Summary
In recent years, food safety problems occurred frequently, and food poisoning cases caused by
pathogens are particularly typical. Fast and accurate food safety detection method for pathogenic
bacteria is a good way to quickly confirm the cause of food poisoning. Therefore, our team chose to
design a detection device for potential pathogenic Escherichia coli (E. coli, EC), Staphylococcus
aureus (S. aureus, SA), Vibrio parahaemolyticus (V. parahaemolyticus, VP) and Salmonella enteritidis
(S. enteritidis, SE) in food. Our team built a test kit equipped with core components for the test:
paper-based sensors, small incubators and quick sampling devices needed during the experiment to
achieve household food safety testing.
During the project, we successfully constructed eight plasmids to express four receptor-binding
proteins [RBPs: TFP (tail fiber protein), gp15, CBD(C-terminal cell-binding domain), and TSP
(tailspike protein)] and their corresponding fluorescently labeled GFP-fusion proteins (GFP-RBPs).
TFP, gp15, CBD, and TSP can specifically bind to the target bacteria EC, VP, SA, and SE,
respectively. GFP-fusion RBPs can trace the binding of RBPs with the bacteria via fluorescence and
are used to validate the specificity of RBPs. RBPs are conjugated with nanomagnetic beads (MBs) to
form magnetic-protein complexes (RBP-MBs) to achieve separation and enrichment for specific target
bacteria by magnetic force. The successful construction of RBP-MBs was verified by calculating the
separation rates of RBP-MBs for different bacteria strains.
RBPs, GFP-RBPs and RBP-MBs of target bacteria
| Bacteria |
RBP |
GFP-RBP |
RBP-MBs |
| EC |
TFP |
GFP-TFP |
MBs-TFP |
| VP |
gp15 |
GFP-gp15 |
MBs-gp15 |
| SA |
CBD |
GFP-CBD |
MBs-CBD |
| SE |
TSP |
GFP-TSP |
MBs-TSP |
Part list
Procedure
Goal 1 : Produce RBPs and identify their activity
1 Plasmid Construction
We designed our functional parts TFP, gp15, CBD, and TSP and respectively cloned them into pET28a
backbone plasmid chemical synthesized by Tsingke Biotechnology Co., Ltd. As mentioned on our design
page, these plasmids are used to express TFP, gp15, CBD and TSP which can respectively bind to and
separate EC, VP, SA, and SE from food samples. For safety concerns, we chose a relatively safe EC
strain K12 MG1655 for laboratory experiments.
We used Gibson assembly method to construct pET28a-TFP, pET28a-gp15, pET28a-CBD, and pET28a-TSP
plasmids.
The gel electrophoresis results (Figure 1 to 3) showed that the gp15 was 561bp, the CBD was 604 bp,
and the TSP was 1993 bp, as expected. In addition, we confirmed the results by sequencing. As for
TFP, we had the plasmid which was already cloned with TFP, so we didn’t electrophorese it.
Figure 1 Nucleic acid gel electrophoresis results of gp15.
Figure 2 Nucleic acid gel electrophoresis results of CBD.
Figure 3 Nucleic acid gel electrophoresis results of TSP.
2 Protein expression test
SDS-PAGE electrophoresis was used to check the expression of four RBPs (TFP: 64.1 kDa, gp15: 25.4
kDa, CBD: 12.8 kDa, TSP: 62.6 kDa). As shown in Figure 4 to 7, the proteins have been successfully
expressed and purified.
Figure 4 Protein SDS-PAGE electrophoresis results of TFP.
Figure 5 Protein SDS-PAGE electrophoresis results of gp15.
Figure 6 Protein SDS-PAGE electrophoresis results of CBD.
Figure 7 Protein SDS-PAGE electrophoresis results of TSP.
3 Binding ability test of RBPs
After the expression and purification of four GFP-RBPs, we measured the absorbance of them by optical
density (OD) values at 562 nm (OD562) and converted OD values into concentration of GFP-RBPs by the
function shown in Figure 8. The concentrations of GFP-RBPs are listed in following table .
Figure 8 A standard curve of optical density (OD) values at 562 nm versus protein
Concentrations of GFP-RBPs
| Fusion protein |
concentration |
| GFP-TFP |
420μg/mL |
| GFP-gp15 |
722μg/mL |
| GFP-CBD |
210μg/mL |
| GFP-TSP |
425μg/mL |
We mixed the GFP-RBPs with relevant bacteria and measured the fluorescence value of the cells with a
microplate reader. As we expected, GFP-TFP had a good specificity toward EC (Figure 9) and so did
GFP-gp15 toward VP (Figure 10), GFP-CBD toward SA (Figure 11), and GFP-TSP toward SE (Figure 12).
These results confirmed that our RBPs could bind strongly to the corresponding target bacteria
respectively.
Figure 9 GFP-TFP binds strongly to the cell surfaces of EC.
Figure 10 GFP-gp15 binds strongly to the cell surfaces of VP.
Figure 11 GFP-CBD binds strongly to the cell surfaces of SA.
Figure 12 GFP-TSP binds strongly to the cell surfaces of SE.
Goal 2 : Find appropriate chromogenic substrates for bacteria and draw Blue
value-concentration curve
1 Literature research
By searching literature, we found out specific enzymes in the four bacteria strains that can catalyze
reactions of certain chromogenic substrates. The chromogenic substrates and their color variation
due to such enzyme catalysis for the four strains are shown below.
Chromogenic substrates for bacteria and their color variation due to
enzyme catalysis
| Name |
Bacteria |
Enzyme |
Color variation |
Concentration |
| CPRG |
EC |
β-gal |
yellow→red |
25mg/mL |
| X-β-glu |
VP |
β-glu |
colorless→blue |
100mg/mL |
| pNPG |
SA |
α-glu |
colorless→yellow |
100mg/mL |
| magenta caprylate |
SE |
esterase |
colorless→violet |
100mg/mL |
2 Experiments of chromogenic reactions
For each chromogenic substrate, we mixed it with colistin and a variety of bacteria strains (SA, SE,
ST, VP, PA and EC) to check if the color reaction could specifically identify target bacteria
strain. A control group (C) containing only chromogenic substrate and colistin was set up for each
experiment. The experiments showed that our paper-based sensors based on these chromogenic
substrates could successfully detect EC, SA, and SE. VP experiment failed at first, and by changing
colistin into polymyxin B and making adjustment of experiment conditions, we also successfully had a
sensor with X-β-glu for detecting VP.
2.1 Experiments for identifying EC, SA and SE
Chromogenic substrates CPRG and pNPG showed specific color changes for EC and SA respectively,
distinguishing the target bacteria from other strains (Figure 13 to 14). Although the specificity of
the chromogenic substrate MC was not specific to SE, the specificity of our paper-based sensor can
still be ensured due to the specific separation of the target bacteria by RBP-MBs before sensor
detection (Figure 15).
Figure 13 Specificity of CPRG to EC
Figure 14 Specificity of pNPG to SA
Figure 15 Specificity of MC to SE
2.2 Experiments for identifying VP
For the experiment of chromogenic substrate X-β-glu for VP detection, we failed at first. No color
change was observed after mixing X-β-glu, bacteria and colistin. We proposed several possible
hypotheses for the failure.
Hypothesis 1: Volume of reactants was not appropriate.
We changed the volume of colistin, X-β-glu and PBS from 0.1μl,0.1μl and 20μl to 0.2μl, 0.5μl and 10μl
to increase the concentration of colistin and chromogenic substrate. This adjustment did not improve
the results.
Hypothesis 2: Reaction time was not enough.
One “failed” tube began to turn blue after one day (Figure 16), we suspected that the reaction
required more time than we expected.
Figure 16 After 1 day, tubes containing X-β-glu and bacteria began to show light blue
Hypothesis 3: Quality of chromogenic substrate was poor.
We tested chromogenic substrates from different companies with bacteria handled with ultrasonication,
and found that X-β-glu from Shanghai Yuanye Bio-Technology Co., Ltd and Shanghai Aladdin Biochemical
Technology Co., Ltd. had better chromogenic ability. But although the chromogenic substrate turned
blue in the tubes, both of them failed to turn blue on the paper base. Eventually, chromogenic
substrate from another company (Bide Pharmatech Co., Ltd.) solved the problem.
Hypothesis 4: The decomposition agent was inappropriate.
The colistin might have a problem in breaking the cytolemma of VP. We tried several other
decomposition agents for VP, including D-cycloserine, metronidazole, ampicillin, kanamycin,
chloramphenicol, thiamphenicol, erythromycin and polymyxin B (Figure 17). Only metronidazole and
polymyxin B succeeded in turning blue in the reaction. But metronidazole turned the control group
blue as well. We found out that it was due to the hydrolysis of X-β-glu, from Shanghai Aladdin
Biochemical Technology Co., Ltd., in metronidazole, a reaction turning the solution blue without
bacteria. Luckily, polymyxin B didn’t make X-β-glu hydrolyze, so we adopted it as the decomposition
agent for VP.
Figure 17 Test of different decomposition agents
Hypothesis 5: The activity of enzyme in some breeds of VP didn’t work well.
To take more strains of VP into comparison, we chose VP49, VP123, and VP17082. The results showed
that VP17082 was the most sensitive breed toward polymyxinB + X-β-glu (Figure 18).
Figure 18 Chromogenic reactions of different breeds of VP
Final solution:
At last we used X-β-glu from Bide Pharmatech Co., Ltd. We mixed 0.5μl of it and 0.2μl of polymyxin B
together with 10μl VP, the solution would turn blue after 2 hours normally (Figure 19). The core
problem of the previous failure was the quality of chromogenic substrate, and we changed the formula
to make sure the bacteria were fully cracked.
Figure 19 Successful results of chromogenic reactions of VP
3 Experiments for determining the concentrations of bacteria
3.1 Chromogenic reactions of different concentration of bacteria
We noticed that the color variation of chromogenic reactions was correlated to the concentration of
target bacteria, so we analyzed RGB values of the color to determine the concentration of target
bacteria in the samples. Using the RGB values, we were able to draw Blue value-concentration curves
to calculate the bacteria concentrations which are helpful for further application of our product in
reality.
We diluted the solution of bacteria (10⁹ CFU/mL) to 10⁸，5×10⁷，10⁷，5×10⁶，10⁶，...... ，10¹ CFU/mL and
added 10μL of diluted solution to a paper-based sensor one by one, then preserved at 37℃ for 2h. Our
experiments showed that the degree of color change of the sensors for each strain was all positively
related to the concentrations of bacterial solution.
When we detected the RGB value, we found that Blue value was the most sensitive value for the
concentration of bacteria, so we adopted Blue value to draw the curve at last. The curves for four
kinds of bacteria are listed below (Figure 20 to 23). Using the blue value, the lower limit in the
detectable concentration for all bacteria could reach around 10⁶ CFU/mL except for EC whose color
changes were quite apparent from the concentration of 7.5×10⁵ CFU/mL
Figure 20 The blue value-concentration curve for EC
Figure 21 The blue value-concentration curve for VP
Figure 22 The blue value-concentration curve for SA
Figure 23 The blue value-concentration curve for SE
Goal 3 : Construction of RBP-MBs and measurement of the separation rates
We combined magnetic beads with RBP using condensation reaction of -NH2 on protein and -COOH on
beads. Then we mixed RBP-MBs with bacteria of 103 CFU/mL and separated the bacteria by magnet. After
oscillation and resuspension, we pasted the solution on LB Broth and counted the number of single
colony before separation (sucked bacteria of 103CFU/mL of the same volume and pasted) (Nbefore) and
after separation (Nafter). The separation rate was calculated by Nafter/Nbefore. The results are
shown below. Our results showed that all RBP-MBs successfully separated target bacteria from
samples, with MBs-TFP, and MBs-gp15 having high capture efficiency over 85% in mixed bacteria
solution and over 95% in purified bacteria solution.
Separation rates of RBP-MBs for target bacteria
