Reactive Oxygen Species as a Promising Theranostic Target

Reactive oxygen species (ROS) are highly reactive chemical compounds that are natural byproducts of metabolism. They are involved in cell growth, differentiation, signaling as well as immune responses^[1]. Low ROS levels are required for proper metabolic function and cell signaling. However, high ROS levels (oxidative stress) can lead to tissue damage through the uncontrolled oxidation of cellular components^[2].

In many organisms, ROS levels spike during signaling of cell death, indicating, for example, an infection. Therefore, selective high-ROS sensing, coupled with a localized and timely secretion of antimicrobials, could be used for early pathogen detection and prevention. In animals, high ROS levels are found in vascular, neurodegenerative and malignant diseases, as well as in aging. Keeping ROS at non-toxic levels by coupling sensing and anti-oxidant secretion could reduce the severity of many human diseases like diabetes or Alzheimer's disease^[3]. Timely localization and quantification of ROS levels could also allow for targeted therapy^[2].

Engineering microbiota as a high-pass ROS filter

We aim to develop a versatile cross-species high-pass ROS sensor platform, harnessing this disease- side effect as a timely and local marker for various health applications.

Which species should we engineer to allow for broad-ranged applications?

A cross-species platform would enable ROS-sensing in many settings. Using microorganisms would allow applications in research and health, microbiota being at the center of many bigger organisms of interest such as humans and plants.

Compatible microbiota agents have been selected for sensor engineering, namely bacteria and yeast, which are both naturally part of many bigger organisms. For bacteria, we have engineered classic E.coli lab strains but also human and plant microbiota strains as they would be closer to the final recipients of our sensor system for health applications. It is crucial to already consider strain variation as the oxidative stress sensitivities and reactivities can differ significantly^[4]. E. coli Nissle 1917 is a well-characterized and engineered human microbiota strain^[5,6], which we engineered as proof of concept for medical applications, whereas, for plants, we chose to focus on Pseudomonas alloputida mt-2 KT2440 and P. protegens Pf5 strains, an oxidative-stress resistant strain and a biocontrol lab strains which were accessible to us^[7-10]. For yeast, we first focused on making our construct work in the classic S. cerevisiae lab strain.

We aim to adapt their natural ROS-sensing ability to suit our purpose.

Which native sensors did we choose to engineer and why?

For bacteria, the main native transcription factor OxyR will be tuned as high-pass ROS sensor ^[11], and, similarly for yeast, Yap1 ^[12-14]. An alternative eucaryotic ROS-sensitive TF from plants called Tga2 will also be tested in yeast, as it is naturally involved in plant immunity in response to pathogen infections called systemic acquired immunity (SAR) ^[15,16].

How do these factors regulate gene expression in a ROS-dependent manner?

1. OxyR forms a dimer of dimers (homotetramer). A change in configuration is induced upon cysteine oxidation by ROS, allowing it to bind to DNA and activate/repress genes depending on the location of the binding site. In the oxyRS regulon , oxyR negatively regulates its own expression and activates OxyS^[11]. We used the OxyR-OxyS-GFP plasmid from the Panke lab (ETH Zürich, Tsvetan Kardashliev) as our base sensor to engineer.
2. Yap1 is trapped in the nucleus by covalent bonding with Gpx3 through an intramolecular disulfide bond inhibiting export, and binds to ARE motifs (AP-1 recognition element), such as in the Trx2 gene coding thioredoxin^[17]. We aim to strongly constitutively express Yap1 and add our own mutated version of the Trx2 promoter with varying numbers of AP-1 motifs to alter sensor sensitivity, inspired by Dacquay and McMillen, 2021^[13].
3. Tga2, however, is characterized mainly in regards to the salicylic acid (SA) -dependent SAR pathway, in which defense and stress-related genes are activated through the As-1-like cis-elements motifs in their promoters where Tga2 binds acting as a repressor and is released upon NPR1 binding^[18]. However, in oxidative stress, NPR1-independent activity has been shown, possibly involving SCL14 protein, but the mechanism remains uncertain^[19]. We will constitutively express Tga2 alone and its promoter the CaMV35S (Cauliflower Mosaic Virus 35S) containing one or multiple As-1 motifs in hopes of observing H2O2-dependent activity^[20].




How can we then modify their sensitivity?

Since the TF is natively expressed (OxyR and Yap1), we chose to focus our strategy on promoter engineering in order to alter TF binding properties and, thus, gene expression^[13,20]. To this end, we focused on convenient GFP-reporter vectors adapted for bacterial and yeast expression, in order to be able to assess H2O2 sensitivity of our new constructs through fluorometric assays.

• Our first approach is to randomly mutate the promoter sequence during an error-prone PCR, and have a high-throughput screening of mutants with flow cytometry (GFP sorting).
• Our second, more tedious but more precise, low-throughput parallel approach is directed mutagenesis on specific bases, computationally predicted to impact binding affinity more or less strongly (see Model page).

Medical and Agritech Applications

Such a sensor could be paired with a detection and/or therapeutic output depending on the aim. As a proof of concept, we chose to focus on 2 diseases, each highly cumbersome: IBD in humans, and fireblight in plants.

Medical application output: catalase/SOD ROS-scavenging enzymes

We aim to create an autonomous theragnostic system for chronic inflammation treatment. The probiotic bacteria equipped with high-ROS-tuned OxyR will be able to adapt the amounts of therapeutic ROS-scavenging enzymes produced, with local and timely precision. Catalase (KatG gene) was chosen as it is commonly used in ROS-scavenging therapeutic engineered bacteria, although the sensing pathway differs ^[6]. Superoxide dismutase (SOD) would also be pertinent, however, more difficult to test since we chose H2O2 as our ROS.

Catalase: 2H2O2 → 2 H2O + O2

SOD: 2O2•−+2H+ → H2O2+O2

Agritech application output: mScarlet fluorescent protein and a bacteriostatic

the agent prevents the growth of bacteria (i.e., it keeps them in the stationary phase of growth)

The sensor will be engineered in epiphytic microbiota to detect high-ROS occurring during plant infections like fireblight, producing a fluorescent glow (mScarlet protein) as infection indicator. In response to detection, a bacteriostatic would be secreted, such as E. amylovora’s (fireblight agent) own siderophore, an iron chelator called desferrioxamine E ^[21], which would allow our competing bacteria to slow down the infection by stealing its vital iron. This translatable platform could also be applied to other high-ROS-producing diseases, like the Xylella fastidiosa infection in grapevines, Pseudomonas syringae on leaves, and Zymoseptoria tritici infection in wheat.

Medical Case-Study: Chronic Inflammation

Inflammation is a critical part of the body’s immune response. It helps in eliminating pathogens and promotes tissue repair and recovery. ‘Chronic’ inflammation however, impairs the healing process, persists over prolonged periods of time and can cause serious damage ^[22,23]. Many prevalent diseases like diabetes, cardiovascular disease, inflammatory bowel disease (IBD), and chronic obstructive pulmonary disease (COPD) are associated with chronic inflammation and are incurable ^[24]. Chronic inflammation is therefore one of the most significant causes of death worldwide ^[25].

Excessive ROS Production – a Hallmark of Chronic Inflammation

One characteristic most sites of chronic inflammation have in common are significantly elevated ROS levels. This is the case for chronic wounds in diabetic patients and is a major factor in IBD ^[26,27]. While low levels of ROS are essential in stimulating effective wound healing, excessive ROS cause cellular damage and impair wound repair by obstructing the shift from inflammatory stage to the proliferative stage of wound healing ^[28]. High ROS levels may even have a causative role by activating signaling pathways that are associated with chronic inflammation ^[29,30].

Our Solution: Local and Timely Targeting of ROS Through Tuned-Antioxidant-Producing Probiotics

Our technology aims to target high ROS levels as both an indicator for chronic inflammation and a problem to be solved. The sensor will be specifically tuned to high concentrations and activate transcription and secretion of antioxidant enzymes to treat the inflammation (catalase, SOD). Due to the high relevance of oxidative stress in chronic inflammation, ROS scavenging is an attractive therapeutic approach. Anti-oxidative enzymes have been shown to help with diabetic wounds, like promising hydrogels loaded with SOD ^[32]. Also in IBD, antioxidants are a promising therapeutic approach ^[33]. We aim to engineer probiotic bacteria, equipping them with a high-ROS-tuned sensor to produce therapeutic ROS-scavenging enzymes with local and timely precision.

Agritech Case-Study: Fighting Fireblight

Farmers worldwide face increasing challenges from weather catastrophes, economic difficulties, and biological hazards. We aim to alleviate the risk and economic loss caused by bacterial infections by helping farmers detect and prevent their spread. Erwinia amylovora is a bacterium that infects and kills fruit trees. While E. amylovora can be treated with antibiotics, it is mostly prohibited in Europe due to environmental concerns ^[34], thus cutting of infected plants remains the only curative solution ^[35]. In 2007, fireblight caused 50 million CHF in damages ^[36].

Early Detection of Bacterial Infections through ROS Sensing

Fireblight is caused by the bacteria Erwinia amylovora that infects and colonizes plant flowers, shoots and leaves ^[37]. E. amylovora, and other Gram negative bacteria like Pseudomonas syringae, inject toxins into the plant’s cells to kill them, promoting the infection of plant tissue ^[37]. Bacterial toxins are transferred through a secretion system (T3SS) that acts like a needle, perforating the plant cell ^{[39, 40]}. This activity, specific to some pathogenic species of bacteria, triggers strong ROS release from the plant as a defense mechanism (Fig. 11). This leads to cell death for both plant cells and sensitive bacteria. Early detection of abnormal ROS levels in plant tissue could thus reveal the presence of pathogenic bacteria before it spreads. However, ROS are produced in many situations, like abiotic stress (drought) or ripening, where H₂O₂ ranges from 0.5 to 1.5 mM ^[41]. Therefore, detection has to take into account the normal fluctuating ROS levels as a threshold. The range should be between 1-10 mM as ROS levels during infection can be at least 5 times higher (Fig. 11).

Our Solution: Competitive Sensor Bacteria, Applicable to ROS-generating Diseases

Currently, the main detection systems are based on sample retrieval and individual screening (PCR ^[43], bacteriophages). Moreover, E. amylovora biocontrol with competing beneficial micro-organisms like P. fluorescens (found in Serenade®) / P. agglomerans, or copper, sprayed on the flowers, stand out as the best modern chemical-free approaches ^{[9, 37, 44, 45]}. Engineered bacteria with a lytic E. amylovora-killing phage have managed to combine detection and biocontrol ^[46] but rely on transduction (deemed unsafe as biopesticide ^[47]), would push pathogens towards resistance ^[48] and are country-strain specific ^[49]. We propose a simpler approach, allowing early detection of any E. amylovora strain infection by engineering bacteria already present in the plant microbiota to detect ROS, acting as preventive and inducible biocontrol agents. Fluorescence would be detected by UV lamps at night and could be automated with drones ^[50,51], allowing early removal of infected flowers before further dissemination.