To investigate MtbFBA and for hit confirmation of the generated virtual hits, the protein was produced in E. coli and purified to apparent homogeneity. E. coli C41 DE3 cells containing fba in the pET-3a vector were inoculated and grown at 37 C until an OD600 of 0.6 and then subjected to cold shock for an hour to induce the production of chaperon proteins. Expression was then induced by the addition of IPTG (Figure 6a). After two hours, the cells were harvested via centrifugation and lysed. MtbFBA was purified from the lysate using a two-step purification strategy: initially passed over an IMAC, Co2+ column, then immediately over a size exclusion chromatography column. Analysis by SDS PAGE showed bands in every lane at ~36 kDa, which corresponds to the molecular weight of the expressed FBA (Figure 6b insert), with a purity of >95% (deduced using Image J) yielding 51.3 mg of FBA from 1 L culture. Which was much higher than expected and higher than the previously reported expression and purification, using E. coli BL21(DE3) cells and a one-step purification strategy (7.67 mg after purification from 1 L culture). (10.1016/j.enzmictec.2008.08.009). As expected, the eluates were most pure, with the highest concentration, in the fractions corresponding to the elution peak seen in the chromatogram produced from SEC (A8-B10; Figure 6b, only SEC chromatogram shown) and were pooled accordingly. The protein stock concentration was measured using a nanodrop, giving a concentration of (0.805 mg/ml).
Essay due? We'll write it for you!
Dynamic Light Scattering
The hydrodynamic size and molecular weight of MtbFBA were estimated by DLS. A stock of FBA in protein buffer (0.46 mg/ml) was run in triplicate and the results were averaged. Initial analysis by Size distribution by intensity obtained two peaks (Figure 7a), though the second peak is not visibly displayed as it represents only 7.7 % of the overall intensity. This secondary peak was expected and corresponds to the glycerol present in the protein buffer (not seen due to scaling). However, the estimated molecular weight of MtbFBA was much greater than expected (199 kDa). Analysis by Size distribution by volume (Figure b), lead to an estimated molecular weight of 143 kDa, which agrees with the expected molecular weight of 144.04 kDa (monomer MW: 36.011 kDa), suggesting that the purified complex is in its expected tetrameric form. The second peak (not shown) contributed to 99.9% of the overall mass of the sample, with a MW of 0.111 kDa, confirming that the presence of the second peak is due to the high concentration of glycerol (actual MW; 92 Da) in the buffer (10 % v/v).
The mean hydrodynamic range of MtbFBA was estimated to be 11.44 2.578 nm and 9.941 2.141 nm for the respective plots (Figure 7a and b), which is lower than expected, indicating that the incorrect viscosity of the sample was used.
To assess the impact of the addition of ligands and metal ions to the activity of MtbFBA, it is necessary to first characterise the kinetics of MtbFBA with only its substrate. Initially, the final well concentration that would give the optimal reaction rate to allow the observation of even small levels of inhibition/activation upon ligand addition was determined. A two times dilution series was utilised (1:2 to 1:32; initial concentration 0.46 mg/ml) to determine the final well concentration (0.18 ) needed for the subsequent enzymatic assays. Using varying concentrations of FBP a Km of 55.59 4.99 M was found with a Vmax of ….. (Figure 8a), which correlate to those of previously published work (https://www.brenda-enzymes.org/enzyme.php?ecno=126.96.36.199&Suchword=&reference=&UniProtAcc=&organism%5B%5D=Mycobacterium+tuberculosis&organism%5B%5D=Mycobacterium+tuberculosis+H37Rv).
Mtb FBA activity and its alteration
To screen the ligands generated from the CODASS structure based virtual screen it is necessary to have a control ligand known to cause inhibition to the protein target. In this instance, we used the previously published non-competitive inhibitor 8-Hydroxyquinoline-2-carboyxlic acid (8-HCA; IC50 of 10 M; 10.1021/bi401022b). However, the inhibitory concentration of a specific ligand is dependent on the conditions used, we therefore re-characterised the IC50 using varying concentrations of 8-HCA. Surprisingly, we found an IC50 of 66.76 M (Figure 8b), which was unexpected as the presence of 10% v/v of glycerol, which is known to cause an inhibitory effect (20 % glycerol (v/v) reduced enzyme activity by 65%) was not present in the published inhibition assay (10.1021/bi401022b). Thus, it would be expected that the IC50 would decrease.
To reinforce the necessity of the divalent Zn2+ ion for FBA function, varying concentrations of Ethylenediaminetetraacetic acid (EDTA) were incubated initially with MtbFBA for 10 minutes. However, this did not produce the expected results, with complete inhibition at high concentrations not reached (data not shown). Therefore, the same concentrations of EDTA were incubated for 50 minutes at 4 C, to preserve the assay mix, and for a further 10 minutes at 25 C before the assay was induced. Providing an IC50 of 623.5 0.033 M (Figure 8c) showing the necessity of a divalent metal ion for activity. To elucidate the nature of MtbFBA activity upon the divalent zinc ion, we conducted an activity assay using varying concentrations of ZnCl2 to supply the divalent metal ion. Interestingly, at low concentrations of ZnCl2 the activity was increased suggesting that the active sites were not fully occupied. However, surprisingly, at high concentrations of ZnCl2, strong inhibition was seen (Figure 9a). To rule out the possibility that the inhibitory effect was not being implemented by the chloride ion, the same assay conditions were utilized but with a ZnSO4 zinc donor, suggesting that it was Zn2+ was causing an inhibitory effect as similar results were observed (Figure 9a). Furthermore, CoCl2 supplementation at the same concentrations used for both zinc supplemented assays, showed no inhibition Figure 9b). Together this suggests that the properties of Zn2+ provide the observed effect.
Completely stripping MtbFBA from its divalent metal ions using EDTA was more complicated than expected. Initial incubation with 10 mM EDTA for 4 hours failed to ablate all activity, after desalting using a rapid desalt column. EDTA separation was observed through the conductivity peak upon elution. Therefore, we re-incubated the desalted sample over night at 4 C in 250 mM EDTA, desalted and incubated a final time for 4 hours in 250 mM EDTA to fully ablate activity. Differing concentrations of zinc were then incubated with EDTA and then subjected to the FBP cleavage assay (figure 10). Providing a change from activation to inhibition above 3.8 M Zn2+.
As citrate has been found in the active site of Methicillin-Resistant Staphylococcus Aureus FBA in its X-ray crystal structure (10.1021/bi501141t.) we were curious to see if citrate had any inhibitory abilities against MtbFBA. Predictably as the reaction mechanism and active site structure for both aldolases are near identical, an IC50 of 55.2 mM was found, indicating binding to the active site.
Aldol condensation assay
The aldol condensation assay utilises the reverse reaction with DHAP and G3P being converted to FBP. This assay was utilised to test the hypothesis that Zn2+ inhibited MtbFBA only in one direction, providing a novel mode of control over the reaction direction. However, it was seen that the inhibitory effect at 10:1 and 200:1 ratios (Zn2+: MtbFBA) was consistent with that seen in the forward reaction (supplementary info …). Also, indicating the observed results from both forward and reverse assays were not exhibited due to the inhibition of the secondary assay enzymes.
Thermal denaturation assay
Thermal denaturation assays utilise a probe that binds to the hydrophobic patches, revealed upon protein denaturation, causing fluorescence as a reporter of denaturation. Three concentrations of MtbFBA were denatured by increasing temperature, determining that 5 M was the ideal concentration for the subsequent TDAs with a melting temperature (Tm) of 59 C (Figure 11).
As it is known that 8-HCA binds to the active site of MtbFBA, we determined the melting temperature of MtbFBA in complex with 8-HCA at varying concentrations. We found an initial decrease in Tm at low concentrations of 8-HCA, which was inverted at higher concentrations. This suggested, as expected, that 8-HCA was acting as a chelating agent at low concentrations, causing the protein to be destabilised and increasing the Tm at high concentrations compared to 100 M 8-HCA (Figure 12a). This suggests that at high concentrations, 8-HCA can bind to the active site, in a non-competitive manner. Due to the unexpected findings that the addition of high concentrations of Zn2+ inhibited MtbFBA we conducted a TDA with the same protein to zinc ratios as in the inhibition assays. Surprisingly, the melting temperature increased at 100 M markedly to 67 C but then started to decrease at higher concentrations (Figure 12b). However, there was very little change with vary Co2+ concentrations (Figure 12c).
Incubating FBA with 1 mM final well concentration of EDTA to strip the Zn2+ ion lead to a decrease in Tm to 56 C whereas incubation with 1 mM citrate lead to a surprising small decrease in Tm to 58 C. On the contrary, incubation with the substrate, FBP, lead to a marked increase to 62 C, showing binding of the substrate (Figure 12b).
In silico virtual screen for compound identification
To be able to perform computational docking of potential ligands using structure based virtual screening, it is vital to have a high-quality X-ray crystal structure of the protein. To this end, the structure 4LV4 was selected from the PDB for screening as it is in complex with the non-competitive inhibitor, 8-hydroxyquinoline-2-carboxylic acid (8-HCA), with a high resolution (2.08 Å) and low R-factors (R-Factor Work/Free: 0.161/0.191). As previously described, 124,732 compounds from the BCC were screened against 4LV4 using the CODASS script by Dr. Douglas Houston (Figure 3) using two docking programs: AutoDock Vina (10.1002/jcc.21334), and AutoDock (10.1002/prot.340080302), which use differing docking methods. After completion of the CODASS run (performed by Dr. Douglas Houston, University of Edinburgh), the generated virtual hits were established and ranked by additional scoring with DSX, X-score 1.3, NNScore 1, NNScore 2 and RFScore 4 and Pose matching from both docking simulations (RMSD 2Å). The resulting virtual hits were selected for visual analysis, leading to the selection of 132 compounds for chemical screening.
Experimental validation of virtual hits
Although virtual screening is a powerful method to identify potential inhibitory ligands against a certain protein, the generated hits must still be validated using a biochemical assay. We utilised the FBP cleavage assay to examine the efficacy of 60 of these virtual hits (supplementary information…) as some of the generated hits had previously been tested to no avail. As the ligands are suspended in DMSO, which has been seen to increase MtbFBA activity, we utilised an MtbFBA control in the presence of 1% DMSO and a positive inhibition control of 75 M 8-HCA. Due to a lack of time, ligands could only be tested once. Therefore, ligands were only taken as hits if they exhibited an inhibitory effect of > 30% or a similar inhibition was seen in E. coli FBA (Debbie Shi, unpublished). After screening of the 60 ligands, no compound exhibited a substantial decrease in activity of MtbFBA (Figure 13) nor produced a comparably sufficient level of inhibition with E. coli FBA (data not shown).
To determine the molecular parameters and gain structural insights into the expressed tetrameric MtbFBA, SEC-SAXS was performed using un-cleaved His6 MtbFBA and cleaved MtbFBA (prepared by Emmaline Stötter).
SEC-SAXS was utilised due to its ability to separate the oligomeric states of the sample being run before data collection is obtained. Both cleaved and un-cleaved samples eluted as one large and small peak. The large peak, presumed to be tetrameric, was used for data collection. Data was collected under one condition in protein buffer at 15 C. The scattering profile determined (Figure 14a) shows that, as expected, the scattering intensity decays with increasing q for both data sets. Guinier analysis showed linearity at small q (Figure 14a insert) allowing the calculation of the radius of gyration (Rg) which was 41.35 0.37 Å which was in close agreement with Rg-real (41.27 0.22 Å) calculated by GNOM analysis. Implementing CRYSOL to run on the previously published X-ray crystal structure (PDB:3ELF), calculated an Rg value of 38.64 Å, suggesting that the MtbFBA sample used for SAXS data collection was slightly less compacted. This could be due to the differing his tag lengths as 3ELF contains His3 residues N-terminal and analysed sample a His6-tag. Though it is also noteworthy that a crystal structure image doesn’t show flexibility. This was further consolidated by the cleaved MtbFBA having an Rg of 40.74 0.21 Å, showing more compactness.
To determine the molecular configuration of the analysed MtbFBA, normalised Krakty analysis was performed as it had the beneficial output that allows for the semi-qualitative assessment of protein ‘unfolded-ness’ (Figure 14b). As there is an initial peak seen at and magnitude this shows the particle obeys Guiniers approximation and this was also seen in the cleaved sample (Figure…). This suggests that cleaved and un-cleaved MtbFBA is a compact protein. However, the secondary peak seen in both dimensionless Kratky plots suggests the complex is multimodular, which can be taken as representing the four monomers in the tetramer. To be able to compute the P(r) distribution it is first necessary to determine the Porod exponent. scÅtters inbuilt flexibility analyses code initially shows the particle to be inflexible as the initial plateau/linear region after the first hyperbolic rise occurs in the Porod-Debye plot. Further solidifying this assumption is the fact the first plateau/linear region in the Krakty-Debye plot has a negative gradient (data not shown). The bi-peaked distribution has a general downward slope at high q, suggesting the protein is folded. Further analysis determined a Porod exponent of 3.26 0.04, qualifying the visual assumption that the MtbFBA tetramer is not flexible but compact. This was also seen on the cleaved sample, producing a Porod exponent of 3.16 0.03. P(r) plots were constructed using GNOM analysis, which provided a Dmax of 123.5 Å, for both analysed samples, with the distribution suggesting the protein should be ‘doughnut’ in shape, containing two peaks (Figure 15). After refinement, the total estimated score of GNOM analysis was 0.9046 and 0.9850 respectively for un-cleaved and cleaved samples. These suggest the solutions found are of excellent quality. Furthermore, the P(r) distribution fit was validated further by statistical analysis, producing a Chi2 of 1.19 and an Sk2 of 0.044 for un-cleaved MtbFBA and 1.01 (Chi2) and 0.06 (Sk2) for cleaved MtbFBA. The calculated Dmax was in good agreement with that calculated from the protein crystal structure (PDB:3ELF), using CYSOL, with an observed Dmax of 124.9 Å.
Molecular envelope construction
Molecular envelope construction was performed on un-cleaved MtbFBA by running 13 DAMMIN runs on the calculated GNOM data under P1 symmetry, initially, as to avoid bias of symmetry. The 13 runs were then averaged with the molecular envelope showing P2 symmetry. Therefore, molecular envelope construction was re-run, running 23 DAMMIN runs under P2 symmetry. The final reconstructed envelope encompasses the whole native-state model, except for four protruding loops, which could be due to their low electron density. The envelope has the same general shape as the native model, following the right-handed nature of the native-state model. There is, fittingly, an obvious density hole missing in the centre of the construction, which holds true to the native state model and agrees with the expected ‘doughnut’ shape seen from the P(r) distribution plot. The reconstructed envelope seems to imply there is some flexibility near the active site loops as the envelope density is not matched by the position of the native model structure. FOXES was utilised to strengthen the conclusion made as the calculated scattering curve from the native model structure was in excellent agreement with the experimental scattering
Disclaimer: This essay has been submitted by a student. This is not an example of the work written by our professional essay writers. You can order our professional work here.
Sorry, copying is not allowed on our website. If you’d like this or any other sample, we’ll happily email it to you.
Your essay sample has been sent.
Want us to write one just for you? We can custom edit this essay into an original, 100% plagiarism free essay.Order now