Results

Four constructs (pMQ3A,B,C,D) were designed using different vanillin biosynthetic gene combinations (Figure #1E see Engineering section for more details). All of the constructs were made in the broad host range expression vector pUS250. The rationale for choosing this cloning vector was 1. The cumate inducer is cheap, non-toxic, and effective, and 2. The plasmids can potentially be easily moved to other hosts apart from E.coli in future to test the system further

The first three genes should be expressed from the cumate-responsive promoter in the vector while the second two genes should be expressed from a second cumate-responsive promoter sequence provided with the C3H gene (Sam5 or HpaBC). The use of an inducible promoter system should minimise the metabolic burden of these plasmids on the E.coli cells during the construction and other stages.

Figure 1.Structure of vanillin biosynthesis plasmids tested here

A mix of colonies were obtained on LB-Km plates after transforming the Golden Gate reactions into E.coli DH10B; most of these were white, providing preliminary evidence for successful cloning since this vector has a blue/white marker function. The overall efficiency of cloning and transformation was variable, with some reactions (pMQ3A,B) yielding dozens of colonies, and others (pMQ3C,D) only a handful.

Figure 2. Example of representative transformation plates.
Blue colonies (putative non-recombinant pUS250 vector) have been boxed.

Several representative clones of each type were patched to new plates, and cultures were regrown in LB-Km broth for plasmid extraction and sequencing. The nanodrop data for the resultant extracted plasmids is shown in Tables 1 and 2 below.

Table 1. Concentrations of pMQ3 plasmid DNA - expt 1.

Table 2. Concentrations of pMQ3 plasmid DNA – expt 2.

The figures below illustrate the comparison between our computational design and the actual genetic constructs made in the wet lab. Sequencing analysis of all the clones revealed that only one clone, pMQ3A Clone 7, was flawless, matching the design exactly. All other clones (pMQ3B, C, and D) exhibited either major deletions or minor point mutations (Figures #3, 4, 5, 6) . A summary of the problems with each plasmid clone are provided in Table 3 below.

Table 3 Summary of plasmid sequencing results from Engineering Cycle 1

Figure 3. Alignment of constructed pMQ3A plasmids with expected sequence

Figure 4. Alignment of constructed pMQ3B plasmids with expected sequence

Figure 5. Alignment of constructed pMQ3C plasmids with expected sequence

Figure 6. Alignment of constructed pMQ3D plasmids with expected sequence

When recombinants were patched from single colonies to new LB-Km plates, it was noticed that some of the pMQ3A and pMQ3C clones produced a brown diffusible pigment (shown for pMQ3C in Figure 7). This was never seen in controls that lacked vanillin biosynthesis genes i.e. DH10B(pUS250) or RARE(pUS250) (see Figure 9)

Based on these results, clone pMQ3A_7 (correct sequence) was set aside for later phenotypic testing, and some of the pMQ3B, pMQ3C, and pMQ3D clones that had only single point mutation errors were targeted for repair efforts – in theory this can be done by amplifying the plasmids with mutagenic primers, then self-ligating the products.

Figure 7. Unexpected pigment production from recombinant E.coli clones

To address the mutations seen in plasmid clones pMQ3B, pMQ3C, and pMQ3D, we designed specific primers to target the inserted or deleted bases and performed a PCR and self-ligation to restore the intended sequence. The results of these PCRs are shown in Figure 8.

In the first attempt at this PCR, no bands of the correct sizes were seen. In the second attempt, a faint band of the correct size was seen for pMQ3B but many other bands were also present in this PCR. In the third attempt, no correct-sized sized bands were seen although a strong band of the wrong size was obtained from pMQ3C.

Figure 8. PCR of vanillin plasmids to repair mutations.

Panel A: First attempt on 7 Aug. (standard protocol) Panel B. Second attempt on 8 Aug. (increased amount of template DNA and used touchdown PCR method) Panel C. Third attempt on 12 Aug. (increased annealing temperature and testing DMSO addition) M: DNA marker, B: PCRs from pMQ3B, C: PCRs from pMQ3C, D: PCRs from pMQ3D

We decided to repeat the original GoldenGate reactions, transformations, and plasmid extractions, hoping to find correct clones of types B, C, D using the correct golden gate protocol (2-step cycling) instead of a short cut method (37C 1 h) and via screening more clones of each type by whole-plasmid sequencing. As a step towards optimising the system, we also tested a different expression host (E.coli RARE) that has been engineered to enable accumulation of aldehydes (e.g. vanillin) (1), and tested the effect of temperature on growth of the different recombinant clones (Figure 9).

Again in this cycle of testing, the brown pigmentation was seen from DH10B(pMQ3C) clones - three of these (11, 15, 20) were retained for further testing and sequencing. An unexpected finding was that we obtained only a handful of transformants of E.coli RARE containing any of the biosynthetic plasmid constructs despite getting good transformation efficiencies with the pUS250 vector with the same batch of competent cells (Figure 9) – this could indicate that the vanillin genes are specifically toxic to the RARE host. Lower temperatures seemed to favour growth of E.coli containing the pUS250 vector but did not have a big effect on the recombinant clones.

Figure 9. Growth of Different Recombinant Clones

Many more new clones of pMQ3B were picked and patched and then grown up for plasmid extractions. Nanodrop data for these plasmids is shown in Table 4. Note that some clone numbers are missing from the set since these cultures did not grow.

Table 4. Concentrations of pMQ3B plasmid DNA – clones 14 to 44

Cultures of pMQ3A_7 and pMQ3C_11,15, 20 transferred into LB-Km broth containing tyrosine (0.5 g/l) and cumate (100 µM) to enable full induction of the biosynthetic enzymes and hopefully produce vanillin! Supernatants of these cultures will be used for LC-MS analysis of metabolites.

After overnight incubation, the 3C cultures are very dark brown, while 3A only slightly brown. Interestingly, the 3A and 3C cultures do smell different to control E.coli (pUS250) although no-one who volunteered to smell-test thought it smelled like vanilla.

Figure 10: LC-MS of pMQ3C_11