Overview
In vivo projects are invariably complicated by the many moving parts in biological systems. It is therefore an important goal in synthetic biology to modularise genetic components and their individual functions, so that different functional parts can be combined in any number of ways without significantly compromising the integrity of individual components.
BEACON’s project was composed of several modular aspects, each serving independent purposes. It is intended that, with the exception of the complete proof of concept, any of the following sections can be read, understood and used without the need for confusing crosstalk.
Transformation of lanM into Methylobacterium extorquens (Mex)
Gibson assembly of the lanM gene into pMW224 was successful. We successfully transformed the lanM-pMW224 plasmid into E.coli, as evident by the positive colony PCR (Figure 1), restriction digest and sequencing results (Figure 2).
The alignment is not 100% identity, but each individual mismatch was assessed for impact on function, and it was found that all mismatches either led to synonymous substitutions or were present outside of annotated features. Therefore, this plasmid was used to transform into M. extorquens.
Following electroporation into Mex, colonies grew on plates with the selectable marker tetracycline, indicating successful insertion. Additionally, there were positive results for the colony PCR using primers for the assembled plasmid.
Transformation of tar into Mex
Gibson assembly of the tar gene into pMW220 was successful; the assembled plasmid is named pTAR001. We succesfully transformed the tar-pMW220 plasmid into E.coli DH5α, as evidenced by the positive colony PCR (Figure 5), restriction digest (Figure 3), and sequencing results (Figure 4). For the restriction digest, enzymes were chosen which cut in both the insert and backbone to yield meaningful band sizes.
As seen from Figure 3, the plasmid in the first lane (pTAR001) has the same results as that of the expected bands for the assembled plasmid which suggests the assembled plasmid has been correctly inserted. This is confirmed by the sequencing results. The plasmids in other lanes show improper band patterns, so we used the plasmid in the first lane for all future experiments. Sequencing confirmed a 100% identity of this plasmid with our intended assembly.
Following electroporation into Mex, colonies grew on plates with the selectable marker gentamycin. Additionally, there were positive results for the colony PCR.
The colony PCR suggests that colonies 3, 4 and 6 contain the pTAR001 plasmid. For future experiments, we chose to use colony 3.
Transformation of lanM + tar into Mex
We then transformed both the pLNM001 and pTAR001 plasmids into Mex. We did this via electroporation. These transformations were reliably successful, and were routinely confirmed by colony PCR (Figure 6).
From figure 6, we can see that several colonies tested had reliably uptaken the plasmid. We chose to use colony 3 as it had grown well and appeared successful on the gel.
Transformation of XoxF into BL21
As seen in Figure 7, the pXOX001 successfully transformed into BL21 cells.
As seen in Figure 8, all the colonies from transformation contained the gene insert xoxF, since the bands aligned with the positive control. Though the length is not that of the expected length, it was later confirmed that this was an issue with the DNA ladder. Therefore we grew the colonies and purified the plasmids by miniprep to send off for sequencing at Full Circle Labs.
Transformation of lutH into M. extorquens
We attempted to transform our pUMP001 plasmid into E. coli with heat-shock:
To test if our transformed bacteria had the lutH insert, we conducted colony PCR.
From the above figure, it can be seen that colonies 1-7 and colony 14 appeared to contain the insert, as the bands somewhat aligned with the positive control. The figure does however show that the amplicons are heterogeneous, which indicated that there may potentially be an issue. We grew the seemingly successful colonies and purified them with miniprep to send the plasmids containing lutH off for sequencing at Full Circle Labs.
The sequencing results in the above figure show that only a part of the lutH gene was inserted into the plasmid, suggesting that the bacteria have ejected parts of the lutH gene because the pumps expressed by this gene causes too much stress for the bacterium, even under leaky, uninduced expression.
We would have chosen a different vector with a weaker promoter, but around this time it became apparent through modelling and WT characterisation that the WT uptake was sufficiently efficient that engineered lutH was unnecessary.
Wild type, tar and lanM+tar Swim plate assay demonstrating chemotaxis
We conducted swim plate assays to test our chemotactically active engineered strains, tar and lanM + tar, for chemotaxis towards aspartate.
| Plate 1 | Plate 2 | Plate 3 | |
|---|---|---|---|
| Wild type |  |
 |
 |
| tar |  |
 |
 |
| lanM + tar |  |
 |
 |
The swim plates in Table 1 are composed of media and 0.2% w/v agar. This was the required concentration to facilitate motility in our strain, and unfortunately leads to the agar splitting upon inoculation, reducing the robustness of our data. However, the results can still be analysed as below.
The more opaque circle in each picture is the 2% w/v aspartate plug.
The two distances in Figure 1, D1 and D2 (Figure 1), were measured digitally on GeoGebra and their ratios were calculated with the formula: and the following results were obtained:
| Wild type | tar | lanM + tar | |
|---|---|---|---|
| Plate 1 | 1.41 | 1.8 | 0.802 |
| Plate 2 | 1.054 | 1.898 | 3.563 |
| Plate 3 | 1.99 | 1.499 | 1.061 |
| Mean | 1.485 | 1.732 | 1.808 |
The split agar significantly impacted the reliability of our data, such that we were unable to obtain statistical significance with the number of replicates we performed. We nonetheless consider the results to be qualitatively promising, and would in future optimise the setup to remove the noise generated by the splitting.
| tar Plate 1 | tar Plate 3 | lan + tar Plate 2 |
|---|---|---|
 |
 |
 |
tar Plate 1, tar Plate 3 and lanM + tar Plate 2 clearly showed that the cells swam towards the aspartate plug indicating a chemotactic response.
Therefore, we conclude that our strains with the inserted gene tar, are likely to show more chemotactic behaviour towards aspartate than the wild type.
Stress plate assay
This assay involves growing serial dilutions of cells on media supplemented with increasing concentrations of Nd. The range selected was intended to represent concentrations around typical environmental conditions linearly, and concentrations above environmental conditions increasing exponentially. This was expected to give the most informative dataset.
Growth is supported at low concentrations of Nd, and is preferred by the WT at 5ppm, so future assays were conducted at 5ppm. Growth is inhibited at concentrations as high as 10ppm, and almost prevented by 25ppm. No growth is observed at higher concentrations (the visible blemishes on the 1500ppm plate are caused by NdCl3 diffusion into the 2µL droplet followed by evaporation and precipitation of the salt).
Later in the project, when we had successful transformants of lanM under inducible expression, we repeated this experiment to demonstrate the difference between the strains. The WT controls for this run of the experiment were unfortunately contaminated, but we present what results we do have for the lanM strain. The lack of a control for these specific plates sheds some doubt on these results, and we would like to repeat this experiment with newly grown colonies.
These results can be quantified by counting colonies at the lowest dilution for which growth can be observed, and back-calculating the number cells that can viably grow in the original OD600 = 1 culture.
These results do not show the full range of stress for LanM, but nonetheless demonstrate a clear success of stress reduction over at least two magnitudes of [Nd].
Neodymium uptake assay
This assay involved incubating cells in late exponential phase overnight with IPTG and Nd. The cells were then collected by centrifugation, lysed by sonication and the resultant sample quantified for Nd by ICP-MS. ICP-MS also facilitates standardising to cell density by quantifying sulfur present – an approximately constant average value per cell in all our strains.
We incubated M. extorquens strains in 5ml media containing 5ppm Nd. We tested WT and inducible expression LanM transformants.
The volume of sample added to 5ml ICP-MS (Inductively coupled plasma mass spectrometry) solution (ultrapure 3.6% HNO3) was calculated such that at 5ppm Nd (as for equilibration by passive transport) the concentration of Nd would be within the sensitive range of ICP-MS: 5ppb-1000ppb.
| Sample | [Nd] in media (ppb) | [Nd] in pelleted sample (ppb) in 5ml HNO3 | [Nd] standardised by WT cell density (ppb) | [Nd] standardised to WT |
|---|---|---|---|---|
| WT | 5000 | 9834 | 9834 | 1 |
| LanM | 5000 | 7162 | 14449 | 1.469 |
Consequently, this result is somewhat unreliable as it lies outside of the calibration curve, but the magnitude remains reasonably reliable.
Clearly, the uptake at this concentration is already extremely efficient even in the WT. Consequently, we decided against continuing to engineer in uptake mechanisms, and focused on stress mitigation for the bioaccumulation aspect of the project.
We do see a noticeable increase in uptake in the lanM strain, which indicates that our idea of using it as a buffer system to sequester neodymium from neodymium-inhibited uptake pathways is functional.
Wild type Mex chemotaxis swim plate characterisation (unsuccessful)
We conducted swim plate assays with wild type Methylobacterium extorquens to determine if the wild type strain already has a chemotactic response to chemoattractants for which E. coli shows positive taxis such as aspartate, methanol and other chemicals such as maltose, serine and galactose.
| Buffer only plug (negative control) | Aspartate plug (or other test attractant) | Methanol plug (positive control) |
|---|---|---|
|
|
|
As seen in the above table, M.extorquens appears not to show chemotactic behaviour towards any of the plugs, including the positive control, indicating an unsuitable experimental setup. We optimised in several ways, including by decreasing the concentration of agar in the plates to better facilitate motility, but we were able to obtain no positive results, even for the positive control methanol.
From this experiment we could not conclude whether the wild type strain of M. extorquens had a chemotactic behaviour or not towards the plugs. Therefore, we redesigned the setup to use a growth supportive media, which yielded results that could be more easily troubleshot.
| Buffer only plug (negative control) | Aspartate plug (or other test attractant) | Methanol plug (positive control) |
|---|---|---|
|
|
|
We saw growth in these plates, but no motility, which was very concerning, because M. extorquens is widely characterised as being motile (Tsagkari E, Sloan W, 2018) (Alessa O et al., 2021). We concluded that the strain we had received from the Tobias Erb lab was likely domesticated such that it had lost motility, which is a documented phenomenon (Barreto H et al., 2020) (Patrick J, Kearns D, 2009). As a result, we began to perform a directed evolution experiment, wherein we took cells from the farthest point they had reached from inoculation, and regrew them on new swim plates.
Repeating this several times (2-3) produced more motile M. extorquens, which displayed an unusual characteristic motility – it would migrate to the edge of the plate, then spread radially from there. This could perhaps be due to a clumping or quorum sensing pathway. This phenotype is not dissimilar to the phenotype for Methylobacterium demonstrated here.
In any case, we reran the growth-supportive swim plates with attractant-in-plug, but saw no motility – we have come to the conclusion that the primary metabolic substrate methanol induces motility, as the cells have more energy available. Methanol could be used in the directed evolution swim plates, but was excluded from the attractant-in-plug taxis plates because it would have interfered with the positive control.
As a result, we ordered the DNA for the aspartate-sensing MCP Tar, and decided to rerun the swim plates with methanol using aspartate as the attractant, as has been detailed in the successful results section.
Lanthanide enrichment proof of concept swim plates
Our proof of concept aimed to see if our bacteria could enrich certain spatial areas in Nd. That is, we want to know if our bacteria disrupt homogeneity of Nd within the media. To this end, we collected samples from 3 distances from the plug, and quantified the Nd in them. Before this, however, we were able to define our statistical null hypotheses.
H0, homogeneity: The concentration of Neodymium in the media is uniformly distributed.
H0, enrichment: The concentration of Neodymium in the media is uniformly distributed, with µ = [Nd] added to plates.
The purpose of H0, homogeneity is to capture whether we see a disruption in the uniformity of Nd within the locations of our samples.
The purpose of H0, enrichment is to capture whether we see a different concentration of Nd than expected by our calculated value for [Nd] within the plates. This is useful if the bacterial and thus neodymium migration patterns are not accurately represented by our sample collection methodology.
We represent the results visually in a histogram-reminiscent graph (Figure 1). (Full data is available supplementarily.) These graphs display the average Nd across the range of distance from the centre of the plug, which we believe to be the most scientifically precise visualisation with our sample collection methodology.
Performing Chi-squared statistical tests on these datasets yields the following results:
µ for H0, enrichment is for these plates calculated as = 4.167 ppm. A concentration of 5ppm was added to the plates, which was subsequently diluted by addition of 1/5 volume of a solubilisation buffer after sample collection.
| Sample | H0, homogeneity | H0, enrichment | Notes |
|---|---|---|---|
| WT | p = 0.707 | p = 0.590 | |
| tar | p > 0.999 | p = 0.533 | |
| lanM tar uninduced | p = 0.987 | p = 0.304 | |
| lanM tar induced | p = 0.181 | p = 0.003 < 0.05 | One replicate appeared to fail – excluding this yields p = 0.0280 < 0.05 for H0, homogeneity. It is poor practice to exclude replicates without cause, so we mention this but do not claim this level of significance. |
We see what we consider to be a likely; and biologically and industrially meaningful disruption of homogeneity, but we do not achieve strict statistical significance by this metric with only duplicates – a limitation of ICP-MS is that the cost prevented us running more samples.
However, it is probable that we miss Nd migration from the very edges of the plate with our collection methodology, because we do see a statistically significant greater set of values than expected across the board in our complete, induced strain. The greater than expected neodymium concentrations in these samples must have come from somewhere, and the most reasonable explanation is that the neodymium came from the edges of the plate.
In our proof of concept, we do also see a clear trend in [Nd], increasing closer to the plug, as expected.
We do not see effective Nd migration in the tar only strain, and only very minor, statistically insignificant migration in the uninduced tar, lanM strain. We hypothesise that this is a consequence of that cells that uptake Nd immediately begin to express Tar, and therefore swim towards the plug, which drastically reduces the amount of Nd they take with them.
We believe LanM acts as a buffer against uptaken Nd, preventing it from mediating an upregulation in tar expression until the LanM is saturated and cannot sequester any additional Nd.
What is ICP-MS?
Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is an advanced analytical technique used to detect and quantify trace elements and isotopes in a variety of samples. It works by converting a liquid sample into an aerosol, which is then introduced into a high-temperature plasma (approximately 6,000 to 10,000 Kelvin) generated by argon gas. In the plasma, the sample undergoes ionization, transforming the atoms into positively charged ions. These ions are then directed into a mass spectrometer, where they are separated based on their mass-to-charge ratio (m/z). A detector measures the intensity of each ion, providing a highly sensitive analysis of the sample's elemental composition with detection limits often in the parts-per-billion (ppb) or parts-per-trillion (ppt) range.
Why are we using it?
Aside from it being a very cool technique (that the chemist in our team really geeked out on), the significance of ICP-MS in our project lies in its ability to accurately measure the uptake of neodymium (Nd) by our genetically engineered cells compared to the wild-type strains. Nd is a rare earth metal that is not abundantly present in biological systems, and detecting its precise concentration is crucial to assessing the efficacy of our bioremediation or bioextraction processes. By using ICP-MS, we can quantify even minute differences in Nd accumulation within our modified cells, providing a clear picture of how genetic modifications impact the ability of our bacteria to bind or absorb neodymium. This comparison is essential for validating whether our engineered strains exhibit enhanced metal uptake over their unmodified counterparts, which is a key indicator of the success of our project.
The analytical process begins by lysing the bacterial cells to release any accumulated Nd into the liquid such that it can become homogenous. Once prepared, the sample is introduced into the ICP-MS, where the Nd ions are ionized in the plasma and directed into the mass spectrometer for separation and detection. Because ICP-MS is highly sensitive and can distinguish between different isotopes, it allows us to specifically quantify Nd in the presence of other elements, even if they are present in much higher concentrations. This specificity is vital for our research, as it enables us to accurately monitor Nd uptake without interference from other components of the bacterial growth medium or residual elements present in the cells.
Moreover, ICP-MS can provide isotope ratio information, which could be used to trace the origin of Nd if desired. In our context, while this feature may not be directly relevant for uptake studies, it could become important if we want to differentiate between Nd sourced from different electronic waste materials, or if we wish to track different Nd supplementation in the lab. The technique's speed and ability to handle complex matrices also make it particularly suitable for processing multiple samples efficiently, which is necessary for comparative studies between wild-type and engineered strains.
Finally, one of the most significant aspects of using ICP-MS for our project is its role in refining and validating our biotechnological approach. By generating robust quantitative data, ICP-MS helps us establish the relationship between genetic modifications and metal-binding efficiency. This enables us to iteratively improve our strains, making ICP-MS not just a tool for measurement but a cornerstone of our experimental validation and strain development process. Through repeated testing and analysis, we can confidently determine whether our engineered cells outperform the wild type, ultimately supporting our goal of developing a sustainable and efficient method for rare earth metal recycling.
Protocol
We collected various types of samples, containing cells throughout our project. Cells were always lysed by sonication to allow neodymium to become homogenous in the samples.
After lysis, the cell samples were allowed to cool to room temperature. In a fume hood, 5 mL of 3.6% nitric acid (HNO₃) was added to each lysed sample to digest the cellular material and release any bound neodymium. The mixtures were vortexed thoroughly and left to incubate at room temperature for 1 hour to ensure complete digestion of the cellular matrix. Following the digestion period, deionized water was added to each sample, bringing the final volume to 5 mL, which resulted in a final nitric acid concentration of approximately 10%. The prepared samples were then transferred to clean ICP-MS vials for subsequent analysis.
Before introducing the samples to the ICP-MS instrument, the system was calibrated using standard solutions of neodymium at known concentrations to establish a baseline for accurate quantification. In addition to the experimental samples, blank samples (3.6% nitric acid in deionized water) and quality control standards were included to ensure reliability and precision throughout the analytical run. The samples were then introduced into the ICP-MS using the automated sample introduction system, where the instrument separated the ions based on their mass-to-charge ratio and quantified the Nd content using isotopic signatures such as 144Nd or 146Nd. The results provided a precise measurement of neodymium concentration within each sample, allowing for direct comparison between the engineered strains and the wild-type bacteria.
Post-analysis, the ICP-MS system was thoroughly flushed with deionized water to remove any residual nitric acid or metal ions, preventing contamination in subsequent runs. All waste generated, including the acid-digested samples and rinsing solutions, was disposed of following appropriate hazardous waste protocols to ensure environmental safety and compliance with laboratory regulations. By following this protocol, the total Nd content per gram of dry cell mass or per gram of sample, as appropriate, was accurately determined, providing a clear assessment of the efficiency of Nd uptake by the genetically modified strains compared to the wild-type. This data is crucial for evaluating the success of our project’s aim to enhance rare earth element recovery through bioengineering.
Explaining our results
In our initial set of ICP-MS measurements, we assessed the neodymium (Nd) uptake capacities of our wild-type (WT) Methylobacterium extorquens (Mex) strain and the genetically modified strain expressing the LanM protein (lanM strain). The ICP-MS analysis revealed a significantly higher concentration of Nd in the lanM strain compared to the wild-type. This result indicates that the LanM protein, which is designed to selectively bind and sequester lanthanides, substantially enhances the bacteria’s ability to accumulate neodymium from the surrounding medium.
The increased Nd concentration in the lanM strain confirms the functional role of the LanM protein in facilitating the uptake of rare earth elements (REEs) such as neodymium. The wild-type strain, lacking this specific protein, showed relatively lower Nd concentrations. This disparity between the two strains is crucial, as it validates our genetic engineering approach to improving REE bioaccumulation and suggests that the LanM protein could be a valuable tool for bioremediation applications.
To confirm the difference in Nd concentrations between strains, we performed calculations to standardize our data. Due to low sensitivity, we were unable to use Sulfur and phosphorus to generate reliable calibration curves (Figure 1). We thus also were not able to use these two elements in our data and get absolute values for Nd concentration per individual cell. In turn, we had to use copper (Cu) due to sensitivity towards it being significantly higher. This should be accurate but prevents the calculation of [Nd] per individual cell, so only relative values per cell density can be obtained.
![copper [He] calibration graph](https://static.igem.wiki/teams/5277/images/micpms11.webp)
![Neodymium [no gas] calibration graph](https://static.igem.wiki/teams/5277/images/micpms12.webp)
Despite not having absolute cell mass, we could still standardize relatively to cell density accurately. The calculations we performed are as follows:
Calculations
Where x = the concentration of 144 Nd [No gas] from graph (ppb):
The same was done for the calibration curve using Cu (see Figure 1). The following results were obtained:
| Samples | 144 Nd [No gas] CPS | Conc. of 144 Nd [No gas] from graph (ppb) | ppb from instrument |
|---|---|---|---|
| WT – 1 | 569672284,8 | 10699,90998 | 9169,448 |
| WT – 2 | 671578990 | 12613,47124 | 10499,281 |
| LanM – 1 | 546952725,2 | 10273,29164 | 8539,971 |
| LanM – 2 | 342690728,2 | 6437,745737 | 5783,934 |
| Samples | 63 Cu [He] CPS | Conc of Cu 63 [He] from graph (ppb) | ppb from instrument |
|---|---|---|---|
| WT – 1 | 177,33 | 7,340408834 | 5,374 |
| WT – 2 | 240 | 9,682424605 | 7,598 |
| LanM – 1 | 109,67 | 4,811913749 | 2,794 |
| LanM – 2 | 138 | 5,870622968 | 3,636 |
We then calculated averages of copper to get the relative ratio of cell density. This was then used to standardise concentrations of Nd by cell density. It is assumed that Cu is directly proportional to cell number and constant between strains.
| Samples | Average Cu (ppb) | Standardised to WT |
|---|---|---|
| WT | 6,486 | 1 |
| LanM | 3,215 | 0,49568301 |
| Samples | [Nd] in pelleted sample (ppb) in 5mL HNO3 | [Nd] standardised by WT cell denisity (ppb) | [Nd] standardised to WT |
|---|---|---|---|
| WT | 9834 | 9834 | 1 |
| LanM | 7162 | 14449 | 1,469 |
XoxF Protein Purification
We put a lot of work into attempting to purify XoxF, but we were unfortunately unsuccessful due to the lack of experience with protein work across the team.
Initially, we were directed to the expression vector pET-28a(+) for E. coli BL21. We successfully assembled and transformed this plasmid through DH5α into BL21. However, upon purification it became apparent that the protein was retained in the insoluble fraction, as can be seen in the SDS gel below.
This gel strongly indicates that the protein is insoluble in our expression system. Unsurprisingly, this prevented it from showing any activity. Had we had more experience with protein work on the team, we would have switched our expression system to a successful purification from the literature at this stage. However, we did not recognise the issue posed by the protein being insoluble.
Therefore, we attempted several alternative conditions to promote solubilisation, including:
- Using lysate directly.
- Reduced concentration of inducer (IPTG).
- Performing the assay in vivo.
Unfortunately, none showed activity – it is likely the protein was aggregated in inclusion bodies in every attempt.
At this stage, the iGEM deadline was fast approaching, so we did not have time to redesign our expression system. We therefore attempted a solubilisation and refolding protocol for inclusion bodies. Unfortunately, we lacked some of the requisite chemicals and materials for efficient refolding, and were therefore unable to produce active purified protein. We did reliably obtain denatured, solubilised protein at a reasonable concentration, as shown by the Nanodrop below (Figure 1).
Neodymium Quantification Assay Using XoxF
Unsurprisingly, without active protein our enzymatic assay could not be performed. We present a calibration curve for our photometrically active chemical DCPIP under the conditions of use (Figure 2), and deeply regret that we cannot present results for the assay itself.
DCPIP does not mix rapidly with water, and therefore requires thorough mixing by pipetting before readings are taken.
The gradient 0.0131 is the molar extinction coefficient, ε, in µM cm-1. This can be converted to ε = 13.1 mM cm-1, which is comparable to that in the literature, as summarised here.