Engineered Marine Biofilms for Ocean Environment Monitoring

Abstract

Marine bacteria offer a promising alternative for developing Engineered Living Materials (ELMs) tailored to marine applications. We engineered Dinoroseobacter shibae to increase its surface-associated growth and develop biosensors for ocean environment monitoring. By fusing the endogenous extracellular matrix amyloidogenic protein CsgA with mussel foot proteins, we significantly increased D. shibae biofilm formation. Additionally, D. shibae was engineered to express the tyrosinase enzyme to further enhance microbial attachment through post-translational modifications of tyrosine residues. By exploiting D. shibae's natural genetic resources, two environmental biosensors were created to detect temperature and oxygen. These biosensors were coupled with a CRISPR-based recording system to store transient gene expression in stable DNA arrays, enabling long-term environmental monitoring. These engineered strains highlight D. shibae's potential in advancing marine microbiome engineering for innovative biofilm applications, including the development of natural, self-renewing biological adhesives, environmental sensors, and "sentinel" cells equipped with CRISPR-recording technology to capture and store environmental signals.

Keywords: Dinoroseobacter shibae; ELMs; biofilm; biosensors; marine bacteria; surface colonization.

1

Biofilm structure in variants expressing dsCsgA and Mfp3 Genes. (A) Schematic representation of the different adhesion modules containing the dsCsgA and Mfp3 genes. (B) SEM images (first row) show an extensive fibrous network in the engineered variant, while TEM images (second row) reveal a notable fibrous network in the engineered strains. Images were taken at magnifications of 9,500× (SEM) and 60,000× (TEM). (C) Percentage of bacteria producing amyloid fibers, calculated from scanning TEM images at 20,000× magnification. Data are presented as mean ± SD (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001).

2

Enhanced biofilm formation in variants expressing dsCsgA and Mfp3 genes. (A) Biofilm formation values measured with crystal violet staining of the dsCsgA-Mfp3 variants compared to the wild-type on polystyrene plates. (B) Schematic representation of the tyrosinase construct, along with western blot analysis showing the OsmY-melC2 fusion protein (53.9 kDa). (C) Biofilm formation of the dsCsgA-Mfp3-A and dsCgsA-Mfp3-SP variants, with or without coincubation with tyrosinase producing strain on polystyrene plates. (D) Biofilm formation of the dsCsgA-Mfp3 strains, with or without coincubation with tyrosinase producing strain, on plates coated with PSX 700 paint. Data are presented as mean ± SD (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001).

3

Natural response adaptation for temperature sensing. (A) MA plot of significantly (FDR < 0.05) upregulated genes. Genes highlighted correspond to canonical homologues related with the heat shock stress response. Dshi_0075 is highlighted as the selected gene. (B) RNA coverage at the Dshi_0075 genomic locus. The red line represents the average of the samples treated with 42 °C for 15 min whereas the gray line is the average control coverage. (C) FbFP fluorescent signal of the temperature biosensor strain after different times of heat sock treatment (5, 15, 30, 42 min) at different temperatures (32, 37, 42 °C). The diagram at the top depicts the genetic design of the temperature sensor architecture. (D) FbFP fluorescence of the temperature biosensor strain after 40 min of exposure to different temperatures. The diagram at the top depicts the genetic design of the temperature sensor architecture. (E) Representative fluorescent TIRF images of a biofilm formed by the wild-type (WT) and the temperature sensor strains (T° Sensor), exposed to either room temperature (25 °C) or heat shock (42 °C).

4

FnrL regulated promoter adaptation for oxygen sensing. (A) Schematics of the two alternative plasmid designs for oxygen sensing pOxy-A and pOxy-B. (B) FbFP fluorescent values of the two different oxygen biosensors for absence and presence of oxygen (*** p < 0.001).

5

Sentinel stores transcriptional information in DNA using Record-Seq. (A) Schematics of the Record-seq recording technology. (B) Length distribution of the acquired spacers. (C) GC content distribution of the acquired spacers. (D) Relative spacer counts of FbFP aligning spacers between the 42 °C treated and the control sample, DESeq2 computed fold-change and adjusted p value are reported at the top.