In system one, we aim at boosting NAD+ levels for its critical role in mitochondrial function. We choose Lactobacillus Plantarum L168 as the chassis and pLDHLH673 as the vector. The gene circuit begins with the promoter ldhl, followed by the PncA gene, whose overexpression enables our bacteria to produce additional NAD+ for the host.

The DNA Level

$Lactobacillus Plantarum$ L168, a strain we obtained from our PI, Xingyin Liu's lab, has been proven to be able to alleviate social behavior deficits in animal models of autism. It carries the PncA gene itself and our aim is to amplify the gene, insert it into the vector, and transform the vector back to the L168 which enables the engineering bacteria to overexpress the PncA encoding protein. Thus, we designed primers, performed PCR, and ran gel to verify the successful amplification of the PncA gene.

As shown in the figure above, after PCR, the bands we obtained are near 600 bp, which was consistent with the target band (621 bp). This further indicated that we have successfully amplified the PncA gene.

We then cut bands out of the agarose gel and purify the DNA samples. Using pLDHLH673 as the vector, we constructed the pLDHLH673-PncA plasmid expression system, the main components of which is the PncA gene.

To prove our construction is successful, we performed enzyme digestion verification on the plasmid, adding restriction endonuclease NdeI and XbaI for double enzyme digestion, and ran the agarose gel to verify.

As shown in the figure above, after enzyme digestion, the lower bands we obtained are near 600 bp, which was consistent with the target band (632 bp). The upper bands we obtained were higher than the marker and were near 5500bp, which was as we expected. This further indicated that have successfully constructed the pLDHLH673-PncA plasmid.

Although the construction was quite successful, we encountered difficulty in the electric transformation step. In fact, our original design used pMG36e as the vector. To amplify the plasmid, we need to transform it into Escherichia coli DH5a. However, every time we performed the transformation, extracted the plasmid, and ran the gel, we found no expected band on the gel. Considering that pMG36e was a low-copy plasmid, we switched our plasmid extraction kit from FastPure Plasmid Mini Kit (Vazyme) to TIANprep Midi Plasmid Kit (TIANGEN). However, the result was the same. Then, we changed the condition during the heat shock of the plasmid but the result showed no change. In this case, we had to change our vector to pLDHLMCS then to pLDHLH673, and finally constructed the target plasmid successfully. By this time, however, with few days left to finish the project, we were trapped in the electric transformation. After careful thinking and consulting our PI, we came to several explanations. First, the plasmid extraction kit we used was ineffective. In the lysing bacteria step, the color of the solution should have become clear after adding the reagent, but it remained unchanged. Second, the chassis bacteria we chose, Lactobacillus plantarum, are gram-positive bacteria that have different physiological characteristics from the gram-negative bacteria commonly used in plasmid extraction.

In fact, in the design stage, we discussed about using the chassis bacteria that are more common in synthetic biology. Due to its capacity to ameliorate social behavior deficiencies in animal models of autism, however, we elected to continue using it. Although we have not fully verified the functionality of this system, we believe that in the future, when given enough time, we can demonstrate its functionality and viability.

Function Verification

In our pre-experiments, we knocked out the pncA gene in Lactobacillus plantarum L168 to determine whether it would produce fewer NAD+.

As shown in the figure above, the bacteria with pncA knocked out produce fewer NAD+ compared with that of Wildtype of LP-L168 supernatant through mass spectrometry assay, which implicated that our experimental design is reasonable. Despite the fact that this result does not directly demonstrate the NAD+-boosting ability of our engineering bacteria, it does provide indirect evidence.

In system two, we aim at eliminating heavy metal ions such as Hg2+ and Pb2+ in the intestine of autistic children. We choose Escherichia coli 1917 as the chassis and pET-28a(+) as the vector. The gene circuit begins with two sensors which can be induced by Hg2+ and Pb2+ respectively and follows the Lpp-OmpA-MT gene, which encodes the protein that can anchor to the bacterial outer membrane and chelate the heavy metal ions. For biosafety, a kill switch that can be induced by arabinose is assembled downstream.

The DNA Level

We constructed the pET28a-MT plasmid, transformed it into Escherichia coli Nissle 1917, and extracted the plasmid to prove the success of our transformation. First, we performed enzyme digestion verification by adding restriction endonucleases SphI and Xhol for double enzyme digestion and running an agarose gel. Second, we performed the plasmid PCR verification, by designing primers to amplify the target segment and running an agarose gel.

As shown in the figure above, after enzyme digestion, the sample from the extracted plasmid ran out bands near 2000 bp, which was consistent with the target band (2399 bp). After plasmid PCR, the sample from the extracted plasmid also ran out bands near 2000 bp, which was consistent with the target band (2399 bp). This further indicated that our plasmid transformation was successful.

The Protein Level

To verify the expression of our functional genes Lpp-OmpA-MT, we induced the bacteria at different concentrations of lead ion salt solution. After lysing the bacteria and performing the centrifugation, we took the supernatant or precipitation at the first or third hour. Then we ran the SDS-PAGE to check MT protein level for all samples

The pre-stained protein ladder is from 10 to 180 kDa. We induced the protein expression at different concentrations of lead ion salt solution (112 mg/L, 61mol/L, and 30.5 mol/L) and used double distilled water as the negative control. As shown in the figure above, the bands we obtained were nearly 40 kDa, which was almost twice the molecular weight we expected. After careful consultation with our PI and searching literature, we found it has been reported that MT easily form a dimer in vivo. Hence, we believed this indicated that Lpp-OmpA-MT proteins would combine and form a dimer in the bacteria, which confirmed the expression of our target protein.

Then, using ImageJ, we quantitatively analyzed the band brightness and obtained the following results: (ImageJ analysis: measure the brightness of corresponding lanes, minus the background brightness, and compare it with the marker to get the protein expression of each lane)

As shown in the figure above, the protein expression was greater in the precipitation than in the supernatant and was greater after three hours than after one, which was consistent with our expectations. Using the above data, we solved the model of inducer-protein expression. More information can be found on the Model page.

Function Verification

To determine whether our engineering bacteria can grow in the presence of heavy metals, we measured the growth curve. The most frequent model for describing the growth kinetics of bacteria is the logistic equation, which is frequently used to fit growth curves based on optical density (OD) in order to comprehend the dynamics of biological systems. Its differential expression is $\dfrac{1}{N}\dfrac{\partial N_t}{\partial t} = r(1-\dfrac{N_t}{K})$, by variation, we can obtain$N_t = \dfrac{K}{1 + \frac{K - N_0}{N_0} \times e^{-rt}}$

We use MATLAB to fit the data and obtain the figure and formulas shown below.

We used untransformed Escherichia coli Nissle 1917 (wild type) as the control. The concentration of Pb(NO$_3$)$_2$ solution is 112 mg/L. As shown in the figure above, the growth of our engineered bacteria will not be affected by heavy metals. Also, at the logarithmic phase, when the bacteria are most sensitive to changes in the environment, our engineering bacteria grew faster than the wild type, which indirectly proved that our engineering bacteria have the ability to eliminate heavy metals.

As to prove that our engineering bacteria can eliminate heavy metals directly, we put the engineering bacteria into the metal salt solution Pb(NO$_3$)$_2$ and cultured them for 2 hours. Then, we took the supernatant and measured the concentration of the metal salt solution using atomic absorption photometer.

We used untransformed Escherichia coli Nissle 1917 (wild type) as the negative control. As shown in the figure above, the transformed bacteria eliminated more heavy metals than the wild type, and the difference was significant($P$$<$0.05). This indicated that our engineered bacteria were able to eliminate heavy metals. In the future, we will adjust the reaction time with the metal salt solution to further optimize the experiment condition. Kill Switch Verification

To validate the function of our arabinose-induced kill switch, we inoculated 3% fresh bacterial solution in the culture medium and added 0 mM and 10 mM arabinose respectively as the control and experimental groups. Every 1 hour, 2 hours, and 4 hours, we draw 100 ul of bacteria to the smear plate and cultured at 37 ℃ overnight. The colony counts are shown below.

As shown in the figure above, in one, two, and four hours, the number of colonies with arabinose induction are all lower than that without arabinose induction. This indicated the effectiveness of our kill switch.

In system three, we aim at increasing the bioavailability of our engineering probiotics. We wrap them in alginate, coat them with whey protein and perform freeze-drying. Microencapsulation enables our engineering bacteria to go through the human digestive tract environment with less damage and fatalities. The experiments are conducted using the engineered probiotics in systems one and two, respectively. For convenience, our results are referred to by their chassis names.

Morphology

To observe the morphology of the microencapsulated bacteria, we took electron microscope photos. Since the morphology observation was considered to be the same for different chassis, we used Escherichia coli as an example.

As shown in the figure above, the black structure is the bacteria and the greyish mist surrounding the bacteria is the coating of sodium alginate.

Then, we investigated the microencapsulation yield and effect of Lactobacillus Plantarum and Escherichia coli, respectively.

Results of System 4