COVID-19 presented a challenge to our team by limiting our access to facilities, reagents, and sampling campaigns. For this reason we were not able to materialize the biosensor yet. However, with the help of the gathered literature we next describe the expected results for each of the 5 phases of our experimental design with their corresponding differences with the biosensor designed by Wang et al. [3]

I. Arsenic toxicity and fitness cost of genetic constructs over bacterial chassis

We expect that the toxicity exerted on the biosensor will be minimal, since the concentrations we plan to use are several orders of magnitude lower than the reported concentrations as toxic to the chassis. Nadra et al. tested a biosensor that uses chromoproteins in concentrations of arsenic as high as 1000 ppb withoutreporting a significant toxicity [1], the same way, Zhuang et al. reported that the bacterial chassis started to be affected by toxic effects of arsenic on concentrations above 1874 ppb. [2]

Figure 1. Arsenic toxicity over bacterial chassis development.

We expect that in our design the As0 and As2 constructs will present a higher metabolic load for the bacteria in comparison with As4 and As5 The metabolic load that genetic constructs will have is closely related to their complexity and to the length of the amplification cascade. Wang et al. noticed a high metabolic load when the cascade was composed of three amplifiers. In contrast, by reducing the use of one amplifier, the construct will represent a lower load. It was also stated that a certain combination of amplifiers (i.e HrpRS-RinA) will result in lower metavolic effort for the microorganism. [3]

Figure 2. Metabolic load of our genetic constructs over bacterial chassis represented as a OD600 decrease.

II. Biosensor quantitative response characterization

The theoretical limit of detection (LOD) of the constructs are As0: ≥ 0.5 ppb , As2: ≥ 3 ppb, As4: ≥ 10 ppb and As5: ≥ 50 ppb, the experiments at this stage are of a semi-quantitative type without the use of instruments so we expect that the LODs we gonna report to be considerably higher, but the data collected will be very helpful for later stages. We also expect the background noise levels to be somewhat higher than those reported by Wang et al. since the sequence of our design does not contain additional ArsR binding sites (ABS) that would help mitigate this. [3] The lyophilization protocol on paper strips that we expect to test makes use of a protective solution that ensures bacterial viability up to 2 months after processing, we expect minimal loss of their detection activity under the appropriate storage conditions. [4]

Figure 3. Paper strip with the 4 biosensors lyophilized over the surface generating traffic light patterns after arsenic exposure.

III. Biosensor quantitative response characterization

At this stage we will work with the construct that presents an optimal response at concentrations ≥ 5 ppb of arsenic, we expect that it will be the As2 or As4 construct that fulfills this characteristic. As mentioned before, we expect that the background noise will be considerably high in comparison to constructs without amplifiers , especially in constructs with transcriptional amplifiers due to the absence of additional ABS sites [3], we also expect that the selected construct will have a linear response within the range of concentrations 5 - 100 ppb of arsenic and that the repeatability, probably lower than traditional methods, will allow the generation of a calibration curve with an acceptable error.

Figure 4. Expected calibration curve for As2/As4 construct response

IV. Biosensor quantitative response validation

In an attempt to test the specificity of the response of our biosensor we plan to expose it to 3 different metals (Fe, Hg, Mg) to observe if there is any type of non-specific response compared to the response to arsenic. We expect that there will be no significantly different type of response from the background noise against these 3 metals. [3] The only reported case of interference in this type of biosensor is antimony which, depending on the system and design employed, can generate even higher responses than arsenic itself. [5]

Figure 5. Specific arsenic response, reporter induction by other metals is negligible.

V. Biosensor response with real samples

Initially, we expect that the concentrations reported by the biosensor will be lower than those reported by spectrometric methods since this type of biosensors do not report total arsenic concentrations, but bioavailable concentration [6] and this heavily depends on several physicochemical factors of the sample and the culture medium. For example, we expect the bioavailability to be lower if the samples have a high concentration of dissolved organic matter. Similarly, if the phosphate concentration is high, the uptake of As(V) will also be significantly reduced [7], which would ultimately result in a considerably lower reported concentration than spectrometric methods. On the other hand, an advantage that our biosensor has when analyzing well water samples is the speciation of arsenic in these samples, as they do not have as much contact with oxygen as surface waters, the speciation of arsenic in these samples is predominantly As (III) and the whole-cell biosensors are especially sensitive to this form because it does not need an extra reduction step. [7]

VI. Tackling the background noise problem:

As mentioned before, one of the principal expected drawbacks is the considerable background noise (leakage) of our biosensor, especially in constructs with transcriptional amplifiers. We propose 2 potential solutions:

• Adding an extra ABS sequence into the inducible promoter pArs: this could be accomplished easily by performing PCR with specially designed primers and considering that the ABS sequences are relatively short (24 bp).

• Coupling with a degradation tag-TEV protease system: this is a special approach that takes certain advantage of natural protein degradation systems of the bacterial chassis. A short TEV cleavage site and a degradation tag is added at the end of the reporter (mRFPViolet) coding sequence, additionally, an arsenic responsive construct, responsible for the TEV protease expression, is coupled. Without arsenic in the medium the reporter is immediately degraded by bacterial proteases because it still has the degradation tag, thus lowering the background noise. With arsenic in the medium, the TEV protease is expressed and cuts the degradation tag allowing normal expression of the reporter.

References