Team:Peking/Measurement
Measurement
Overview
The overall state of bacteria has proved to be one of the most essential factors that affect operation of functions of genetically engineered bacteria. Herein, we developed a CRISPR-based replication interference system to bridge the synthetic parts and bacterial overall states. A rounded characterization system, including multiple measurement methods, is well developed as a full-scale quantitative description of E. coli general states, while three parameters are chosen as measure: cell growth, cell morphology and irrelative protein productivity. Besides common biochemical methods, unique 4-channel and multi-view microfluidic system is developed for observation of the cell division and cell morphology. Another designed microfluidic chip enables us to directly record cell adhesion with different flow velocity. Nucleo-cytoplasmic ratio is measured using a mature nucleoid staining method followed by observation under laser scanning confocal microscope. Each approach we utilized is highly repeatable and portable and we helped Tsinghua iGEM with observation of cell division using our microfluidic system. Most of observed phenomenon from one method can be proved by another, to ensure the reliability of our results. High robustness and precision of our measurement system helped us build a quantitative insight into the inner link among different parameters of E. coli, which provides a potential characterization and evaluation framework to extend to any artificial bacteria system.
Appropriate measurements reveal hidden but valuable information
When studying cell growth, the most widely used parameter is cell growth rate. In a narrow sense, growth rate is characterized by absorbance at 600-nm light (OD600). While we usually believe the doubling time calculated from OD600 is a good representation of growing of bacteria concentration, it actually reflects the accumulation of total biomass, and only focuses on population-level behaviors.
Throughout June and July, we took effort in characterizing the effect of our system with OD600, measured by microplate readers. The extremely frustrating fact was that little differences were observed between experiment groups and control groups (Figure 1A). Very fortunately, we came up with the idea of verifying the state of those bacteria under microscope(Figure 1B). Observation of the astonishing morphology of experiment group inspired us to re-consider the methods we took.
Figure 1.Biomass of CRISPRri-implanted bacteria (A) Time-course OD600 of bacteria with CRISPRri targeted to M+ box under different arabinose concentration, tested by microplate reader.(B) Microscope imaging of bacteria under our system'scontrol. Significant morphology was observed.Bacteria were incubated for 8h in M9 culture with 0% and 0.5% Ara respectively.
Ever since realizing the significance of characterizing single cell behavior in our project, imaging becomes one of our prioritized measurements.
We first cultured cells in LB in ordinary way, extracting a portion of the solution once two hours, taking pictures of the cell under traditional microscope. Since we realize that cell shape elongation signifies compromise in cell number, considering the unchanged OD600, hemocytometer (counting chamber commonly used to calculate the number of blood cells) were also used to give quantitative description of total cell number (Figure 2).
Figure 2.Cell number change of bacteria with CRISPRri targeted to M+ box under induction of different arabinose concentration. Cell counting is finished in hemocytometer every 2 hours.
While these experiments yielded valuable information, the measuring process requires long-time engagement of labor, and the number of timepoints tested were of strong limitation. Conversation with professor Luo (Peking University) in the field of bacteria physiology enlightened us to consider more advanced and automated devices, including microfluidic chips.
Figure 3A.Structure of microfluidic chip used for microscopic imaging. There are 4 main channels on one chip and thus 4 groups of bacteria can be tested each time. Small chambers are arranged along the both sides of main channel, which serves as a microscopic view. 8 view fields are taken for each group, which are at the side of the same main channel. Structural parameters are shown on the left while a real microscopic image of a main channel is shown on the right.
Utilization of microfluidic chips largely facilitated our measurements, since it provided many advantages. First, it enables constant experiment environment, including nutrient contents and antibiotics concentration, preventing measurement from being disturbed by nutrient consumption or reduction in the potency of antibiotics. Second, all repeat groups are under the effect of identical external conditions, which are convenient for multiple sets of repetitive experiments. Third, height of the chamber (1.2 μm) is slightly larger than the diameter of E. coli, aligning the bacteria to form a single layer in the chambers, which is convenient for observing and tracing each single cell. Fourth, fluidic inputs and imaging of each microscopic view is completely automated, enabling our collection of larger mass of data.
Exploiting all these advantages, we generated abundant data (Figure 3B).
Figure 3B.Microscopic GIFs of bacteria transformed with CRISPRri system targeted to different boxes from microfluidic system. Transformed Top 10 strain is transferred to M9 medium in the ratio of 1:10 after overnight cultivation in LB medium. About 2 hours after transferring, bacteria in its log phase is precipitated by 5000-rpm centrifuging for 4 min and is re-suspended by M9 medium arabinose. Re-suspended bacteria are injected into the chip and observed and recorded continuously for 10 hours, under constant flow of 1 mL/16 hours.
We used imageJ for image processing to automatically count cell number in the chamber of mircofluidic chips.(Figure 3C) Spatial filtering,background subtraction,automatic segmentation bsed on threshold adjusting were employed in sequence during the image processing.
Based on these image processing methods, we generated large mass of quantitative results from these raw data. Key results include increase in cell number doubling time and length which corresponds to inducer concentration (Figure 3D).
Figure 3C.3D.3E.(C) Image processing with imageJ.(D)Cell number doubling time accessed from image processing.(E)B)Morphology change can be tested by flow cytometry. Elongation of cell body can be observed through increase of SSC-A. Slight SSC increase in PolyA group might be attributed to cell stress led by overexpression of dCas9. R1+ group shows a remarkable increase in SSC-A as the arabinose concentration increases. However, flow cytometry is argued to be indirect test of cell morphology change. SSC change is not absolutely positively correlated to cell shape elongation and flow cytometer actually has a upper limit for cell size which is around 150 μm. Therefore, flow cytometry is a qualitative but high-throughput characterization of cell morphology.
Full-scale measurements generate insights into the inner link between multiple parameters
Bacterial physiology is reflected by a set of inter-correlated parameters, including morphology, cell number doubling time, biomass accumulation rate, nucleo-cytoplasmic ratio, and irrelated protein productivity, etc. These are seen as the outputs of our DNA replication interference system. The challenge for us is that, to fully characterize their responses to CRISPRri as well as their relationships, we need to choose effective and efficient measurements for each of them. In the last section, we mentioned that information regarding cell number doubling time and cell morphology can be acquired with microfluidic chips; biomass accumulation rate can be read from OD600 increments in microplate readers.
Another parameter we concern about is nucleo-cytoplasmic ratio. We performed DAPI staining to visualize the distributions of nucleoids in single cell under laser scanning confocal microscope (LSCM). LSCM is an optical imaging technique for increasing optical resolution and contrast of a micrograph by means of using a spatial pinhole to block out-of-focus light in image formation. Since different portion of a filamentous cell may be on different focal plane, traditional microscopic imaging can lead to omittance of stained nucleoids. That is why we consider LSCM as the most appropriate instrument to meet our needs. In this way, we found a decrease in average nucleo-cytoplasmic ratio when treated with CRISPRri targeted to OriC (Figure 4).
Figure 4.(A)The mechanism of LSCM (https://imb.uq.edu.au/facilities/microscopy/hardware-software/confocal-microscopes)(B)Nucleoid staining followed by imaging under laser scanning confocal microscope. DAPI is used to stain the nucleoids in E. coli with a working concentration of 10 μg/mL. About a minute after mixing bacteria with DAPI, the medium is replaced by PBS through precipitation-resuspension process. After washing for three times, the bacteria are available for microscopic imaging. Z-axis scanning for 2 μm with 0.2 μm each step overcomes the imaging difficulty caused by rise and fall along the long cell body. Nucleo-cytoplasmic ratio is calculated by the total number of nucleoids being divided by cell length. Sample number N1 = 6 for R1+ group and N2 = 24 for control group. Different sampling number coincide with different cell density in solution for each group.
Production of irrelated protein—in our experiment, GFP expression (under J23119 promotor on a high-copy-number plasmid, co-transformed with CRISPRri system)—was measured with flow cytometry (Figure 5A). Production of indigo (relying on FMO enzyme on a plasmid, co-transformed with CRISPRri system) was characterized with the absorbance at 620nm by spectrophotometer (Figure 5B).
Figure 5.(A)Indigo production of E. coli with or without CRISPRri.(A) GFP production per cell mass unit. CRISPRri system targeted to M+ is co-transformed with a constantly-expressed GFP plasmid. Fluorescent intensity and OD600 are both measured by microplate reader. The transformed strain is transferred to M9 medium after overnight cultivation. Three hours after transferring, the bacterial solution is diluted into medium with different arabinose concentration. The ratio of FI to OD600 shown here is at the timepoint of 6 hours.(B)Indigo production of E. coli with or without CRISPRri. Quantitative results of samples
All of these parameters of E. coliare taken into consideration because they stand for different features of evaluating the cell state. A rounded characterization system, including multiple measurement methods, is well developed as a full-scale quantitative description of E. coli general states.
Various measurements guarantee credibility and repeatability
Although we use various methods to interrogate different physiological parameters, it should be noted that information acquired with different measuring methods are not irrelevant or exclusive to each other. In fact, they can be mutually supportive.
For example, OD600 measured with microplate readers indicate unaffected biomass accumulation despite changes in sgRNA sequence and inducer concentration (Figure 6A). Meanwhile, results from microfluidic system reveal the same conclusion, as total covering areas of bacteria in each chamber (normalized according to covering area at time zero) are quite similar (Figure 6B).
Figure 6.Biomass of CRISPRri-implanted bacteria (A) Time-course OD600 of bacteria with CRISPRri targeted to M+ box under different arabinose concentration, tested by microplate reader. (B) Normalized covering area of M+ group bacteria in a microfluidic chip chamber, under flow of M9 medium with different arabinose concentrations. Covering area at the beginning is treated as 1. 8 chambers is measured for each group.
We intentionally carried out various measurements to verify a certain conclusion on different levels. For instance, to demonstrate that ssrA tag do cause reduction in the potency of dCas9, we verified that dCas9-sfGFP with ssrA tag accumulate slower and that effects of CRISPR interference with dCas9-ssrA is gentler (Figure 7A), effects of CRISPR replication interference with dCas9-ssrA targeting to R1+ box is also gentler (Figure 7B), and that reversibility of CRISPRri using dCas9-ssrA is better(Figure 7C) (figures regarding the reversibility of CRISPRri using dCas9 without ssrA is not shown because it behaves poorly).
Figure 7. Characterization of CRISPRri-ssrA system. A. Comparison between dCas9 and dCas9-ssrA system by expression level and CRISPRi effect on mRFP fluorescence. B. Comparison of effect on cell growth between CRISPRri and CRISPRri-ssrA, both targeted to R1+ box. C. Reversibility of CRISPRri-ssrA system targeted to R1+ box. Hollow arrows stand for removal of arabinose while solid black arrows stand for addition of arabinose.
For another instance, we characterize cell morphology change with more than one experimental equipment. While microscopic pictures have direct visual effects(Figure 8A), scientists can sometimes be accused of possible selection of desirable fields of view. Therefore, we analyzed the sample with flow cytometry, which further validated cell elongation(Figure 8B).
Figure 8.(A)Microscope imaging of bacteria under our system'scontrol. Significant morphology was observed.Bacteria were incubated for 8h in M9 culture with 0% and 0.5% Ara respectively.(B)Morphology change can be tested by flow cytometry. Elongation of cell body can be observed through increase of SSC-A. Slight SSC increase in PolyA group might be attributed to cell stress led by overexpression of dCas9. R1+ group shows a remarkable increase in SSC-A as the arabinose concentration increases. However, flow cytometry is argued to be indirect test of cell morphology change. SSC change is not absolutely positively correlated to cell shape elongation and flow cytometer actually has a upper limit for cell size which is around 150 μm. Therefore, flow cytometry is a qualitative but high-throughput characterization of cell morphology.
In short, we try to make results of experiment able to stand the test by using a series of different measuring method.
Novel measurements can be derived from specially designed experiment
Apart from those well-characterized measuring methods and devices, we also made modifications to or designed some special experiment equipment to suit our needs.
One device is used to measure the relative adhesiveness of long and short cells. We suppose one of the biggest advantages for long-shape-type cell is that its membrane area per cell increases, thus allowing stronger interaction between a cell and other interface. To test this assumption, we constructed a uniquely designed microfluidic chip device. Description! (Figure 9A) During our experiment, bacteria which have been cultured in medium with inducer are incubated in the chamber of our device for about 10 minutes. Mixture of long and short cells are supposed to adhere to the glass surface with different affinity. Washing medium is injected into the chamber from the injection pump, washing away floating cells. Then, fluids with gradually increasing velocity flow over the surface. Pictures are taken throughout the process, recording the portion of long cells to normal cells in the remaining cells. We found ratio of long cells to normal cells greatly increases as the flow gets faster (Figure 9B).
Figure 9. Measurement of cell adhesiveness. (A) A uniquely designed microfluidic chip as a platform to measure cell adhesiveness. The preparation workflow is as follows: (a) Use a puncher to make two holes at proper locations, according to the structural parameters shown here. (b) Stick a double-sided adhesive tape to fully cover the area of cover glass and the tape covering the view field is removed (the dotted box). (c) Add the cover glass onto the glass slide. (d) Fasten the slide with two clamps and heat the slide in 200 degree Celsius in vacuum. (e) The chip is available after natural cooling. (B) A simple experiment is designed to test the cell adhesiveness. E. coli transformed with CRISPRri system targeted to M+ box and induced by 0.50% arabinose M9 medium. Then the elongated M+ group cells is mixed with normal-sized PolyA group to obtain a long and short cell mixture. The mixed bacterial solution is injected into the chip and stand for 20 min to ensure the sedimentation of most cells. After that, the injection pump constantly provides the liquid buffer with a settled flow velocity. The initial velocity is 1 mL/4 hrs, and is slowly developed by 2-fold. Microscope would record the process until the flow velocity reaches the maximum of the injection pump. We defined a cell to be long if it takes more than 80 pixels in a microscopic graph. We counted the ratio of long cells to normal cells every time we develop the flow velocity. Initial ratio is normalized to 1.
Another specially designed experiment tool is for the visualization of effect of quorum sensing, which is displayed on spatial-level through a donor/receiver system. The donor cells, which merely express and release AHL, would activate the GFP expression of receiver cells through AHL diffusion. This is validated on the agar plate, by dropping the donor cells in the center and receiver cells around them with different distances. We found a progressive decrease in fluorescence as the receiver cells locate farther from central AHL donor. (Figure 10A) Then GFP gene is exchanged for dCas9 with companion of constantly-expressed single guide RNA to establish the qs-CRISPRri system. Through the donor-receiver system, we realized spatial-level control of cell growth (Figure 10B). Receiver colonies located nearer to the donor grow much slower than farther ones, which is observed in either dropping or smearing plate.
These measurements really display the creativity of our team.
Figure 10.(A)The donor-receiver split quorum sensing system of CRISPRri system. Upper figure is the gene circuit of donor and receiver cells of qs-CRISPRri system. For the figure below, a donor-receiver split quorum sensing GFP system enables fluorescent intensity control on spatial scale. Left figure is the sketch map of how donor and receiver bacteria is dropped onto the agar plate. Right figure is the real picture of agar plate under illumination of blue light. The location of the white arrow is the donor cells. Receiver cells are marked by number one to six.(B)Spatial-level growth control of donor-receiver split qs-CRISPRri system. For both two agar plates, the colony located at the center is the donor cells.
Portable measurements allow transplantation into other experiments
Portability is reflected in two aspects. The first one is the measuring methods we use, and the second one is our methods for processing measured data.
Our measurement of cell physiological parameters in microfluidics can be applied to other experiments about cell division, cell cycle, etc. We were glad to collaborate with iGEM team of Tsinghua. They wanted to collect some data on cell asymmetric division, and showed their interest upon hearing of our microfluidic chips system. We helped them image their engineered bacteria. We believe our whole set of measurement and subsequent image processing have strong portability, since it feature in exceptional stability, autonomy, and ability to generate large mass of temporal data with relatively low input of human resource.