Team:NEFU China/Results

Experiment

Magnetic bacteria AMB-1 has long intrigued researchers due to its ability to synthesize intracellular magnetic particles “Magnetosome”. However, Magnetosome has not been developed commercially, mainly because AMB-1 is difficult to cultivate and a high yield of Magnetosome has not been achieved.

We used Mms13 protein as a fusion partner to express any protein of interest on the membrane of Magnetosome. In order to achieve this, it is crucial important to culture AMB-1 and purify Magnetosome on a large scale. It has been reported that a pulsed magnetic field can enhance Magnetosome formation. So we design a device, the pulsed magnetic field generator (see device part) for AMB-1 incubation. Additionally, we used ultracentrifugation and electromagnet separator (see device part) for Magnetosome purification, then the membrane of Magnetosome was analyzed by transmission electron microscopy (TEM).

Fig.1 (A) TEM image of AMB-1 incubated for 4 days (negative control). (B) TEM image of AMB-1 incubated under a pulsed magnetic field for 4 days. (C)

TEM image of Magnetosome after purification by ultracentrifugation and electromagnet separator.

These results showed that a pulsed magnetic field enhanced the growth of AMB-1 and Magnetosome formation (Fig.1). It might exceed the chain of Magnetosome and change the homogeneity of the Magnetosome particles. Moreover, the culture time was reduced from 6-10 to 4-5 days. In addition, the membrane surrounding the Fe3O4 granules indicates that this purification procedure did not damage the phospholipid bilayer of the Magnetosome.

All these results demonstrated that, the culture method using a pulsed magnetic field generator is simple, rapid, continuous, and highly efficient for AMB-1 incubation and large-scale purification of Magnetosome, without break the membrane of Magnetosome.

Plasmids construction

Eleven expression vectors were constructed for genetic engineering. We transformed vectors into host cell E. coli by heat shock. The positive colonies were identified by colony PCR or mini-prep plasmid DNA for restriction enzyme identification. The figures show that all the vectors were constructed successfully.

The Hunter System：We constructed four vectors

Name Function
pET-14b-Pmsp3-EGFP Determine whether the Pmsp3 promoter can drive gene expression in AMB-1.
pET-14b-Pmsp3-Mms13-EGFP Prove whether Mms13 and EGFP can be successfully expressed in AMB-1, and observe Mms13 anchored on the membrane of Magnetosome.
pGex-4T-1-Spycatcher As the engineered AMB-1 bacteria grow very slowly, so we use Spycatcher-GST instead to confirm the interaction between Spycatcher and Spytag .
pET-14b-Pmsp3-Mms13-Spycatcher The recombined Mms13 and Spycatcher protein conjugates to Spytag and the complex can be simply co-purified by a magnet.

Fig.2 Colony PCR

The prey system: We constructed two vectors

Name Function
pET-14b-Spytag-TEV-Amilcp It conjugates to Spycatcher and they can be simply co-purified by a magnet.
pET-14b-Amilcp-intein-Spytag As the TEV protease is still present in the final elution of the recombinant protein while using the TEV system. We use intein to replace the TEV site, which can be cleaved by DTT.

Fig.3 Colony PCR

Gradient protein construction for modeling

In order to test the binding rate of Spycatcher and Spytag with different protein molecular weight we constructed 5 vectors successfully: 3×Z domain, 6×Z domain, 9×Z domain, 12×Z domain, 15×Z domain respectively.

Name Function
pET-14b-3×Z domain These are our gradient proteins and these parts will be used to test the binding rate of Spytag and Spycatcher with the protein we want to purify with different molecular weight due to the simplicity of adjusting the size of SPA Z domain.
pET-14b-6×Z domain
pET-14b-9×Z domain
pET-14b-12×Z domain
pET-14b-15×Z domain

Fig.4 Restriction enzyme identification

Functional Identification

We use Electromagnet separator designed by us (see device part) to purify Magnetosome. As shown in Fig.5, The electromagnet produces magnetic force, and all Magnetosomes were adhered to one side of the tube. This result proved that our device could work as expected.

Fig.5 Magnetosomes were attracted by Electromagnet separator

We use MagneGST-pull down to detect whether Spycatcher could bind with Spytag-TEV-Amilcp (STA) fusion protein from the Rosseta’s lysate, and then analyzed by SDS-PAGE. A covalent complex corresponding to the expected molecular weight (67KD) of Spycatcher-STA could be observed clearly. We can conclude that Spytag could form a high specific recognition and covalent conjugation between Spycatcher.

Fig.6 SDS-PAGE analysis of whether Spycatcher could pull down Amilcp fused with Spytag or not. (STA: 27KD, Spycatcher: 14KD, GST tag: 26KD)

Additionally, in order to know whether the Spycatcher, Spytag and blue pigement protein can bind together and localized on the magnetic beads, we used a microscope to detect the blue pigment. As the result was shown in Fig.7 a blue pigment signal can be detected on the surface of magnetic beads.

Fig.7 Observe the surface of magnetic beads through microscope (400×)

Fig.8 SDS-PAGE analysis of Spycatcher pull down 3×Z domain fused with Spytag at different incubation time. (3×Z domain: 24KD, Spycatcher: 14KD, GST tag: 26KD)

Fig.9 SDS-PAGE analysis of Spycatcher pull down 6×Z domain fused with Spytag at different incubation time. (6×Z domian: 46KD, Spycatcher: 14KD, GST tag: 26KD)

Fig.10 SDS-PAGE analysis of Spycatcher pull down 9×Z domain fused with Spytag at different incubation time. (9×Z domian: 67KD, Spycatcher: 14KD, GST tag: 26KD)

Modeling & Proof

Modeling

Introduction

Theoretically, many factors can influence the interaction between Spycatcher and Spytag, including temperature, pH, buffer types and protein molecular weights. To simplify the modeling process, we just take protein molecular weight as a major variable factor because other factors are normally controllable.

In this part, our models can not only help determining the relationship between molecular weight and optimal incubation time, but also guide us in prioritizing important factors.

Ordinary differencial equation

We used the available data to create a kinetic equation for second order reaction. After a series of data transfer and fitting, we can create a graph that mirrors the correlation between molecular weight and incubation time.

1.The construct of equation

The interaction between Spycatcher and Spytag should follow the second order reaction kinetics, so we created a differential equation. It is listed as below:
According to this equation, we can get following additional equations:
In those equations:
a—— the initial concentration of reactant
x—— the concentration of reacted Spycather
t—— the reaction time
λ——the binding rate between Spycatcher and Spytag
k——the second order rate constant

Obviously, we can build the correlation between the protein concentration and reaction time. In order to simplify the description, we reciprocally use free Spycatcher’s concentration and reaction time as two variables. As a result, equation (1) shows a linear relation between 1/c and t, where c represents the concentration of free Spycatcher.

2.Experimental Data

Step 1: (calculate additional data)

According to the experimental data using Spytag tagged proteins with different molecular weights, we observed that data show a linear relationship fitting the Equation (1). This proved our prediction. Thus, we used the least square fitting to generate the following graphs (Fig.1, 2 and 3) and specific linear equations (Equation 3, 4 and 5).

Fig.1 The relationship between 1/c and t of Spytag-3×Z domain were plotted against time in straight lines.

Fig.2 The relationship between 1/c and t of Spytag-6×Z domain were plotted against time in straight lines.

Fig.3 The relationship between 1/c and t of Spytag-9×Z domain were plotted against time in straight lines.

These graphs in Fig.1, 2 and 3 for the relationship between 1/c and t of proteins with different molecular weights were plotted against time in straight lines. Then, adding arbitrary concentrations into the equation under each graph, we can get the corresponding time. This is intended to calculate additional data for the reactions of Spytag-3×Z domain/Spytag-6×Z domain/Spytag-9×Z domain with Spycatcher in triplicate at 37 °C for 1, 3, or 5 min, shown as the linear part of the reaction.

Step 2: (find out the optimal incubation time)

We then put data of different molecular weights into a graph that describes the correlation between binding rate of Spycatcher and Spytag, and time for three different molecular weights. In the following graph (Fig.4), we can find out optimal time for different molecular weights.

Fig.4 This graph descirbes the conjugation of Spycatcher and Spytag with incubation time.

From the graph, we can find that conjugation of Spycatcher and Spytag increases in proportion to incubation time, which fits with Equation (2). Take 50% as a standard binding rate, we can calculate that Spytag-3×Z domain, Spytag-6×Z domain and Spytag-9×Z domain’s optimal incubation time that are approximately 5, 7.8 and 12.6min, respectively.

3.Fitting

Obviously, the relationship between protein molecular weight and optimal incubation time is not linear. Based on data distribution, we predict that the trend or our data should fit in a polynomial equation. Due to quick proportional increase between molecular weight and incubation time in the beginning and tends to saturate with the time, we choose quadratic polynomial as the basic equation to fit our data.

Fig.5 This graph describes the relationship between protein molecular weight and optimal incubation time.

We noticed that this graph fits well with the experimental data. (R-square: 1, SSE: 5.099e-27)
Meanwhile, we can get an equation that describes the relationship between protein molecular weight and optimal incubation time:
Using the graph and equation above, we can easily predict the optimal incubation time in future for any recombinant protein.

Analysis

The results demonstrated above are under the condition that external environment is controllable. In actual practice, many factors can affect the process of protein purification. We need to determine which is the key factor and focus on optimizing it. For this purpose, we introduce the method called sensitivity analysis that deals with the change of model results when a parameter changes. By comparing the “sensitivity coefficient” of each parameter, we can prioritize important factors of purification.

A common method is the Extended Fourier Amplitude Sensitivity Test (EFAST). It computes the “main effect” contribution of each input factor to the variance of the output, a quantity that later indicated as “sensitivity coefficient”. If one of the factors (i.e. the ith factor) has strong influence on the output, the oscillations of y at frequency wi shall be of high amplitude. This is the basis for computing a sensitivity measure, which, for factor, is based on the coefficients of the corresponding frequency and its harmonics. It has been widely used and has less calculation. The main formula of this method is shown as following:

Using the four steps we can determine the factor that should be optimized firstly.

Conclusion

In summary, we have generated equation and graphs that can be used to easily predict the optimal incubation time for any recombinant protein.

Future application

The AMB-1 culture and prokaryotic expression processes of our experiments have lasted for nearly two months. We tested various approaches, but the growth rate for AMB-1 bacteria remains low. To date, we have completed the transformation of ten vectors constructed by us into E. coli, and designed two devices, by which we can enhance AMB-1 bacterial growth and facilitate Magnetosome purification. AMB-1 bacterial transformants grow very slowly and thus we have not use them to express Spycatcher due to limited time. To test the principle of our design, we first constructed and expressed GST-Spycatcher and used Glutathione-conjugated magnetic agarose to purify Spytag-tagged proteins. Employing this system, we have expressed proteins with different molecular weights to test their purification efficiencies, and established a mathematic model to represent the correlation between protein molecular weight and incubating time to reach maximal binding of Spycatcher and Spytag.

We will optimize medium components and culturing conditions of AMB-1 in order to produce Spycatcher-linked Magnetosomes in an efficient way and then in a large scale. In our future studies, we will integrate the essential genes for AMB-1’s Magnetosome production into E. coli that can grow very quickly and thus can be a better host bacterial strain than AMB-1 to produce Magnetosomes. In the system using Magnetosome, we will evaluate the effects of other parameters on purification efficiency, including temperature, pH and incubation time. These assessments will help us to expand our purification system to a large scale of production. At this point, we will consult biotechnological manufactures and scientific researchers for comments and advice in order to improve our protein purification system. The success of these planned research experiments and related processes demonstrated above will lead to the invention of a protein purification kit.