Design

Aiming at specific detection and sterilization of S. aureus, we designed an antimicrobial peptide complex using the V8 protease cleavage site as a linker between our antimicrobial peptides. V8 protease is an extracellular protease of S. aureus that cleaves peptide bonds exclusively on the aspartate (D) and glutamic acid (E) amino residues. In our antimicrobial peptide complex, we combine one kind of antimicrobial peptide. The main antimicrobial peptide we planned to use the human AMPs, such as hBD-3, HNP-1, or LL-37.

Redesign

However, IDT company can’t produce repetitive sequences. Interestingly, we found that two different AMPs have the synergistic effect for better sterilization in the literature, so we redesign that with two different AMPs. Besides, we selected an AMP called lysostaphin to be the main AMP component, which can hydrolyze the cell wall of S. aureus. We also combined it with other AMP, at this time, not only hBD-3, HNP-1, and LL-37, but Histatin 5 and Ranalexin became the AMP component. Different from the first design, we tried to use different AMP and nucleic acid to increase the randomness, making our sequences work.

Fig.3 The evolution of AMPC design (a) first designed AMPC (b) redesigned AMPC

Build & Test

We put the bacterium into 37°C incubator. However, we found that the protein expression was not obvious, so the sample was changed to be incubated at 16°C to facilitate stable expression of the protein. According to the SDS-PAGE analysis results, we divided protein expression into three categories, soluble , insoluble and not induced. Therefore, we try to purify the protein by the method of solubilization of inclusion body to break this barrier, so that it could reach the bacteria and function.

Learn

At the purification part, we tried hard to get the proteins and do the disc diffusion assays. We have tried eight proteins and found that protein No.3 and No.7 being the most effective. Next, we focus on protein No.3 and No.7, cultivating and purifying them repeatedly. After that, we adjusted their concentration and redesigned the experiment in order to enhance the antimicrobial effect.

See Result - Microneedle page and Result - AMPC page for detail information.

Design

Our model purpose was to predict the effective time of our system. That is, we’re trying to see how long the AMP will be effective after the AMPC release from the microneedle and be cleaved by V8 protease. We’ve tried to get the reaction rate constant of the reaction of that AMPC to become single AMPs, the minimum inhibitory concentration of different single AMP, and assume a value to the degradation rate of a single AMP according to experiments. Last, we transfer the chemical reaction to a differentiated equation and use python code to draw the drug concentration-time curve.

We designed the formula and assumed two different reaction rate constants for the different site. we assumed k_1 is the first V8 protease cleavage site, and k_2 is the second one.

Preference sites are represented EL, ED, EF, and other sites are EG, EA, and EK.
(’’EL’’ means E and L are P1 and P1’ sites respectively.)
Next, we would need two more variables for our model. One is the MIC values of each AMP, the other one is the degradation rate of a single AMP.

Bulid

MIC Values

We found the minimum inhibitory concentration of each AMP, we check whether the MIC is reached to judge whether our drug is efficient.

Degradation Rate

Because we couldn’t find more related research about degradation rate, we refer to the research paper that LL37’s half-life time is 1 hour in cell as the AMPs half-life time. Then, set the parameter in the differential equation to draw the concentration-time curve.

Python Code

In fact, we’ve completed the code and tried to set the default value randomly based on the 2021 CCU_Taiwan team’s result and some papers before we get the real data of each reaction rate from the experiment to see if the result is reasonable or not. However, we think there’s another way to simulate realistic circumstances. We then talked with seniors and redesigned our model. Full source code can be found in our GitLab and more design details can be seen on the Model page.

Learn

Che-Kang Chang who is NTHU 2016 member advised us to apply the enzyme kinetics in our model, making our model more persuasive. We looked into some tutorials and examples to start redesigning our work.

Redesign

We modified our model design and applied the Enzyme Kinetics formula.

Enzyme Kinetics has four hypotheses:

1. [ES] is constant. So k_1[E][S] = k_2[ES] + k_3[ES] always holds.
2. All enzyme concentration [E_t] = exist alone [E_f] + [ES]
3. The reaction rate is decided by k_3[ES].
4. The maximum reaction rate V_max is the situation in which all the enzymes are turned into [ES]. V_max = k_3[E_t]

Keeping the MIC values and the assumption of degradation rate in the original version, we modified more in the reaction rate constant part among the three main variables in our model.

Bulid

We referred to the paper about V8 protease to set the k_3. Because [S] is much larger than [E], we consider that k_3 is much larger than k_2 and assume the k_2 value is 0. Therefore, we can get the k_1 value by definition K_max = (k2+k3)/k1.

In the new version of python code, the skeleton was basically the same but was changed the equations and had three reaction rate constants this time. Please check out the full source code in our GitLab.