Expression plasmid and ACC1 knockout

We created an expression plasmid that enables S. cerevisiae to take up external malonate as well as establishes a novel pathway for malonyl-CoA synthesis. To achieve this, the plasmid consists of the integrated matB gene from R. leguminosarium bv. trifolii under the control of constitutive promoters on the backbone (pNK26) carrying an ampicillin resistance for selection in E. coli and the TRP1 gene for selection in S. cerevisiae. The plasmid was assembled via Gibson assembly reaction and Yeast assembly method.

For amplification of the matB and mae1 gene we wanted to use codon optimized, synthesized nucleotide sequences. Since we had problems amplifying the received nucleotide sequence which delayed the progress of this project, we decided to reorder the sequence of the codon optimized matB and amplify the mae1 gene using genomic DNA from S. pombe 2698(lov)972 h- strain. The amplified matB fragment had a total length of 1615 bp and possessed short overhangs homologous to the backbone and the promoter region for the assembly reaction that were added by using specially designed primers (Fig.1A). Mae1 was amplified using genomic S. pombe DNA. The amplified fragment had a total length of 1398 bp with overhangs homologous to the backbone and the promoter region (Fig.1B). The bidirectional promoter region consisting of the constitutive promoters pPGK1 and pTEF1 had a total length of 1409 bp and the backbone carrying the marker genes had a total length of 5962 bp. For both amplifications we used pNK26 as template.

For assembling the fragments and creating the final expression plasmid we performed Gibson assembly as well as Yeast assembly to directly introduce the expression plasmid into the S. cerevisiae strain CEN.PK113-7D. The Gibson assembly product was transformed into E. coli NEB Turbo. The transformation was confirmed by colony PCR and analytical restriction using the restriction enzyme XhoI (Fig.1D). Due to limitations in time, we were not able to transform the expression plasmid into S. cerevisiae CEN.PK113-7D.

In addition to establish the novel malonyl-CoA synthesis pathway, a deletion of the acc1 gene in S. cerevisiae is required to disable yeast to produce malonyl-CoA from acetyl-CoA [1]. We created a knockout construct consisting of the his3 gene flanked by a 200 bp nucleotide sequence homologous to the upstream area of the E. coli acc1 gene and a 200 bp nucleotide sequence homologous to the downstream area of the E. coli acc1 gene. We were able to amplify the fragments via Gibson assembly and cloned them into a pUC19 backbone, which was subsequently transformed into E. coli NEB Turbo and verified performing an analytical digest using the restriction enzyme PstI. A knockout of acc1 could not be performed due to its lethality and lacking of the expression plasmid which could compensate for the absence of malonyl-CoA.

Operon plasmid

The operon consists of genes that will direct the flux into the beta alanine pathway in E. coli towards malonic acid production [2]. In the end, the plasmid consists of panD from C. glutamicum, aspA from E. coli, ppc from E. coli, yneI from E. coli and mae1 from S. pombe. The operon was designed to be controlled by the constitutive promoter BBa_J23108. For the terminator, we chose the double terminator BBa_J61101. The backbone consists of the E. coli origin p15A ori and the chloramphenicol resistance gene under its native cat promoter as the marker-gene (part of pACYC184). Additionally, we integrated two different spacer sequences between adjacent genes to avoid identical overhangs for the Gibson assembly reactions. To realize the design, we had to assemble the parts via Gibson assembly reaction since the yeast assembly methods would require additional sequences in the plasmid for selection and replication. We planned to pre-assemble some parts of the final plasmid, in order to avoid Gibson assembly reactions with more than four fragments. We had the templates for the amplifications of panD and pa0132 codon optimized and synthesized for E. coli. The promoter for the operon was part of the synthesized sequence for panD. With this we were able to amplify the first two parts of the operon via PCR. The whole fragment including the promoter, the panD gene, one of the two spacer sequences between adjacent genes and a small overhang for the Gibson assembly reaction had a length of 600 bp (Fig.1C). The synthesized sequence for the codon optimized pa0132 gene also possessed overhangs for adjacent genes for Gibson assembly. The sequence had a total length of 1499 bp (Fig.1A). The last part of the pre-assembled plasmid Gibson1 was the E. coli native aspA gene consisting of 1437 bp (Fig.1A). Those three fragments were cloned into a pUC19 backbone and transformed into E. coli NEB Turbo. The plasmid was verified via an analytical digest using Kpn1. Subsequently it was used as template for the amplification of the combined fragment for the promoter panD, aspA and pa0132 with a total length of 3471 bp (Fig.2A).

For the second pre-assembled plasmid Gibson2 we amplified the E. coli native ppc gene using genomic DNA from E. coli strain MG1655. The fragment consisted of 2650 bp carrying the ppc gene as well as upstream and downstream overhangs homologous to the adjacent genes (Fig.2C). The overhangs contained one of the two different spacer sequences and were created using specially designed primers. The second fragment for the pre-assembled plasmid Gibson2 was amplified using genomic DNA from E. coli MG1655 strain. The fragment had a length of 1432 bp and consisted of the coding area of the E. coli native yneI gene with overhangs homologous to adjacent genes (Fig.1A). For the last part we used genomic DNA from S. pombe 2698(lov)972 h- strain to amplify mae1 with a length of 1397 bp (Fig.1B). The three fragments were assembled by Gibson assembly reaction and cloned into a pUC19 backbone with an ampicillin resistance gene. The transformation was verified by colony PCR. It was later used for the amplification of the combined fragment of ppc, yneI and mae1 with a total length of 5480 bp (Fig.2B).

For the final Gibson assembly we amplified the terminator BBa_J61101 using DNA from the distribution plates and the backbone using the plasmid pACYC184 as template. The Gibson assembly product was transformed into the E. coli strain NEB Turbo, which was plated on chloramphenicol containing media.

To verify the presence of the construct a 2352 bp long fragment containing the backbone, the promoter and the panD gene was amplified via PCR (Fig.3A) and a 8939 bp long fragment containing the backbone, the terminator BBa_J23108, the combined fragment of the second pre-assembled plasmid and pa0132 (Fig.3B).

Malonte that was secreted into the supernatant was detected by UPLC-MS. Both the mutant E. coli and WT were measured through a modified protocol from Schwander et al. [3]. We compared the detected malonate levels related to the optical density of the control strain NEB Turbo with the transformed strain carrying our assembled operon plasmid. For the NEB Turbo transformed with the operon plasmid the ion count normalized to the OD600 was at 5906.52 which is doubled the amount of the control (Fig.4). Both samples were treated the same. Future work should include further validation of the system's functionality in addition to these promising results.

Protein-based dependency

Our initial attempt was to use a previously described system, in which the E. coli protein YebF is fused to the N-terminal end of the protein of interest [5]. After transport of the fusion protein to the membrane due to a signal sequence from YebF, it will get cleaved while the protein of interest gets exported to the extracellular space. We amplified yebF with an additional His-tag performing colony PCR from the E. coli genome and ligated it into a pET28b vector under the control of an IPTG-inducible promoter. Even though the construct was verified through PCR, induction with different concentrations of IPTG showed no expression of the protein on SDS-page and via Western Blot. We suspected this could be due to the unknown cleavage mechanism that is part of the YebF export system. Hence, we fused the common reporter gene ß-glucunronidase to YebF and cloned it into pET28b, respectively. Again, induction with IPTG led to no detectable expression. The second approach on establishing a functional protein export was by utilizing the flagellum assembly machinery of E. coli. The flagellin encoding gene fliC is regulated by an 172 bp signal sequence upstream of the protein coding region, containing promoter as well as RBS, and has been shown to be suitable for protein export in an E. coli strain with knocked out FliC and FliD [6]. Both genes are essential for the tightly regulated flagellar assembly and a knockout should result in unregulated expression of a protein fused to the fliC signal sequence. Therefore, we cloned the signal sequence, which was amplified from the E. coli genome to an His-tagged mRFP (BBa_J04450). The fusion protein got ligated into a tet-inducible pASK vector, since strong overexpression led to interference with the existing promoter of the signal sequence. Simultaneously, we tried to knock out fliC and fliD using the E. coli strain AB330, which is capable of homologous recombination using an amplified kanamycin resistance with 50 bp overhangs homologous to fliC and fliD. Even though the knockout was successful and could be verified indirectly through the formation of adhesive E. coli cells lacking the flagellum the culture did not grow sufficiently in the remaining time to transform it with the construct. Another attempt to show protein expression was transforming the construct into BL21 and inducing it with different concentrations of anhydrotetracycline. Again, no expression could be detected on either SDS-page or via Western Blot, which should be due to the aforementioned tight regulation of FliC expression in an E. coli strain with functional genomic fliC.