN92 AN154 AN156 AN191 PBBl PBB2

PBB3 Phage XNK370

-

Menaquinones

’

iso-Chorismic acid

2,3-Dihydro2,3-dihydroxybenzoic acid

MATERIALS

-

(17) (17) (17) (17) This study This study

This study (31)

2,3-Dihydroxybenzoic acid

and menaquinones.

Growth under iron deficiency was carried out in a medium (24) containing the following components (amount per liter): NaCl(5.8 g), KC1 (3.7 g), CaC1,.2H,O (0.15 g), MgC12.6H,0 (0.10 g), NH&l (1.1 g), Na$O, (0.142 g), KH2P0, (0.272 g), Na,succinate (5 g), thiamine HCl (0.004 g), trishydroxymethylaminomethane (12.1 g), and the L-amino acids (0.04 g) Arg, Trp, Tyr, Phe, and Pro. a,a’-Dipyridyl (10m4M) was added as an iron chelator. The medium was adjusted to pH 7.4 with concentrated HCl. When PBB2 or PBB3 was grown under iron deficiency the medium again contained ampicillin. A preculture contained 150 ml medium and was incubated at 37°C for 36 h on a rotary shaker (150 rpm). The main culture (1 liter) was inoculated with the complete preculture and incubated at 37°C for 24 h on a reciprocal shaker at 70 strokes per minute. Collection of cells. After incubation cells were collected by centrifugation (4”C, 5000 rpm, 20 min). The spent medium was decanted and the pellet washed (KH,PO, buffer, pH 7.O,O”C, 0.02 M) and again centrifuged to yield (g liter ’ fresh wt) 8.0 (E. coli AN154), 7.0 (PBB2), 8.0 (E. coli AN92), and 6.0 (PBB3) when grown aerobically in TSB medium; or 2.0 (E. coli AN92), 1.5 (I?. coli AN154), 2.0 (PBB3), and 2.0 (PBB3) when grown anaerobically; or 3.0 (E. coli AN154), 2.5 (PBB2), 2.0 (E. coli AN92), and 2.0 (PBB3) when grown under iron deficiency. This procedure Determination of isochorismate synthase activity. was carried out as described (7). Enzymes metabolizing isochorismic acid to 2,3-dihydroxybenzoic acid were removed by protamine sulfate precipitation prior to incubations. This procedure was carQuantitative estimation of menaquinones. ried out as described (7). TnlO mutagenesis (31). E. coli MC4100 was grown in 20 ml YM medium [lo g liter-’ tryptone (Difco), 5 g liter-’ NaCl, yeast extract (O.Ol%), maltose (0.2%, added after autoclaving)] until an optical density of 0.36 at 546 nm was reached. The suspension was centrifuged (8000 rpm, 10 min) and the pellet resuspended in 1 ml YM medium. A freshly prepared lysate of phage hNK370 (m.o.i. = lOi phages per 3 X lo7 cells) was added. After absorption at 37°C for 60 min the cells were plated onto LB agar (1% tryptone, 1% NaCl, 0.5% yeast extract, sodium pyrophosphate 2.5 mM) containing tetracycline HCl(15 Fg 100 ml I, Sigma). Cells growing on these plates were selected as follows. Selection Selection for menaquinone and enterobactin deficiency. for menaquinone deficiency (18) was carried out on plates containing (per liter) casamino acids (0.5 g), KH,PO, (5.44 g), K,HPO, (10.49 g), (NH,)$OI (2.0 g), MgSO,, 7H,O (0.05 g), CaC12 (0.5 g), glycerol (0.04 M), Na,fumarate (0.04 M), and agar (16.0 g, Difco Bacto). After autoclaving, 15 pg tetracycline HCl and 0.125 g FeSO,. 7H,O were added in a filter-sterilized solution. Strains were incubated anaerobically on these plates for 24 h at 37°C under a Hz/CO, atmosphere in a Becton Dickinson Gaspak system. Selection for enterobactin deficiency was carried out on agar plates (16 g liter-‘) containing the components listed above for growth under iron deficiency.

ROLE

OF e&C

IL

100

200

300

400

500

time(min)

FIG. 2. Growth curves of E. coli wild-type MC4100 (a) and of E. coli mutant PBBl (b) in the presence of isochorismic acid (4 pmol liter ‘) (c) and chorismic acid (4 pmol liter ‘) (d). Absorbance at 546 nm of a fivefold diluted sample of the culture broth was taken as an indication of growth. The cultures were grown aerobically.

Genonic library. Genomic DNA of E. coli MC4100 was prepared by W. Leinfelder, Munich, FRG, following the methods of Marmur (32) and Saito and Miura (33). After partial digestion (Sau3a) the DNA fragments were packaged in vitro into X phages according to the methods of Hohn and Collins (34). Selection for a recombinant plasmid with the entC gene. Selection for a clone carrying a recombinant plasmid with the entC gene was carried out first under conditions of iron deficiency and subsequently under anaerobiosis using the media described above. E. coli AN154 was used as an indicator strain. Other methods. Transformations were carried out according to the method of Hanahan (35). Electrophoresis was done for 12 h at 80 V on preparative 1% agarose gels (Biozym) in 1% TBE electrophoresis buffer. Plasmidpreparation was done according to the method of Birnboim and Doly (36).

RESULTS

AND

333

GENE

tatively determined (Table II). Enterobactin formation of the wild-type strain was also evident because it formed orange halos on chrome azurol S (CAS) agar plates (37). The activity of the isochorismate synthase determined in crude protein extracts of E. coli MC4100 was 1.7 nmol isochorismic acid/mg protein/60 min. With this strain the lower limit of detection was established as 0.02 nmol isochorismic acid/mg protein/ 60 min. After infection of an E. coli MC4100 cell suspension with phage XNK370, 2208 tetracycline-resistant clones were selected. These clones were screened for their inability to grow on iron-deficient medium (indicating a block in the enterobactin pathway) and under anaerobic conditions, with fumarate as an electron acceptor (indicating a block in menaquinone biosynthesis). To avoid selection for mutants blocked in the shikimate pathway before chorismic acid, IA-amino acids (Arg, Trp, Tyr, Phe, Pro) were omitted from the medium during selection for growth under iron deficiency. Thus, a strain was obtained with an intact shikimate pathway but with a block in both enterobactin and menaquinone biosynthesis, suggesting that the isochorismate synthase gene (e&C) was mutated (compare Fig. 1). This strain was designated E. coli PBBl. The growth curve of mutant PBBl is shown in Fig. 2b. The strain produced neither detectable amounts of menaquinones (Table II) nor enterobactin [no orange halos on chrome azurol S (CAS) agar]. Moreover isochorismate synthase activity was not detectable in crude protein extracts. Menaquinones were not detected in mutant PBBl supplemented with chorismic acid. However, addition of isochorismic acid to the medium resulted in menaquinone formation (Table II). Addition of isochorismic acid also enhanced growth in this mutant (compare curves c and d in Fig. 2). The levels of menaquinones in PBBl supplemented with isochorismic acid were lower than those in the wild-type strain MC4100. This may be due

DISCUSSION

Mutants E. coli AN154 and AN191 were believed to be blocked between chorismic and isochorismic acids, yet produced menaquinones (16, 17). This had been taken as evidence that chorismic acid is the branch point for menaquinone biosynthesis. These mutants were obtained after treatment of E. coli K12 derivatives with N-methyl-N’-nitro-N-nitrosoguanidine (NNG). It is known that NNG mutants can be leaky (31), and indeed E. coli AN154 and AN191 were shown to have very low isochorismate synthase activity and a low level of menaquinones (7). We therefore decided to select a mutant after TnlO mutagenesis of E. coli MC4100 wild-type strain. Before mutagenesis this strain was characterized: The growth curve of this wild-type strain is shown in Fig. 2a. The strain produced menaquinones which were quanti-

TABLE

II

Amount of Menaquinones Present in E. coli Wild-Type Strain MC4100 and in E. coli Mutant PBBl after Growth in the Presence or Absence of Chorismic or Isochorismic Acid” Amount (b&g wet wt)

E. coli strain

Cell paste extracted (g wet wt)

MK8

DMKS

MC4100 PBBl PBBl + isochorismic acid PBBl + chorismic acid

6.0 5.0 5.5 4.5

158 n.d. 46 n.d.

18 n.d. 6 n.d.

“The E. coli strains were grown in 1 liter medium. Chorismic or isochorismic acid was added at a concentration of 4 rmol liter ‘. n.d. = not detectable.

334

KAISER

EcoRV

AND

EcoRV

PVUII

PVUII Sma+-

HpaI--

I Eco RV

PVUII

--I- PVUII

lkb a)

b)

FIG. 3. Comparison of restriction maps of the enterobactin gene cluster (a) according to Ref. (24) and (b) this study. Vectors (pBR328 for a and pHC79 for b) are not shown.

to the instability of isochorismic acid or to its restricted uptake into the E. coli cells. These observations confirm our earlier results (7, 8) which indicated that isochorismic acid rather than chorismic acid is the immediate precursor of menaquinones. They show, in addition, that the entC gene is involved in menaquinone biosynthesis. This gene encodes isochorismate synthase which has also been implicated in enterobactin biosynthesis (3, 4, 6, 25-27). The gene is part of the enterobactin gene cluster. To obtain further evidence in support of our conclusions we have cloned the entC gene. E. coli AN154, which has a very low isochorismate synthase activity (7), was chosen as an indicator strain. This strain grows slowly under anaerobic conditions in the presence of nonfermentable substrates or under iron deficiency (data not shown). E. coli AN154 was made recA and tetracycline resistant (tet’) by Pl transduction from strain E. coli JC10240 [recA-s-1:: TnlO] (38). The E. coli mutant AN154 recA tetf was subsequently infected with phages containing the cosmids of pHC79. The cosmids in turn carried DNA fragments (20-40 kb) of genomic DNA of E. coli MC4100 ligated into the BumHI site of this vector. Recombinant clones were screened either for growth on minimal medium containing the iron chelator cy,cr’dipyridyl or for production of menaquinones under anaerobic conditions with fumaric acid as electron acceptor. A clone (PBBS, Table I) was found which grew under both conditions. The recombinant cosmid (pAEK1) isolated from this clone contained a 24-kb DNA fragment with two EcoRI restriction sites. To decide which of the two EcoRI fragments contained the entC gene the isolated DNA from this cosmid clone was digested with EcoRI, subsequently ligated with T4 DNA ligase, and retransformed into the indicator sbrain E. coli AN154. A new strain (PBBS, Table I) with plasmid pAEK2 was obtained. This plasmid carried a 9-kb DNA insertion as

LEISTNER

part of the enterobactin entC gene.

gene cluster which included the

i. In contrast to the recipient strain (AN154), PBB2 and PBB3 were able to grow well under iron deficiency and anaerobiosis with fumarate as an electron acceptor. ii. Restriction site mapping of our plasmid showed that the pattern of restriction sites is very similar to that of a previously published map of the enterobactin gene cluster (23, 24) (Fig. 3). However, the location of one EcoRI site is different on both maps for reasons which we do not understand at this time. iii. The plasmid (pAEK2) contained in strain PBB3 complemented E. coli AN156 entA. Both entA and entC are part of the enterobactin gene cluster (23-26). PBB3, PBBS, as well as reference strains (E. coli AN92 and AN1541 were now grown either in TSB medium or under iron deficiency. The isochorismate synthase activity was determined and found to be significantly enhanced under iron deficiency when compared to the normal TSB medium (Fig. 4). This observation was expected (2526) and confirmed that the cosmid (pAEK1) of PBB2 and the plasmid (pAEK2) of PBB3 contained the entC gene. When strains AN154, AN92, PBBB, and PBB3 were now grown under anaerobic conditions, the activity of the isochorismate synthase was also significantly stimulated (Fig. 4). In particular PBB3 with a part of the enterobactin cluster and the entC gene on the plasmid had a greatly enhanced activity when compared to the recipient strain which lacks this plasmid (E. coli AN154). This shows that the entC gene or the entC gene together with genes of the enterobactin gene cluster respond not only to conditions (iron deficiency) which trigger the forma-

11 TSB medium 10

i Q-

8765-

FIG. 4. Activity of the isochorismate synthase in protein extracts of different bacterial strains grown in standard medium (TSB) or under iron deficiency or anaerobiosis.

ROLE

OF e&C

tion of enterobactin but also to anaerobiosis which stimulates menaquinone biosynthesis (39). We conclude therefore that not only isochorismic acid and isochorismate synthase but also the gene (entC) that encodes this protein are involved in both enterobactin and menaquinone biosynthesis. It is evident that nucleotide sequences are part of our plasmid, carrying the enterobactin genes, which influence menaquinone biosynthesis. It has recently been shown (25-28) that the nucleotide sequence of the entC gene exhibits 53% homology to the pabB and trpE genes. This is in agreement with the observation that the polypeptides encoded by these genes metabolize chorismic acid [either to isochorismic acid (entC) or topara-aminobenzoic acid (pabB) or anthranilit acid (trpE)]. Their similarity is also evident from the finding that the enzymes controlling the synthesis of these compounds are difficult to separate (4-O). Since the polypeptides encoded by the pabB and trpE genes are parts of dimeric proteins, this feature may also apply to the entC gene product (26,27,41). ACKNOWLEDGMENTS This work was supported by a grant from the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. We thank Dr. A. Bock, Dr. F. Zinoni, Dr. A. Birkmann, and Dr. W. Leinfelder, Institut fur Mikrobiologie, Universitat Munchen, and Dr. B. Wiedemann, Dr. P. Heisig, Dr. C. Korfmann, and Dr. C. Kliebe-Frisch, Medizinische Mikrobiologie, Bonn, for their help and advice. Strains E. coli AN92 AN154, and AN156 were kindly supplied by Dr. I. G. Young, Canberra, Australia. We thank Mrs. W. Garvert for her technical assistance and Mrs. I. Wahl for typing the manuscript. We are grateful to Dr. A. Oaks and Dr. I. D. Spenser, McMaster University, Hamilton, Canada, for helpful discussions.

REFERENCES 1. Weiss, U., and Edwards, J. M. (1980) The Biosynthesis matic Compounds, Wiley, New York. 2. Haslam, don.

E. (1974) The Shikimate

Pathway,

of Aro-

Butterworth,

Lon-

3. Young, I. G., Batterham, T. J., and Gibson, Biophys. Acta 177,389-400.

F. (1969) Riochim.

4. Young, I. G., and Gibson, F. (1969) Biochim. 401-411.

Biophys. Acta 177,

5. Marshall,

B. J., and Ratledge,

C. (1972) Biochirn.

Biophys.

Acta

264,106-116. 6. Winkelmann,

G., van der Helm, D., and Neilaads, J. B. (Eds.) (1987) Iron Transport in Microbes, Plants and Animals, VCH Verlagsgesellschaft, Weinheim.

7. Weische, A., Johanni, M., and Leistner, Biophys. 256,212-222. 8. Weische, Biophys. 9. Emmons, Biophys.

E. (1987) Arch. Biochem.

A., Garvert, W., and Leistner, E. (1987) Arch. Biochem. 256,223-231. G. T., Campbell, I. M., and Bentley, R. (1985) Biochen. Res. Commun. 131,956-960.

335

GENE

10. Shaw, D. J., Guest, J. R., Meganathan, R., and Bentley, R. (1982) J. Bacterial. 152, 1132-1137. 11. Kolkmann R., and Leistner, E. (1987) 2. Noturforsch. C 42, 542-

552. 12. Kolkmann,

R., and Leistner,

1207-1214. 13. Simantiras,

M., and Leistner,

E. (1987) 2. Naturforsch. E. (1989) Phytochemistry

C 42,

28,1381-

1382. 14. Leete, E., and Bodem, G. 8. (1976) J. Amer. Chem. Sot. 98,63216325. 15. Meganathan, R. (1981) J. Biol. Chem. 256,9386&9388. 16. Young, I. G. (1975) Biochemistry 14,399-406. 17. Young, I. G., Langman, L., Luke, R. K. J., and Gibson, F. (1971) J. Bacterial. 106, 51-57. 18. Newton, N. A., (lox, G. B., and Gibson, F. (1971) Biochim. Biophys. Acta 244, 155-166. 19. Guest, J. R. (1979) J. Gen. Microbial. 115, 259-271. 20. Laird, A. J., Ribbons, D. W., Woodrow, G. C., and Young, I. G. (1980) Gene 11,347-357. 21. Laird, A. ,J., and Young, I. (1980) Gene 11, 359-366. 22. Fleming, T. P., Nahlik, M. S., and McIntosh, M. A. (1983) J. Bacteriol. 156, 1171-1177. 23. Pickett, C. L., Hayes, LaD., and Earhart, C. F. (1984) FEMS Microbiol. Lett. 24, 77-80. 24. Schrodt Nahlik, M., Fleming, T. P., and McIntosh, M. A. (1987) J. Bacterial. 169, 4163-4170. 25. Elkins, M. F., and Earhart, C. F. (1988) FEMS Microbial. Lett.

56,35-40. 26. Ozenberger,

B. A., Brickman, T. J., and McIntosh, M. A. (1989) J. Bacterial. 171, 775-783. 27. Schrodt Nahlik, M., Brickman, T. J., Ozenberger, B. A., and McIntosh, M. A. (1989) J. Bacterial. 171, 784-790. 28. Elkins, M. F., and Earhart, C. F. (1988) FEMS Microbial. Lett.

56,35-40. 29. Bachmann, B. J. (1983) Microbial. Reu. 47,180-230. 30. Csonka, L. N., and Clark, A. (1980) J. Bacterial. 143,529~530. 31. Kleckner, N., Roth, J., and Botstein, D. (1977) J. Mol. Biol. 116, 125-159. 32. Marmur, G. (1961) J. Mol. Biol. 3,208-218. 33. Saito H., and Miura, K. G. (1963) Biochim. Biophys. Acta 72,619629. 34. Hohn, B., and Collins, B. (1980) Gene 11,291-298. Practical Approach. 35. Hanahan, D. (1983) in DNA Cloning-A (Glover, D. M., Ed.), Vol. I, pp. 109-135, IRL Press, Washington, DC. 36. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: Alkaline Lysis Method, pp. 368-369, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 37. Schwyn, B., and Neilands, d. B. (1987) Anal. Biochem. 160,47-

56. 38. Ihara, M., Oda, Y., and Yamamoto, wt. 30,33-35. 39. Bentley,

R., and Meganathan,

K. (1985) ITEMS Microbial.

R. (1982) Microbial.

Reu. 46, 241L

280. 40. Johanni, M., Hofmann,

P., and Leistner, E. (1989) Arch. Biochem. Biophys. 27 1, in press. 41. Liu, J., Duncan, K., and Walsh, C. T. (1989) J. Bacterial. 171, 791-798.