The role of isochorismate hydroxymutase genes entC and menF in enterobactin and menaquinone biosynthesis in Escherichia coli

Biochimica et Biophysica Acta 1425 (1998) 377^386

The role of isochorismate hydroxymutase genes entC and menF in enterobactin and menaquinone biosynthesis in Escherichia coli Claudia Dahm, Rolf Mu«ller, Gaby Schulte, Karsten Schmidt, Eckhard Leistner * Institut fu«r Pharmazeutische Biologie, Rheinische Friedrich-Wilhelms-Universita«t, Nussallee 6, D-53115 Bonn, Germany Received 15 May 1998; revised 23 July 1998; accepted 4 August 1998

Abstract Klebsiella pneumoniae 62-1, a triple mutant impaired in aromatic amino acid biosynthesis (Phe3 , Tyr3 , Trp3 ), excretes chorismic acid into the culture broth. When transformed with plasmids harbouring Escherichia coli genes entC or menF the mutant excretes a mixture of both chorismic and isochorismic acid indicating that not only entC but also menF encodes an isochorismate hydroxymutase (isochorismate synthase, EC 5.4.99.6) enzyme. These enzymes catalyze the first step in enterobactin or menaquinone biosynthesis, respectively. Although both gene products (EntC and MenF) catalyze the same reaction, they play distinct roles in the biosynthesis of menaquinone (MK8) and enterobactin. An E. coli mutant (PBB7) with an intact menF but a disrupted entC produced menaquinone (MK8) but no enterobactin, whereas a mutant (PBB9) with an intact entC but a disrupted menF produced enterobactin and only a trace of menaquinone (MK8). When both menF and entC were disrupted (mutant PBB8) neither menaquinone (MK8) nor enterobactin was detectable. Our previous assumption that entC is responsible for both menaquinone and enterobactin biosynthesis is inconsistent with these mutant studies and has to be revised. The presence in the promoter region of menF of a putative cAMP receptor protein binding site indicates that menF is regulated differently from entC. The menF gene was overexpressed as a fusion gene and its product (6UHis-tagged MenF) isolated. The enzyme catalyzed the formation of isochorismic from chorismic acid and as opposed to a previous publication also the reverse reaction. The enzyme was characterized and its kinetic data determined. ß 1998 Elsevier Science B.V. All rights reserved. Keywords: Isochorismic acid; Menaquinone; Enterobactin; (Klebsiella pneumoniae 62-1)

1. Introduction Chorismic and isochorismic acids play roles in the terminal sequences of the shikimate pathway. Both compounds represent metabolic branch points and are interconverted by an enzyme named isochorisAbbreviations: bp, base pair; EntC and MenF, enterobactin speci¢c and menaquinone speci¢c isochorismate hydroxymutases; nt, nucleotide(s) * Corresponding author. Fax: +49 (228) 733250; E-mail: [email protected]

mate hydroxymutase or isochorismate synthase (EC 5.4.99.6) [1,2] (Fig. 1). The enzyme is encoded by entC [3] which forms part of a gene cluster responsible for the synthesis of a siderophore called enterobactin. Enterobactin [1] as well as menaquinones [4,5] are derived from isochorismic acid. A new open reading frame (ORF) (menF) homologous to entC was discovered recently [6] and assumed to be an isochorismate synthase isogene. This assumption was based on a complementation and gene inactivation experiments. Additionally, MenF showed 24.2% identity and 42.4% similarity

0304-4165 / 98 / $ ^ see front matter ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 8 ) 0 0 0 8 9 - 0

BBAGEN 24683 14-10-98

378

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

(CLUSTALW 1.6) to EntC. The encoding menF gene is clustered with menaquinone genes and is located directly upstream from the menD gene [6]. Our observations on the role of menF have now been extended. menF has been overexpressed and the kinetic data of the encoded enzyme determined. A menF3 mutant has been generated and characterized together with the entC3 single and entC3 /menF3 double mutants described earlier [6]. Our data show that menF is functionally related to menaquinone biosynthesis. The menF gene has also been overexpressed by Daruwala et al. [7] and the encoded protein characterized. Some of their data are not in agreement with our observations. 2. Materials and methods 2.1. General techniques Competent Escherichia coli and Klebsiella pneumoniae 62-1 cells were prepared as described by Nishimura et al. [8]. Chromosomal DNA from bacterial strains was isolated by the method of Kohlbrecher et al. [9]. Other DNA manipulations were based on standard procedures [10]. 2.2. Bacterial strains E. coli Y1089 [10], E. coli XL1 Blue [11], E. coli MC 4100 [12], E. coli XL1 Blue MRFP Kan [13], E. coli B121(DE3)pLysS [14] and K. pneumoniae 62-1 (formerly Enterobacter (syn. Aerobacter) aerogenes 62-1), ATCC 25306 [15] have been described in the literature. K. pneumoniae 62-1 and its recombinants were stored at 380³C on glass beads as described by Su«ssmuth et al. [16]. This ensured maximum genetic stability. The sequence of menF is given in Mu«ller et al. [6]. The sequence has been corrected and extended (EMBL Databank Z50849). 2.3. Construction of mutants from E. coli Y1089 Mutagenesis of entC and/or menF genes in E. coli Y1089 yielding PBB7 (entC3 ) or PBB8 (entC3 , menF3 ) has been described [6]. Construction of

PBB9 (menF3 ) by mutation of menF in E. coli Y1089 was carried out as described for PBB8 except that E. coli Y1089 rather than E. coli PBB7 was subjected to mutation. The proper position of the deletion in menF was veri¢ed by PCR as described [6]. 2.4. Construction of plasmids The following recombinant plasmids were constructed using pBGS8 [17] (compare Fig. 3). pCR16.1 was constructed from a PCR fragment (1.4 kb) obtained with genomic DNA from E. coli MC 4100 as template and primers CD3 (5P-GTG AAG CTT TTA ACA GGG AGA GGT C-3P) and CR4 (5PTAT GGA TCC GTA ATG ATG CGA CTC AT3P). The resulting fragment was treated with T4 DNA polymerase, cut with BamHI and ligated into the SmaI-BamHI site of pBGS8. (The insert starts 28 base pairs (bp) upstream from the GTG-start and contains the Shine-Dalgarno sequence of menF but not the promoter sequence.) pCR 1-5.1 was constructed from a PCR fragment obtained with genomic DNA from E. coli MC 4100 as template and primers CR5 (5P-ACC CCC GGG ATG GTG ATG AAG AAG CTG TCG TC-3P) and CR4. The 1.2 kb fragment was cut (SmaI-BamHI) and the resulting fragment ligated into pBGS8 after hydrolysis with the same combination of enzymes (pCR1-5.1 is pCR1-6.1 shortened by 172 bp at the 5P end). pCR1-9.1 was constructed from a PCR fragment obtained with genomic DNA from E. coli MC 4100 as template and primers CD3 and CR4. The resulting fragment was treated with T4 DNA polymerase, cut with HindIII and ligated into the HindIII-EcoRV site of pBluescript II KS(3) [18]. This intermediary plasmid was named pCR1-6. A 191 bp fragment (155 bp of this fragment are part of menF) was subsequently removed by digestion with EcoRI. The shortened and linearized plasmid was religated and named pCR1-9. The plasmid pCR1-9.1 was obtained by digestion of pCR1-9 with HindIII and treatment with the Klenow fragment of E. coli DNA polymerase I in the presence of dNTPs, followed by restriction with PstI and ligation into SmaI-PstI digested pBGS8. pCL1-16.1: a 1.5 kb fragment was ampli¢ed with primers CD13 (5P-TAG GAA TTC TGC TGG CAC GCC GTT A-3P) and CR4 and genomic DNA from

BBAGEN 24683 14-10-98

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

E. coli MC 4100. The ampli¢cation was carried out with Pfu DNA polymerase (Stratagene), generating blunt ends. The blunt-end PCR fragment was cut with BamHI and ligated into SmaI-BamHI digested pBGS8. (The insert starts 100 bp upstream from the GTG-start and contains promoter sequence, putative cAMP receptor protein (CRP) binding site and Shine-Dalgarno sequence.) pCR1-10.1 was constructed from a PCR fragment obtained with genomic DNA from E. coli MC 4100 as template and primers CD1 [6] and CR4. The resulting fragment was treated with T4 DNA polymerase, cut with BamHI and ligated into the SmaI-BamHI site of pBGS8. (The insert starts 538 bp upstream from the GTG-start and contains promoter sequence, putative CRP binding site, Shine-Dalgarno sequence and elaB (EMBL accession No. U58768). These plasmids, containing parts of menF, the complete menF gene or additional sequences upstream from menF, were transformed into cells of the chorismate overproducing strain K. pneumoniae 62-1. The resulting recombinant Klebsiella strains were grown in 50 ml cultures as described by Schmidt and Leistner [19] over 21 h. Accumulation of isochorismic acid in the culture broth was determined by HPLC. 2.5. Construction of 6UHis-tagged menF The menF gene was generated by PCR using primers CD15 (5P-CAG ATC TGC ATA TGC AAT CAC TTA CTA CGG CGC TGG-3P) and CR4. Plasmid pCR1-20 [6], carrying the genomic menF gene, served as a template. Ampli¢cation was carried out in a Trio-Thermoblock (Biometra) using Pfu DNA polymerase (Stratagene). The resulting fragment (1.25 kb) was isolated by gel electrophoresis and then ligated into the vector pCR-Script SK(+) [20] by employing the pCR-Script SK(+) Cloning Kit (Stratagene). After transformation into E. coli XL1 Blue MRFP Kan and reisolation of the recombinant plasmid, menF was removed using BglII-BamHI. The fragment was then ligated into the expression vector pRSETB [21] (Invitrogen) linearized with BamHI and dephosphorylated. The resulting plasmid pCL1-15 was checked for the correct orientation of the insert and transformed into E. coli B121(DE3)pLysS (Stratagene).

379

2.6. Puri¢cation of 6UHis-tagged MenF E. coli Bl2l(DE3)pLysS harboring pCL1-15 was grown in LB medium (50 ml). After induction (4 h) with IPTG (0.5 mM) at an OD600 of 0.5 cells were harvested by centrifugation (4000Ug, 15 min, 4³C) and stored at 320³C. Subsequent puri¢cation steps were carried out at 0^4³C. The pellet was thawed while kept on ice at room temperature and suspended in bu¡er I (NaH2 PO4 50 mM, pH 8.0; NaCl 300 mM; L-mercaptoethanol 20 mM). After sonication (Branson soni¢er, 8 times 15 s pulses, duty cycle 40%, 40% output, 3 mm microtip) and centrifugation (10 000Ug, 30 min) the solution was ¢ltered (Syringe ¢lter 0.2 Wm). The fusion protein was isolated by a¤nity chromatography on a Ni-NTA-agarose (Qiagen) column following the manufacturer's instruction. The protein was electrophoretically homogeneous. 2.7. Enzyme assay ^ forward reaction (chorismic to isochorismic acid) The incubation mixture contained (¢nal volume 100 Wl) chorismic acid (40^60^80^100^200^400^ 600^800^1000 WM), MgCl2 (1 mM), Tris-HCl (pH 7.5; 200 mM), dithiothreitol (1 mM), EGTA (1 mM) and protein (2 Wg). The mixture was incubated for 1.5 min at 37³C and the reaction then terminated by addition of cold MeOH (100 Wl). The mixture was directly analyzed by HPLC (see below). The following amounts of isochorismic acid were observed: 18.2^23.3^29.2^33.9^54.2^64.2^68.2^ 79.3^79.7 pmol s31 . 2.8. Enzyme assay ^ reverse reaction (isochorismic to chorismic acid) The reaction mixture was identical to that described for the forward reaction except that isochorismic acid (7.3^10.9^14.5^18.1^21.8^29.1^36.3 WM) was used as substrate [19]. The mixture was incubated for 30 min at 37³C and the reaction terminated by addition of cold MeOH (100 Wl). The mixture was directly analyzed by HPLC (see below). The following amounts of chorismic acid were observed: 0.11^0.21^0.26^0.31^0.37^0.47^0.60 pmol s31 .

BBAGEN 24683 14-10-98

380

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

2.9. Identi¢cation of reaction products Isochorismic acid was identi¢ed by UV spectroscopy [22] cochromatography with an authentic sample and thermic conversion to salicylic acid (100³C, 15 min). Chorismic acid was identi¢ed by cochromatography with an authentic sample and thermic conversion to 4-hydroxybenzoic acid (100³C, 15 min) [23]. Chorismic acid was also collected, lyophilized, solved in H2 O and used as substrate for the enzymic conversion to isochorismic acid with MenF as catalyst. 2.10. Quantitative determination of chorismic and isochorismic acid HPLC analysis was carried out on a Multospher RP 18 (4U250 mm, particle size 5 Wm) column. Solvents used were A: H2 O-HCOOH (99:1 v/v) and B: H2 O-HCOOH-MeOH (99:1:100 v/v) with a gradient: 0^5 min 100% A, 5^20 min 80% (v/v) A and 20% (v/v) B, 25^30 min 50% (v/v) A and 50% (v/v) B, 30^35 min 100% (v/v) A. The UV detector was set to 278 nm and the £ow rate to 1 ml min31 . Retention times: isochorismic acid 13.8 min; chorismic acid 19.4 min. 2.11. Determination of menaquinones and enterobactin Isolation of menaquinones was carried out as described [6] but modi¢ed: crude and dried menaquinone fractions were dissolved in 1 ml acetonitrile and then analyzed by HPLC (compare [24]) (Multospher RP 18, 4U250 mm column, particle size 5 Wm). Solvent: acetonitril-isopropanol (85:15, v/v), £ow rate 1 ml min31 ; detection: 245 nm. Authentic menaquinones (Ho¡mann La Roche) were used as a refer-

ence. Retention times for demethylmenaquinone 8 (DMK 8) and menaquinone 8 (MK8) were 13.3 min and 21.2 min, respectively. Enterobactin was determined by the method of Schwyn and Neilands [25]. 3. Results Transformation of the triple mutant K. pneumoniae 62-1 with plasmids carrying genes involved in chorismic acid metabolism has been repeatedly shown to a¡ect the pattern of compounds excreted by the mutant [6,19,26,27]. Thus, the recombinant mutant is a valuable tool to directly correlate a gene with the substrate and product of its encoded enzyme. Whereas K. pneumoniae 62-1 harboring pBGS8 without the isochorismate synthase gene excretes chorismic acid alone (Fig. 2a), introduction of an isochorismate synthase gene (entC on pBGS8, i.e. pKS3-02) [19] results in the appearance in the culture broth of a mixture of chorismic and isochorismic acid (Fig. 2b). A similar result is observed after introduction into K. pneumoniae 62-1 of pCR1-10.1 (i.e. pBGS8 carrying menF). Again, the recombinant mutant excretes both chorismic and isochorismic acid (Fig. 2c). This shows clearly that both EntC and MenF are isoenzymes, which equilibrate chorismic and isochorismic acids. The yield in isochorismic acid is lower, however, in the case of menF (50 mg l31 ) when compared to entC (220 mg l31 ). This may re£ect a di¡erence in the stability or kinetic properties of both enzymes or a di¡erence in the synthesis, stability or turnover of the mRNAs involved. In order to probe the role of entC and menF in enterobactin and menaquinone biosynthesis, selective in-frame mutations of either entC (PBB7) or menF

Fig. 1. Metabolism of chorismic and isochorismic acid to either menaquinones or enterobactin.

BBAGEN 24683 14-10-98

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

381

Fig. 2. Time-dependent accumulation of chorismic acid (A) and isochorismic acid (B) in cultures of K. pneumoniae 62-1. (a) K. pneumoniae 62-1 with plasmid pBGS8 (control); (b) K. pneumoniae 62-1 with plasmid pKS3-02 containing entC ; (c) K. pneumoniae 62-1 with plasmid pCR1-10.1 containing menF. Data points are means of two independent experiments. F, growth; R, chorismic acid; b, isochorismic acid. The HPLC trace (left) was obtained 17.5 h after inoculation.

(PBB9), or both entC and menF (PBB8) were carried out. The genes were replaced by menF or entC carrying a 363 bp (menF) or 336 bp (entC) in-frame deletion. The exchange was mediated by plasmid pMAK705 [28] carrying either menFv363 or entCv336. In every case the proper positions of the resultant genomic deletions were veri¢ed by PCR in which oligonucleotides homologous to border regions of menF or entC were incubated with genomic DNA before and after the replacement reaction [6].

In each case the proper position of the deletions was evident from the appearance of shortened DNA fragments of the expected size. Production of enterobactin by the mutants was determined on Chromazurol S agar plates [25] whereas menaquinone (MK8) formation was monitored by HPLC. A preliminary account of part of the mutant studies has been published [6]. A more complete description of data is given in Table 1. The data show that maximum production of me-

BBAGEN 24683 14-10-98

382

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

Table 1 Enterobactin and menaquinone formation of E. coli Y1089 and its derived mutants (PBB7, PBB9 and PBB8) Strains

Y1089

PBB7 (entC3 )

PBB9 (menF3 )

PBB8 (entC3 , menF3 )

Enterobactin formation Menaquinone (MK8) (Wg g31 wet weight)

Yes 10.3

n.d. 9.3

Yes 0.4

n.d. n.d.

Formation of enterobactin was checked on Chromazurol S agar plates [25]. Menaquinone (MK8) production was determined by HPLC. Limit of detection 6 0.1 Wg g31 wet weight. n.d., not detectable.

naquinone (MK8) depends on the presence of an intact menF gene. While mutant PBB9 (menF3 ) produces only a trace of menaquinones (MK8), it is not detectable in mutant PBB8 (entC3 , menF3 ). The trace of MK8 in PBB9 may be explained by a cross-feeding from EntC. By contrast, enterobactin synthesis depends on the presence of an intact entC gene. No enterobactin synthesis is seen in mutants PBB7 (entC3 ) or PBB8 (entC3 , menF3 ). Production of menaquinone was restored to 1.9 Wg/g wet weight when o-succinylbenzoic acid (10 mg l31 ) was added to a culture of PBB8 (entC3 , menF3 ). Similarly, enterobactin production was seen in the same mutant when 2,3-dihydroxybenzoic acid (10 mg l3l ) was added to the broth. Thus formation of both menaquinones and enterobactin was restored in the mutant when intermediates between isochorismic acid and menaquinone or enterobactin were fed to the double mutant PBB8 (entC3 , menF3 ). The K. pneumoniae mutant 62-1 was also used to test plasmids harboring inserts of varying length for their ability to sustain isochorismic acid production (Fig. 3). The inserts were constructed by PCR using oligonucleotide primers targeted to the menF region as indicated in Fig. 3. The inserts were ligated into plasmid pBGS8 and the recombinant plasmids introduced into the chorismate excreting K. pneumoniae 62-1 strain. The recombinant strains produced varying amounts of chorismic and isochorismic acid (Fig.

3). The experiment with plasmid pCR1-10.1 and pCL1-16.1 indicated that the unknown reading frame (elaB) upstream from the menF region (EMBL accession No. U58768) did not in£uence isochorismate production. The function of elaB is unknown. Isochorismate production was diminished when the sequences between elaB and menF (pCR1.-6.1) were omitted. Experiments with plasmids pCR1-5.1 and pCR1-9.1 are control experiments which help to de¢ne the functional role of the 5P and 3P bounds of the menF gene. Either no (pCR1-9.1) or only a trace (pCR1-5.1) of isochorismic acid was detectable when menF was shortened at either the 5P or 3P ends by 155 or 172 base pairs, respectively. Subsequently the region upstream from the start codon was sequenced because a signi¢cant di¡erence in the isochorismate production (compare pCL1-16.1 and pCR1-6.1) indicated that a regulatory sequence may be present. Indeed, a putative binding site for a CRP was found homologous to a previously published sequence [29] as well as the wild type 335 and 310 promoter sequence and a Shine-Dalgarno sequence (Fig. 3). The menF gene was overexpressed as a fusion gene and the resulting 6UHis-tagged MenF isochorismate hydroxymutase protein isolated using a Ni-NTAagarose column. The puri¢ed fusion protein was incubated with either chorismic or isochorismic acid and found to catalyze the formation of the corresponding structural isomer in both cases. Thus, the

Table 2 Properties of isochorismate hydroxymutase fusion protein (6UHis-tagged MenF) Property

Forward reaction (chorismic acidCisochorismic acid)

Reverse reaction (isochorismic acidCchorismic acid)

pH optimum Km (WM) Vmax (pmol s31 ) Turnover number (min31 )

7.5 166.9 91.7 144.9

7.5 119.0 2.4 3.8

BBAGEN 24683 14-10-98

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

Fig. 3. Position of a putative CRP binding site in the promoter region of menF and isochorismate excretion into the culture broth of recombinant K. pneumoniae 62-1 strains harboring plasmids with either an intact menF gene (pCR1-6.1), an intact menF gene and its upstream sequences (pCL1-16.1, pCR1-10.1) or a shortened menF gene (pCR1-5.1, pCR1-9.1). Oligonucleotide primers mentioned in Section 2 are CD1, CD3, CR4, CR5, CD13. Restriction sites: H III, HindIII; C I, ClaI; S I, SalI; E V, EcoRV; E I, EcoRI; P I, PstI. SD, Shine-Dalgarno.

forward (chorismic to isochorismic acid) as well as reverse (isochorismic to chorismic acid) reaction can be observed. Maximum activity of the protein was observed at pH 7.5 and at 37³C. The reaction was dependent on the presence of Mg2‡ and the rate remained constant within a range of 0.8^4.0 mM. The kinetic data of the enzyme were determined and are listed in Table 2. 4. Discussion We have earlier postulated that entC is responsible for both enterobactin and menaquinone synthesis in E. coli MC 4100 [30]. However, this model is neither supported by the data obtained from experiments with E. coli Y1089 and presented in Section 3 of this paper nor by the data of Kwon et al. [31].

383

Our previous conclusions were mainly derived from two observations. First, the entC gene responds not only to iron deprivation [32] but also to conditions which stimulate menaquinone formation (namely anaerobiosis) [30]. This has been con¢rmed [31] but cannot be considered as direct proof for a metabolic relation between entC and menaquinones. Second, a mutant of E. coli MC 4100 (PBB1) blocked in isochorismate synthesis did not produce menaquinones. Careful reexamination of this mutant has, however, revealed that it does produce low quantities of menaquinones. Thus, our previous menaquinone determination may have been in error. Alternatively, we cannot exclude the possibility that this mutant changed its phenotype. We conclude that there is a menaquinone-speci¢c and an enterobactin-speci¢c isochorismate synthase gene in E. coli. This conclusion is in agreement with the mutant studies listed in Section 3 (Table 1) and with those of Mu«ller et al. [6] and Kwon et al. [31]. Moreover, the situation in Bacillus subtilis is similar. Two isochorismate hydroxymutase encoding genes (menF and dhbC, the latter corresponding to entC in E. coli) were detected [33]. Since our last publication on menF [6] we have improved the sensitivity of our HPLC method and found that a trace of menaquinone (MK8) is detectable in mutant PBB9 (entC‡ , menF3 ) (Table 1). By contrast, PBB7 (entC3 , menF‡ ) does not produce any enterobactin. Thus, entC restores menaquinone de¢ciency to a very minor extent whereas menF is unable to compensate for the loss of entC in mutant PBB7. This corresponds very well with the situation in B. subtilis mutant RB 1255 [33]. The results show very clearly that the two isochorismate hydroxymutase genes are responsible for either enterobactin (entC) or menaquinone (menF) biosynthesis, respectively. Neither enterobactin nor menaquinone is detectable in the double mutant PBB8 (entC3 , menF3 ). The sequence for the newly discovered menF gene has been reported twice [6,34]. Both sequences, however, were redetermined and found to contain reading errors. This is con¢rmed by the sequence data of the complete E. coli genome now available (DDBJ/ EMBL/GenBank databases accession No. D90857 AB 001340). Our corrected sequence of menF including its 5P upstream sequence is accessible under EMBL Z50849. The calculated molecular weight of

BBAGEN 24683 14-10-98

384

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

the encoded MenF protein is 48 737 Da rather than 37 667.5 Da [6] or 48 777 Da [34]. The sequence by Daruwala et al. [34] lacks one and contains two wrong amino acids. Data shown in Fig. 3 indicate that the sequence upstream from the GTG start codon of menF is essential for full MenF activity. Plasmids pCL1-16.1 and pCR1-6.1 di¡er by 72 nucleotides (nt) and by the amount of isochorismate excreted into the culture broth by recombinant Klebsiella mutants carrying these plasmids. The Shine-Dalgarno sequence is found 6 nt upstream from the GTG start codon (Fig. 3). For the entC gene this distance is only 4 bp. It has been noted [35] that this is at the short extreme of all such distances known in E. coli. Typical distances are 7^9 nt. Thus, the distance between the start codon and the Shine-Dalgarno sequence is low also in the case of the menF gene. The upstream sequence of menF is also characterized by a putative CRP binding site (TGTGA-N6 -TAACA). The consensus sequence given by Saier et al. [29] is TGTGA-N6 -TCACA. cAMP and the dimeric CRP interact with speci¢c sequences in cAMP-CRP responsive promoters and may function as a positive or a negative e¡ector, thereby promoting or preventing transcription in response to the availability or absence of a utilizable carbon source [29]. Under anaerobic conditions and in the presence of non-fermentable substrates as well as fumarate as terminal electron acceptor menaquinone is required as an essential electron transporting compound [36]. Under these conditions also cAMP levels are high [37] which in turn may stimulate transcription of menF. Indeed addition of cAMP to a mutant impaired in cAMP synthesis increases the menaquinone level [24]. We conclude that a relation between the putative cAMP-CRP binding site upstream from menF, the cAMP level in E. coli, isochorismate hydroxymutase activity and menaquinone biosynthesis are plausible. Thus, the menF and entC genes seem to be di¡erently regulated, with the entC gene being under control of the ferric uptake regulatory (Fur) protein and a binding site called `iron box' [32]. The isochorismate hydroxymutase enzyme has been characterized after isolation from di¡erent sources. The enzyme (EntC) is known from E. coli [35], from Flavobacterium K3ÿ15 [38] and from an anthra-

quinone producing cell suspension culture of Galium mollugo (Rubiaceae) [39]. The enzymes from these sources share some features but are not identical in every respect. Thus, in every case enzyme activity depends on the presence of Mg2‡ and both the forward (chorismate to isochorismate) and reverse reaction (isochorismate to chorismate) can be observed. Moreover, the Km values for isochorismate are lower in every case when compared to the Km values for chorismate. There is also a signi¢cant di¡erence in the Km values for chorismate (E. coli EntC: 14 WM, Flavobacterium: 350 WM, Galium mollugo: 807 WM) and for isochorismate (E. coli EntC: 5 WM, Flavobacterium 254 WM, G. mollugo 475 WM) [35,38,39] in di¡erent organisms. The data for the E. coli MenF enzyme as reported by Daruwala et al. [7] and by us (cf. Section 3) are in good agreement as far as pH optimum, optimum Mg2‡ concentration and the Km value for chorismate are concerned (compare Section 3 and Table 2). Our observations and those of Daruwala et al. [7] di¡er, however, in that we clearly observe not only a forward but also a reverse reaction. It is shown in Table 2 that the Vmax and the turnover number for the reverse reaction are rather low when compared to the forward reaction. This, and the fact that the concentration coe¤cient of chorismate is also low [1], may be the reason why the reverse reaction escaped detection in the case of Daruwala et al. [7]. The analytical data presented by Daruwala et al. [7] in Fig. 5 are not convincing. The peaks are rather broad indicating that the analytical system employed is unsuitable to detect a minor amount of chorismic acid. Acknowledgements This work was assisted by a grant from the Fonds der Chemischen Industrie. We thank Dr. B.E. Ellis, Vancouver, for helpful suggestions. References [1] C.T. Walsh, J. Liu, F. Rusnak, M. Sakaitani, Molecular studies on enzymes in chorismate metabolism and the enterobactin biosynthetic pathway, Chem. Rev. 90 (1990) 1105^ 1129.

BBAGEN 24683 14-10-98

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386 [2] A. Kaiser, E. Leistner, The role of isochorismic acid in primary and secondary metabolism, World J. Microbiol. Biotechnol. 8, (Suppl. 1) (1992) 92^95. [3] B.A. Ozenberger, T.J. Brickman, M.A. McIntosh, Nucleotide sequence of Escherichia coli isochorismate synthetase gene entC and evolutionary relationship of isochorismate synthetase and other chorismate-utilizing enzymes, J. Bacteriol. 171 (1989) 775^783. [4] A. Weische, W. Garvert, E. Leistner, Biosynthesis of o-succinylbenzoic acid II: properties of o-succinylbenzoic acid synthase, an enzyme involved in vitamin K2 biosynthesis, Arch. Biochem. Biophys. 256 (1987) 223^231. [5] A. Weische, M. Johanni, E. Leistner, Biosynthesis of o-succinylbenzoic acid I: cell free synthesis of o-succinylbenzoic acid from isochorismic acid in enzyme preparations from vitamin K producing bacteria, Arch. Biochem. Biophys. 256 (1987) 212^222. [6] R. Mu«ller, C. Dahm, G. Schulte, E. Leistner, An isochorismate hydroxymutase isogene in Escherichia coli, FEBS Lett. 378 (1996) 131^134. [7] R. Daruwala, D.K. Bhattacharyya, O. Kown, R. Meganathan, Menaquinone (vitamin K2 ) biosynthesis: overexpression, puri¢cation, and characterization of a new isochorismate synthase from Escherichia coli, J. Bacteriol. 179 (1997) 3133^3138. [8] A. Nishimura, M. Morita, Y. Nishimura, Y. Sugino, A rapid and highly e¤cient method for preparation of competent Escherichia coli cells, Nucleic Acids Res. 18 (1991) 381^386. [9] D. Kohlbrecher, R. Eisermann, W. Hengstenberg, Rapid isolation and e¤cient puri¢cation of bacterial genomic DNA for PCR ampli¢cation (Perkin Elmer Cetus), Ampli¢cations 4 (1990) 30. [10] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: a Laboratory Manual, 2nd edn., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. [11] W.O. Bullock, J.M. Fernandez, J.M. Short, A high e¤ciency plasmid transforming rec A Escherichia coli strain with galactosidase selection, Biotechniques 5 (1987) 376^378. [12] M.J. Casabadan, Transposition and fusion of the lac genes to selected promoters in Escherichia coli using bacteriophage lambda and mu, J. Mol. Biol. 104 (1976) 541^555. [13] B. Jerpseth, A. Greener, J.M. Short, J. Viola, P. L Kretz, New restriction minus derivatives of XL1-Blue E. coli cells, Strategies 6 (1993) 24. [14] F.W. Studier, A.H Rosenberg, J.J Dunn, J.W. Dubendor¡, Use of T7 RNA polymerase to direct expression of cloned genes, in: D.V. Goeddel (Ed.), Methods of Enzymology, Academic Press, San Diego, CA, 1990, pp. 60^89. [15] M.J. Gibson, F. Gibson, Preliminary studies on the isolation and metabolism of an intermediate in aromatic biosynthesis: chorismic acid, Biochem. J. 90 (1964) 248^256. [16] R. Su«ssmuth, J. Eberspa«cher, R. Haag, W. Springer, Stammhaltung und Konservierung von Mikroorganismen, in: R. Su«ssmuth (Eds.), Biochemisch-mikrobiologisches Praktikum, Thieme, Stuttgart, 1987, pp. 31^50. [17] B.G. Spratt, P.J. Hedge, S. te Heesen, A. Edelman, J.K.

[18]

[19] [20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

385

Broome-Smith, Kanamycin-resistant vectors that are analogues of plasmids pUC8, pUC9, pEMBL8 and pEMBL9, Gene 41 (1986) 337^342. M. A Alting-Mees, J.M. Sorge, pBluescriptII: multifunctional cloning and mapping vectors, Methods Enzymol. 216 (1992) 483^495. K. Schmidt, E. Leistner, Microbial production of (+)-transisochorismic acid, Biotechnol. Bioeng. 45 (1995) 285^291. G.L. Costa, M.P. Weiner, Polishing with T4 or Pfu polymerase increases the e¤ciency of cloning of PCR fragments, Nucleic Acids Res. 22 (1994) 3259^3260. F.W. Studier, B.A. Mo¡at, Use of bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes, J. Mol. Biol. 189, (1) (1986) 113^130. I.G. Young, T.J. Batterham, F. Gibson, The isolation, identi¢cation and properties of isochorismic acid. An intermediate in the biosynthesis of 2,3-dihydroxybenzoic acid, Biochim. Biophys. Acta 177 (1969) 389^400. I.G. Young, F. Gibson, C.G. MacDonald, Enzymic and nonenzymic transformations of chorismic acid and related cyclohexadienes, Biochim. Biophys. Acta 192 (1969) 62^72. G. Unden, Di¡erential roles for menaquinone and demethylmenaquinone in anaerobic electron transport of E. coli and their fnr-independent expression, Arch. Microbiol. 150 (1988) 499^503. B. Schwyn, J.B. Neilands, Universal chemical assay for the detection and determination of siderophores, Anal. Biochem. 160 (1987) 47^56. R. Mu«ller, A. Wagener, K. Schmidt, E. Leistner, Microbial production of speci¢cally ring-13 C-labelled 4-hydroxybenzoic acid, Appl. Microbiol. Biotechnol. 43 (1995) 985^988. R. Mu«ller, M. Breuer, A. Wagener, K. Schmidt, E. Leistner, Bacterial production of trans-dihydroxycyclohexadiene carboxylates by metabolic pathway engineering, Microbiology 142 (1996) 1005^1012. C.M. Hamilton, M. Aldea, B.K. Washburn, P. Babitzke, S.R. Kushner, New method for generating deletions and gene replacements in Escherichia coli, J. Bacteriol. 171 (1989) 4617^4622. M.H. Saier Jr., T.M. Ramseier, J. Reizer, Regulation of carbon utilization, in: F.C. Neidhardt (Ed.), Escherichia and Salmonella, Cellular and Molecular Biology, American Society of Microbiology, Washington, DC, 1996, pp. 1325^ 1343. A. Kaiser, E. Leistner, Role of the entC gene in enterobactin and menaquinone biosynthesis in Escherichia coli, Arch. Biochem. Biophys. 276 (1990) 331^335. O. Kwon, M.E. S Hudspeth, R. Meganathan, Anaerobic biosynthesis of enterobactin in Escherichia coli: regulation of entC gene expression and evidence against its involvement in menaquinone (vitamin K2 ) biosynthesis, J. Bacteriol. 178 (1996) 3252^3259. T.J. Brickman, B.A. Ozenberger, M.A. McIntosh, Regulation of divergent transcription from the iron-responsive fepB-entC promoter-operator regions in Escherichia coli, J. Mol. Biol. 212 (1990) 669^682.

BBAGEN 24683 14-10-98

386

C. Dahm et al. / Biochimica et Biophysica Acta 1425 (1998) 377^386

[33] B.M. Rowland, H.W. Taber, Duplicate isochorismate synthase genes of Bacillus subtilis: regulation and involvement in the biosyntheses of menaquinone and 2,3-dihydroxybenzoate, J. Bacteriol. 178 (1996) 854^861. [34] R. Daruwala, O. Kwon, R. Meganathan, A new isochorismate synthase speci¢cally involved in menaquinone (vitamin K2 ) biosynthesis encoded by the menF gene, FEMS Lett. 140 (1996) 159^163. [35] J. Lui, N. Quinn, G.A. Berchthold, C.T. Walsh, Overexpression, puri¢cation, and characterization of isochorismate synthase (EntC), the ¢rst enzyme involved in the biosynthesis of enterobactin from chorismate, Biochemistry 29 (1990) 1417^ 1425.

[36] J.R. Guest, Anaerobic growth of Escherichia coli K12 with fumarate as terminal electron acceptor. Genetic studies with menaquinone and £uoroacetate-resistant mutants, J. Gen. Microbiol. 115 (1979) 259^271. [37] G. Unden, A. Duchenne, On the role of cyclic AMP and the Fnr protein in Escherichia coli growing anaerobically, Arch. Microbiol. 147 (1987) 195^200. [38] P.M.M. Schaaf, L.E. Heide, E.W. Leistner, Properties of isochorismate hydroxymutase from Flavobacterium K3ÿ15 , J. Nat. Prod. 56 (1993) 1294^1303. [39] C. Ledu«c, I. Birgel, R. Mu«ller, E. Leistner, Isochorismate hydroxymutase from a cell-suspension culture of Galium mollugo L., Planta 202 (1997) 206^210.

BBAGEN 24683 14-10-98