The Escherichia coli highly expressed entD gene complements the pfaE deficiency in a pfa gene clone responsible for the biosynthesis of long-chain n-3 polyunsaturated fatty acids

The Escherichia coli entD gene, which encodes an Sfp-type phosphopantetheinyl transferase (PPTase) that is involved in the biosynthesis of siderophore, is available as a high-expression ASKA clone (pCA24N::entD) constructed from the E. coli K-12 strain AG1. In E. coli DH5α, pCA24N::entD complemented a pfaE-deficient clone that comprised pfaA, pfaB, pfaC and pfaD, which are four of the five pfa genes that are responsible for the biosynthesis of eicosapentaenoic acid derived from Shewanella pneumatophori SCRC-2738. Sfp-type PPTases are classified into the EntD and PfaE groups, based on differences between their N-terminal-domain structures. Here, we showed that all Sfp-type PPTases may have the potential to promote the biosynthesis of long-chain n-3 polyunsaturated fatty acids.


Introduction
In some marine bacteria and eukaryotic microorganisms, long-chain n-3 polyunsaturated fatty acids (n-3 PUFAs), such as the eicosapentaenoic acid (EPA) and the docosahexaenoic acid (DHA), are synthesized de novo via a polyketide biosynthesis pathway (Metz et al., 2001;Orikasa et al., 2006a, b, c). Five genes (pfaA, pfaB, pfaC, pfaD and pfaE) that are involved in the biosynthesis of EPA or DHA have been cloned from bacteria (Tanaka et al., 1999;Allen & Bartlett, 2002;Orikasa et al., 2004). Similar eukaryotic genes have been cloned from heterokont algae such as Schizochytrium (Metz et al., 2001) and Ulkenia (Luy et al., 2009), although their gene structures are different from those of bacteria. The bacterial gene structures and domain structures of all pfa genes that are essential for the biosynthesis of EPA and DHA are well conserved (Okuyama et al., 2007).
Among the five pfa genes, the pfaE gene encodes an Sfp-type phosphopantetheinyl transferase (PPTase) of approximately 30 kDa, which catalyses phosphopantetheinylation via transfer of the 4 0 -phosphopantetheine prosthetic group from coenzyme A to a conserved serine residue in the carrier proteins, thus converting these proteins from their inactive 'apo' forms into their active 'holo' forms (Gehring et al., 1998). Orikasa et al. (2006b) classified Sfp-type PPTases into two groups: the first includes PPTases that are involved mainly in the biosynthesis of n-3 PUFAs, while the second includes PPTases that are involved principally in polyketide and/or nonribosomal peptide synthesis. The Sfp-type PPTases have three conserved domains: P1, P2 and P3 (Weissman et al., 2004). The P1 and P3 domains are responsible for coenzyme-A binding and domains P2 and P3 are responsible for Mg 21 binding (Reuter et al., 1999;Chirgadze et al., 2000). However, the PPTases that are required for the biosynthesis of n-3 PUFAs (i.e., PfaEs) are different from the other Sfp-type PPTases in some aspects: the P1 domain at their N terminus can be separately recognized as P1a and P1b in PfaE and is highly conserved among PfaEs. Moreover, PfaEs have an additional conserved P0 domain (L/VRxL/VLS) (where x is a nonconserved amino acid) upstream of P1a (Orikasa et al., 2006a).
The second representative group of PPTase includes the EntD protein of Escherichia coli, which is responsible for the synthesis of the siderophore enterobactin (Hantash et al., 1997). Interestingly, the genome of Photobacterium profundum SS9, which is an EPA-producing deep-sea bacterium, includes only one Sfp-type PPTase gene that was categorized into this second group (the EntD type; Sugihara et al., 2008). These findings suggest that this Sfp-type PPTase of P. profundum (SS9 PPTase) may be involved in the production of EPA, together with the other pfa genes (pfaA, pfaB, pfaC and pfaD) (Allen & Bartlett, 2002) located in the P. profundum SS9 genome (Vezzi et al., 2005).
Previously, we provided evidence that the SS9 PPTase gene complemented a pfaE-deficient pfa gene clone, pDHA3, which carried only pfaA, pfaB, pfaC and pfaD derived from the DHA-producing Moritella marina MP-1 (Sugihara et al., 2008). However, there is no evidence that pfaE is replaced with the E. coli entD gene. In the past, E. coli entD was considered as not being responsible for the biosynthesis of n-3 PUFAs, as neither EPA nor DHA was detected in any E. coli recombinant cells that carried vectors harbouring pfaE-deficient pfa genes prepared from Shewanella pneumatophori SCRC-2738 (Orikasa et al., 2004), M. marina MP-1 (Tanaka et al., 1999;Orikasa et al., 2006a, b) and P. profundum SS9 (Allen & Bartlett, 2002).
To elucidate whether pfaE is replaced with entD, we used the ASKA clone pCA24N<entD, which is a plasmid that expresses entD at high levels. This clone was obtained from the cloning vector collection of the E. coli Strain National BioResource Project (http://www.shigen.nig.ac.jp/ecoli/ strain/top/top.jsp). In this study, pCA24N<entD was coexpressed with pEPAD,1,2,3, which was a pWE15 cosmid clone carrying an EPA biosynthesis gene cluster that lacked pfaE from S. pneumatophori SCRC-2738 (Orikasa et al., 2004).

Bacterial strains and culture conditions
The bacterial strains and vectors used in this study are listed in Table 1. Escherichia coli DH5a recombinant cells were precultivated in Luria-Bertani (LB) medium supplemented with the indicated antibiotics at 37 1C for 16 h under shaking at 160 r.p.m. Portions of the culture were then transferred to the same medium and grown at 20 1C for 72 h, for EPA production.

Plasmid preparation and transformation
The ASKA library is a comprehensive E. coli K-12 ORF plasmid library in which one gene was cloned into each E. coli strain via gene cloning at the Nara Institute of Science and Technology (Kitagawa et al., 2005). The E. coli strain K-12 carrying pCA24N<entD was obtained from the National BioResource Project. The ASKA clone library is based on the E. coli K-12 strain AG1 and individual genes were cloned into the pCA24N vector (see Table 1).
Escherichia coli K-12 cells carrying pCA24N<entD were grown at 30 1C for 16 h in LB medium. pCA24N<entD was isolated using the mini-prep method and was used to transform E. coli DH5a cells carrying pEPAD1,2,3 using the heat-shock method. The transformed E. coli DH5a cells were grown in LB medium containing ampicillin at 50 mg mL À1 and chloramphenicol at 30 mg mL À1 at 20 1C for 72 h with shaking.
Fatty-acid analysis and sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) of proteins Transformed E. coli DH5a cells were collected by centrifugation. The precipitated cells were washed and were then directly subjected to methanolysis using 10% v/v acetyl chloride in methanol at 100 1C for 1 h. The resulting fatty-acid methyl esters were analysed by gas-liquid chromatography

Results and discussion
Coexpression of pCA24N<entD with pEPAD1,2,3 in E. coli DH5a cells pCA24N<entD was used to transform E. coli DH5a cells carrying pEPAD1,2,3. GC/MS analysis of fatty-acid methyl esters prepared from E. coli DH5a cells that carried pCA24-N<entD plus pEPAD1,2,3 revealed the presence of an unknown peak with a retention time of 30.2 min (Fig. 1a), which was not detected in E. coli DH5a host cells carrying only pEPAD1,2,3 (Fig. 1b). The retention time of the unknown peak was the same as that of the methyl ester of authentic EPA (data not shown). The GC/MS profile of the unknown peak shown in Fig. 1c was typical of methyleneinterrupted PUFAs, and analysis of the fragmentation profile using a program from the National Institute of Standard and Technology Databases (http://www.nist.gov./srd/htm) indicated that the profile of this unknown component was closest to that of EPA. Based on these results, this compound was identified as EPA methyl ester. The content of EPA was 9.2 AE 0.2% of the total fatty acids from cells grown at 20 1C for 72 h. PUFAs other than EPA were not detected.
Expression of the EntD protein in E. coli DH5a cells Figure 2 shows the SDS-PAGE profiles of E. coli DH5a cells carrying either pEPAD1,2,3 or pEPAD1,2,3 plus pCA24N<entD, in the presence or absence of IPTG. A significantly denser Fig. 1. Gas chromatograms of total fatty-acid methyl esters prepared from Escherichia coli DH5a recombinants, and electron-impact MS of the unknown fatty-acid peak. Escherichia coli carrying pEPAD1,2,3 plus pCA24N<entD (a) and E. coli carrying pEPAD1,2,3 (b). Mass spectrum of the unknown peak with a retention time of 30.2 min detected in a (c). Peaks A and B are those of cis-vaccenic and 3-hydroxyl tetradecanoic acids, respectively. Heneicosanoic acid (21 : 0) was used as an internal standard. Escherichia coli DH5a cells carrying pEPAD1,2,3 and cells carrying pEPAD1,2,3 plus pCA24N<entD were cultivated at 20 1C for 72 h in LB medium containing ampicillin at 50 mg mL À1 and ampicillin at 50 mg mL À1 and chloramphenicol at 30 mg mL À1 , respectively. band of 26 kDa was detected only in recombinant cells carrying pEPAD1,2,3 plus pCA24N<entD in the presence and absence of IPTG (lanes 3 and 4, indicated by arrows). Although the intensity of this band was slightly stronger in cells treated with IPTG than that observed in cells not treated with IPTG, it is evident that pCA24N<entD can be highly expressed without induction by IPTG. There is no information regarding whether or not ASKA library plasmids are expressed at a low temperature without an inducer (see Kitagawa et al., 2005). However, it is interesting to note that the EPA biosynthesis gene cluster from Shewanella oneidensis MR-1 cloned under the lacZ promoter on a high copy number plasmid, pBluescript SK(1), was highly expressed in the absence of IPTG in E. coli (Lee et al., 2008), which was assessed by the high content of EPA produced at 20 1C. Thus, the inducer (IPTG)-independent leaky expression of the ASKA library plasmid would be due to the low-temperature effects on this plasmid.
The entD gene sequence encodes a predicted protein of 23 579 Da. The size detected for the induced band (EntD; 26 kDa) corresponds to the protein with a His tag and four and five spacer amino-acid sequences at its N and C termini, respectively (Kitagawa et al., 2005). Native EntD was not detected in cells carrying only pEPAD1,2,3 (lanes 1 and 2 of Fig. 2). According to Armstrong et al. (1989), no native band of EntD was detected in E. coli strains by SDS-PAGE, unless it was overexpressed in the T7 promoter-directed highexpression system. The present results suggest that PfaE can be replaced by significantly higher levels of EntD. An undetectable level of expression of the native entD gene product of host E. coli DH5a cells (lanes 1 and 2 of Fig. 2) was insufficient to complement pEPAD1,2,3 lacking pfaE. The difference in the N-terminal domain structure between EntD and PfaE, and the addition of a His tag and of spacer amino-acid sequences to EntD would affect its affinity for its substrates, i.e. coenzyme A and/or a conserved serine residue in carrier proteins (such as acyl carrier proteins). This would be the most relevant reason for the partial replacement of PfaE with high levels of EntD. However, we have no idea how the addition of a His tag and of spacer amino-acid sequences to native EntD affects the structure and the catalytic activity of the Pfa enzyme complex. The pfaE from the EPA biosynthesis genes is compatible with that from the DHA biosynthesis genes (Orikasa et al., 2006a, c). It should be noted that the Sfp-type PPTases responsible for the biosynthesis of siderophores (and probably other polyketide compounds) and those responsible for the biosynthesis of n-3 PUFAs from terrestrial and marine bacteria, respectively, are partially compatible.
The PPTase involved in the production of EPA in P. profundum SS9 is an EntD-type enzyme (see above and Sugihara et al., 2008). This suggests that the PfaA-D proteins of this bacterium do not need high levels of EntD to synthesize EPA in an E. coli recombinant. Considering that the entD gene is expressed under low-iron conditions (Armstrong et al., 1989), no production of EPA in the E. coli recombinant grown in nutrient broth (2216 Marine Medium, Difco; Allen & Bartlett, 2002) might be caused by lack of the EntD protein.
The dense band detected below EntD corresponded to chloramphenicol acetyltransferase derived from the pCA24N vector, as assessed from its amino-acid sequencing.