Summary
As a team in the Foundational Advance track developing a new chassis, Mycobacterium smegmatis, and investigaing a relatively unexplored area of synthetic biology, namely soil synthetic biology, measurement became a central and critically important piece of every aspect of our project.
Measurement in soil synthetic biology experiments, compared to laboratory experiments in a flask or tube, faces many challenges in accuracy and quantitation. First, bacteria tend to attach to soil particles, and the opacity and autofluorescence of these soil particles interferes with fluorescence-based methods, such as spectrometry and fluorescence microscopy. Separation procedures are often unable to capture the whole bacterial community, and are biased towards specific taxa and metabolic states. Thus, our project focused on conducting multiple measurement approaches for each variable that was examined. In particular, after determining that microscopy has the potential to provide many advantages for monitoring engineered constructs in fieldable soil synthetic biology applications, we extended iGEM’s plate reader and flow cytometer calibration protocols to microscopy, and wrote a guide to quantitative microscopy and a guide to soil microscopy for future iGEM teams to use.
Our contributions and achievements in Measurement include the following:
New Chassis Development: Foundational Measurement Contributions
Soil SynBio-Specific Measurement: a Multimodal Approach for Analyzing Engineered Constructs in Soil Microcosms
For detailed information on each of these contributions, click on the links above or scroll down to paragraphs below.
Mycobacterium smegmatis OD to CFU Standard Curve
As part of W&M’s 2023 iGEM project, we developed Mycobacterium smegmatis mc2155 as a novel synthetic biology chassis. M. smegmatis is both genetically tractable and native to soil environments, making it ideally suited for soil bioengineering applications. However, in order for M. smegmatis to be a fully-fledged chassis, researchers need a detailed understanding of the organism’s metabolism, which includes accurate growth curves. We developed our own standard curve to facilitate accurate comparison between optical density values and cell counts. This serves both to develop necessary details about M. smegmatis for its use as a chassis across soil synthetic biology, and to improve the precision of our soil microcosm experiments, allowing us to report approximations of inoculant cell counts that were added to the microcosms.
Microbial growth curves enable researchers to characterize the normal metabolic activity of different bacteria and measure the effect of variables such as media type and antibiotic concentration on cellular growth and metabolism (Tonner et al., 2017). Moreover, by performing optical density (OD) and standard plate count measurements on a given organism, researchers are able to plot standard curves relating the OD of the cultured bacteria to a given cell concentration. Such simple procedures are critical first steps towards the development of novel chassis for soil synthetic biology (de Lorenzo et al., 2021).
Mycobacterium smegmatis is a non-pathogenic soil bacterium often utilized as a model organism for Mycobacterium tuberculosis (Sparks et al., 2023). Distinguished by its fast growth in laboratory conditions and abundance of tools for genetic engineering, M. smegmatis is an extremely promising candidate for soil synthetic biology applications. It possesses a well-characterized and annotated genome, a suite of tools for genetic engineering, an abundance of bacteriophages, and a thorough description of its transcriptome and biochemistry (Bardarov et al., 2002; Deshayes et al., 2007; Gallien et al., 2009; Huff et al., 2010; Mohan et al., 2015; Li et al., 2017; Murphy et al. 2018; Lammens et al., 2020; Sundarsingh T et al., 2020; Armianinova et al., 2022). Further, it is a native soil bacterium, harboring strategies to survive hypoxic environments and avoid antimicrobials released by other soil organisms. Research exploring M. smegmatis’s native capabilities have shown its ability to uptake toxins like carbon monoxide, metabolize hydrogen gas, and produce a range of enzymes resisting oxidative stress (Sparks et al., 2023).
However, juxtaposed with the understanding of its structure and metabolism is the considerable lack of consistency seen in growth curves of M. smegmatis in the scientific literature. Some of this is, of course, due to differences in conditions, such as nutrient availability and initial cell concentrations across experiments. However, whether comparing nutrient broth or minimal media, significant disparities remain. More information on these literature discrepancies can be found below.
Given the wide range of relationships observed in the literature between OD values and cell counts, we conducted our own experiments, monitoring M. smegmatis mc2155 growth over time using both OD600 to measure turbidity and standard plate counting. This allows us to (A) interpret data from other experiments using OD600, as well as (B) construct our soil microcosm experiments with precise inoculation of bacteria.
1. Obtain 7 tubes for each culture (a total of 21 tubes), and arrange them in groups of 7 on your tube rack. Label tubes "unt 1", "unt 2", or "unt 3" at the top, including numbers 0-6 (0 is for the aliquot that is not diluted, 1 for the first dilution, etc.). Repeat for other bacteria being assayed if relevant. This is mainly personal preference, so do what you prefer to stay organized and know which tube is which.
2. Fill all tubes except the aliquot tubes (labeled "0") with 900 uL of 1X PBS.
3. Make sure you have a clean 96-well plate and lid. Know beforehand which columns you will be adding the samples, making sure to pick columns where the plate and the lid are clean.
4. Once ready, carefully grab each bacterial culture and arrange them in the hood from left to right, with culture 1 on the left and culture 3 on the right. This is purely for organization.
5. Aliquot 800 uL from each culture into the proper tubes labeled "unt# 0" (Here # can be either 1, 2, or 3 for the culture number).
6. Return the cultures to the shaking incubator.
7. Grab the media blanks that are in the no-bacteria 4C fridge in a yellow tube rack, near the middle section of the fridge.
8. Pipette samples into the plate, doing only one well for each media blank. Do three per culture in the columns of your choice.
9. Read the plate at three separate OD wavelengths using the microplate reader: 595 nm, 600 nm, and 640 nm.
10. Once done, begin the dilutions. Perform six 10-fold dilutions for each culture, transferring 100 uL from the left tube into the right one, making sure to mix the dilution after pipetting by turning the tube upside down seven times. Switch pipette tips between each dilution as well.
11. Label 7H9-agar plates with “7H9”, the sample, the measurement hour, the dilution, your initials, and the date.
12. Using 7H9 plates, plate the appropriate dilutions determined by preliminary growth experiments. First add 5-10 sterile beads to the plate, and add 100 uL of the sample to the plate. Swirl the plate back and forth without making a circular motion.
13. Once you are finished, place the plates in the incubator at 37 C and make sure to save the plate reader data to a spreadsheet. Name the file Date_MsmegGrowth_MeasurementHour_Initials.
14. Bleach the plate you used, and rinse/place to dry any other plates that have already been bleached. This makes sure you have a good plate for the next measurement.
M. smegmatis untransformed and pMLCherry genome-integrated showed higher maximum specific growth rates than pCherry. Triplicate cultures of M. smegmatis untransformed demonstrated a maximum specific growth rate of 0.178 h-1 and a doubling time of 3.88 hours, both of which agree with the range in the literature for each parameter (0.1-0.31 h-1 and three to four hours, respectively).
With this data, we were able to accurately report the concentration of the bacterial suspensions which we inoculated our soil microcosms with simply by measuring the OD600 of the cultures and relating it using the linear fit equation. This enhances the replicability of our research by letting investigators add similar cell titers to future microcosm experiments, enabling comparisons between their results and ours. Moreover, we add to the literature on the kinetics of M. smegmatis growth in laboratory conditions. This basic work was in response to the profound variation of standard curve relationships seen in the literature. While we could have utilized these values, measuring the characteristics in-house leads to the highest accuracy for reporting our cell count values. Still, given our emphasis on rigorous measurement this year, we are aware of the significant drawbacks of standard plate counting for cell count estimation, namely its high level of variability, concerns with clumping, and underestimation of total cells in the sample (Beal et al., 2020). Future characterizations of M. smegmatis cell counts in relation to optical density should use indirect methods as well, potentially using silica microspheres the size of M. smegmatis (Beal et al., 2020). However, the standard plate count method is particularly advantageous for developing a chassis given its sensitivity to non-viable cells, yielding approximate counts for metabolically active cells (Beal et al., 2020). As such, we hope this data, along with several other aspects of W&M’s 2023 iGEM project, can spread the use of M. smegmatis throughout synthetic biology. A detailed explanation of results can be found on the results page.Mycobacterium smegmatis Promoter Library
Currently, bioengineers and synthetic biologists focus their research on only a few well-understood bacterial species, namely, E. coli and B. subtilis. However, these species only thrive in a few very specific environments. Thus, there is a huge need to develop additional model systems if we want to push synthetic biology outside of the laboratory (Adams 2016). Mycobacterium smegmatis is a non-pathogenic species of the Mycobacterium genus (members include M. tuberculosis and M. leprae). Fast-growing and native to soil, M. smegmatis has the potential to be a successful host for the emerging field of soil synthetic biology.
Nine promoters were characterized using the methods mentioned above. In order to characterize and quantify promoters in a precise and reproducible way for SynBio application, more scrutiny than merely classifying promoters into “weak” and “strong” is needed. By adopting calibrated fluorescence units (Beal 2019, 2022), constructing a dual-channel reporter system (Rudge 2015), and measuring relative promoter strengths to a reference promoter (Kelly 2009, Guiziou 2016), we characterized these nine promoters, spanning a range of strengths from strong to weak, in a quantitative and reproducible manner. We hope this promoter library could lay the groundwork for utilizing M. smegmatis as a fieldable chassis and a preceding guide for the synthetic biology community on how to develop and characterize novel chassis in the future.
Potential promoters were extracted from M.smegmatis transcriptome profiling data from two published studies (Li 2017, Martini 2019). We identified consensus TSS predictions between these two studies, and we constructed a DataFrame with the complete genome sequence and annotation from NCBI and potential promoters identified from this analysis. In most circumstances, the 50 bp upstream of these TSS were collected as a potential promoter sequence. The strength of these promoters is estimated according to their average log-phase RPKM. Candidates were selected to cover a wide range of strengths based on their corresponding average log phase RPKM (Reads Per Kilobase per Million mapped reads) values.
For measurements using an indirect reporter (fluorescence proteins), it is inevitable that the outputs will be impacted by global variation, such as metabolic burden introduced by the plasmid, slight differences in growth conditions, and the varying initial state of the inoculated cell. Measurements using a single reporter would require these extrinsic factors to be taken into consideration, but collecting these variables could be difficult and even unrealistic. Therefore, we opted to utilize this ratiometric characteristic, where the control promoter acts as a concurrent reference of the effect of these extrinsic factors on the overall gene expression pattern (Rudge 2015); this approach has been demonstrated in studies to be more reliable than single-channel reporter systems.We chose to use a bi-fluorescent construct that contains two transcriptional units in the opposite direction. One transcription unit serves as a control, where psmyc (a characterized M. smegmatis strong constitutive promoter) is located upstream of mCherry (red fluorescence). The other transcription units serve as the test unit, where each uncharacterized promoter will be inserted upstream of sfGFP (green fluorescence).
Figure 1. Golden Gate Assembly was used to build each promoter construct. pSUM8 (backbone) consists of 2 BsaI cut-site flanking a placeholder upstream of the RBS sequence and mCherry coding region. For each Golden Gate reaction, both the test promoter sequence (previously cloned into another vector) and the placeholder were excised out. The promoter sequence with the compatible sticky ends eventually gets ligated into pSUM8 to complete the backbone to pSUM8.XX.
We used the following constructs, which we engineered:
Table 1. Mycobacterium smegmatis promoter library.
After construction, these circuits were sequence confirmed and then transformed into Mycobacterium smegmatis mc2155. For detailed protocol, see Experiments page.
Figure 2. Fluorescence normalization equation.
The fluorescence of the constructs at each time point was normalized to the corresponding negative controls’ fluorescence and the media’s OD600. We chose to select the green fluorescence profile of untransformed M. smegmatis as the background green fluorescence for the control transcription unit, and we used the red fluorescence profile of M. smegmatis transformed with single-promoter backbone pSUM8 as the negative control for the red fluorescence measurements of the test transcription unit, since the backbone pSUM8’s red fluorescence potentially inflates the test promoters’ activity measurements. Raw Fluorescence measurements are normalized using the following steps:
1. Locally estimated scatterplot smoothing (LOESS) models were built using the fluorescence measurements over time of the corresponding negative control and were used to predict background fluorescence at different time points.
2. The distribution of background OD600 for both SESOM(soil-extracted solubilized organic and inorganic matter(soil-extracted solubilized organic and inorganic matter)+7H9 and 7H9 is not symmetrical, thus the medians of all blank media OD600 were selected as the universal background OD600 for normalization.
3. The fluorescence of the constructs at each time point was normalized to the negative controls’ fluorescence predicted by the corresponding LOESS model.
4. The OD600 was normalized using the blank OD600 median.
5. The normalized OD600 were converted to CFU using the M.smegmatis-specific standard curve we built.
6. The normalized fluorescence output were than normalized with the corresponding CFU.
7. All fluorescence measurements were then converted to either MEFL(molecules equivalent of fluorescein)/CFU (colony forming units) or MESR(Molecules of Equivalent Sulforhodamine 101)/CFU (Beal 2022).
Figure 3. OD and Raw Fluorescence readings results. LOESS models constructed for Fluorescence Normalization. Connected Black dots represent fluorescence measurements of the negative controls. n=3. Green/Red dots represent the predicted outputs of the corresponding LOESS model.
The ratiometric characteristics were calculated by taking the ratio of normalized red fluorescence/normalized green fluorescence.
Fig 4. Rationmetric fluorescence calculation.
Where:
Gt = sfGFP fluorescence measurement of the transformed (engineered) culture
Gu = sfGFP fluorescence measurement of the untransformed culture predicted by a LOESS Model constructed using UNT green fluorescence over time
Dt = optical density measurement of the transformed (engineered) culture
Du = optical density measurement of blank media
Rt = mCherry fluorescence measurement of the transformed (engineered) culture
Ru = mCherry fluorescence measurement of the pSUM8 backbone culture predicted by a LOESS Model constructed using pSUM8 red fluorescence over time
Nine promoters were characterized using the methods mentioned above. These results can be found more extensively on our Parts page.
Figure 5. Promoter Strength of nine promoters relative to the reference promoter psmyc. n=3. Error bars show ±1 standard deviation from the mean for each group.
In order to characterize and quantify promoters in a precise and reproducible way for SynBio application, more scrutiny than merely classifying promoters into “weak” and “strong” is needed. By adopting calibrated fluorescence units (Beal 2019, 2022), constructing a dual-channel reporter system (Rudge 2015), and measuring relative promoter strengths to a reference promoter (Kelly 2009, Guiziou 2016), we characterized these nine promoters, four of them being previously-uncharacterized, in a quantitative and reproducible manner. We hope this promoter library could lay the groundwork for utilizing M. smegmatis as a fieldable chassis and a preceding guide for the synthetic biology community on how to develop and characterize novel chassis in the future.
Fluorescence Microscopy Calibration & Standardization
iGEM’s InterLab study has aimed to encourage and evaluate the use of an absolute unit, specifically Molecules of Equivalent Fluorescein (MEFL) per cell, for fluorescence readings among iGEM teams (Beal et al., 2022; Beal et al., 2021; Beal et al., 2018, Beal et al., 2016). In this part of our project, we aimed to extend this same calibration method to fluorescence microscopy. Similar to the concept of plate reader and flow cytometry calibration, imaging the intensity of a slide with a known concentration of a fluorescent dye along with your sample would allow you to convert microscopy fluorescence intensity data to MEFL (molecules of equivalent fluorophore) units (Kedziora et al., 2011).
Due to the many components of fluorescent microscopy that introduce error and noise, quantitation and standardization in fluorescent microscopy has proved difficult compared to other fluorescence-based instruments. Ideally, a standard could account and correct for differences between systems and day to day, as well as make data communicable between labs. Standards also allow for samples with large differences in fluorescence intensity, requiring differing acquisition parameters, to be normalized to the same unit and compared (Souchier et al., 2004).
Fluorescent microspheres, or “beads”, and fluorescent plastic slides are common internal microscopy standards (Sanjani et al., 2017). Beads and fluorescent slides can be imaged day to day to validate proper set-up, operation, and system performance (Poon & Lansdorp, 2001; Koren et al., 1990). For example, microspheres of differing intensities allow for the evaluation of the linearity of a microscope (Swedlow, 2013; Bour-Dill et al., 2000). Fluorescent slides of uniform illumination can be used to correct for inhomogeneous illumination, referred to as flat-field or shading correction (Souchier et al., 2004; Webb & Brown, 2013; Stack et al., 2011; Model & Burkhardt et al., 2001). Slides of distributed fluorescent microspheres can also be used for flat-field correction (Poon & Lansdorp, 2001). Fluorescent microspheres that emit in multiple wavelengths can be used for channel registration/chromatic shift correction (Bour-Dill et al., 2000; Stack et al., 2011; Jost & Waters, 2019).
Additionally, since fluorescent microspheres intended for calibration have consistent fluorescent intensities according to the manufacturer, they have been used for sample fluorescence intensity normalization (Cunha-Reis et al., 2011; Richani et al., 2014; Suidiman et al., 2014; Ziolkowski et al., 2017; Poon & Lansdorp, 2001; Bezakova et al., 2001). However, they are more commonly used for characterizing new imaging or analysis methods (Kim & Naemura, 2015; Barlow & Guerin, 2007; Lee et al., 2014), as opposed to use as “everyday” standards, likely due to their high cost.
There have been many research efforts to not only normalize to an absolute unit, but to determine in situ protein counts from the use of standards in fluorescent microscopy. These calibrants contain a known amounts of molecules or proteins, such as MotB, an E. coli protein with 22 molecules per motor (Leake et al., 2006), viral particles with 120 EGFPs per assembled capsid (Charpilienne et al., 2001), or yeast cells expressing known amounts of Cse4p-GFP (Joglekar et al., 2008).
Unfortunately, both fluorescent microspheres and plastic slides designed for microscopy are often quantified by the manufacturer in relative units, instead of absolute units such as MEFL. While there are fluorescent microspheres available that are quantified in MEFL units, they are typically designed for flow cytometry use, and none have been validated for microscopy use. Additionally, beads are very expensive and are not feasible for use along with every image acquisition, especially in high-throughput imaging, which is necessary in many soil microscopy applications.
Slides of concentrated or diluted fluorophore solutions provide an alternative to plastic fluorescent slides and microspheres. The use of fluorophores allows the researcher to normalize fluorescence intensity measurements to absolute MEFL units. Model & Burkhardt et al., 2001, demonstrated the ability for concentrated fluorescent dye solutions to be used for calibration and shading correction in fluorescence microscopy, with sodium fluorescein and rhodamine B solutions (Model & Burkhardt et al., 2001). Standard slides have also been prepared by embedding the fluorescein dye in a uniform polymer film, polyvinyl alcohol, and has been shown to have high reproducibility and ability to perform shading correction (Zweir et al., 2004; Kedziora et al., 2011).
Slides composed of fluorescent dyes are able to perform shading correction, normalization, and validation of consistent system performance (Model & Burkhardt et al., 2001; Zweir et al., 2004). By normalizing to absolute units, microscopic calibration with dye solutions allows for interinstrument and InterLab comparison, due the flexibility and accessibility of dyes. Researchers are also able to use a specific fluorophore and concentrations to match the acquisition parameters you use for sample data collection.
An additional problem is the lack of clarity in published methods, as many papers include that they calibrated their microscope with a certain calibration tool (fluorescent microspheres, stained fluorescent slides, etc.), but do not elaborate on what specifically they are calibrating or how (Ichihara, 2019; Gardner et al. 2000; Moll et al., 2001). This poses a challenge to undergraduate iGEM teams, as new researchers often do not know what may be “obvious” to experts in the field.
A Nikon Eclipse Ti2 inverted microscope with the Nikon N Plan Apo 100X/1.45 oil, infinity to 0.17 WD 0.13. The sCMOS pco.edge camera was used, with no gain or binning. All images were acquired at an exposure time of 100ms, a DIA Iris Intensity of 23.0, and a Sola Illuminator Voltage of 15.0, unless otherwise stated. All images were taken at a magnification of 100X with the Nikon N Plan Apo 100X/1.45 oil, infinity to 0.17 WD 0.13.objective. Fluorescent dichroic mirror and filter cube wavelengths are as follows: DAPI mirror is >=400nm. Excitation is from 340 to 380 nm. Emission is from 435 to 485 nm. FITC mirror is >=505 nm. Excitation is 465 to 495 nm. Emission is 515 to 555 nm. TRITC mirror is >=565 nm. Excitation is from 528 to 553 nm. Emission is from 590 to 650 nm. Cy5 mirror is >= 625nm. Excitation is from 625 to 650 nm. Emission is peak at 670 nm. The Sola Light Engine from Lumencor was used for fluorescence illumination, while diascopic illumination was from the Nikon LED lamphouse with high-powered LED, Ti2-D-LHLED. A motorized stage was used. NIS-Elements AR 5.0201 64-bit, ImageJ 1.54d, and Google Sheets, was used for all analysis.
Images were collected over multiple days, and at varying concentrations. This experiment aimed to mimic real usage of standards, so new solutions were created when old solutions ran out, at varying time points. Sulforhodamine 101 powder was dissolved in 1X PBS to varying final concentrations, which were validated by triplicate plate reader measurements that were fit to the plate reader calibration standard curve, following the InterLab study protocol. Solutions were stored at 4C, protected from light.
Images were collected in a 3 X 3 grid, with points separated by 3 mm. Slides were always focused with transmitted light in the top left corner, in order to account for potential photobleaching during focusing. Images were collected using the NIS elements “Capture Multipoint Automatically” function, with active shutter closed during stage movement. Fisherbrand Frosted Microscope Slides, catalog No. 12-550-343 slides were used. 22 X 22 mm cover glass slips from Carolina Biological or 25mm X 50 mm Fisherfinest Premium Superslip coverslips. If larger coverslips were used, then 18.5 ul of solution was imaged. If 18 * 18 mm coverslips were used, then 5 ul of solution was imaged. For this experiment, slides were imaged in the TRITC channel.
Reproducibility
Reproducibility was calculated among all images for each slide, and among all images taken across the experiment. Reproducibility was defined as (1 - Standard deviation/mean)*100% in terms of average image intensity. Inhomogeneous illumination, or shading, was not corrected for in the evaluation of the reproducibility of dyed slides, since the average intensity across the whole image should not differ between images due to inhomogeneous illumination.
The average reproducibility of multiple images of the same Sulforhodamine 101 dye slide, or multiple Inspeck beads on the same slide, was calculated. For the Inspeck beads, average reproducibility was found to be 77.59%, while the average reproducibility of the Sulforhodamine 101 slides was found to be 54.43%. This indicates that within one slide, multiple images will result in differing intensities, for both images of inspeck beads and Sulforhodamine 101 solution. The variability of the Sulforhodamine 101 slides is greater than that of the Inspeck beads, but it is in a reasonable range, given that the Inspeck beads are considered to be uniform and consistent.
Reproducibility as a function of solution concentration, normalized to Inspeck bead reproducibility, was plotted, and is shown below (Figure 6). A linear fit was performed, and it was determined that there was limited correlation between concentration and reproducibility.
Figure 6. Solution concentration of Sulforhodamine 101 standard slides., as determined by the plate reader standard curve, vs. normalized reproducibility.
Time was evaluated as an influencing factor on reproducibility of a set of images for one slide, and a linear fit was performed to determine this. It was concluded that there was not a strong correlation between date and reproducibility (See fig. 2).
Figure 7. Time vs normalized reproducibility for Sulforhodamine 101 standard slides.
Reproducibility for images of Inspeck beads was determined after flat-field correction, as beads did not cover the slide uniformly. After flat-field correction based on Ted Pella., Inc. Fluorescence Reference slides, the reproducibility of beads are found below. Flat field correction was performed in ImageJ, by averaging the sample and reference images, and then dividing the sample image by the reference image, and scaling by a factor to maximize significant figures. After determining that Inspeck beads were resistant to photobleaching, field of views (FOVs) were refocused before every image capture, in order to reduce effects of poisson noise, and the known variance in fluorescent intensity as a function of focus depth for dyed fluorescent microspheres.
Flat field correctionSince Ted Pella Inc. fluorescent reference slides had very consistent distributions and average intensities over 30 images in different locations on the slide, this slide was used as a reference for flat field correction. If the dye slides can accurately perform flat field correction, then the distribution of pixel intensities should match that of the fluorescent reference slides.
The value of each pixel in the resulting image should be exactly the same. Deviations from this were measured by the average standard deviation of the image. This is compared to the standard deviation between inspeck bead intensities after being corrected in the same manner (by the fluorescent reference slide). Due to the strong intensity of this slide, acquisition parameters had to be changed to a 1.5 ms exposure time, and a Sola Illuminator Voltage of 5.0.
This was evaluated by dividing the pixel intensities of an average image of 9 images of one dye slide by the average intensity of all 30 images of the fluorescent reference slide. Averaging was performed in ImageJ by Image->Stacks->Z Project with “Projection Type” as “Average Intensity.” Image division was performed with the Calculator Plus Pug-In in ImageJ, with the Division operation. I1 was the image being corrected, and i2 was the reference image. K1 was set to 10,000.0 in order for data to fall within the middle pixel intensity of the 16-bit image outputs. For dyed slide images, the intensity mean and standard deviation values of the resulting image were obtained by the Analyze->Measure function in ImageJ. For bead images, resulting TIF files were exported and opened in NIS Elements as a multipoint file. Bead ROIS were selected with the Auto Detect ROI Tool, and intensity values were obtained with the “Automated Measurement Results” window. The results of flat field correction are outlined in the table below.
Figure 8. Flat-field correction results of Sulforhodamine 101 standard slides and red InSpeck beads, with Ted Pella Inc. red fluorescent reference slide used as reference.
This is somewhat expected, as ThermoFisher Scientific claims for the beads to be within a certain range. That is, between 0.62% – 1.6% for the 1% beads, and between 1.9% – 4.8% for the 3% beads, relative to the 100% beads (ThermoFisher Scientific, 2022). This result also validates the use of fluorescent dye standard slides as microscopy standards, as they produce a relatively uniform image, compared to inspeck beads, after flat-field correction, indicating their ability to perform flat-field correction on sample images.
The correlation between concentration of dye slides and resulting fluorescent intensity from images was evaluated. Ideally, there would be a linear relationship, leveling off at saturation. However, the correlation between the two variables found was weak, with an R2 of 0.451, as shown in Figure 4.
Figure 9. RFU vs concentration for Sulforhodamine 101 standard slides.
A residual plot was produced, in order to determine whether a pattern existed in the distance from the expected intensity value (determined in the linear fit in figure 4) based on concentration. This chart is shown in Fig. 5. The R2 of 0 and visual inspection of distribution of points about zero suggest that a linear regression was the correct fit for this data, and that there was no pattern in deviation from the linear fit based on concentration.
Figure 10. Residual plot of RFU vs Concentration of Sulforhodamine 101 slides (Figure 9.).
Our team has hypothesized many potential causes for the low linear fit between concentration and fluorescent intensity value measured from our imaging system. One potential source of error could be pipetting error, due to the precision necessary to make perfect concentrations. Another explanation is an inhomogeneous distribution of Sulforhodamine 101 particles on the slide, which has been suggested to occur at low concentrations. This is also supported by our low reproducibility values across images from the same slide. Additionally, Sulforhodamine 101 particles may not be fully dissolved in the 1X PBS solution, which is supported by our low reproducibility in fluorescence intensity across slides that should be at the same concentration.
These results indicate that Sulforhodamine 101 fluorescent dye standard slides have the potential to act as an MEFL standard for fluorescence microscopy, but more work must be done to improve the reproducibility of fluorescence intensity measurements across different slides of the same concentration.
Direct Microscopic Enumeration and Fluorescent Intensity Quantitation
Microscopy was chosen as an important means of measurement for this project due to the variety of advantages that it holds over other instruments when working with soil. These include the ability for direct observation, which allows for more accurate estimations of bacterial abundance compared to traditional culture methods. Direct observation also limits the need for bacterial extraction procedures to be performed on soil samples, which disturbs native metabolic states and results in biased bacterial estimations.
Additionally, because we had many bacterial constructs constitutively expressing fluorescent proteins, we were curious as to whether or not we could recognize our bacteria in the soil through direct observation, without staining or labeling. This would allow researchers to monitor their construct in real world environments, in real time, and with minimal preprocessing and equipment. Active metabolic staining, with CTC and DAPI, was also performed, as well as characterizing the fluorescent intensity of engineered bacterial cells.
In preliminary tests of the ability for our fluorescent microscope to detect bacteria in soil, positive controls were imaged. M. smegmatis pCherry3 culture was known to have high fluorescent levels, through high plate reader measurements, normalized to OD. To prepare test slides, turbid M. smegmatis pCherry3 culture was resuspended in PBS or NFW, and was added at 1 ml per gram to sterile soil, leaving a final concentration of ~108 CFU/g soil. This solution was diluted in 1mL of PBS. Solution was not vortexed, in order to not lyse freshly added cells. Slides were prepared as described on the "Experiments" page.
Analysis was generally split between enumeration and fluorescence intensity throughout all experiments. Segmentation was performed first in order to digitally select cells, and then the number of cells per image was recorded and the average fluorescence intensity per cell recorded.
Preliminary image analysis was performed. Initially, thresholding was attempted based on background subtraction and shape restrictions, without fluorescent intensity thresholding, in order to count untransformed bacteria in the same manner as transformed bacteria. Briefly, background subtraction was performed first, by selecting the largest possible ROI of a uniform, non-soil containing section of the image. The NIS elements “Segment Tight Borders'' function was then utilized, with a threshold level of 20. Next, counts were then obtained using the “Automated Measurement Results” window, and TRITC intensity was obtained by the average of mean intensities of all the counted selections. A statistically significant result was obtained when comparing bacterial counts between negative control and positive control slides, when a Mann-Whitney U test was performed.
For the bottle microcosm experiment, samples were prepared as described in our slide preparation protocol, which can be found on the Experiments page. Briefly, ~0.100 g of soil was diluted in 1mL PBS, and 5ul of this solution was allowed to dry on a coverslip, and then sealed. Only the following groups were imaged: Sterile Soil and M. smegmatis pCherry3, Sterile Soil and M. smegmatispMLcherry, Sterile Soil and M. smegmatisUntransformed, and Sterile Soil.
Initially, thresholding was attempted based on background subtraction and shape restrictions, without fluorescent intensity thresholding, in order to count untransformed bacteria in the same manner as transformed bacteria. Thresholding was determined based on preliminary images taken under the same conditions, in order to select for all M. smegmatis cells, including untransformed. Preliminary measurements indicated an average TRITC intensity for untransformed M. smegmatis cells is ~200 RFU, pMLcherry average TRTIC intensity was found to be ~1000 RFU, and pCherry3 average TRTIC intensity was found to be anywhere from ~20,000 RFU to 65,535 RFU. However, this thresholding method proved unsuccessful in giving accurate counts, as soil particle morphology and autofluorescence in both TRITC and FITC channels does not consistently differ from that of untransformed Mycobacterium smegmatis. This was determined by visual inspection of inaccuracies, as well as no statistical difference at alpha=0.05 found between counts obtained from images of groups containing bacteria, verified by bacterial plate counts, and negative control groups (sterile soil), when running a Mann-Whitney U test. This was also attempted with a threshold selection for pMLcherry bacteria, but no statistical significance between resulting sample counts and sterile soil counts was found. This was further compounded by the fact that negative control, or sterile soil, eventually was not sterile, as shown by bacterial CFUs after plating of the negative control. See results page for explanations as to why sterility was not maintained. The below images illustrate our challenges in determining bacterial counts from soil sample images. All images were captured and treated in the same way, with red intensity increased by the same amount in each image for easier visualization.
Figure 11. Binary results of differing thresholding methods on images from Bottle Microcosm experiment samples.
We believe our difficulty in distinguishing bacteria in these realistic samples is due to the low numbers of bacteria expected per image. With an initial inoculation bacterial concentration of 1E8 CFU/(g dry soil), and assuming our sample preparation protocol captures all bacteria in the soil sample, we would expect approximately 15 bacteria per image. Calculations are as follows:
Figure 12. Calculation of the number of bacteria expected to be imaged in one FOV.
Next, in order to be able to reliably count M. smegmatis pCherry3, a TRITC fluorescence intensity threshold of 3000 RFU to 65535 RFU was set in the NIS-Elements “Automated Measurement” window. Size restrictions from 0 um to 5um for TRANS, and 1 um to 5 um for TRITC were also set. Separate 4X in the TRITC channel was also applied. When comparing groups within a single time point, statistical significance between images containing bacteria, confirmed by plating results, and sterile soil images, was found from a Mann-Whitney U Test. Results of this analysis method on Sterile Soil and M. smegmatis pCherry3, Sterile Soil and M. smegmatis pMLcherry (integrated), Sterile Soil and M. smegmatis untransformed, and Sterile Soil control images are shown below:
Figure 13. Direct microscopy counts per image from Bottle Microcosm Experiment. Bacterial count per image results after red fluorescence thresholding designed to count Mycobacterium smegmatis pCherry3 cells.
Krustal-Wallis tests for each image for each group, at every time point separately, were run. For images with large artifacts, data was not collected, meaning every group did not have 9 images. For each Krustal-Wallis test, the minimum amount of images across groups was used, with data from later images not being redacted if necessary. The following results were obtained:
Figure 14: Table contains p-values from Krustal-Wallis test on direct microscopic counts on data across four groups at each time point, a = 0.05. Green indicates statistical significance, and blue indicates no statistical significance.
We then conducted Mann-Whitney U tests on groups of interest at each time point that resulted in significance in the Krustal-Wallis test. Results are as follows.
Figure 15. Table contains results of Mann-Whitney U Tests on groups of interest at each time point, after statistical signficance found at said time point in Krustal-Wallis test.
Only M. smegmatis in sterile soil, and sterile soil counts were found to be statistically different, and at every time point. This makes sense, based on our counting method, which was based on fluorescence intensity thresholding. We set the lower threshold higher than typical TRITC fluorescence intensity values of M. smegmatis pMLCherry (integrated) and M. smegmatis untransformed, but lower than typical TRITC fluorescence intensity values for M. smegmatis pCherry3. This statistical insignificance is representative of the limitation of the analysis method, not of biological results,since we know that M. smegmatis untransformed in sterile soil and M. smegmatis pCherry3 in sterile soil microcosms contained some level of bacteria, confirmed by plating.
One approach to improve the fieldability of soil synthetic biology that our team took was integrating our gene of interest into the bacterial genome, instead of having the gene contained within a plasmid. This improves the safety of releasing engineered bacteria into the environment, since the engineered genetic information cannot be transferred to the surrounding bacterial community via horizontal gene transfer.
However, integration of fluorescent protein coding genes produces a smaller number of fluorescent proteins in the membrane compared to fluorescent coding plasmids, due to much lower copy number than the plasmid. While integrated genes are only present in one genomic location, multiple plasmids, along with their genes, may be present in one cell. If the same promoter is used before the gene of interest in the plasmid and the genome, then this will result in more frequent transcription of the plasmid gene compared to genomic gene.This causes the total fluorescent output per cell to be lower, and potentially indistinguishable from background, which in this case is soil. On the other hand, a higher copy number plasmid adds more burden to a cell, potentially reducing cell fitness. In realistic environmental scenarios, bacteria transformed with a high copy number fluorescent plasmid, as well as a gene of interest, may not be able to express the gene of interest at necessary levels, or may lose fitness altogether and be unable to survive. Genomic integration significantly reduces burden on the cell, but also reduces fluorescent output levels. Realistic fluorescence values and realistic numbers of bacteria in the soil make bacteria difficult to distinguish via fluorescent microscopy without further labeling or other manipulation. We hypothesize that the reduced fluorescence of our integrated bacteria to our M. smegmatis pCherry3 plasmid bacteria played a large role in our difficulties distinguishing bacteria in soil sample images.
In order to reduce autofluorescence of soil particles in microscopy images, extraction utilizing the nycodenz density gradient was attempted on samples from the bottle microcosms. However, visible soil particles remained in the final step before plate reader reading, effectively making any readings unreliable. Additionally, red fluorescent intensity plate reader values of negative control samples were higher than for “spiked” positive control samples, invalidating data. Modifications to the procedure were evaluated, such as centrifugation at 20C and 4C, and both large and small volumes of soil samples.When performing nycodenz extraction on bottle microcosm 1 groups, sample plate reader values were lower than negative control values. Given that nycodenz has been reported to lose large percentages of cells, we hypothesize that our starting levels of cells left too little cells after nycodenz extraction. We were also concerned that separation and extraction procedures were biased and limited metabolic data extracted through fluorescent intensity. When images were taken after nycodenz extraction, remaining soil particles were seen.
In our team’s attempts to use nycodenz to isolate bacteria from soil, we were often unable to entirely remove soil from samples, which made plate reader results uninterpretable. When we were able to fully remove visible soil chunks from our samples, OD and fluorescence signals of spiked samples were often higher than positive control samples, suggesting contribution of remaining soil particles to measurements. This may have been due to the type of soil used, since differing temperatures, moisture content, and sand-silt-clay ratios has not been robustly evaluated as a factor in the efficiency of nycodenz separation.
Microscopy Guides
To support our aim of improving the fieldability of soil synthetic biology, we determined that microscopy provided various advantages in collecting data from soil samples, and that iGEM guidelines for the use of microscopy were necessary. We have written a Guide to Quantitative Microscopy as well as a Guide to Soil Microscopy, both aimed at iGEM teams.
Many challenges were encountered in our use of microscopy in soil synthetic biology experiments, but we hope that our results inform future researchers in realizing the full potential of microscopy for soil synthetic biology.
Direct Plating of Bacteria and Phage
Ratiometric Fluorescence of Mycobacterium smegmatis Promoters in Soil-like Conditions
In addition to the work done to characterize the promoter library in 7H9, (which can be found above), we tested the ratiometric fluorescent activity of three promoter constructs grown in soil-like conditions compared to in lab-like conditions.
However, briefly, three promoter constructs were characterized in 4 different media types, spanning both laboratory settings and soil-mimicking conditions. When comparing cultures grown in typical lab conditions (7H9) vs. in lab conditions supplemented with soil extract (SESOM7H9), we observe that two out of the three promoter constructs had a statistically significant difference in fluorescence expression between the two growth conditions, but none of the constructs had significant differences in the red/green ratiometric characteristic. Using the Friedman test, the comparison of the red/green ratio curves for M. smegmatis-pSUM8.06-pwmyc, and M. smegmatis-pSUM8.11-MSMEG2784 across different growth conditions resulted in a p-value of 0.11 and 0.11, which suggest potential different promoter activity dynamics for these two promoters tested across the two different media. SESOM and SESOM+ were unable to sustain bacteria growth and likely caused the cells to enter dormancy. We’ve shown that by successfully rescuing the constructs from SESOM and SESOM+ cultures using both 7H9 plates and liquid media. These results highlight the stark difference in bacteria metabolism when experiencing different growth conditions and a lack of transferability between data obtained under lab conditions and system performance under field situations. We again stress the necessity of testing in-field to be included in the design-build-test cycle for the SynBio community.
qRT-PCR for Soil SynBio
qRT-PCR is a very sensitive procedure due to the amplification of undetectable amounts of target DNA to detectable levels (Smith 2009). This sensitivity makes it a promising technique for soil SynBio, where bacteria can exist at very low relative abundances and DNA extractions can be inefficient.
However, qRT-PCR on DNA is limited in the information it provides about soil microbes, as DNA of both alive and dead microbes can be amplified and counted. Additionally, due to the sensitivity of this method, our team dealt with contamination issues in our no-template control. qRT-PCR requires extensive optimization and takes great effort to avoid contamination, often necessitating DNA-free PCR hoods as contamination-preventing equipment. Finally, it is also a very expensive procedure, requiring specialized equipment and expensive master mixes and probes.
Preliminary results: As a proof of concept, our team extracted DNA from our 18 spatial microcosms (n=3), at both the center (where bacteria and phage were inoculated) and 5 inches from the center over five different time points across four weeks. TaqMan qRT-PCR was performed on several of these samples as a proof of concept for this technique in our lab, and we plan on performing more qRT-PCR, to generate a complete dataset, after more optimization. We dealt with contamination, as DNA was detected in our biological negative controls as well as in the no template control (NTC) of many of our experiments.
Spatial Microcosm qRT-PCR Preliminary Results: As stated above, we dealt with chronic contamination in our no template controls due to the sensitivity of qRT-PCR as an assay. Biologically, there should be no amplification at all in any of these NTC wells; however, the standard in the field is that a CT score above 35 is acceptable for a NTC (Li 2021, Lin 2023). This was not always the case for our data. For this reason, we are only presenting preliminary data as a proof of concept for the experiments that we plan to perform after this iGEM season. The data we will be reporting is from our engA (endogenous control) and Origin of Replication (GOI) assays, on all 9 sterile soil microcosms, from samples taken at the inoculation site (center) on Day 4 and Day 7.
We want to make clear that this data is being analyzed as a proof of concept to demonstrate how the DNA we have extracted will be used in future analysis, and not to make definitive claims about our microcosms. To help account for this DNA-contamination, we have analyzed our data relative to our biological control microcosms (microcosms with no Mycobacterium smegmatis or Kampy added).
Relative quantification (RQ) calculations can be found on the results page. Results are outlined below.
Figure 16. Relative quantification (RQ) results.Although qRT-PCR is inherently relative, the absolute amount of starting material can be approximated through the creation of a standard curve. In order to compare the results of our qRT-PCR to the results of other measurement techniques performed in our lab, we generated a standard curve for our housekeeping gene (using an engA geneblock synthesized by IDT), which provides information about the number of bacteria (copy number of the genome) initially in the sample.
Figure 17. engA qRT-PCR Standard Curve.
In the graph above, blue points represent the average CT value of triplicate technical replicates for 8 serial dilutions of the engA gene. Error bars are the standard deviation of technical replicate CT values, but are not visible due to consistency between technical replicates. This standard curve was used to generate following linear equation (r2 = 0.99), relating CT to the logarithmic copy number for our engA assay:
Figure 18. Linear equation (r2 = 0.99), relating CT to the logarithmic copy number of engA assay.
However, a limitation of qRT-PCR, particularly for soil extractions, is that the DNA extraction efficiency is not 100%, so the absolute amount of starting material in the qRT-PCR reaction may be more indicative of the efficiency of the extraction than the actual amount of starting material present. That is why, for the bulk of our analysis, we present normalized and relative results.
Multimodal Measurement
Multimodal measurement is necessary in soil synthetic biology measurements, in order to counter the inaccuracy and error introduced by soil particles in wet lab procedures. Multimodal measurement was kept in mind throughout our whole project and exercised when possible, and data from multiple measurement sources was evaluated and compared.
Below outlines the different measurement methods used for each type of information collected:
We were interested in characterizing the relationship between the results obtained from different measurement methods. Here, bacterial counts obtained from plating and qRT-PCR analysis are shown, after visual counting and fitting to our Ct standard curve, respectively. For visualization, RT-PCR values are scaled by a factor of 1000. Interestingly, qRT-PCR results show a relative larger change in bacterial numbers over time than plating measurements, for both M. smegmatis pCherry3 and M. smegmatis pMLcherry (integrated) samples. This could be explained by the limited extraction efficiencies of qRT-PCR, or conversely potential overestimation of bacterial counts from plating due to contamination.
Figure 20. Bacterial counts per g dry soil results from qRT-PCR and plating.We were also interested in comparing the bacterial counts obtained from both bacterial plating and direct microscopy for the bottle microcosm experiment. More information about how these results were obtained can be found on the experiments and results page, as well as the “Microscopy Measurement” section of the Measurement page. A comparison of these results are shown below.
Figure 21. Bacterial counts from plating and direct microscopy of bottle microcosm experiment samples. Y-axis is shown on a log-scale.
As expected, plating counts are generally lower than direct microscopic counts, likely due to the ability of direct microscopic observation to enumerate all bacteria, while culturing only allows for “culturable” bacteria to be counted. However, both groups show a similar trend, with an exponential increase in bacterial counts over time.
Seventh Annual International InterLaboratory Study
Additionally, we participated in the Seventh Annual International InterLaboratory Study, in order to contribute to improving data comparability between labs and instruments. We also utilized the calibration standard curves throughout our experiments, to determine the concentrations of microscopy standards, to convert RFU to absolute units with the plate reader, and for the creation of an OD to CFU standard curve for Mycobacterium smegmatis.References
Adams, B.L. (2016). The Next Generation of Synthetic Biology Chassis: Moving Synthetic Biology from the Laboratory to the Field. ACS Synth. Biol. 5, 1328–1330. 10.1021/acssynbio.6b00256.
Armianinova, D.K., Karpov, D.S., Kotliarova, M.S., and Goncharenko, A.V. (2022). Genetic Engineering in Mycobacteria. Mol Biol 56, 830–841. 10.1134/S0026893322060036.
Ashelford, K.E., Day, M.J., and Fry, J.C. (2003). Elevated Abundance of Bacteriophage Infecting Bacteria in Soil. Appl Environ Microbiol 69, 285–289. 10.1128/AEM.69.1.285-289.2003.
Ayala-Torres, C., Hernández, N., Galeano, A., Novoa-Aponte, L., and Soto, C.-Y. (2014). Zeta potential as a measure of the surface charge of mycobacterial cells. Ann Microbiol 64, 1189–1195. 10.1007/s13213-013-0758-y.
Bardarov, S., Bardarov, S., Pavelka, M.S., Sambandamurthy, V., Larsen, M., Tufariello, J., Chan, J., Hatfull, G., and Jacobs, W.R. (2002). Specialized transduction: an efficient method for generating marked and unmarked targeted gene disruptions in Mycobacterium tuberculosis, M. bovis BCG and M. smegmatis. Microbiology 148, 3007–3017. 10.1099/00221287-148-10-3007.
Barker, P.E. (2003). Cancer Biomarker Validation: Standards and Process: Roles for the National Institute of Standards and Technology (NIST). Annals of the New York Academy of Sciences 983, 142–150. 10.1111/j.1749-6632.2003.tb05969.x.
Barlow, A.L., and Guerin, C.J. (2007). Quantization of widefield fluorescence images using structured illumination and image analysis software. Microsc. Res. Tech. 70, 76–84. 10.1002/jemt.20389.
Beal, J., Baldwin, G.S., Farny, N.G., Gershater, M., Haddock-Angelli, T., Buckley-Taylor, R., Dwijayanti, A., Kiga, D., Lizarazo, M., Marken, J., et al. (2021). Comparative analysis of three studies measuring fluorescence from engineered bacterial genetic constructs. PLoS ONE 16, e0252263. 10.1371/journal.pone.0252263.
Beal, J., Farny, N.G., Haddock-Angelli, T., Selvarajah, V., Baldwin, G.S., Buckley-Taylor, R., Gershater, M., Kiga, D., Marken, J., Sanchania, V., et al. (2020). Robust estimation of bacterial cell count from optical density. Commun Biol 3, 1–29. 10.1038/s42003-020-01127-5.
Beal, J., Haddock-Angelli, T., Baldwin, G., Gershater, M., Dwijayanti, A., Storch, M., De Mora, K., Lizarazo, M., Rettberg, R., and with the iGEM Interlab Study Contributors (2018). Quantification of bacterial fluorescence using independent calibrants. PLoS ONE 13, e0199432. 10.1371/journal.pone.0199432.
Beal, J., Haddock-Angelli, T., Gershater, M., De Mora, K., Lizarazo, M., Hollenhorst, J., Rettberg, R., and iGEM Interlab Study Contributors (2016). Reproducibility of Fluorescent Expression from Engineered Biological Constructs in E. coli. PLoS ONE 11, e0150182. 10.1371/journal.pone.0150182.
Beal, J., Telmer, C.A., Vignoni, A., Boada, Y., Baldwin, G.S., Hallett, L., Lee, T., Selvarajah, V., Billerbeck, S., Brown, B., et al. (2022). Multicolor plate reader fluorescence calibration. Synthetic Biology 7, ysac010. 10.1093/synbio/ysac010.
Bezakova, G., Helm, J.P., Francolini, M., and Lømo, T. (2001). Effects of Purified Recombinant Neural and Muscle Agrin on Skeletal Muscle Fibers in Vivo. The Journal of Cell Biology 153, 1441–1452. 10.1083/jcb.153.7.1441.
Brazma, A., Krestyaninova, M., and Sarkans, U. (2006). Standards for systems biology. Nat Rev Genet 7, 593–605. 10.1038/nrg1922.
Brokaw, A.M., Eide, B.J., Muradian, M., Boster, J.M., and Tischler, A.D. (2017). Mycobacterium smegmatis PhoU Proteins Have Overlapping Functions in Phosphate Signaling and Are Essential. Front. Microbiol. 8, 2523. 10.3389/fmicb.2017.02523.
Charpilienne, A., Nejmeddine, M., Berois, M., Parez, N., Neumann, E., Hewat, E., Trugnan, G., and Cohen, J. (2001). Individual Rotavirus-like Particles Containing 120 Molecules of Fluorescent Protein Are Visible in Living Cells. Journal of Biological Chemistry 276, 29361–29367. 10.1074/jbc.M101935200.
Cunha-Reis, C., El Haj, A.J., Yang, X., and Yang, Y. (2013). Fluorescent labeling of chitosan for use in non-invasive monitoring of degradation in tissue engineering: Fluorescent chitosan for monitoring of degradation. J Tissue Eng Regen Med 7, 39–50. 10.1002/term.494.
de Lorenzo, V., Krasnogor, N., and Schmidt, M. (2021). For the sake of the Bioeconomy: define what a Synthetic Biology Chassis is! New Biotechnology 60, 44–51. 10.1016/j.nbt.2020.08.004.
Deshayes, C., Perrodou, E., Gallien, S., Euphrasie, D., Schaeffer, C., Van-Dorsselaer, A., Poch, O., Lecompte, O., and Reyrat, J.-M. (2007). Interrupted coding sequences in Mycobacterium smegmatis: authentic mutations or sequencing errors? Genome Biology 8, R20. 10.1186/gb-2007-8-2-r20.
Dhanasekaran, S., Doherty, T.M., and Kenneth, J. (2010). Comparison of different standards for real-time PCR-based absolute quantification. Journal of Immunological Methods 354, 34–39. 10.1016/j.jim.2010.01.004.
Gallien, S., Perrodou, E., Carapito, C., Deshayes, C., Reyrat, J.-M., Van Dorsselaer, A., Poch, O., Schaeffer, C., and Lecompte, O. (2009). Ortho-proteogenomics: Multiple proteomes investigation through orthology and a new MS-based protocol. Genome Res 19, 128–135. 10.1101/gr.081901.108.
Gardner, T.S., Cantor, C.R., and Collins, J.J. (2000). Construction of a genetic toggle switch in Escherichia coli. Nature 403, 339–342. 10.1038/35002131.
Gonzalez-y-Merchand, J.A., Zaragoza-Contreras, R., Guadarrama-Medina, R., Helguera-Repetto, A.C., Rivera-Gutierrez, S., Cerna-Cortes, J.F., Santos-Argumedo, L., and Cox, R.A. (2012). Evaluation of the cell growth of mycobacteria using Mycobacterium smegmatis mc2 155 as a representative species. J Microbiol. 50, 419–425. 10.1007/s12275-012-1556-0.
Greening, C., Villas-Bôas, S.G., Robson, J.R., Berney, M., and Cook, G.M. (2014). The Growth and Survival of Mycobacterium smegmatis Is Enhanced by Co-Metabolism of Atmospheric H2. PLOS ONE 9, e103034. 10.1371/journal.pone.0103034.
Guiziou, S., Sauveplane, V., Chang, H.-J., Clerté, C., Declerck, N., Jules, M., and Bonnet, J. (2016). A part toolbox to tune genetic expression in Bacillus subtilis. Nucleic Acids Res, gkw624. 10.1093/nar/gkw624.
Huff, J., Czyz, A., Landick, R., and Niederweis, M. (2010). Taking phage integration to the next level as a genetic tool for mycobacteria. Gene 468, 8–19. 10.1016/j.gene.2010.07.012.
Ichihara, K., Yamazaki, T., Miyamura, S., Hiraoka, M., and Kawano, S. (2019). Asexual thalli originated from sporophytic thalli via apomeiosis in the green seaweed Ulva. Sci Rep 9, 13523. 10.1038/s41598-019-50070-x.
Joglekar, A.P., Salmon, E.D., and Bloom, K.S. (2008). Counting Kinetochore Protein Numbers in Budding Yeast Using Genetically Encoded Fluorescent Proteins. In Methods in Cell Biology (Elsevier), pp. 127–151. 10.1016/S0091-679X(08)85007-8.
Jost, A.P.-T., and Waters, J.C. (2019). Designing a rigorous microscopy experiment: Validating methods and avoiding bias. Journal of Cell Biology 218, 1452–1466. 10.1083/jcb.201812109.
Kang, J., Xu, L., Yang, S., Yu, W., Liu, S., Xin, Y., and Ma, Y. (2013). Effect of Phosphoglucosamine Mutase on Biofilm Formation and Antimicrobial Susceptibilities in M. smegmatis glmM Gene Knockdown Strain. PLoS ONE 8, e61589. 10.1371/journal.pone.0061589.
Kedziora, K.M., Prehn, J.H.M., Dobrucki, J., and Bernas, T. (2011). Method of calibration of a fluorescence microscope for quantitative studies: METHOD OF CALIBRATION OF A FLUORESCENCE MICROSCOPE FOR QUANTITATIVE STUDIES. Journal of Microscopy 244, 101–111. 10.1111/j.1365-2818.2011.03514.x.
Kelly, J.R., Rubin, A.J., Davis, J.H., Ajo-Franklin, C.M., Cumbers, J., Czar, M.J., De Mora, K., Glieberman, A.L., Monie, D.D., and Endy, D. (2009). Measuring the activity of BioBrick promoters using an in vivo reference standard. J Biol Eng 3, 4. 10.1186/1754-1611-3-4.
Kim, B., and Naemura, T. (2015). Blind Depth-variant Deconvolution of 3D Data in Wide-field Fluorescence Microscopy. Sci Rep 5, 9894. 10.1038/srep09894.
Koren, E., Franzen, J., Fugate, R.D., and Alaupovic, P. (1990). Analysis of cholesterol ester accumulation in macrophages by the use of digital imaging fluorescence microscopy. Atherosclerosis 85, 175–184. 10.1016/0021-9150(90)90109-V.
Kutter, E. (2009). Phage Host Range and Efficiency of Plating. In Bacteriophages Methods in Molecular Biology., M. R. J. Clokie and A. M. Kropinski, eds. (Humana Press), pp. 141–149. 10.1007/978-1-60327-164-6_14.
Lammens, E.-M., Nikel, P.I., and Lavigne, R. (2020). Exploring the synthetic biology potential of bacteriophages for engineering non-model bacteria. Nat Commun 11, 5294. 10.1038/s41467-020-19124-x.
Leake, M.C., Chandler, J.H., Wadhams, G.H., Bai, F., Berry, R.M., and Armitage, J.P. (2006). Stoichiometry and turnover in single, functioning membrane protein complexes. Nature 443, 355–358. 10.1038/nature05135.
Lee, J.-S., Wee, T.-L. (Erika), and Brown, C.M. (2014). Calibration of Wide-Field Deconvolution Microscopy for Quantitative Fluorescence Imaging. J Biomol Tech 25, 31–40. 10.7171/jbt.14-2501-002.
Li, X., Mei, H., Chen, F., Tang, Q., Yu, Z., Cao, X., Andongma, B.T., Chou, S.-H., and He, J. (2017). Transcriptome Landscape of Mycobacterium smegmatis. Front. Microbiol. 8, 2505. 10.3389/fmicb.2017.02505.
Martini, M.C., Zhou, Y., Sun, H., and Shell, S.S. (2019). Defining the Transcriptional and Post-transcriptional Landscapes of Mycobacterium smegmatis in Aerobic Growth and Hypoxia. Front. Microbiol. 10, 591. 10.3389/fmicb.2019.00591.
Mohan, A., Padiadpu, J., Baloni, P., and Chandra, N. (2015). Complete Genome Sequences of a Mycobacterium smegmatis Laboratory Strain (MC2 155) and Isoniazid-Resistant (4XR1/R2) Mutant Strains. Genome Announcements 3, e01520-14. 10.1128/genomeA.01520-14.
Moll, J., Barzaghi, P., Lin, S., Bezakova, G., Lochmüller, H., Engvall, E., Müller, U., and Ruegg, M.A. (2001). An agrin minigene rescues dystrophic symptoms in a mouse model for congenital muscular dystrophy. Nature 413, 302–307. 10.1038/35095054.
Murphy, K.C., Nelson, S.J., Nambi, S., Papavinasasundaram, K., Baer, C.E., and Sassetti, C.M. (2018). ORBIT: a New Paradigm for Genetic Engineering of Mycobacterial Chromosomes. mBio 9, 10.1128/mbio.01467-18. 10.1128/mbio.01467-18.
Poon, S.S.S., and Lansdorp, P.M. (2001). Quantitative Fluorescence In Situ Hybridization (Q‐FISH). CP Cell Biology 12. 10.1002/0471143030.cb1804s12.
Richani, D., Sutton-McDowall, M.L., Frank, L.A., Gilchrist, R.B., and Thompson, J.G. (2014). Effect of Epidermal Growth Factor-Like Peptides on the Metabolism of In Vitro- Matured Mouse Oocytes and Cumulus Cells1. Biology of Reproduction 90. 10.1095/biolreprod.113.115311.
Rudge, T.J., Brown, J.R., Federici, F., Dalchau, N., Phillips, A., Ajioka, J.W., and Haseloff, J. (2016). Characterization of Intrinsic Properties of Promoters. ACS Synth. Biol. 5, 89–98. 10.1021/acssynbio.5b00116.
Rutledge, R.G. (2003). Mathematics of quantitative kinetic PCR and the application of standard curves. Nucleic Acids Research 31, 93e–993. 10.1093/nar/gng093.
Sharifi-Sanjani, M., Meeker, A.K., and Mourkioti, F. (2017). Evaluation of telomere length in human cardiac tissues using cardiac quantitative FISH. Nat Protoc 12, 1855–1870. 10.1038/nprot.2017.082.
Smeulders, M.J., Keer, J., Speight, R.A., and Williams, H.D. (1999). Adaptation of Mycobacterium smegmatis to Stationary Phase. Journal of Bacteriology 181, 270–283. 10.1128/jb.181.1.270-283.1999.
Smith, C.J., and Osborn, A.M. (2009). Advantages and limitations of quantitative PCR (Q-PCR)-based approaches in microbial ecology: Application of Q-PCR in microbial ecology. FEMS Microbiology Ecology 67, 6–20. 10.1111/j.1574-6941.2008.00629.x.
Souchier, C., Brisson, C., Batteux, B., Robert-Nicoud, M., and Bryon, P.-A. (2004). Data reproducibility in fluorescence image analysis. Methods Cell Sci 25, 195–200. 10.1007/s11022-004-2383-4.
Sparks, I.L., Derbyshire, K.M., Jacobs, W.R., and Morita, Y.S. (2023). Mycobacterium smegmatis: The Vanguard of Mycobacterial Research. J Bacteriol 205, e00337-22. 10.1128/jb.00337-22.
Stack, R.F., Bayles, C.J., Girard, A.-M., Martin, K., Opansky, C., Schulz, K., and Cole, R.W. (2011). Quality Assurance Testing for Modern Optical Imaging Systems. Microsc Microanal 17, 598–606. 10.1017/S1431927611000237.
Stephan, J., Mailaender, C., Etienne, G., Daffé, M., and Niederweis, M. (2004). Multidrug Resistance of a Porin Deletion Mutant of Mycobacterium smegmatis. Antimicrob Agents Chemother 48, 4163–4170. 10.1128/AAC.48.11.4163-4170.2004.
Sudiman, J., Sutton-McDowall, M.L., Ritter, L.J., White, M.A., Mottershead, D.G., Thompson, J.G., and Gilchrist, R.B. (2014). Bone Morphogenetic Protein 15 in the Pro-Mature Complex Form Enhances Bovine Oocyte Developmental Competence. PLoS ONE 9, e103563. 10.1371/journal.pone.0103563.
Swedlow, J.R. (2013). Quantitative Fluorescence Microscopy and Image Deconvolution. In Methods in Cell Biology (Elsevier), pp. 407–426. 10.1016/B978-0-12-407761-4.00017-8.
T, J.A.S., J, R., Rajan, A., and Shankar, V. (2020). Features of the biochemistry of Mycobacterium smegmatis, as a possible model for Mycobacterium tuberculosis. Journal of Infection and Public Health 13, 1255–1264. 10.1016/j.jiph.2020.06.023.
Taylor, D.L. (1988). Chapter 13 Basic Fluorescence Microscopy. In Methods in Cell Biology (Elsevier), pp. 207–237. 10.1016/S0091-679X(08)60196-X.
ThermoFisher Scientific. (2022).Certificates of Analysis Lot# 2513636, Catalog number I7224. Retrieved from https://www.thermofisher.com/order/catalog/product/I7224
Tonner, P.D., Darnell, C.L., Engelhardt, B.E., and Schmid, A.K. (2017). Detecting differential growth of microbial populations with Gaussian process regression. Genome Res 27, 320–333. 10.1101/gr.210286.116.
Wang, L., and Hoffman, R.A. (2017). Standardization, Calibration, and Control in Flow Cytometry. CP Cytometry 79. 10.1002/cpcy.14.
Webb, D.J., and Brown, C.M. (2012). Epi-Fluorescence Microscopy. In Cell Imaging Techniques Methods in Molecular Biology., D. J. Taatjes and J. Roth, eds. (Humana Press), pp. 29–59. 10.1007/978-1-62703-056-4_2.
Ziolkowski, P.A., Underwood, C.J., Lambing, C., Martinez-Garcia, M., Lawrence, E.J., Ziolkowska, L., Griffin, C., Choi, K., Franklin, F.C.H., Martienssen, R.A., et al. (2017). Natural variation and dosage of the HEI10 meiotic E3 ligase control Arabidopsis crossover recombination. Genes Dev. 31, 306–317. 10.1101/gad.295501.116.
Zwier, J.M., Van Rooij, G.J., Hofstraat, J.W., and Brakenhoff, G.J. (2004). Image calibration in fluorescence microscopy. J Microsc 216, 15–24. 10.1111/j.0022-2720.2004.01390.x.