Overview
Our project incorporated measurements across multiple scales—ranging from macroscopic visual observations to microscopy, molecular, and “omic” (DNA-Seq and RNA-seq) analyses. This approach not only provided a comprehensive evaluation of experimental outcomes but also promoted accessibility, allowing investigators with varying resources to contribute to research in synthetic biology within aquatic systems. Perhaps most importantly and uniquely, we include a type of measurement not typical for iGEM teams, but increasingly important for thorough measurement in projects—namely bioinformatic mining of existing “omic” data relevant to the project. Below, we outline these different levels of measurement.
Macroscopic Visual Measurement
Visual analysis methods offer a simple and accessible measurement option that is less resource-intensive and often easier to conduct on a large scale than instrument-based alternatives.
Cyanophage Assays for Harmful Algal Bloom Remediation:
To efficiently search for M. aeruginosa cyanophages in bulk, we used two visual measurement methods: liquid assays and traditional agar phage plaque assays. Procedures for M. aeruginosa phage isolation require several modifications to traditional phage isolation procedures, accounting for the bacterium’s photosynthetic requirements and slow growth rate. [For more information, see the M. aeruginosa manual in our water SynBio guidebook]
Figure: Liquid assays in 35 mm dishes (left) and traditional agar plaque assays (right)
A) Liquid cyanophage assays: 4 ml M. aeruginosa was cultured with 1 ml sample filtrate (typically collected via 0.22 um filtration of enriched environmental water samples) in 35 mm dishes and monitored over several weeks for bacterial clearing indicative of phage infection. Cleared assays were filtered (0.22 μm) and filtrate was used to conduct new assays assessing the putative phage filtrate’s re-infection potential.
Figure: Evidence of bacterial clearing over time in an assay “infected” with enriched sample filtrate (left) versus a corresponding control “infected” with NFW. By day 3, the enrichment assay had begun yellowing, and by day 4 had almost fully cleared. The control retained approximately the same coloration throughout the experiment.
B) Agar cyanophage plaque assays: 1 ml of highly concentrated M. aeruginosa (concentrated from an initial 30 ml via centrifugation at 1500 x g) was mixed with sample filtrate and plated on BG-11 plates. Microcystis lawns were visually monitored for phage plaque formation. Plaques were resuspended in liquid Microcystis culture and monitored for clearing. A plate containing prominent plaques was flooded and lysate harvested and used for more reinfection assays.
Figure: Example setup for cyanophage enrichments, liquid assays, and agar assays.
Colony Counts to Assess A. baylyi Survival in Harmful Algal Bloom Microcosms:
We used colony counts to quantify the survival of the chassis species A. baylyi in microcosms containing lakewater as compared with control cocultures. Colony count spread plates are easy to create and can supplement higher-level measurements such as genomics-based abundance estimates. Our 16S data indicated that the A. baylyi percent abundance decreased over the course of the microcosm experiment, but percent abundances alone cannot conclusively prove that the population declined (rather, they suggest only that the proportion of A. baylyi decreased over the course of the experiment). The reduction in A. baylyi colony formation indicates that the A. baylyi abundance decreased as the population declined.
Figure: Spread plates approximately 1 day post-inoculation (left) versus 7 days post-inoculation (right). The top row of plates in each image were spread using samples from lakewater microcosms, the bottom row from control cocultures.
Phage Plaque Assays with Mycobacterium smegmatis:
To prepare high-titer phage stocks for downstream applications outside standard laboratory conditions, we conducted a series of experiments aimed at both estimating and amplifying phage titers. Phage CrimD, Phage Neighly, and Phage Raid were filtered and serially diluted (10⁻² to 10⁻¹⁰) in phage buffer, then combined with Mycobacterium smegmatis and plated using top agar overlays. Following overnight incubation, plaque assays were used to determine dilution ranges and estimate titers based on plaque-forming units (PFU). Plates exhibiting webbed lysis patterns were subsequently flooded with phage buffer and filtered to generate high-titer lysates. Final titers ranged from approximately 10¹⁰ to 10¹³ PFU/mL. Notably, CrimD and Neighly produced webbed lysis at 10⁻¹⁰ dilutions, while Raid achieved similar lysis at 10⁻⁸, indicating robust infectivity and effective clearance of the mycobacterial lawn.
Figure: Plaque assays with Phage CrimD, Phage Neighly, and Phage Raid at 10^-6, 10^-8, and 10^-10 dilutions.
Bacterial Titer Assay from Pipe Microcosm Effluent:
Quantifying bacterial populations within biofilms is challenging due to the structural complexity of the extracellular matrix and strong bacterial adhesion. To circumvent this, we measured bacterial titers from the microcosm effluent. Biofilms of Mycobacterium smegmatis, transformed with a red fluorescent protein (RFP) plasmid conferring hygromycin resistance, were grown on PVC pipes under nonsterile conditions and inserted into custom 3D-printed adaptors simulating a showerhead. Pipes were flushed daily with tap water, and effluent samples were plated on hygromycin-containing media to estimate bacterial load. Following several days of biofilm establishment, phage was introduced via the tap water, and daily effluent sampling continued. No significant reduction in hygromycin-resistant bacterial titers was observed post-phage treatment, suggesting limited phage efficacy against M. smegmatis within established biofilms under flow conditions.
Figure: Bacterial titer assay from time point 1 on hygromycin selective plates.
Observing morphology of B. subtilis biofilms:
We cultured all of our 6 strains (Natural strain B. subtilis 3610, lab strain B. subtilis 168, and the 4 engineered variants, which produces different biofilm) in different induce levels and on different media to evaluate their biofilm formation and colony morphology, Natural strain B. subtilis 3610 developed a distinct brown biofilm, while among all the modified strains only the B. subtilis SinI+ strain showed biofilm-like folds and wavy edges; all other engineered strains formed smooth, non-biofilm colonies. Biofilm formation occurred on MSgg medium but not on LB, confirming MSgg as the appropriate biofilm assay medium. Applying the inducer xylose directly into the MSgg medium yielded more uniform and structured colonies than spreading it on the surface of agar, which caused uneven distribution of inducer concentration and uneven growth. However, 0.03% and 0.2% xylose had minimal effect on biofilm formation, suggesting these concentrations differences were too low to make difference in inducing.
Figure: Different colony/biofilm morphology of our six Bacillus subtilis strains cultured on MSgg and LB agar with 0.03% xylose.
Motility assay for Bacillus subtilis:
The motility of a B. subtilis strain was assessed using semi-solid agar plates (0.3%, 0.7%, and 1.5% agar). The ability of spreading on 0.3% agar reflects that this strain has flagellum-dependent swarming ability, while spreading on 0.7% agar indicates EPS-dependent sliding ability. We compared the spreading of our six B. subtilis strains (natural strain B. subtilis 3610, lab strain B. subtilis 186, and the four engineered variants) on 0.3% and 0.7% agar with their spreading on 1.5% agar to test if they have any motility. All plates have 0.2% xylose inducer in the agar. Our results suggested that only the natural strain B. subtilis 3610 has both swarming and sliding motility, B. subtilis 168 and B. subtilis 168 SinI+ variant showed strong swarming. B. subtilis 168 BslA+ and TapA-SipW-TasA+ variants displayed moderate swarming. B. subtilis 168 TasA variant showed neither swarming nor sliding.
Figure: B.subtilis 168 SinI+ on 0.3%, 0.7%, 1.5% agar plates with 0.2% xylose.
Microscopy
In order to more closely assess biological processes in ways that are not visually clear with the naked eye, we employed microscopy techniques, such as scanning electron microscopy and digital microscopy with the HIROX RV2000.
Mycobacterial biofilm surface analysis:
Mycobacterial biofilms, composed mainly of sugars and polysaccharides, were examined to assess structural changes after phage treatment. We created four microcosms—two treated with phage and two untreated—and analyzed biofilm samples 36 hours post-treatment using a Hirox RV 2000 microscope, which created a 3-dimentional image of the biofilm structure. The HIROX software assigned colors ranging from blue to red, with red indicating taller surfaces and blue lower surfaces. Mycobacteria expressing RFP were identified by red fluorescence. Biofilm height and surface area coverage on PVC pipe were compared between treated and untreated samples.
Figure: The images above show the pseudocolors assigned by the HIROX RV 2000 microscope. The image on the right shows a biofilm surface on a PVC pipe that was not treated with phage. The image on the left shows a biofilm surface on a PVC pipe that was treated with phage.
Scanning electron microscopy: corrosion preventing ability of B. subtilis
To investigate the effects of different Bacillus subtilis strains on the corrosion behavior of steel in seawater microcosms, we used scanning electron microscopy (SEM) to examine the microstructures of steel surfaces after bacterial treatments. Steel specimens were immersed in seawater microcosms inoculated with different strains of B. subtilis (the natural strain, lab strain, and several engineered variants) and retrieved after 3 and 7 days for SEM imaging. SEM observations were performed at 10 kV with magnifications ranging from 100× to 50,000×.The results showed that the steel specimens in the seawater control (didn’t inoculate any B. subtilis) formed a thick oxide layer with distinct corrosion textures. Steel surfaces treated with the natural strain B. subtilis 168 showed distinct corrosion lines interspersed with deposits of irregularly shaped compounds. Steel treated by engineered strains, SinI+ and BslA+, exhibited parallel-aligned inorganic crystal structures with anisotropic morphologies. These microstructural variations suggest that different B. subtilis strains exert different influences on corrosion behavior in marine environments.
Figure: Scanning electron microscopy images of steel surfaces treated with different Bacillus subtilis strains
Molecular
M. aeruginosa Growth Curves:
To better understand M. aeruginosa’s growth under our specific lab conditions and to inform our culture setup, we conducted a 56-day growth curve for M. aeruginosa strain LE3 (UTEX 3037) using daily fluorescence (absorbance 659 nm, emission 700 nm) and optical density measurements at 730 nm (OD730) as proxies for cell counts. Fluorescence absorbance and emission values were selected to approximate M. aeruginosa’s natural chlorophyll a fluorescence.
Figure: Average M. aeruginosa growth curves based on OD730 measurements (left) and fluorescence at 659-700 nm (right).
Measurements were collected in technical triplicates from six 50 ml M. aeruginosa sub-cultures. Three of the six sub-cultures (biological triplicates) were cultured with in-house BG-11 media, the remaining three in pre-made UTEX BG-11 media under approximately the same conditions. All measurements were collected using a standard plate reader.
B. subtilis Growth Curves:
To evaluate the ability of growing for the wild-type strain (B. subtilis 3610), the laboratory strain (B. subtilis 168), and the four engineered variants, we subcultured their overnight inoculations into 96-well plates containing LB medium with xylose at induction levels ranging from 0.001% to 2% (v/v). Growth curves were recorded at 24 and 36 hours using a plate reader, enabling comparisons across strains and induction conditions. We discovered that varying xylose concentrations had minimal impact on growth rate. The four engineered strains showed similar growth rates to B. subtilis 168, all lower than the wild-type B. subtilis 3610.
Figure: 36 hours B. subtilis 3610 growth curves based on OD600 with xylose induction levels ranging from 0.001% to 2%
Genomic
Genomic data contains species abundance information that can clarify chassis survival and competitor species dynamics across timepoints or conditions, enabling comparative assessment of effective deployment and survival potential in simulated water environments versus in vitro conditions.
16S amplicon sequencing in particular provides a rapid way to obtain relevant relative abundance estimates at the genus or species level. We assessed chassis survival in aquatic microcosms by quantifying relative abundance via 16S sequencing and analysis.
16S Species Abundance Measurements: B. subtilis Seawater Microcosms
We used 16S data to evaluate the survival of Bacillus subtilis strains and the microbial population within seawater microcosms. The objective was to determine whether different strains: the wild-type natural strain, the lab strain, and the genetically engineered variants, could maintain a stable presence in marine environments or be outcompeted by other bacteria. Seawater microcosms were prepared by inoculating varying concentrations of the strains. After 14 days of culturing, we amplified the 16S DNA of the samples via PCR, and purified using the NEB PCR purification kit. Purified samples were sequenced by Plasmidsaurus to determine species-level relative abundances. Results showed that the wild-type strain B. subtilis 3610 exhibited extremely low survival, whereas the engineered strains SinI+ and BslA+, demonstrated persistence. In addition, the inoculum concentration of B. subtilis influenced the community composition: Ochrobactrum dominated in microcosms inoculated with 50 mL of B. subtilis per 150 mL seawater, while Acinetobacter was significantly more abundant in those inoculated with 1 mL of B. subtilis per 200 mL seawater.
Figure: Relative abundance of bacterial species in seawater microcosms inoculated with different concentrations of Bacillus subtilis 3610.
16S Species Abundance Measurements: A. baylyi Lakewater Microcosms
We used 16S data to quantify the survival of chassis A. baylyi in microcosms containing lakewater as compared with control cocultures. We collected and flash-froze microcosm and coculture samples at multiple experimental timepoints, extracted total genomic DNA via phenol chloroform isoamyl alcohol (PCI) extraction from frozen sample pellets, amplified 16S fragments via PCR, and purified PCR products using the NEB PCR cleanup kit. Purified samples were sequenced and relative species abundances estimated by Plasmidsaurus. In conjunction with visual assessment of A. baylyi colony counts (a “lower level” measurement), relative abundance estimates indicate that A. baylyi declined over time and was outcompeted by native lakewater species.
16S Species Abundance Measurements: M. smegmatis Pipe Microcosms:
16S rRNA gene sequencing of the biofilm revealed a predominance of Acinetobacter species, along with a diverse array of other bacterial taxa. Although no sequences were specifically classified as Mycobacterium, the detection of Actinobacteria suggests the possible presence of related genera, warranting further taxonomic resolution. The biofilm was cultivated under non-sterile conditions to emulate the microbial complexity of household water systems, and the presence of multiple species was anticipated. However, it is important to note that 16S sequencing is limited to bacterial identification and does not capture the full microbial community, including fungi, archaea, or viruses, thereby providing only a partial view of the biofilm’s ecological composition.
Transcriptomic
Transcriptomic data allows high-level analysis of chassis behavior and functioning. We performed differential gene expression analysis on RNA sequencing results to compare the behavior of engineered bacteria and chassis strains in simulated real world environments versus traditional laboratory conditions. Our analysis revealed differential gene expression across microcosm and control groups, indicating that engineered bacteria function differently when deployed in realistic environments.
Transcriptomic Data Analysis Methods Summary: A baylyi and B. subtilis Microcosms
We extracted RNA from relevant samples using the NEB Monarch Spin RNA Isolation Kit. We analyzed RNA sequence data through the Galaxy platform, where we removed adapters with Trim Galore!, aligned reads against the relevant reference genome (RefSeq accession NC_005966.1 for A. baylyi microcosm experiments and RefSeq accessions NC_000964.3 and CP020102.1 for 168-based B. subtilis strains and for B. subtilis 3610, respectively) via HISAT2, sorted outputs with Samtools, generated counts with HTSeq, and conducted differential gene expression analysis with DESeq2.
Transcriptomic Analysis: A. baylyi Lakewater Microcosms
We used RNA sequencing to analyze A. baylyi gene expression under simulated lakewater conditions (also inoculated with M. aeruginosa) versus in control cocultures with M. aeruginosa, where the chassis’s remediation potential is typically tested. We extracted RNA from frozen sample pellets collected 1, 5, and 7 days post-inoculation. We observed significant differential expression of a number of genes, including upregulation under microcosm conditions of genes encoding an NAD(P)/FAD-dependent oxidoreductase, an ATP-dependent Clp protease ATP-binding subunit, and downregulation of encoding LyrR family transcriptional regulators.
Transcriptomic Analysis: B. subtilis Seawater Microcosms
We performed RNA sequencing to compare Bacillus subtilis gene expression under simulated seawater microcosm conditions and standard laboratory culture. Three high-density microcosms were prepared using engineered B. subtilis 168 SinI+, B. subtilis 168 BslA+, and the wild-type B. subtilis 3610. Transcriptomic analysis results indicate significant differential expression of a large number of genes, including upregulation of metabolic genes and flagellar assembly genes under microcosm conditions.
Bioinformatic Data Mining
In order to expand past traditional measurement techniques, we mined the potential wealth of information available in existing databases and literature. We employed AI and our own RNA-Seq analysis pipelines to perform a comprehensive meta-analysis of RNA-Seq data and identified commonly differentially expressed genes and their associated pathways between laboratory and natural conditions. This allowed us to extract key design principles for circuits in order to allow – eventually — for the safe and effective deployment of synthetic constructs in real-world aquatic environments.
We created Python scripts to extract metagenomic sampling data from a wide range of metagenomic sites. This data was a vital tool in characterizing species presence and survival in previously sampled locations. We obtained quantitative data on both abiotic and biotic factors, which allow correlation determination between species and their sampling environment factors. By reviewing these results, we found that accounting for an environment's complete context–including both its abiotic and biotic features– is essential for understanding chassis survivability when deployed in a real-world aquatic environment.