Contents

CycA Characterisation

In order to characterise the function of the amino acid transport protein CycA, we attempted to measure the difference in alanine uptake over time by E. coli BL21 DE3 pLysS, engineered with either an empty pX1900 plasmid backbone (control strain), or the cycA gene present in this backbone (CycA strain). In order to calculate alanine uptake per cell, we additionally measured cell density (cells/ml) via imaging flow cytometry. Results indicated that total net alanine uptake per cell was actually reduced in the CycA strain versus the control strain, most likely suggesting CycA is being expressed (using up alanine and other metabolic resources) but is not functional, so does not contribute to elevated alanine uptake.

In performing this experiment, we acquired a set of cell density measurements with corresponding OD₆₀₀ data, for both E. coli strains. This data, and data previously collected by our PI Dr Chloe Singleton, allowed us to assess the accuracy of estimating cell density from standard curves of OD₆₀₀ against the concentration of silica microspheres, as was investigated in the iGEM 2018 Interlab study. We were also able to use this data to produce our own standard curve of OD₆₀₀ against cell density measurements, which other teams will be able to use in the future to rapidly estimate cell density from OD₆₀₀ data to a reasonable degree of accuracy.

Firstly, to determine the tolerance of both strains to extracellular alanine and thus, a suitable alanine concentration for the extracellular medium in our alanine uptake assay, we incubated the control and CycA strains in a range of alanine concentrations from 5.8 (blank LB) up to 100 mM. As is indicated below in figure 1, there is no obvious difference in the growth of either strain across the range of extracellular alanine concentrations studied. From this, we concluded that we could use the uppermost concentration of extracellular alanine, 100 mM, in the growth medium for our alanine assay.

We then cultured both strains in undoped and doped LB ([Ala] = 5.8 and 100 mM respectively), measuring OD₆₀₀ and cells/ml (using an Imagestream imaging flow cytometer) every hour across a 7 hour incubation period, as well as harvesting pellets for the final alanine assay. Figure 2 below shows the variation in cells/ml for each condition with time. It is apparent that the form of growth is essentially the same for the control strain at both alanine concentrations. In contrast the growth for the CycA strain seems to vary between the 2 alanine concentrations. In undoped LB, the CycA strain does not appear to reach stationary phase within the time range measured over, suggesting expression of the protein has delayed the onset of this phase of growth. Whereas, in doped LB, the CycA strain appears to reach stationary phase at an earlier time point than the other conditions and may in fact enter death phase, with the final cells/ml dropping from t=6 to t=7.

It is possible that the combination of expression of likely non-functional and misfolded CycA proteins as inclusion bodies, and the associated metabolic strain is sufficient to trigger cell death by t=7. Alternatively, the apparent decrease in cells/ml may instead result from a change in cell morphology in response to suboptimal conditions (lack of available resources, accumulation of extracellular waste etc), reducing the proportion of live cells present which are identified correctly by the Imagestream.

In order to determine alanine concentration in the harvested pellets, we used the Merck-distributed alanine assay kit. First a standard curve was plotted of absorbance at 570 nm, which is proportional to the concentration of pyruvate (into which alanine has been enzymatically converted) bound by an alanine probe, against known concentrations of alanine standards. We then measured the absorbance of samples collected at t = 3, 5 and 7 hours after inoculation, in turn determining their relative total alanine concentrations, and normalised them by cells/ml, to calculate alanine uptake per cell shown below in figure 3.

Alanine uptake for each strain is higher for the [Ala] = 100 mM condition, as is to be expected. However, alanine uptake is greater for the control strain than the CycA strain, at both alanine concentrations. The most logical explanation for this is that CycA, whilst being expressed, is non-functional, resulting in investment of metabolic resources into the transcription and translation of the protein, without a coupled increase in alanine uptake. This would divert resources (ATP, transcriptional and translational machinery etc) away from synthesis of other porins responsible for alanine uptake. Moreover, CycA translation itself uses significant amounts of intracellular alanine, with alanine accounting for 10% of the AA sequence of the protein (47/470 residues). The simplest explanation for the non-functionality of CycA is that the protein is misfolding, resulting in the formation of inclusion bodies.

The strength of the conclusions we can draw is severely limited by the amount of data we have been able to collect. Whilst we managed to culture in triplicate both strains in undoped and doped LB, we were not able to perform repeats for the alanine tolerance experiment. Moreover, without monitoring OD₆₀₀ at sufficiently high concentrations of extracellular alanine to limit growth of E. coli, we were not able to put an actual upper bound to the alanine concentration at which our 2 strains could grow. Finally, we were significantly limited in the amount of samples we could run the alanine detection assay for, by the availability of the necessary reagents. Consequently, we could only analyse one sample per condition (out of 3), preventing us from assessing the reliability of conclusions drawn regarding alanine uptake per cell.

Accuracy of silica microsphere OD₆₀₀ calibration

To determine the accuracy of calibrating OD₆₀₀ measurements using silica microsphere suspensions, with an aim to estimate cells/ml, we interpolated a standard curve for OD₆₀₀ against microspheres/ml, produced by our PI, Dr Chloe Singleton, shown below in figure 4. A strong linear correlation is obtained between OD₆₀₀ and beads/ml with an R² value of 0.9921. This confirms the linearity of the relationship between the 2 quantities across an OD₆₀₀ range from 0 to 0.8, enabling interpolation of objects/ml data within this range.

We then compared plots of OD₆₀₀ and interpolated microspheres/ml results against time, with plots of cells/ml (using Imagestream data from the CycA characterisation experiment), shown in figure 5 below. The form of the plots of OD₆₀₀ and microsphere ‘beads’ per ml against time are identical, which is logical given the direct interpolation of beads/ml values from OD₆₀₀ measurements. In contrast, the form of the plot of cells/ml against time substantially differs from the other two plots, reflecting the inability of OD₆₀₀, and thus interpolated beads/ml measurements, to ignore the contribution of non-live cell objects to absorbance. Whilst the cells/ml plot corroborates the OD₆₀₀ and beads/ml plots in so far as when log and stationary phases are entered, the variation in growth after stationary phase is entered (t>5 hours) between the 4 conditions measured over is not accurately captured in the beads/ml plot.

The beads/ml plot suggests growth in all 4 conditions for the final 2 hours is approximately uniform, however, actual cells/ml data shows growth was much greater for the CycA strain in undoped LB, which does not actually appear to reach stationary phase within the time period measured over, than the other 3 conditions. Moreover, as is more clearly illustrated in figure 2, cell count/ml may have actually decreased from t=6 to t=7 hours after inoculation for the CycA strain in doped LB. The decrease in cells/ml observed in the Imagestream data for this condition, and the variation in the rate of slowdown of growth for the other 3 conditions, is not accurately captured in the beads/ml approximation. This most likely results from the fact that microsphere estimates from OD₆₀₀ data cannot distinguish between live cells and other factors contributing to absorbance, including aggregates of cell debris from lysed cells, which can be identified and filtered out of flow cytometry data on an Imagestream through examination of the relevant object images.

The percentage inaccuracy of the microsphere calibrated estimate of objects/ml as a function of time is plotted below in figure 6. The form of the curve for each condition is approximately the same. Percentage difference (between Imagestream cells/ml data and approximated beads/ml data) increases with time during lag phase from t=0 to a maximum at t=2-3 hours after inoculation. Percentage difference then decreases with time during log phase from t=3 to a minimum at t=5 hours. Finally, percentage difference then either plateaus, as for the control strain in doped LB and CycA strain in undoped LB, or increases in stationary phase, as for the other 2 strains. It is a point of note that the 3 phases of growth (excluding death phase) are clearly demarcated by turning points in the plots of percentage difference against time. The variation in growth between all 4 conditions from t=6 to t=7 hours is reflected in the fact that the percentage difference varies across this time period differently for the 4 conditions.

These results illustrate clearly the limitations of estimating cell count via microsphere calibrated OD₆₀₀ measurements. Measuring change in OD₆₀₀ across an incubation period cannot capture changes in cell morphology in response to environmental stressors, nor cell lysis in the eventuality of the resultant debris aggregating and contributing to overall light scattering. Results also exemplify the advantages provided by use of Imagestream flow cytometry, which not only measures objects per ml by analysing light scattering, it also yields images of each identified object. These images allowed us to examine whether any changes in cell morphology over time were present, and provided reference points when gating the signal region which denotes objects of interest.

Strong conclusions cannot however be drawn from this data alone. Whilst the reliability of data obtained is verified by the collection of 3 replicates, the reproducibility of our conclusions has not been tested. We hope future iGEM teams will attempt to reproduce our results and further assess the degree of accuracy of microsphere calibrated OD₆₀₀ measurements of live cell count.

Estimating cell density from OD

In the process of generating data for measuring alanine uptake, we obtained OD₆₀₀ measurements using a Tecan Infinite M PLEX plate reader and corresponding cells/ml measurements using an Imagestream image flow cytometer for 4 different strains of E. coli BL21 DE3 pLysS across a 7 hour incubation period. Linear regression of OD₆₀₀ as a function of log₁₀(cells/ml) (given OD₆₀₀ is a logarithmic metric of light attenuation) identifies the former is directly proportional to the latter across an OD₆₀₀ range from 0 to 0.9. As shown below in figure 7 below, the goodness of fit of the linear regression is at a maximum value when data up until t = 5 hours after inoculation is included in the statistical analysis. However, for all datasets with a final time point from 2 up until 7 hours after inoculation (the complete incubation period), R² values above 0.93 are obtained. This indicates that across the OD₆₀₀ range studied, log₁₀(cells/ml) and thus cells/ml can be interpolated from OD₆₀₀ measurements with a reasonable degree of accuracy.

As indicated in figure 8 below, the goodness of fit for a linear correlation is clearly superior for the up to t=5 hours dataset than the up to t=7 dataset. However, the latter dataset covers OD₆₀₀ values up to approximately 0.9, versus 0.5 for the t=5 dataset, significantly increasing the range over which cells/ml can be interpolated from the standard curve, rather than being extrapolated. Therefore, when looking to estimate cells/ml from OD₆₀₀ using the standard curves obtained below, a trade-off must be struck between the superior fit of the correlation for the t=5 dataset, versus the practicality of having to produce samples with OD₆₀₀ values which lie within (or in close proximity to) the narrower range measured over.

To assess the degree of disparity in estimates of cells/ml from the 2 standard curves, cells/ml values were interpolated from both curves above for OD₆₀₀ values ranging from 0.05 to 0.9 and the percentage difference between them was calculated. As shown below in figure 9, across an OD₆₀₀ range from 0.05 to 0.2, the percentage difference in cells/ml values obtained using the different standard curves is sufficiently small (less than 10%) to have little effect for experimental applications which require only a low degree of accuracy (e.g. identifying the correct volume of cell culture on which to perform a GeneJet RNA extraction which requires approximately 1 x 10⁹ cells).

However, for OD₆₀₀ values greater than 0.2, percentage differences up to -42% are yielded. This indicates that selecting the appropriate standard curve for estimating cells/ml is relatively important, and depends on whether overall goodness of fit or a desire to avoid extrapolation is prioritised. As is indicated below (with data points for which cells/ml is extrapolated rather than interpolated circled in blue for the t=5 standard curve, and orange for the t=7 curve), a greater degree of extrapolation is required when using the narrower t=5 dataset, so the t=7 standard curve may be more appropriate for analysis of higher OD₆₀₀ datasets.

Overall, the results obtained indicate that OD₆₀₀ can provide a good basis for estimating cells/ml of a sample, with standard curves indicating a strong direct proportionality between them for OD₆₀₀ values ranging from 0.05 to 0.85. Using the data gathered, as well as the equations for the linear correlations of the 2 standard curves plotted, future iGEM teams working with BL21 DE3 and its derivatives will be able to rapidly estimate cells/ml for samples from easily obtained OD₆₀₀ values recorded using the Tecan M PLEX plate reader (specifically using black clear bottom Corning 96 well plates without the lid on). This will provide a rapid and cheap way of estimating cells/ml for a variety of wet lab applications (e.g. normalising uptake of a substrate to assess uptake per cell).

This method is preferable to microsphere based calibration because it relies on real-world image flow cytometry data of E. coli cells rather than silica beads approximating the optical properties of these cells. By interpolating standard curves of OD₆₀₀ against cells/ml data, extraneous factors which reduce the validity of using OD₆₀₀ itself as a measure of cells/ml, such as the contribution of aggregated cell debris, are accounted for. This method is also a useful alternative to image flow cytometry itself, given the expense and lack of availability of the equipment, in addition to the expertise required to operate it. We hope future iGEM teams, with access to the necessary equipment, will expand the applicability of estimating cells/ml by interpolating standard curves of OD₆₀₀ against image flow cytometry data by producing similar standard curves using other strains of E. coli and species of model organism (e.g. Saccharomyces cerevisiae), as well as recording OD₆₀₀ using other 96 well plates and plate readers, and recording corresponding cells/ml data using other flow cytometers.

Fibroin-Graphene Films

Producing SF-rGO bionanocomposite films from silkworm cocoons

The majority of our preliminary experiments succeeded in forming a graphene and fibroin film, however they often remained clumpy, making it difficult to assess their conductive properties, particularly using methodologies such as the four point probe. Literature suggests this coagulation is due to sericin binding proteins that ‘glue’ the protein fibres together ^[1]. The silkworm cocoons were boiled in Na₂CO₃ in a degumming process to remove the sericin binding proteins according to previous literature ^[2]. Softening of the structure occurred but an evenly dispersed protein solution was not obtained. For these reasons, we did not obtain any data from these early material prototypes. This led us to use silk fibroin solution instead, as the protein is already pre-dispersed.

Producing SF-rGO bionanocomposite films from fibroin solution

Scanning electron microscopy

The surface morphology of rGO/fibroin composite was observed under a scanning electron microscope (SEM) (Zeiss Gemini 500) with an acceleration voltage of 1 kV. Before the observation, the surface of the samples were coated with a sputter coater equipped with a gold target to increase electrical conductivity of the samples.

From these images it can be observed that an even distribution of graphene and fibroin is seen in the composite, as there is no visible segregation of material. This evidence shows that the method we used to homogenise our biocomposite films was successful. However, the images also show a slightly uneven surface. Further analysis would confirm the cause of this, but we speculate it could be from coagulation of protein on a microscopic scale. Potential effects of this on the composite’s biocompatibility are unknown.

Conductivity

Films produced from fibroin solution were visually smoother, enabling more reliable conductivity analysis. Resistance was measured using a multimeter, giving conductivity values of up to 33-45 Scm^-1. Silk fibroin alone is not conductive ^[3], so this increase is promising. Furthermore, this falls into the semiconductor range (between 10^-8-10²Scm ^-1)^[4].

In comparison to polymer scaffolds produced by previous literature, the conductivity is high. Ideal conductivity to mimic nerve conduction is around 5.6 mS cm^-1

Raman spectroscopy

These films were later analysed using Raman spectroscopy with the aim to give information on the structure and insight of the binding of graphene and fibroin - potentially highlighting any changes between their isolated and biocomposite forms.

Raman spectra graphs obtained are detailed below. The graph uses the average of our collected data after background subtraction. Raman spectra can suffer from fluorescence interference which can be orders of magnitude more than the Raman peaks, leading to a broad background under the Raman peaks. Distorted background results can lead to incorrect integration or peak height determination, which are central to many spectral applications. Hence it is necessary to separate the Raman peaks from the background ^[5].

Conclusions and discussion from Raman data

The peaks for the rGO Raman spectra at 1572cm^-1 and 1342cm^-1 correspond to the characteristic G and D bands of graphene respectively. The G band is indicative of sp²-hybridised carbon-carbon bonds in a hexagonal lattice and the D band corresponds to disorder in the graphene structure ^[6].

The Raman spectra from the rGO/SF biocomposite retains the G and D bands observed in the rGO spectra, demonstrating the presence of rGO in the composite. However, in the prescence of fibroin there is a noticeable increase in the relative intensity between the two peaks. The relative increase in absorbance intensity of the G peak may be attributed to the binding of SF with the graphene, although individual fibroin peaks will be difficult to identify in the composite ^[6]. Alternatively, this intensity increase could be due to potential layering of the rGO ^[7]. The peak in the blank rGO spectra at ~1100cm^-1 may be due to the presence of nanocrystalline sp³ carbons in distortions of the rGO; the disappearance of this peak in the composite spectra may suggest even dispersion of the graphene during the formation of the composite ^[8].

Details of the system and parameters for Raman analysis are as follows: Raman spectra were collected using a confocal Raman spectrometer (Alpha 300R, WITec GmbH), equipped with a UHTS 300 VIS spectrometer and a thermoelectrically cooled CCD detector (down to -62 °C), with a 600 g mm^-1 grating. For excitation, a 532-nm Nd:YAG laser was focused onto the sample using a Zeiss 50 x objective (NA 0.7), with a maximum power of 15mW at the sample. Between 5 - 10 single Raman spectra were recorded for each sample, each with an exposure time of 1 second and twenty accumulations. WITec Project 4.0 software was used to average the spectra and remove fluorescent background using the 'shape function'.

References

• [1] Kunz RI, Brancalhão RM, Ribeiro LD, Natali MR. Silkworm sericin: Properties and biomedical applications. BioMed research international. 2016 Nov 14;2016.
• [2] Liang B, Fang L, Hu Y, Yang G, Zhu Q, Ye X. Fabrication and application of flexible graphene silk composite film electrodes decorated with spiky Pt nanospheres. Nanoscale. 2014;6(8):4264-74.
• [3] Dong X, Liu Q, Liu S, Wu R, Ma L. Silk Fibroin Based Conductive Film for Multifunctional Sensing and Energy Harvesting. Advanced Fiber Materials. 2022 Mar 23:1-9.
• [4] Fielding LA, Hillier JK, Burchell MJ, Armes SP. Space science applications for conducting polymer particles: synthetic mimics for cosmic dust and micrometeorites. Chemical Communications. 2015;51(95):16886-99.
• [5] H Ruan, LK Dai. Automated Background Subtraction Algorithm for Raman Spectra Based on Iterative Weighted Least Squares. Asian Journal of Chemistry. 2011 Aug 10;23(12): 5229-5234.
• [6] Liang B., Fang L., Hu Y., Yang G., Zhu Q., Ye X. Fabrication and application of flexible graphene silk composite film electrodes decorated with spiky Pt nanospheres. Nanoscale (6), March 2014.
• [7] Wall M., The Raman Spectroscopy of Graphene and the Determination of Layer Thickness, Thermo Fisher Scientific, 2011
• [8] Paillard B., On the origin of the 1100cm^-1 Raman band in amorphous and nanocrystalline sp3 carbon. EPL, 2007 Jan