Team:Poitiers/Results
Results
For the production of our molecules of interest, we worked with the laboratory Qingdao Institute of Bioenergy and Bioprocess Technology, Chinese Academy of Sciences, Qingdao, China. The latter modified an E. coli BL21 strain to optimize the production of free fatty acids (FFA) and Hydroxylated fatty acids (HFA). For this, we first characterized the strain in order to validate the modifications and its production.
First, we tried to prove the presence of cytochrome CYP102A1 by performing induction and analysis of protein expression on the SDS page. In the well 5, corresponding to a 24 hours induction, a band between 100 and 135 kDa is observed. This band isn’t present into the none induced condition. The cytochrome cloned into the strain at a size of 119 kDa which corresponds to what is observed. The vector is present and the protein is produced.
Fig 1: Expression of the recombinant enzymes in engineered E. coli.
The control corresponds to the T0 of the induction, so the addition of IPTG. The marker used is the BlueEye Prestained marker PS-104 (Jena Bioscience). 12% acrylamide gel, in reducing conditions.
GC results to characterize the octanoic and decanoic acid production by BL21/ΔfadD/pE-A1’tesA&pA-acc induction
BL21/ΔfadD/pE-A1’tesA&pA-acc strain was grown for 24h in six LB + chloramphenicol + kanamycin media. Three of them were induced by the injection of IPTG. After 24h of induction, we wanted to analyze this strain’s production of octanoic and decanoic acids. In the reference article (Cao Y. et al, 2016), the researchers esterified the free fatty acids (FFA) and hydroxy fatty acids (HFA) produced by this strain with sulfuric acid/methanol (1:99) to generate fatty acids or HFAs methyl esters (FAMEs or HFAMEs) in order to perform a GC analysis. These FAMEs and HFAMEs were then extracted with hexane before analysis.
On both the induced and non induced conditions, slight peaks appear around 3.157 min and 3.755 (D, E, F, G, H and I). The octanoic acid and decanoic acid standards prepared at 1g/l, extracted in hexane after a methyl-esterification also present peaks approximately at the same retention time (respectively 3.166 min and 3.760) (B and C).
This means that the peaks observed in the non induced and the induced conditions (D, E, F, G, H and I) represent a presence of octanoic and decanoic acids in these samples. However, the area of these peaks only varies between 1 and 4 with no real difference between the induced and non-induced conditions, meaning a very low presence of these molecules.
Moreover, in the pure hexane analysis, there is also a presence of a slight peak at the same retention time as the decanoic acid, which means that the decanoic peaks observed in the samples could also be due to the hexane solvent.
Fig 2: GC results after 24h of IPTG induction on BL21/ΔfadD/pE-A1’tesA&pA-acc strain.
X axis:Time (min). Y axis: Response (pA). Red arrows: octanoic acid peak. Orange arrows: Decanoic acid peak. A: Result of the GC analysis of pure hexane. B: Result of the GC analysis of octanoic acid standard at 1g/l obtained after esterification. C: Result of the GC analysis of decanoic acid standard at 1g/l obtained after esterification. D, E and F: Results of the GC analysis of Non Induced BL21/ΔfadD/pE-A1’tesA&pA-acc at T+24h of culture. G, H and I: Results of the GC analysis of Induced BL21/ΔfadD/pE-A1’tesA&pA-acc at T+24h of culture. The spectrum B shows a high peak at 3.166 min, representing the spectrum of the octanoic acid standard at 1g/l after an esterification. The spectrum C shows a high peak at 3.760 min, representing the spectrum of the decanoic acid standard at 1g/l after an esterification. The spectrums D, E, F, G, H and I are almost identical. They all present a very slight peak at the same retention time as the octanoic and decanoic acid. However, the spectrum of pure hexane in which each sample is extracted presents the same peak at the same retention time as the decanoic acid as the others samples. So the decanoic acid peaks of the samples may be due to the hexane solvent.
To make sure the extraction and GC analysis protocols of are effective in order to apply them on our modified strain, and that the strain at least produces a bit of octanoic and decanoic acids, more analysis are already scheduled.
In order to train on the extraction and analysis of 2-nonanone by GC and HPLC analysis, we also worked with the Lab Operations Associate Joint BioEnergy Institute (JBEI) Lawrence Berkeley National Laboratory. The latter modified an E. coli DH1 strain to optimize the production of Methyl Ketones. For this, we first characterized the strain in order to validate these modifications and its Methyl Ketone production.
First, we tried to prove the presence of the inducible Mlut_11700 (Acyl-CoA oxidase), FadB (Fatty acid oxidation complex subunit alpha) and FadM (long chain acyl CoA thioesterase) and ‘TesA (Acyl ACP thioesterase) genes by performing an induction and analysis of protein expression on SDS page.
These genes are respectively present on the pEG530 and the pEG855 plasmids.
Fig 3: pEG530 and pEG855 maps.
The ‘TesA, Mlut_11700 and fadB genes are on the pEG530, mediated by a lac operator. This plasmid grants a resistance to chloramphenicol. The FadM gene is on the pEG855, also mediated by a lac operator. This plasmid grants resistance to kanamycin.
Fig 4: SDS PAGE analysis of EGS895 proteins production induction.
NI: Non-Induced EGS895 strain. I: Induced EGS895 strain. M: Molecular weight marker (kDA). Green arrows: Bands of induced proteins. Red arrows: Areas where we are supposed to have a band of induced proteins. A) 8% polyacrylamide gel. B) 12% polyacrylamide gel. The strain was cultivated overnight, and then put into a new LB broth in 2 different tubes. At T+5h, one of them was induced by adding 0.5mM of IPTG. At T+25h, T+48h and T+73h after culture and induction, an induction band appears near 75 kDA. Plus, for these conditions, bands at 15 and 20 kDA are supposed to appear, but doesn’t.
On both the 8 and 12% gels (A and B), a band around 75 kDa is present in the T+25, 48 and 73h induced samples, but not in the non-induced ones. This means that a protein which has a molecular weight near 75 kDa is overproduced after an induction. We know that these bacteria are supposed to overproduce the Acyl-CoA oxidase (Mlut-11700) and the Fatty acid oxidation complex subunit alpha (FadB) (E-G.Goh et al, 2012) as shown on the table below. these proteins have molecular weights very close to 75 kDa, as well as very close to each other (respectively 77 and 79 kDa, cf table below). So this band could be a merge of the bands corresponding to the Acyl-CoA oxidase and the Fatty acid oxidation complex subunit alpha.
However, there is no band appearing on the induced samples at approximately 20 kDA on the 8% gel (A) and at 15 kDA on the 12% gel (B). But the Acyl ACP thioesterase (‘TesA) which has a molecular weight of 20 kDa, and the long-chain acyl-CoA thioesterase (FadM), which has a molecular weight of 15 kDa, are supposed to be overproduced after induction with IPTG (E-G.Goh et al, 2012). This could be due to the fact that, even if overproduced, the yield of production of these proteins (E-G.Goh et al, 2012) isn’t enough for us to see a band with an SDS PAGE. Plus, the coloration could be too weak to see this band. We should have added a positive control to the last well of the gels to make sure of this.
Table 1: Molecular weights of the EGS895 proteins’ produced by inducible genes
We were more interested in the FadM gene presence, so to make sure that our EGS895 strain has indeed the FadM gene on its plasmid, we decided to make a Colony PCR.
4 EGS895 colonies were then analyzed by Colony PCR to check the presence of the non-native FadM gene, which is supposed to be located on the pEG855 plasmid of the strain (E-G.Goh et al, 2012). The amplification products were then analyzed thanks to an agarose gel shown below.
Fig 5: Agarose gel of a Colony PCR of 4 EGS895 colonies.
C1/2/3/4: Colony PCR products of EGS895 colonies, where FadM was amplified. L: 1 kb Plus DNA Ladder NEB. After amplification of the FadM gene (located on the pEG855 plasmid) thanks to a colony PCR on 4 EGS895 colonies, an agarose gel allowed us to show an amplification band on 3 of the 4 colonies. The amplified fragments are approximately 0.5 kb long.
The agarose gel showed a band at the expected size of FadM fragment (423 bp) on the C1, C3 and C4 samples (E-G.Goh et al, 2012). So the FadM gene was amplified in the strain. To make sure that the amplified gene could not be the native gene, the primers used in the Colony PCR were designed to attach on both the FadM gene and its surrounding restriction sites on the pEG855 plasmid (cf pEG855 map). So these results show that the FadM gene is present on the pEG855 plasmid in 3 of the 4 EGS895 colonies sampled. The fact that C2 doesn’t have a 0.5 kb band could either be due to the fact that the bacteria of this colony doesn’t have the gene in the plasmid or no plasmid at all or because of the PCR mix which could be badly done. Indeed, we can incriminate the mix PCR for now because we didn’t do a positive control. So even if we don’t know yet if the FadM protein is overproduced after an IPTG induction, we know at least that the FadM gene is certainly present in this strain.
Another verification of the EGS895 strain by sequencing the DNA content of the strain was originally planned. The experiments were delayed so far because of a lack of sequencing facilities during the summer.
GC results to characterize the 2-nonanone production by EGS895 induction
Decane and hexane overlays test
In the reference article (E-B.Goh et al, 2012), researchers used decane as a tool to extract the 2-nonanone from the culture medium after induction with IPTG.
A first study of the GC spectrum of pure decane (A) and 2-nonanone mixed at 10 g/l in decane (B) has shown that, with the same GC heating program as in the referenced article, the peak representing the decane spectrum appears between 5.5 and 7.5 min (A and B). But the peak representing the 2-nonanone spectrum (B) appears at 7.710, which is not very far from the peak representing the decane spectrum. We were afraid that in the following experiments, the 2-nonanone peak would not be distinctive enough from the decane peak. We then searched for a potential replacement for the decane.
A second study was launched to study the GC spectrum of pure hexane (C) and 2-nonanone mixed at 1 g/l in hexane (D). Indeed, hexane is also a molecule which could be used as a tool to extract the 2-nonanone from the culture medium, but which wasn’t used by the researchers in the reference article.
This second study showed that, with the same GC heating program as in the referenced article, the peak representing the hexane spectrum appears between 1.2 and 1.5 min (C and D). But the peak representing the 2-nonanone spectrum (D) appears at 7.402, which is further from the peak representing the hexane spectrum than it is from the decane spectrum.
So hexane could indeed be a good replacement for the decane as an overlay to extract the 2-nonanone. We then just needed to make sure that the hexane indeed allowed us to trap 2-nonanone in a culture medium.
Fig 6: Calibration of 2-nonanone detection with GC analysis with two extraction solvents.
X axis: Time (min). Y axis: Response (pA). Red arrows: 2-nonanone peaks. A) Pure decane detection by GC analysis. B) Decane + 10 g/l of 2-nonanone detection by GC analysis. C) Pure hexane detection by GC analysis. D) Hexane + 1 g/l of 2-nonanone detection by GC analysis. On the one hand, decane spectrum, shown in A appears between 5.5 and 7.5 min and is present in the spectrum B too. The difference between the spectrums A and B is the presence of a peak at 7.710 min in the spectrum B, representing the spectrum of the 2-nonanone standard at 10 g/l in decane. On the other hand, the spectrum of the hexane, shown in C appears between 1.2 and 1.5 min, also present in the spectrum D. The difference between the spectrums C and D is the presence of a peak at 7.402 min in the spectrum D, representing the spectrum of the 2-nonanone standard at 1 g/l in hexane. The peak representing the hexane spectrum is more separated from the 2-nonanone peak than the peak representing the decane spectrum.
2-nonanone trapping test in hexane
We injected various concentrations of 2-nonanone standards in a culture medium with Non-induced EGS895 and with a hexane overlay. After an mixing of the media and an extraction of the hexane overlays, a GC analysis was performed.
For each concentration of 2-nonanone in the media (1, 0.5 and 0.1 g/l), a peak appears near 7.5 min (B, C and D) corresponding to the spectrum of the 2-nonanone. Plus, the area under this peak varies with the 2-nonanone concentration added in the media. Indeed, the peaks corresponding to 1g/l, 0.5g/l and 0.1g/l of 2-nonanone have an area of 25291, 11788 and 5493.4 respectively (A).
Fig 7: GC analysis of hexane overlays extracted from culture media supplemented with 2-nonanone.
X axis: Time (min). Y axis: Response (pA). Red arrows: 2-nonanone peaks. A: Pure hexane detection by GC analysis. B: Hexane + 1g/l of 2-nonanone added in culture medium detection by GC analysis. C: Hexane + 0.5g/l of 2-nonanone added in culture medium detection by GC analysis. D: Hexane + 0.1g/l of 2-nonanone added in culture medium detection by GC analysis. The spectrum of the Hexane, shown in A, is present in every spectrum. The difference between the spectrums A, B, C and D is the presence of a peak at respectively 7.551, 7.509 and 7.480 min in the spectrums B, C and D representing the spectrum of the 2-nonanone standard at respectively 0.1, 0.05 and 0.01% in the medium. The less the 2-nonanone was concentrated in the medium, the less the response of the 2-nonanone peak was high.
The calibration curve obtained is linear, with a correlation (R²) of 0.99 (B ).
Fig 8: Calibration curve of several concentrations of 2-nonanone added in culture media detected by GC analysis.
A: Time (min). Y axis: Response (pA). Red arrows: 2-nonanone peaks. A: Calibration table of the 2-nonanone peaks. The concentrations were originally calculated in % in the software. For each percentage of 2-nonanone, a concentration (in g/l) was calculated. The areas of each peak, corresponding to different concentrations of 2-nonanone added in the culture media, are automatically calculated by the software. B: Calibration curve obtained with the calibration table (A). X axis: Amount of 2-nonanone added in the culture media (%). Y axis: Area of the 2-nonanone peaks. No error bars because there isn’t enough results to insert them. The curve obtained is linear, with a correlation (R²) of 0.99028
2-nonanone can indeed be extracted with a hexane overlay, and the concentration of 2-nonanone detected in the overlay by GC analysis is strictly proportional to the concentration of 2-nonanone present in the culture media.
Thus hexane can be used to extract 2-nonanone from a culture medium.
Now, all there is to do is to induce EGS895 bacteria and study the composition of the hexane overlay to see if 2-nonanone is indeed produced by this strain after an induction.
Optimisation of the GC program
The first GC program lasted 25 minutes for each run, whereas the 2-nonanone was detected between 7.5 and 8 minutes. 17 minutes left was way too long, so the idea was to shorten the program, without altering the precision of the 2-nonanone detection.
After tweaking the heating parameters of the program, a new program was designed, which now only lasts 8min 30 seconds for each run.
In Gas Chromatography, the analysed gas is usually composed of several molecules. These molecules are carried in the column by an inert carrier gas (mobile phase, and will have more or less affinity with the column surface (stationary phase). Compounds that have a higher affinity for the stationary phase spend more time in the column and thus elute later and have a longer retention time (Rt) than samples that have a higher affinity for the mobile phase. In GC, the temperature of the oven (and thus of the column) is adjusted to change the elution times. Separations in GC are based on volatility because higher boiling point substances may attach to the column if the temperature is low, thus they are not eluted or take a long time to elute (JoVE Science Education Database, 2019).
So increasing the oven temperature helps to speed up the process of elution of the 2-nonanone.
When moving the heating rate from 15°C/min to 30°C/min, the 2-nonanone retention time moved from 7.5-8 min to 4.5-5 min. Because in our study we are just interested in the 2-nonanone, the heating rate of the program could be increased after the 2-nonanone eluted without compromising the quality of the results. That is why the heating rate moves from 30°C/min to 100°C/min after 4 minutes in the second heating program. Again, this increase of oven temperature rate helped to speed up the process.
Fig 9: Evolution of the heating program of the GC device.
X axis: Time (min). Y axis: GC oven temperature (°C). Red arrows: Retention time of the 2-nonanone with these specific programs. Program 1: First heating program of the GC device, coming from the reference article (E-B.Goh, 2011). Program 2: Second heating program of the GC device, optimised to allow the analysis of 2-nonanone, quicker than the original program (8.5 min instead of 25 min). The origin temperature, as well as the heating rate has been increased to allow the 2-nonanone to exit quicker the column.
Presence of 2-nonanone after 24h of induction
EGS895 strain was grown for 24h in two LB + chloramphenicol + kanamycin media. One of them was induced by the injection of IPTG. After the IPTG addition, the extraction solvent (hexane) was added on top of both of the media. The 2-nonanone being less dense than the media (2-nonanone density, PubChem), but a bit more dense than the extraction solvent (Hexane density, PubChem), 2-nonanone stays trapped in between the LB and the Hexane phase, which will allow its extraction. The hexane phase was then extracted with a Pasteur pipette, purified and the samples were analyzed by Gas Chromatography.
On both the induced and non induced conditions, a slight peak appears at 4.847 min (C and D). The non-induced one has an area of 46.767 whereas the induced one has an area of 106.140. The 2-nonanone standard prepared at 1g/l in hexane has also a peak approximately at the same retention time (4.869 min) (B).
This means that the peaks observed in the non induced and the induced conditions (C and D) represent a presence of 2-nonanone in these samples. Due to the fact that we didn’t have yet a calibration curve of 2-nonanone, we, unfortunately, couldn’t measure the concentration of 2-nonanone in these samples, but thanks to the areas under the peaks, we were at least able to suppose that there is more 2-nonanone present in the induced sample than in the non induced one.
On the spectrum of the non induced bacteria, there is also a peak at 4.110 min which does not appear on the spectrum of the induced bacteria (C).
This peak could be a precursor product for the production of 2-nonanone.
All these results aren’t significant because we don’t have enough results to measure the significativity.
Fig 10: GC results after 24h of IPTG induction on EGS895 strain.
X axis: Time (min). Y axis: Response (pA). Red arrows: 2-nonanone peaks. A: Result of the GC analysis of pure hexane. B: Result of the GC analysis of hexane + 1g/l of 2-nonanone. C: Result of the GC analysis of non induced EGS895 at T+24h of culture. D: Result of the GC analysis of induced EGS895 at T+24h of culture
The spectrum of the Hexane, shown in A, is present in every spectrum. The difference between the spectrums A and B is the presence of a peak at 4.869 min in the spectrum B, representing the spectrum of the 2-nonanone standard at 1g/l in hexane. On spectrum C, there is an additional peak at 4.110 min and a very slight peak at 4.847 min, which is almost the same retention time as the 2-nonanone standard. On spectrum D, there is a slight peak at 4.847 min, which is also almost the same retention time as the 2-nonanone standard.
Presence of 2-nonanone after 72h of induction
An induction for 72h of culture was needed to see if more 2-nonanone was produced if the strain was induced longer, and if the peaks at 4.110 min and 4.847 min on non induced bacteria (C) were still present on non induced bacteria after 72h of culture (C). The same method as previously was used. A peak at 4.863 min, which is approximately the same retention time as the 2-nonanone at 10g/l in hexane (4.841) (B) only appears on the spectrum of the induced bacteria (D). This peak has an area of 2366.3, which is 22 times higher than that of the induced bacteria after 24h of culture (D).
Again, this peak corresponds to the presence of 2-nonanone in the culture medium of the induced bacteria. This time, there is no 2-nonanone in the induced bacteria, and there is 22 times more 2-nonanone accumulated in the medium induced after 72h of culture compared to the medium induced after 24h of culture. This means that more 2-nonanone accumulates when the strain is induced for a longer time.
The non-induced bacteria present almost the same spectrum as that of the pure Hexane (C and A). This time, there is no peak that could be representing a precursor of 2-nonanone.
Again, all these results aren’t significant because we don’t have enough results to measure the significativity. That is why we needed to redo these experiments several times to obtain more results.
Fig 11: GC results after 72h of IPTG induction on EGS895 strain.
X axis: Time (min). Y axis: Response (pA). Red arrows: 2-nonanone peaks. A: Result of the GC analysis of pure hexane. B: Result of the GC analysis of hexane + 1% of 2-nonanone. C: Result of the GC analysis of Non Induced EGS895 at T+72h of culture. D: Result of the GC analysis of Induced EGS895 at T+72h of culture.
The spectrum of the Hexane, shown in A, is present in every spectrum. The difference between the spectrums A and B is the presence of a peak at 4.841 min in the spectrum B, representing the spectrum of the 2-nonanone standard at 10g/l in hexane. The spectrums C and A are almost identical. On spectrum D, there is a peak at 4.863 min, which is almost the same retention time as the 2-nonanone standard.
2-nonanone GC Standard Curve for quantitative results
For the following experiments, the optical density at 600nm of each sample was measured and a 2-nonanone standard curve (B) was used to obtain the quantity of 2-nonanone for every sample, the R² of this curve being 0.99822 (B).
Fig 12: Calibration of 2-nonanone concentration detection with a GC analysis and a calibration curve.
A: Calibration table of the 2-nonanone peaks. The concentrations were originally calculated in % in the software. For each percentage of 2-nonanone, a concentration (in g/l) was calculated. The areas of each peak, corresponding to different concentrations of 2-nonanone added in the culture media, are automatically calculated by the software. B: Calibration curve obtained with the calibration table (A). X axis: Amount of 2-nonanone added in the culture media (%). Y axis: Area of the 2-nonanone peaks. No error bars because there isn’t enough results to insert them. The curve obtained is linear, with a correlation (R²) of 0.99822
Multiple cultures and inductions with negative results
To obtain enough positive results of 2-nonanone present in induced samples and studied by GC analysis to have significant results, several other cultures and inductions had to be made. Unfortunately, we weren’t able to obtain more 2-nonanone positive results after an induction.
Several tests were made to understand and find the origin of the problem, but none of them allowed us to find the origin of the problem.
Eventually, we contacted the writer of our reference article (Harry Beller)(E-G.Goh et al, 2011) to find a solution. He came up with the idea that hexane as an extraction substrate could cause some problems because it is more toxic to bacteria than decane. Plus, hexane is more volatile than decane, increasing the risk to lose 2-nonanone because of this evaporation. He also added that the quantity of 2-nonanone produced by the induced bacteria could be just below the detection limit of the GC device.
For the next experiment, EGS895 inductions with hexane and decane overlays were used to verify these hypotheses.
Extraction solvent change
For this experiment, the optical density was followed before induction, at the induction moment and after the induction. The results revealed a difference in the bacterial density when hexane or decane are used as an extraction substrate.
Fig 13: Bacterial density of bacterial cultures grown for 72h with IPTG induction and different extraction substrates used.
X axis: Time (h). Y axis: Bacterial density (bacterium/ml). NI: Non induced samples. I: Induced samples. H: Hexane overlay. D: Decane overlay. Red arrow: Moment of IPTG of injection and overlays adding. No error bars because there isn’t enough results to insert them. Each of these samples was EGS895 bacteria grown on a LB medium + 25μg/ml chloramphenicol + 50μg/ml kanamycin. After 5 hours of culture (T+5h), some samples (I1H, I2h, I1D and I2D) were induced by adding 0.5mM of IPTG in their culture medium. NIH, I1H and I2H received 1ml of Hexane as an overlay, whereas NID, I1D and I2D received 1ml of Decane as an overlay.
Indeed, the bacterial density of the culture media receiving a decane overlay is almost twice as dense as the culture media receiving a hexane overlay (2.10^9 bacterium/ml for the media receiving a decane overlay versus 1.10^9 bacterium/ml for the media receiving a hexane overlay). Plus, the bacterial density of the culture media which received a hexane overlay has lowered after the adding of hexane, which is not the case for the decane. This lowering of the population could be due to a bacterial death.
Thus these results seem to corroborate the information given by the writer of the referenced article (Harry Beller), that hexane could be toxic for bacteria.
We hoped that the GC results of this test will show the presence of 2-nonanone in the induced media which have received a decane overlay, or at least more 2-nonanone in these samples compared to those with a hexane overlay. Unfortunately, that wasn’t the case.
Fig 14: GC results after 72h of IPTG induction on EGS895 strain with hexane extraction.
X axis: Time (h). Y axis: Response (pA). Red arrow: 2-nonanone peak. A: Result of the GC analysis of pure hexane. B: Result of the GC analysis of hexane + 2g/l of 2-nonanone. C: Result of the GC analysis of Non Induced EGS895 at T+72h of culture with hexane extraction. D: Result of the GC analysis of Induced EGS895 at T+72h of culture with hexane extraction. The spectrum of the Hexane, shown in A, is present in every spectrum. The difference between the spectrums A and B is the presence of a peak at 3.75 min in the spectrum B, representing the spectrum of the 2-nonanone standard at 2g/l in hexane. The spectrums C and A are almost identical. On spectrum D, no peak can be observed at the same retention time as the 2-nonanone standard. However, a huge peak at approximately 2.8 min can be observed, but doesn’t correspond to 2-nonanone..
Fig 15: GC results after 72h of IPTG induction on EGS895 strain with decane extraction.
X axis: Time (h). Y axis: Response (pA). Red arrow: 2-nonanone peak. A: Result of the GC analysis of pure decane. B: Result of the GC analysis of decane + 2g/l of 2-nonanone. C: Result of the GC analysis of Non Induced EGS895 at T+72h of culture with decane extraction. D: Result of the GC analysis of Induced EGS895 at T+72h of culture with decane extraction. The spectrum of the decane, shown in A, is present in every spectrum. The difference between the spectrums A and B is the presence of a peak at 4.447 min in the spectrum B, representing the spectrum of the 2-nonanone standard at 2g/l in decane. The spectrums C, D and A are almost identical. On spectrum D, no peak can be observed at the same retention time as the 2-nonanone standard.
Indeed, no 2-nonanone peak could be observed in the induced samples extracted with decane. The yield of production of the 2-nonanone in this strain may be too low to be detected at each analysis.
A better 2-nonanone producing strain (EGS1895) was ordered to the same laboratory which provided us the EGS895, but we didn’t received it yet.
As soon as we receive it, we will be able to redo these analyses to obtain significant results on the 2-nonanone analysis by Gas Chromatography, so that we will be prepared for when we achieve to modify a bacterial strain to produce 2-nonanone.
JoVE Science Education Database. Analytical Chemistry.(2019) Gas Chromatography (GC) with Flame-Ionization Detection In : JoVE, Cambridge, MA.
2-nonanone density, PubChem
Hexane density, PubChem
For the production of our molecules of interest, we worked with the Biorenewables Research Laboratory Iowa State University. The latter modified an E. coli MG1655 strain to increase its inherent fatty acids production to overproduce octanoic acids. Our idea was to make this strain produce octanoic acids to extract them and supplement the colture medium of our own modified strain to increase the origin pool of octanoic acid. We first had to characterize this strain in order to validate its production.
GC results to characterize the octanoic and decanoic acid production in ZEFA TE10 culture
After 24h of a culture of this strain in a LB medium without antibiotics, 1ml of the culture was extracted, and the fatty acids were esterified into fatty acid ethyl ester (FAEE) for them to be analyzed by Gas Chromatography. The FAEE were then extracted with Hexane and were analyzed by GC. Unfortunately, the HP-5 column that we usually used was unusable, thus we had to use the only column remaining, a BPX-70, which is used to study the Fatty acid methyl esters (FAMEs).
Fig 16: GC results after 24h of culture of ZEFA TE10 strain and ethyl esterification of fatty acids.
X axis: Time (h). Y axis: Response (µV). A: Result of the GC analysis of pure hexane (peak on the left) and of 1g/l of octanoic and decanoic acid standards (peaks on the right). B: Result of the GC analysis of Non-induced and Induced ZEFA TE10 at T+24h of culture after ethyl esterification of the FFAs an hexane extraction. The spectrum of the hexane, shown in A, is present in every spectrum. The spectrums of octanoic and decanoic acid standards present a response like a merge of two peaks. The retention time of the octanoic and decanoic acid standards only differ by 1 second. The spectrums of the samples are very similar to the spectrums of the standards.
As shown on this GC spectrums, the ZEFA TE10 sample seem to present the same spectrum than the octanoic acid and decanoic acid standards, meaning that the ZEFA TE10 might indeed produce one of these molecules. However, the retention time of the standards only differ by 1 second, preventing us from being able to differentiate the presence of one or the other of these molecules in the samples. It also seems that the peaks observed in the samples and the standards could in fact be a merge of two peaks. In any case, a good GC peak is not supposed to look like that. The ethyl-esterification of fatty acids method used here could be at the origin of this issue. a better protocol should be use next.
Finally, the molecules of interest are detected very early in the analysis, meaning that the BPX-70 column doesn’t retain our molecules and can’t be used to analyse the presence of octanoic and decanoic acids. A new HP-5 column must be bought next in order to redo this experiment.
The coding sequence of the gene was recovered after contact with the authors of the publication “Overexpression, Purification, and Biochemical Characterization of the Thermostable NAD-dependent Alcohol Dehydrogenase from Bacillus stearothermophilus”. This enzyme has been selected because it has a better activity than E. coli and also a better activity for the substrate.
Thanks to Geneious software we have designed a gene with a promoter, a terminator and a tag. The promoter and terminator are those of pBAD inducible to arabinose. This allows a controlled expression of the synthetic gene to avoid any effect of toxicity. In addition, arabinose is an inexpensive inducer and very present in the laboratories of our university. We then decided to add a tag, in one condition in N-Ter position and another in C-ter position. Both conditions are tested to see which one does not disrupt the enzymatic activity of the protein. The c-myc tag was chosen because we have anti c-myc antibodies that allow purification and detection of the protein in the host strain.
Fig 17: Photography of designed gene realized with the Geneious software.
A: Design of ADR N-ter flanked with a tag c-myc and 6-His and restriction sites EcorI, XbaI, SpeI and PstI + pBAD promoter. This is a 5'-3'-end sequence.
B: Design of ADR C-ter flanked with a tag c-myc and 6-His and restriction sites EcorI, XbaI, SpeI and PstI + pBAD promoter. This is a 5’-3’-end sequence.
Following the design of the synthetic genes, they are amplified by PCR thanks to the design of primers upstream and downstream of the sequence. After amplification of the synthetic gene some samples are purified, other not, the amplicons are digested with restriction enzymes EcoRI and PstI. Similarly for the cloning vector pSB1A3 according to the protocol described above. The inserts (ADR N-ter and ADR C-ter) are then ligated into the plasmid.
The PCR products as well as the digestion products are loaded on 0.8% agarose gel. Well 2 is a negative control. In wells 3 and 4 respectively the ADR gene tagged in N-ter and the ADR gene tagged in C-ter amplified by PCR. However, it is observed that the bands observed correspond to the expected sizes of the ADR genes, so 1600 pb.
A
Fig 18: Photography of designed gene realized with the Geneious software.
A: Design of ADR N-ter flanked with a tag c-myc and 6-His and restriction sites EcorI, XbaI, SpeI and PstI + pBAD promoter. This is a 5'-3'-end sequence.
B: Design of ADR C-ter flanked with a tag c-myc and 6-His and restriction sites EcorI, XbaI, SpeI and PstI + pBAD promoter. This is a 5’-3’-end sequence.
Following the design of the synthetic genes, they are amplified by PCR thanks to the design of primers upstream and downstream of the sequence. After amplification of the synthetic gene some samples are purified, other not, the amplicons are digested with restriction enzymes EcoRI and PstI. Similarly for the cloning vector pSB1A3 according to the protocol described above. The inserts (ADR N-ter and ADR C-ter) are then ligated into the plasmid.
The PCR products as well as the digestion products are loaded on 0.8% agarose gel. Well 2 is a negative control. In wells 3 and 4 respectively the ADR gene tagged in N-ter and the ADR gene tagged in C-ter amplified by PCR. However, it is observed that the bands observed correspond to the expected sizes of the ADR genes, so 1600 pb.
Fig 19: Electrophoresis gel photography following loads of N-ter and C-ter ADR PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is NEB 1 kb Plus DNA Ladder.
The products of digestion are also loaded on a gel. In wells 2 and 3, respectively the ADR N-ter and ADR C-ter genes digested with the restriction enzymes EcoRI and PstI. In well 4, the cloning plasmid pSB1A3 digested with the same enzymes. Genes migrated to expected sizes of 1600 pb. Plasmid also migrated to expected size of 2200 pb. It is important to note, however that agarose gel migration does not verify the effectiveness of digestion. Indeed, since the restriction sites are at the end of the sequences, only a few base pairs have been removed on either side. The resolution of an agarose gel does not make it possible to observe the size of the fragments so precisely. This step makes it possible to ensure that we did not have a loss of DNA during experiments.
Fig 20: Electrophoresis photography following loads on agarose gel 0.8% of enzymatic digestion products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus Ladder. Lane 1 corresponds to the marker, lane 2 to the digested N-ter,ADR lane 3 to the digested C-ter ADR and lane 4 to the digested plasmid pSB1A3.
Finally, the ligations were performed. To model the construction, we used the Geneious software which allowed us to assemble the sequence of the plasmid pSB1A3 as well as that of ADR N-ter or ADR C-ter inserts at the cloning site. We can then know the size of the construction and visualize the restriction site present on the latter.
Fig 21: Modelization of the ligation between ADR N-ter/pSB1A3 and ADR C-ter/pSB1A3 with Geneious software.
This map shows the pBAD promoter and its terminator flanking the coding sequence of the ADR protein. Also present in N-ter or C-ter are 6-His and c-myc tag. Finally, in the plasmid is present and ampicillin resistance cassette.
The thermocompetent E. coli JM109 bacteria are then transformed and clones are obtained. Different volumes of transformed bacteria are spread on Petri dish with selective medium. The number of clones obtained is consistent with the proportion of bacteria spread on the petri dishes.
Fig 22: Clones on a selective LB medium (+ ampicillin 100 µg/mL) following the transformation of thermocompetent cells with the pSB1A3-ADR ligations and a control plasmid.
A: Clones obtained from pSB1A3 N-ter ADR ligations. B: Clones obtained from the pSB1A3 C-ter ADR ligations. C: Clones obtained from a control plasmid pSB1C3-Red (iGEM parts).
After bacterial transformation, colony PCR is performed with the forward primer of the ADR gene and a reverse primer of the plasmid. 12 clones of each condition (ADR N-ter/pSB1A3 and ADR C-ter/pSB1A3) are tested. The PCR products are loaded on 0.8% agarose gel.
Wells 1 to 12 show PCR products on clones transformed with ADR N-ter/pSB1A3. Clones 4, 5, 10, 11 and 12 have the right profile, an insert-vector fragment of 1800 pb. Wells 13 to 24 show PCR products on clones transformed with ADR C-ter/pSB1A3. Clones 13, 21 and 22 have the right profile, an insert-vector fragment of 1800 pb.
Fig 23: Electrophoresis photography following loads on agarose gel 0.8% of colony PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus Ladder (left in the figure). Lane 1 to 12 corresponds to colony PCR performed on ADR N-ter/pSB1A3 ligation, lane 13 to 24 corresponds to colony PCR performed on ADR C-ter/pSB1A3.
Other clones do not show bands. They probably did not integrate the ligation products. Clones with the right profile are returned to liquid culture and minipreparations are performed. Enzymatic digestion is carried out with BamHI and PstI restriction enzyme. The expected band sizes are 2300 and 1400 pb. Wells 1 to 5 comprise clones 4, 5, 10, 11 and 12 transformed with ADR N-ter/pSB1A3. Wells 7 to 9 contain clones 13, 21 and 23 transformed with ADR C-ter/pSB1A3. All present the right profile of digestion. This experiment therefore confirms the plasmid constructs.
Fig 24: Electrophoresis photography following loads on agarose gel 0.8% of colony PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus Ladder (left in the figure). Lane 1 to 12 corresponds to colony PCR performed on ADR N-ter/pSB1A3 ligation, lane 13 to 24 corresponds to colony PCR performed on ADR C-ter/pSB1A3.
In order to avoid any risk of point mutation, sequencing is performed with the plasmid primer.
After sequencing, induction is performed on the thermocompetent bacteria JM109. The objective is to verify if the cloned gene leads to the production of a protein. The expected size of the ADR protein is 40 kDa. A very strong expression of the ADR protein is observed at this size when the pBAD promoter is induced with arabinose. The gene has therefore been correctly cloned into the strain and the protein is produced.
Fig25: SDS Page 8% photography following the induction of JM109 with arabinose after 4 hours of culture.
Coloring with coomassie blue. The lane 1 to 4 correspond to induce or non induce cultures transformed with ADR N-ter/pSB1A3. Lane 6 to 8 correspond to induce or non induce cultures transformed with ADR C-ter/pSB1A3.
The last step will consist in evaluating the enzymatic activity of the protein in vitro.
In order to produce the molecule of interest 2-nonanone, we worked with the Lawrence Berkeley National Laboratory, USA which is working on biofuels and modified E. coli strain and obtain a production of 2-nonanone. This production is possible using free fatty acids as substrate.
TesA - Acyl ACP thioesterase
Here we present the cloning of thioesterase I (TesA), an enzyme involved in the synthesis of free fatty acids in E. coli.
Thanks to Geneious software we have designed a gene with a promoter, and a tag. This part doesn’t have a terminator because its produced to create a composite part with other gene involved in 2-nonanone synthesis. The promoter will therefore be associated with the design of the last gene of the composite part. The promoter is inducible to arabinose. This allows a controlled expression of the synthetic gene to avoid any effect of toxicity. In addition, arabinose is an inexpensive inducer and very present in the laboratories of our university.
This part is already exciting with number. But we decided to improve it by adding a 6-his tag. This allows to purify and detect the protein in the host strain by using Ni-NTA columns.
Fig 26: Photography of designed gene realized with the Geneious software.
Design of TesA C-ter flanked with a tag 6-His and restriction sites EcorI, XbaI, SpeI and PstI + pBAD promoter. This is a 5'-3'-end sequence.
Following the design of the synthetic gene, It is amplified by PCR thanks to the design of primers upstream and downstream of the sequence. After amplification of the synthetic gene, sample is purified, the amplicons are digested with restriction enzymes EcoRI and PstI. Similarly for the cloning vector pSB1A3 according to the protocol described above. The insert (TesA) is then ligated into the plasmid.
The PCR product as well as the digestion products are loaded on 0.8 % agarose gel. In well 2, the TesA tagged with 6 his in C-ter amplified by PCR. The most intense band observed corresponds to the size expected for TesA around 900 pb. Another band, this time very weak, is visible below 400 pb. This band may be due to a specific pairing of the primers.
Fig 27: Electrophoresis gel photography following load of TesA PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is NEB 1 kb Plus DNA Ladder.
The products of digestion are also loaded on the gel. In well 2 we see the purified PCR TesA product. There is little DNA loss here, which is encouraging. Wells 3 and 4 respectively show the digestion of the plasmid and the TesA gene by the restriction enzymes EcoRI and PstI. This is to form cohesive ends between the two. We obtain bands at the expected sizes, about 2200 pb for the plasmid and 900 pb for the synthetic gene TesA.
It is important to note, however that agarose gel migration does not verify the effectiveness of digestion. Indeed, since the restriction sites are at the end of the sequences, only a few base pairs have been removed on either side. The resolution of an agarose gel does not make it possible to observe the size of the fragments so precisely. This step makes it possible to ensure that we did not have a loss of DNA during experiments.
Fig 28: Electrophoresis photography following loads on agarose gel 0.8% of enzymatic digestion products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus Ladder (left in the figure). Lane 1 corresponds to the marker, lane 2 to the purified PCR product, lane 3 to the digested pSB1A3 plasmid and lane 4 to the digested TesA synthetic gene.
Finally, the ligation are performed. To model the construction, we used the Geneious software which allowed us to assemble the sequence of the plasmid pSB1A3 as well as that of ADR N-ter or ADR C-ter inserts at the cloning site. We can then know the size of the construction and visualize the restriction site present on the latter.
Fig 29: Modelization of the ligation between TesA/pSB1A3 with Geneious software.
This map shows the pBAD promoter upstream the coding sequence of the ADR protein. Also present in C-ter the 6-His tag. Finally, in the plasmid is present an ampicillin resistance cassette.
The thermocompetent E. coli JM109 bacteria are then transformed and clones are obtained. Different volumes of transformed bacteria are spread on Petri dish with selective medium. The number of clones obtained is consistent with the proportion of bacteria spread on the petri dishes.
Fig 30: Clones on a selective LB medium (+ ampicillin 100 µg/mL) following the transformation of thermocompetent cells with the pSB1A3-TesA ligation and a control plasmid.
A&B: Clones obtained from pSB1A3-TesA. C: Clones obtained from a control plasmid pSB1C3-Red (iGEM parts).
After bacterial transformation, colony PCR is performed with the forward primer of the TesA gene and a reverse primer of the plasmid. 24 clones are tested. The PCR products are loaded onto 0.8% agarose gel.
Clones 1, 3, 4, 5, 6, 7, 8, 9, 12, 13, 14, 15, 17, 18, 19, 21, 23 and 24 have the right profile, an insert-vector fragment of 1100 pb. Wells 2 and 11 show nothing so they probably did not integrate the ligation products.
Wells 10, 16, 20 and 22 seem to have incorporated the aspecific band obtained after PCR on the synthetic gene.
Fig 31: Electrophoresis photography following loads on agarose gel 0.8% of colony PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus Ladder. Lane 1 to 10 corresponds to colony PCR performed on TesA/pSB1A3 ligation, lane 11 to 24 corresponds to colony PCR performed on TesA/pSB1A3.
Clones with the right profile are returned to liquid culture and minipreparations are performed. In order to avoid any risk of point mutation, sequencing is performed with the plasmid primer.
After sequencing, induction is performed on the thermocompetent bacteria JM109. The objective is to verify if the cloned gene leads to the production of a protein. The expected size of the TesA protein is 20 kDa. An expression of the TesA protein is observed at this size when the pBAD promoter is induced with arabinose. The gene has therefore been correctly cloned into the strain and the protein is produced.
Fig 32: SDS Page 12% photography following the induction of JM109 with arabinose after 4 hours of culture.
Coloring with coomassie blue. The lane 1 to the marker. The lane 2 to 5 correspond to induce (I) or non induce (NI) cultures transformed with TesA/pSB1A3.
The next step consists in evaluating the enzymatic activity of the TesA protein in vitro. An acid value (AV) was performed after a fatty acid extraction with hexane on bacterial culture media. Two conditions were used as our pBAD promoter is inducible by arabinose: induced and non induced samples. The volume of potassium hydroxide (KOH) in milliliters to neutralize the free acidity in our samples was measured thanks to an indicator dye: the phenolphthalein. The acid value, which represent here the quantity of KOH needed to neutralize the free fatty acids existent in the sample, was then calculated. It exhibited a difference between the induced sample and the non-one. The non-induced sample presented a smaller AV (0.03 mg) compared with the induced one (3.5 mg) which confirms that the induction leaded to an active TesA 6-His which increased the production of free fatty acids in the media.
Fig 33: Acidic value of TesA 6-His induced and non-induced fatty acids extracted samples.
Transformed bacteria with TesA 6-His gene and pBAD promoter were cultured until DO600nm = 0.5 then induced or not with 0.2% arabinose for 4 hours. Fatty acids were then extracted from our bacterial culture media with hexane and acidic value was then measured.
The final step consists in evaluating the enzymatic activity of the TesA protein by checking the ability of the strain to produce octanoic and decanoic acid.
After 24h of a culture of this strain in an LB medium with antibiotics, 1ml of the culture was extracted, and the fatty acids were esterified into fatty acid ethyl ester (FAEE) for them to be analyzed by Gas Chromatography. The FAEE were then extracted with Hexane and were analyzed by GC. Unfortunately, the HP-5 column that we usually used was unusable, thus we had to use the only column remaining, a BPX-70, which is used to study the Fatty acid methyl esters (FAMEs).
Fig 34: GC results after 24h of culture of JM109/pSB1A3’TesA strain and ethyl esterification of fatty acids.
X axis:Time (min). Y axis: Response (µV). A: Result of the GC analysis of pure hexane (peak on the left) and of 1g/l of octanoic and decanoic acid standards (peaks on the right). B: Result of the GC analysis of Non-induced and Induced JM109/pSB1A3’TesA at T+24h of culture after ethyl esterification of the FFAs a hexane extraction. The spectrum of the hexane, shown in A, is present in every spectrum. The spectrums of octanoic and decanoic acid standards present a response like a merge of two peaks. The retention time of the octanoic and decanoic acid standards only differ by 1 second. The spectrums of the samples are very similar to the spectrums of the standards.
As shown on this GC spectrums, the JM109/pSB1A3’TesA sample seem to present the same spectrum than the octanoic acid and decanoic acid standards, meaning that the JM109/pSB1A3’TesA might indeed produce one of these molecules. However, the retention time of the standards only differs by 1 second, preventing us from being able to differentiate the presence of one or the other of these molecules in the samples. It also seems that the peaks observed in the samples and the standards could, in fact, be a merge of two peaks. In any case, a good GC peak is not supposed to look like that. The ethyl-esterification of fatty acids method used here could be at the origin of this issue. a better protocol should be used next.
Finally, the molecules of interest are detected very early in the analysis, meaning that the BPX-70 column doesn’t retain our molecules and can’t be used to analyze the presence of octanoic and decanoic acids. A new HP-5 column must be bought next in order to redo this experiment.
FadM - Long-chain acyl CoA thioesterase
PCR amplification
Following the design of the synthetic gene, it is amplified by PCR thanks to the design of primers upstream and downstream of the sequence.
Fig 35: Electrophoresis photography following loads on agarose gel 0.8% of PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus DNA Ladder. Lane 1 corresponds to the marker, lane 2 to the amplified FadM product.
Enzymatic digestion and ligation in pSB1C3
After amplification of the synthetic gene, sample is purified, the amplicons are digested with restriction enzymes EcoRI and PstI. Similarly for the cloning vector pSB1C3. The insert (FadM) is then ligated into the plasmid.
Fig 36: Design of FadM/pSB1C3 with Geneious software.
This map shows the pBAD promoter and its terminator flanking the coding sequence of the FadM protein. A Flag tag is also present in C-ter. Finally, in the plasmid is present and chloramphenicol resistance cassette.
Cloning into thermocompetent cells JM109
The thermocompetent E. coli JM109 bacteria are then transformed and clones are obtained.
Fig 37: Clones on a selective LB medium (+ chloramphenicol 25 µg/mL) following the transformation of E. coli thermocompetent cells with the FadM/pSB1C3 ligations.
PCR colony screening
After bacterial transformation, colony PCR is performed with the forward primer of FadM gene and reverse primer hybridizing into the plasmid. The PCR products are loaded on 0.8% agarose gel.
Fig 38: Electrophoresis photography following loads on agarose gel 0.8% of colony PCR products.
The migration was performed at 100 volts for 30 minutes in TAE 1X. The marker used during the migration is the NEB 1 kb Plus Ladder (left in the figure). Lane 1 to 10 corresponds to colony PCR performed other manipulations not study here, lane 1& to 20 corresponds to colony PCR performed on FadM/pSB1C3 ligation.
Clones 11, 12, 14, 16, 18, 19 and 20 have the right profile, an insert-vector fragment of 1200 pb.
The last step consists in evaluating the enzymatic activity of the protein in vitro.
Mlut_11700 - Acyl CoA oxidase
PCR amplification
Following the design of the synthetic gene, it is amplified by PCR thanks to the design of primers upstream and downstream of the sequence.
Enzymatic digestion and ligation in pSB1C3
After amplification of the synthetic gene, sample is purified, the amplicons are digested with restriction enzymes EcoRI and PstI. Similarly for the cloning vector pSB1C3. The insert (Mlut_11700) is then ligated into the plasmid.
Fig 39: Design of Mlut_11700/pSB1C3 with Geneious software.
This map shows the terminator corresponding to the pBAD, flanking the coding sequence of the Mlut_11700 protein. A tag is also present in N-ter. Finally, in the plasmid is present and chloramphenicol resistance cassette.
Cloning into thermocompetent cells JM109
The thermocompetent E. coli JM109 bacteria are then transformed and clones are obtained.
Fig 40: Clones on a selective LB medium (+ chloramphenicol 25 µg/mL) following the transformation of E. coli thermocompetent cells with the Mlut_11700/pSB1C3 ligations.
PCR colony screening
After bacterial transformation, colony PCR is performed with the forward and reverse primer hybridizing into the plasmid. The PCR products are loaded on 0.8% agarose gel.
Our molecules of interest, 2-nonanone, 4-oxo-octanoic and 4-oxo-decanoic come from the chemical signature of Asian hornet and will be used to substitute the current food bait. As we want them to be produced by synthetic biology, we need an analytical method to determine if our strain produces them or not. One of these analytical methods is HPLC, which allows identifying a molecule by its absorption spectrum. As the absorption spectra of these molecules were unknown, we created a databank of these spectra from the chemical standards we bought. To be more precise in our analysis, we also bought two fatty acids; octanoïc acid (C8) and decanoïc acid (C10), precursors of our molecules of interest, and add their spectra to the bank.
The principle of the HPLC is to separate the components of a sample on a stationary phase (HPLC column). This sample will be carried by a mobile phase named eluent through the column and the different compounds contained by the solution will be separated depending on their affinity with the column and the eluent.
Depending on these characteristics, the different components will migrate through the column more or less quickly. Higher eluent (in our case, acetonitrile) concentration will allow unbinding the molecules with higher affinity with the column. A step at 100% of the eluent will wash the column and unbind all the different molecules from the column.
The HPLC machine we use scan the spectrum between 200 nm and 600 nm of the sample’s compounds during the analysis time. At the end of the test, we obtain a 3D plot for each analysis, which represents the absorbance in time of the sample, for each wavelength. These 3D plots allow isolating peaks that correspond to the different compounds of the injected samples. Each peak is analyzed by the software (Millenium) to obtain the corresponding spectrum.
With this method, we collected our standard’s spectrum to create a databank. Indeed, this databank will be useful to determine if our molecules of interest are produced or not in bacterial samples.
Even if each spectrum corresponds to a compound, with this kind of analysis (HPLC), three kinds of unspecific peaks may be observed;
The observed delay of 5 min is justified by the flow of 0.5 mL/min and the pipes’ length. Indeed, each pipe of the machine is tagged with the volume it contains. In that way, the injection volume migrates in approximately 5 min. Same thing for the 100% acetonitrile grade at 20 min, which arrives at the detector around 25 min.
Here are the spectrum and 3D plot obtained for each standard we analysed at different concentrations in 100% acetonitrile. These spectra will be used later to match samples analysis and determine if our molecules of interest (2-nonanone, C8, C10, 4OOA and 4ODA) are present in the bacterial samples.
Fig 41: HPLC 3D plot of the 2-nonanone standard, viewing of absorbance in each wavelength according to retention time in acetonitrile HPLC analysis.
100µl of the standard at 100 g/L in pure HPLC grade acetonitrile are injected. The stationary phase is a C18 column Luna Omega LC column from phenomenex. The mobile phase is a gradient with acetonitrile and Water from 30% acetonitrile to 100%, at a flow rate of 0.5 ml/min.
Three peaks are observed, two of them around 22 min, the last one at 25 min. Peak 1 at 205 nm/22 min may be unspecific, by the precision wavelength range of the machine. Peak 2 at 243.3 nm/22 min appears in the majority of the 2-nonanone tests and seems specific to this compound. Peak 3 at 25 min corresponds to the washing time and may be unspecific. All these peaks are selected and analyzed with the software. As thought, only peak 2 presents a usable spectrum:
Fig 42: Absorbance spectrum of the 2-nonanone standard’s, the absorbance intensity is shown for each wavelength between 200 and 600 nm.
Extracted spectrum from the previous 3D plot. The peak 2 around 22 min has been isolated with the Millenium software.
This peak presents a maximum absorbance of 1,049 at 243.3 nm at 21.658 min and appears in the majority of 2-nonanone tests. This spectrum is saved in the data bank of specific absorbance peaks we are creating. The left part (around 201-205 nm) is due to the machine and is unspecific.
Fig 43: HPLC 3D plot of the 4OOA standard, viewing of absorbance in each wavelength according to retention time in acetonitrile HPLC analysis.
100 µl of the standard at 10 g/L in pure HPLC grade acetonitrile are injected. The stationary phase is a C18 column Luna Omega LC column from phenomenex. The mobile phase is a gradient with acetonitrile and water from 30% acetonitrile to 100%, at a flow rate of 0.5 ml/min.
Three peaks are observed, the first one around 5 min, the two last around 22 min. Peak 1 at 205 nm/5 min and peak 2 may be unspecific, by the precision wavelength range of the machine. Peak 3 at 276,5 nm/22 min appears in the majority of the 4OOA tests and seems specific to this compound. A plateau is observed at 205 nm and beyond during the whole test, it is explained by a lack of precision of the machine at these wavelengths. All these peaks are selected and analyzed with the software. As thought, only the peak 3 presents a usable spectrum:
Fig 44: Absorbance spectrum of the 4OOA standard’s, the absorbance intensity is shown for each wavelength between 200 and 600 nm.
Extracted spectrum from the previous 3D plot. The peak 3 around 22 min has been isolated with the Millenium software.
This peak presents a maximum absorbance of 0.239 at 276.5 nm at 22.58 min and appears in the majority of 4OOA tests. This spectrum is saved in the data bank of specific absorbance peaks we are creating. The left part (around 201-205 nm) is due to the machine and is unspecific.
Fig 45: HPLC 3D plot of the 4ODA standard, viewing of absorbance in each wavelength according to retention time in acetonitrile HPLC analysis.
200µl of the standard at 100 g/L in pure HPLC grade acetonitrile are injected. The stationary phase is a C18 column Luna Omega LC column from phenomenex. The mobile phase is a gradient with acetonitrile and water from 30% acetonitrile to 100%, at a flow rate of 0.5 ml/min.
Five peaks are observed, at 2 min, 20 min, 23 min, 23 min, and 25 min. Peak 1 at 205 nm/2 min, peak 2 at 201 nm/20 min, peak 3 at 201 nm/23 min and peak 5 at 201/25 min may be unspecific, by the precision wavelength range of the machine. Peak 4 at 276,5 nm/23 min appears in the majority of the 4ODA tests and seems specific to this compound. All these peaks are selected and analyzed with the software. As thought, only peak 4 presents a usable spectrum:
Fig 46: Absorbance spectrum of the 4ODA standard’s peak 4, the absorbance intensity is shown for each wavelength between 200 and 600 nm.
Extracted spectrum from the previous 3D plot. The peak 4 around 22 min has been isolated with the Millenium software.
This peak presents a maximum absorbance of 0.331 at 276.5 nm at 23,295 min and appears in the majority of 4ODA tests. This spectrum is saved in the data bank of specific absorbance peaks we are creating. The left part (around 201-205 nm) is due to the machine and is unspecific.
Fig 47: HPLC 3D plot of the C8 standard, viewing of absorbance in each wavelength according to retention time in acetonitrile HPLC analysis.
100 µl of the standard at 100 g/L in pure HPLC grade acetonitrile are injected. The stationary phase is a C18 column Luna Omega LC column from phenomenex. The mobile phase is a gradient with acetonitrile and Water from 30% acetonitrile to 100%, at a flow rate of 0.5 ml/min.
Four peaks are observed, at 5 min, 22 min, 22 min, and 25 min. Peak 1 at 201 nm/5 min and peak 3 at 201 nm/22 min may be unspecific, by the precision wavelength range of the machine. Peak 2 at 224,5 nm/22 min appears in the other C8 tests and seems specific to this compound. Peak 4 at 25 min corresponds to the washing time and maybe unspecific. All these peaks are selected and analyzed with the software. As thought, only peak 2 presents a usable spectrum:
Fig 48: Absorbance spectrum of the C8 standard’s peak 2, the absorbance intensity is shown for each wavelength between 200 and 600 nm.
Extracted spectrum from the previous 3D plot. The peak 2 around 22 min has been isolated with the Millenium software.
This peak presents a maximum absorbance of 1.14530 at 224.5 nm at around 22,218 min. This spectrum presents also a second absorbance value about 0.1 at 271.7 nm. This spectrum appears in other C8 tests and the profile seems specific to the compound. This spectrum is saved in the data bank of specific absorbance peaks we are creating. The left part (around 201-205 nm) is due to the machine and is unspecific.
Fig 49: HPLC 3D plot of the C10 standard, viewing of absorbance in each wavelength according to retention time in acetonitrile HPLC analysis.
100 µl of the standard at 100 g/L in pure HPLC grade acetonitrile are injected. The stationary phase is a C18 column Luna Omega LC column from phenomenex. The mobile phase is a gradient with acetonitrile and water from 30% acetonitrile to 100%, at a flow rate of 0.5 ml/min.
One peak is observed 25 min, time of the usual washing time peak. For each C10 test, we observed similar results in wavelength and retention time. But it doesn’t seem really specific. More tests are needed. This peak is selected and analyzed with the software. The obtained spectrum seems usable:
Fig 50: Absorbance spectrum of the C10 standard’s peak, the absorbance intensity is shown for each wavelength between 200 and 600 nm.
Extracted spectrum from the previous 3D plot. The peak has been isolated with the Millenium software.
This peak presents a maximum absorbance of 1,83 at 229.2 nm at 25,943 min and is similar to others spectra obtained in C10 tests. For unknown reasons, for this molecule, the various obtained spectra presents some differences of wavelength and retention time which still close to others’ ones. These different spectra are added to the data bank.
With the various peaks we obtained by these analysis, we created a data bank of specific absorbance peaks.
Fig 51: Extract from the “iGEM Acetonitrile” data bank of the standard’s spectrum characteristics.
This data bank gathers all recorded peaks’ characteristics: retention time, wavelength of maximum absorbance. The second wavelength of the spectra isn’t recorded by the software; the column “Lambda2” and “Max A2” have been added by the team on a separate document.
This data bank can be used to identify the various compounds of a bacterial sample by matching the different peaks obtained for an injection:
Fig 52: Matching of Nonan-2-one sample at 100 g/L in methanol against the “iGEM Acetonitrile” spectrum data bank. The absorbance intensity is shown for each wavelength between 200 and 600 nm.
This peak identified in a methanol corresponds to 3 other peaks in Acetonitrile conditions; the four absorb at 243.4nm. The peaks in acetonitrile conditions appear latter than in methanol (around 21.5 min against 8.2 min in methanol), but have a stronger absorbance thanks to the low noise floor of the acetonitrile
The creation of this spectrum data bank will be used later to analyse bacterial samples and determine if the bacteria produce molecules of interest of the project. In this test, we confirmed the possibility to identify the same molecule in different conditions, here in methanol.