Project Overview

Part 1 – Cloning

Design of gene construct

With two main pathways identified for glyphosate degradation, the analysis of enzymes involved in the glyphosate reducing step were of highest interest. Sequences encoding for two glyphosate dehydrogenases, goxA and goxB and one C-P-cleaving multimeric enzyme complex, phnGHIJK were obtained from NCBI genebank^1,2,3. Codons were modulated for expression in Bacillus subtilis using IDT codon optimizer tool.

Regulation of high-level protein production within the bee gut microbiome needs to be considered for the choice of promoter. Here, the strong constitutive promoter P_veg would provide constant protein expression during all growth phases. However, continuous expression may lead to increase in stress-level for the cells, inhibiting the recombinant protein production process. To enable better control over this process, a construct under IPTG-inducible promoter P_hysp is introduced. A constant feed of inducer would be required for application in Apis mellifera making it not feasible as a probiotic. To test the proof of concept, it’s effects will still be analysed. Furthermore, a third promoter was acquired. The strong stationary phase promoter, P_ylb exhibits protein expression when late log phase/early stationary phase is reached⁴. Hereby, glyphosate degrading proteins are only produced upon sufficient growth meaning colonisation of bees gut microbiome. When come into contact with other non-target organisms, this regulatory response creates an additional safety mechanism.

Besides different promoter regions, each gene of interest (GOI) was flanked upstream with an 5’UTR-region including a Bacillus subtilis specific ribosome binding site (RBS) as well as two stop codons and a terminator sequence downstream of the gene. For restriction-based cloning, 5’-HindIII and 3’-SalI (for pCW7) and NheI (for pCW101) restriction sites were introduced. The inserts as seen in Figure 1 were synthesised by IDT and TWIST Bioscience.

Design of vector

Two strategies for recombinant protein production were followed – chromosomal integration and plasmid expression. Specific functional units are considered in the design of the vector plasmid. The low copy-number plasmid pCW7 carries two ori allowing replication within E. coli and B. subtilis and thus suitable as a shuttle vector. However, homologous recombination requires a common set of integration sites present in chromosomal DNA and plasmid. Here, pCW101 inhabits amyE integration sites used for chromosomal integration in B. subtilis. Both vectors were kindly provided by Claes von Wachenfeldt.

Restriction and Ligation

Traditional cloning was the method of choice for the introduction of our GOI into the designed vector. Therefore, digestion of the synthesised GOI carrying specific restriction sites as well as the vector was performed. Cloning into pCW7 was based on creating 5’- and 3’-end sticky ends by using HindIII and SalI, respectively. A different 3’-end was created for cloning into pCW101. Although carrying the same restriction site, their close proximity was assumed to create improper overhangs thus reducing ligation specificity. NheI was therefore used for cloning work into pCW101 backbones. Ligation reaction was performed with T4 ligase (See further protocols).

With the described method, transformants for goxB inhabiting the ylb and hysp promoter inside the integrative pCW101 plasmid were obtained. To evaluate these constructs, colony PCR with backbone specific primers should allow amplification of integrated fragments. Figure 3 A shows clear distinguishable bands for several clones for pCW101-goxB-ylb and pCW101-goxB-hysp at around 1600 bp (size of fragment + 260 bp generated by primers binding inside the vector backbone). Sequence similarity of these transformants was further tested by sequencing. The constructs analysed showed 100% alignment (Data not shown) – providing proof for successful cloning.
However, no transformants could be obtained with traditional cloning for goxA or phnGHIJK fragments. Further, ligation into the pCW7 backbone did not produce any constructs whatsoever. Shifting to the exonuclease based cloning method called Gibson Assembly® was expected to increase cloning efficiency (See Engineering success) . Here, transformants of correctly assembled pCW101-goxB-veg as well as one pCW7-goxB-ylb fragment could be obtained (Figure 3 B&C). Sequencing results concluded homologous sequence similarity for pCW101-goxB-veg (Data not shown). However, sequencing of pCW7-goxB-ylb disclosed many missense mutations to be present inside the gene and was therefore not considered for further analysis.

Figure 3| Agarose images of Colony PCR performed in E. coli transformed with goxB-constructs. (A)pCW101-goxBylb and pCW101-goxBhysp (D3). Both fragments were cloned via restriction and ligation using 5'-HindIII and 3'SalI as restriciton sites. (B) goxB-veg cloned into pCW101 via Gibson Assembly. (C) goxB-ylb cloned into pCW101 via Gibson Assembly.

Part 2 - Chromosomal integration into Bacillus subtilis

For our designed probiotic, Bacillus subtilis was chosen as the optimal expression host. Transformation of the constructed plasmids into our target organisms was performed on the basis of self-induced competence. The generated constructs replace α-amylase encoding genes inside the Bacillus chromosome via homologous recombination. Upon integration, the strain loses the ability to degrade starch, a characteristic that can be utilised to determine correct chromosomal incorporation.

This analysis method was used to determine successful goxB transformants. Figure 4A shows cells of non-transformed Bacillus. Areas surrounding these colonies appear clear implying starch hydrolysis. Within the second half of starch agar plates seen in Figure 4 B-C, the dark coloured medium indicates the presence of bacteria with correctly integrated fragments. Colonies growing on the upper part however show similar behaviour as non-transformed Bacillus. It is assumed that these cells either failed to integrate our GOI or low transformation efficiency resulted in low to no uptake of recombinant plasmid.

Protein Expression

To develop an active glyphosate degradation system, overexpression of the goxB encoding gene is necessary. Therefore, overall protein production and thus the functionality of the introduced promoter was assessed via SDS. Cultivation of recombinant Bacillus was hereby performed in minimal media supplemented with glucose and glyphosate. To simulate growth conditions similar to the gut microbiome of bees, cells were incubated at 35°C – the average temperature of bees body temperature within the beehive⁵.

Samples at different time points were taken and analysed via SDS. Figure 5 shows two distinguishable bands with sizes around 20 and 55 kDa visible in all lanes. Between 40 and 50 kDa, smeared areas indicate the presence of further proteins.

Overexpression of glyphosate dehydrogenase under different promoters is expected to result in high intense bands with a molecular weight of 46 kDa. However, neither a band is visible in this area nor a distinction between WT and goxB carrying strains can be made.

To eliminate the possibility of deviant separation of either the marker or protein sample, bands at 20 and 55 kDa were analysed via mass spectrometry and screened for potential proteins within the NCBI database. Ladder was identified as vegetative catalase, an extracellular enzyme, secreted upon oxidative stress⁶. Furthermore the second protein can be connected to alkyl hydroperoxide reductase C – another enzyme playing a crucial role in the protection of cells against oxidative stress⁷. Additionally the smeared area between 40 and 50 kDa was inspected for evidence of target protein production. However, no peptides responding to glyphosate dehydrogenase could be identified.

Moreover, samples of the supernatant after 24 h cultivation were taken and further studied for proteins secreted into the media. Here, only bands corresponding to the oxidative stress reducing proteins were recognised (Data not shown). However, this does not exclude potential target protein secretion as the supernatant was highly diluted and thereby strongly reduced the intensity of a possible low expressed goxB.

Part 3 – Glyphosate Degradation

Minimal Inhibitory assay

Glyphosate strongly inhibits bacterial growth. To determine, the lowest concentration on which 100 % of growth is inhibited, an minimal inhibitory concentration (MIC) assay was performed. Here, B. subtilis 1S129 WT next to engineered strains carrying the goxB encoding gene were analysed. Cells from overnight cultures grown in minimal media supplemented with glucose and antibiotics (OD₄₅₀~0.5-07) were plated on minimal media plates with RoundUp©. RoundUp© is a commonly used herbicide made of concentrated glyphosate (360 g/L) dissolved in potassium salt (411 g/L). This chemical was supplemented in different concentration to agar plates (50 mg/L to 4 g/L). Additionally, IPTG was further added to strains with goxB under IPTG-inducible P_hysp.

Table 1| Minimal inhibitor concentration (MIC) of engineered B. subtilis 1S129 with goxB on different concentrations of Roundup^Tm (Glyphosate 360 g/L). Cells were incubated at 35°C for 5 days. n=2

	WT	PCW101	goxB-veg	goxB-ylb	goxB-hysp
MIC [g/L]	0.5	0.5	1	0.5	0.5

For a time period of 5 days, plates were inspected for growth. After the first day, colonies on plates with up to 200 mg/L glyphosate concentrations were detected. Growth was hereby visible for all tested strains. However, further incubation resulted in enhanced growth for B. subtilis strain carrying goxB under veg promoter up to 1 g/L. All other variants were inhibited by this glyphosate concentration. In comparison to the WT, only goxB-veg produced an altering MIC thus indicating that the introduced glyphosate dehydrogenase contributes to greater cell survival. Based on these results B. subtilis 1S129 seem to generally carry a high tolerance against glyphosate, a highly beneficial characteristic for usage as probiotic in bees. Limited glyphosate resistance has further been reported for some B. subtilis species⁸. Nevertheless, such high tolerance was reported for the first time in this strain.

Growth experiment on glyphosate

Regular OD measurements of cells incubated with glyphosate should give insight about the individual growth behaviour of each variant. For the selection of glyphosate concentration 200 mg/L were used - the MIC for B. subtilis after 24 h. After inoculation of cultivation media, cultures were incubated for 2 hours to reach the start of log-phase. Glyphosate in the form of RoundUp© was then added. Note that for variant goxB-hysp, IPTG was added first and incubated for another 2 hours to induce protein expression before glyphosate supplementation.

Difference in growth between strains w/o and w/ glyphosate were expected. However, from Figure 6, a clear distinction between each growth curve is difficult. Furthermore, B. subtilis WT expresses slightly enhanced growth compared to organisms with integrated goxB coding genes. Expression of these proteins may increase the cell's stress level therefore slowing down its growth. As no difference between growth on glyphosate compared to no glyphosate cultivation is visible, the cells metabolism seem not to be affected by addition of this chemical. The slight alteration in regard to the WT strain can be reduced to the metabolic burden generated by recombinant protein expression. However, this hypothesis could not be supported by generated SDS results (see Figure 5)

Observation of glyphosate degradation using UV-vis

As described above, glyphosate does not seem to impact growth of the engineered bacteria – making it a great probiotic host. To further study the degradation efficiency of the introduced glyphosate dehydrogenase, samples of culture supplemented with glyphosate (RoundUP©) were taken every 24 h. Cultivation was performed under the same conditions as defined above. External glyphosate concentrations were measured via UV-vis spectrophotometry⁹.

In order to obtain absorbance, glyphosate had to undergo an additional derivatisation process based on FMOC-Cl. This light absorbing compound reacts with glyphosates’ amino-group making glyphosate detectable.

For correlation of measured absorbance to glyphosate concentration, a standard curve was created (Figure 7). From this linear regression, the glyphosate concentration in the supernatant could be determined. Expressed glyphosate dehydrogenase should break down assimilated glyphosate, thus reducing the external glyphosate concentration inside the media. With a starting value of 200 mg/L, a steady decline over time was expected. However, Figure 8 shows the contrary. Even firstly observed reductions after 24 h cannot explain the elevated level of glyphosate above the threshold of 200 mg/L after 48 h – in all samples. Furthermore, to exclude general degradation of glyphosate at cultivation conditions, a control of glyphosate in minimal media without cells was analysed. Hereby, no glyphosate could be detected within these samples (Data not shown). This strongly indicates that the measured absorbance did not represent the actual glyphosate concentration.

In order to make sense of these results, one has to look back at the derivatisation reaction (Figure 9). The utilized derivatisation agent FMOC-Cl binds to the amino group – a functional group not only present in glyphosate. With peptides and proteins consisting of amino acids carrying this functional group, interaction with secreted proteins is very likely. Hereby, Bacillus subtilis in particular is known for its many secretory pathways. Furthermore, Yamane et al. described a strong connection between protein secretion and environmental changes¹⁰. By cultivation on glyphosate, this effect might be induced. The increase in absorbance would thus correlate towards growth related peptide or protein secretion. However, this does not explain the low measured absorbance in samples incubated with only glyphosate and media. One explanation might lay in different stock solutions employed for standard curve and degradation analysis. For the standard curve highly pure analytical glyphosate was used. All degradation experiments however were performed with RoundUp©. The concentration of glyphosate (360 g/L) in this product was not further confirmed and relied only on the package description. Alterations in the starting glyphosate content would have further influenced the amount within the culture, giving non reliable results.

Discussion and conclusion

The focus of this project rested in the construction of different glyphosate degrading systems in non-spore forming Bacillus subtilis. Three enzymes under the control of different promoter functions should be analysed and compared in terms of degradation efficiency under conditions similar to bees gut microbiome. Therefore, integration of these protein encoding genes into Bacillus chromosomal DNA was the main approach. Here, we successfully cloned goxB into an integrative vector and transformed it into E. coli to achieve high plasmid replication. Transformation of these obtained recombinant plasmids into Bacillus generated three engineered strains with goxB + promoter integrated into their chromosomal DNA. However, subsequent expressions did not produce the corresponding protein. By changing the detection method from SDS to western blot, the presence of our target protein might be defined when using highly sensitive antibodies. Nevertheless, protein expression might be generally suppressed. The lac operon is still present inside the integrated fragment even with new introduced promoters. Potential lac repressor binding sites within the promoter regions of P_veg and P_ylb might contribute to inhibited protein expression. Nevertheless, the construct carrying P_hysp should still produce our target protein. Thus, sequencing of the integrated genes for mutations inside the promoter might provide more insight on why the system failed to produce any proteins that could be detected.

Even though protein expression could not be directly confirmed, analysis of degradation activity through the introduced glyphosate dehydrogenase may provide evidence for its production. Several methods should screen for differences in behaviour of the engineered bacteria. MIC assay provided evidence for increased survival within goxB-veg carrying Bacillus at higher glyphosate concentrations. However, studying glyphosate related effects on growth did not give a clear distinction between native and altered strains. Furthermore, activity evaluation on glyphosate based on spectrophotometric measurements did not produce any reliable results. Glyphosate unspecific binding of the derivatization agent FMOC-Cl can be considered as the methods bottle-neck. Highly sensitive chromatographic methods such as LC/MS-MS might generate more selective results⁶. However, by relying on derivatization agents binding to the amino group, undesirable binding to secreted proteins and peptides will always accompany these outcomes. Thus, a reliable detection method needs to be established to determine glyphosate breakdown of cell-based systems.

Nevertheless, from enhanced MIC for goxB-veg integrated in B. subtilis alone, no proof of a functioning glyphosate degradation system can be concluded.

Results

What we achieved in the lab (or didn't)