Overview

  In the PACOmega project, our initial idea was to generate an alternative resource of docosahexaenoic acid (DHA) for cardiovascular disease (CVD) prevention. Orikasa et al. showed that the co-expression of pfa A, pfa B, pfa C, pfa D and pfa E from Moritella marina promoted DHA production in E. coli (Orikasa et al., 2006). Accordingly, we designed and built the biobricks for DHA production. However, several studies published this year suggested that eicosapentaenoic acid (EPA), but not DHA, is the main effector in Omega-3 polyunsaturated fatty acid (PUFA) mediated CVD prevention (Mason et al., 2022; Sherratt et al., 2022). To re-design the biobricks for EPA expression without changing the main theme, we explored more references. Fortunately, we learned that swapping the original pfa B gene with that from EPA-producing Shewanella pneumatophori can switch DHA production to EPA production (Orikasa et al., 2009). Therefore, we modified the original DHA synthesis system to a switchable PUFA synthesis system by exchanging the origin of pfa B gene.

  Furthermore, we decided to enhance the PUFA production in our switchable design. Accordingly, we searched the literature and found that ectopic expression of acetyl-CoA carboxylase (ACC) enhances biogenesis of fatty acid (Davis et al., 2000). Therefore, we designed a co-expression system of Acc genes with pfa genes to enhance PUFA production.

  Finally, since the function of ACC is to produce malonyl-CoA, the raw material of fatty acid, we designed the addition of the chemical cerulenin to block the competing fatty acid synthesis pathway and in turn enhance the biogenesis of PUFA (Wan et al., 2016; Giner-Robles et al., 2018).

  In the deep-sea bacteria Moritella marina and Shewanella pneumatophori, the pfa genes are translated into protein and cooperate as a megasynthase to synthesize PUFA by the polyketide synthase (PKS) pathway. The pfa A-D genes encode proteins containing many functional domains for the PKS pathway, such as acyl carrier protein (ACP), 3-Ketoacyl-ACP reductase (KR), 3-Ketoacyl synthase (KS), 3-Hydroxydecanoyl-ACP dehydratase (DH), and acyltransferase (AT). The pfa E gene encodes phosphopantetheinyl transferase (PPTase).

  These functional domains in pfa genes are involved in the extension and modification of the intermediates and formation of the final PUFA products. In brief, the PKS pathway starts from the activation of ACP by pfa E (PPTase). After that, the AT domain will transfer the malonyl-CoA, which is the raw material of PUFA, to the ACPs. The backbone of malonyl-ACP will be further elongated by KS and AT domains, while KR, DH and ER domains are responsible for modification of the intermediate products. Finally, the PUFA will be produced (Gao, Wang & Tang, 2010; Mindrebo et al., 2020).

The functional domains:
KS: Catalyzes a decarboxylative Claisen-like condensation reaction to extend the carbon chain of polyketide.
DH: Reduces hydroxyl groups to enoyl groups.
KR: Reduces β-ketone groups to hydroxyl groups.
ER: Reduces enoyl groups to alkyl groups.

  In Moritella marina and Shewanella pneumatophori, the pfa gene cluster encodes five open reading frames, namely pfa A, pfa B, pfa C, pfa D and pfa E. Each gene contains one or more functional domains to perform carboxylation, dehydration, and reduction of the ketoacyl group and double bonds of the carbon chain in the PKS pathway. In Moritella marina, the pfa genes work together to generate DHA, while Shewanella pneumatophori generates EPA. The exact functions of pfa A-D genes are unclear. To the best of our knowledge, the pfa B gene is the key molecule to determine the final product (e.g. EPA and DHA) (Orikasa et al., 2009).

  The co-expression of pfa A, pfa B, pfa C, pfa D and pfa E genes from Moritella marina can produce DHA in E. coli. However, the total length of pfa genes is almost 20 kb, which makes it difficult to express all in one vector. Therefore, we decided to clone these genes into two different vectors to avoid the potential expression problem (Rosano and Ceccarelli, 2014).

  Among all the pfa genes, pfa A and pfa C are the longest (~8 kb and ~6 kb, respectively). Accordingly, we cloned the largest subunit, pfa A into pColdI vector with pfa D (1.6 kb) for expression, while the pfa C gene is cloned into pSTV28 vector with pfa B (2.6 kb) and pfa E (0.8 kb) genes. With this design, these two pfa-expressing vectors will have similar size and avoid the problem of transformation efficiency caused by plasmid size (Rosano and Ceccarelli, 2014).

  Because of the size limitation, it is too costly or impossible to order biobricks of pfa A (~8 kb) and pfa C (~6 kb) from the company. To keep the cost in a reasonable range, we decided to clone the endogenous pfa A, pfa C and pfa D genes from Moritella marina by PCR, while pfa B, for DHA production (Moritella marina), pfa B', for EPA production (Shewanella pneumatophori) and pfa E genes were ordered.

  To clone pfa A and pfa D into pColdI vector, we first divided the pfa A gene into two parts, pfa A1 (4 kb) and pfa A2 (4 kb), since the length of the pfa A gene makes PCR amplification difficult. The restriction enzymes (REs) for cloning are carefully selected to avoid destroying pfa A or pfa D genes during cloning. Ribosome binding sites (RBS) are introduced in front of the pfa A and pfa D by PCR. Finally, we selected NdeI and EcoR1 for cloning pfa A1, EcoRI and SacII for cloning pfa A2, as well as SacII and XbaI for pfa D.

  To clone the pfa B, pfa C and pfa E genes into pSTV28 vector, REs BamHI and SacI were selected for pfa C cloning, and the RBS were introduced by PCR. The purchased pfa B, pfa B' and pfa E genes were codon-optimized to avoid RE digestion of BamHI and SacI. Finally, we selected EcoRI and BamHI for cloning the pfa B gene, and SacI and HindIII for cloning the pfa E gene. Since the pfa B and pfa B' are both codon-optimized, it is easy to switch these two pfa B genes using the same REs.

  All the sequences of the pfa genes and designed clones are provided in the basic and composition parts.

  To enhance EPA production, we decided to express Acc genes from Corynebacterium glutamicum, AccBC (1.8 kb), AccD1 (1.6 kb) and AccE (0.3 kb), in E. coli. The protein translated from AccBC, AccD1 and AccE genes could form the functional acetyl-CoA carboxylase (ACC) to produce malonyl-CoA. The reaction includes two steps. In the first step, the ATP-dependent biotin carboxylase (ACCBC) catalyzes the transfer of a carboxyl group to biotin, which is linked to the biotin carboxyl carrier protein (ACCE). In the second step, the transcarboxylase (ACCD1) transferred the carboxyl group from carboxybiotin to acetyl-CoA to form malonyl-CoA.

  In our ACC expression design, the biobricks of Acc genes are all ordered and codon-optimized for expression in E. coli. For convenience, we ordered the AccBC gene with RBS as one biobrick, while the AccD1 and AccE were ordered with their RBSs as the other biobrick. The REs flanking the AccBC biobrick are NdeI and EcoRI, while the REs flanking the AccD1 and AccE biobrick are EcoRI and XbaI. These two biobricks will be cloned into pET28a vector with the selected REs.

  In E. coli, the endogenous fatty acid synthesis (FAS) pathway is responsible for the fatty acid synthesis. In the FAS pathway, FabH initiates the fatty acid synthesis by condensing the acetyl-CoA and malonyl-ACP (Acyl carrier protein) into 3-Ketoacyl-ACP. The 3-Ketoacyl-ACP is further elongated by FabB and FabF.

  The function of the chemical cerulenin is to inhibit FabB and FabF, two main synthases in the FAS pathway. Furthermore, previous reports showed that the exogenous pfa system was not inhibited by cerulenin in E. coli. Thus, addition of the chemical cerulenin to the E. coli which co-express the pfa genes and Acc genes should increase the PUFA production.