How to study bacteria adaptation to stress and environmental changes by microfluidics?

This application note explains how to study bacteria adaptation to stress and environmental changes with pressure-driven flow controlled microfluidics such as the effect of antibiotics on bacterial growth. It aims to introduce the reader to an experimental method allowing the study of the effect parameters on bacterial growth. This technical note was written in collaboration with Dr. Bianca Sclavi (LCQB, UMR 7238, CNRS, Sorbonne Université, Paris) and Ilaria Iuliani (ilaria.iuliani@gmail.com), PhD candidate. 

Single cell, time-lapse imaging measurements in a controlled environment

Figure 1: A sample movie of typical mother machine experiment. This is a field of view (FOV), during an experiment up to 20 different FOV are acquired as a function of time. Growing and dividing bacteria can be seen filling the channels. The bacteria at the bottom of the channels are called mother bacteria. Courtesy of the authors.

Bacteria, such as Escherichia coli, can rapidly respond and adapt to changes in their environment affecting nutrient availability or to the presence of specific stressors such as the presence of sublethal concentrations of antibiotics.

These rapid response mechanisms result in changes in growth rate, cell size and gene expression.

However, sometimes there is a heterogeneous response in the population, where some of the cells respond differently than the others, which can result in a subpopulation that can better respond to renewed exposure to the stress at the cost of increased gene expression and/or slower growth rate.

This means that while some cells adapt quickly, within one or two hours, others have much longer times of adaptation that can last several hours. To study the dynamics of this kind of cellular response it is necessary to have a tight temporal and spatial control of the cells’ environment.

These experiments can last up to 72 hours in order to allow us to observe some of the slower adaptation mechanisms and in order to include several switches between different concentrations of nutrients.

Microfluidic chip resistance can be very different between one chip to another. Moreover, during the experiment the resistance can increase due to the formation of clogs. For this reason, it is very important to have a pressure-driven flow to have good reproducibility and long experiments.

Thanks to the efficient interface, we can easily check the experiments remotely. This can be very useful particularly for experiments that can last more than 1 day.

**********

If you wish to estimate the shear stress, microfluidic flow resistance applied in your system for a specific flow rate or chip design, please refer to Elveflow’s microfluidic calculator and its selection of application notes to help you perform your estimation:

• What is a microfluidic calculator?
• Viscosity conversion
• Shear stress calculator
• How to calculate the flow rate with microfluidics 
• Microfluidic resistance calculator

Application

The detailed experimental protocol presented here can be used for other biological applications such as:

• Mother machine experiments
• Single cell dynamics of bacterial growth
• Mechanics of cell growth
• Gene regulation and antibiotic resistance studies

Advantages of microfluidics

• Fine control of the liquid flow and pressure
• Rapid switching between up to 8 different growth media
• Temporal programming of the sequence of media switch steps

Material & methods

The microfluidic chip

Figure 2: Simplified schematic representation of the PDMS microfluidic chemostat device used for for the study of bacteria adaptation to stress and environmental changes [2]

To easily image the changes in single cell size and gene expression we have used a PDMS microfluidic device where the bacterial cells are growing in 2000 parallel channels that are about 20 µm long, 0.8 µm wide and 1.1 µm deep (Long et al., 2013, 2014) [1] .

These channels are flanked by larger channels (50 µm wide and 20 µm deep) where the growth medium flows and brings nutrients to the bacteria at the same time as washing away waste and metabolites.

This allows for the bacteria to grow in a constant environment to achieve what is called balanced growth.

Balanced growth is difficult to achieve in bacteria growing in flasks or on an agar pad, as the nutrients at a certain point run out and metabolites can accumulate decreasing the pH and slowing down growth.

List of components

Microfluidic setup diagram for bacteria adaptation study

Figure 3: Schematic representation of the microfluidic setup employed for the study of bacteria adaptation to stress and environmental changes comprising an OB1 pressure-driven flow controller with two channels, flow sensor, Mux Distribution and microfluidic chip.  Courtesy of the authors.

The above microfluidic setup composed of a pressure-driven flow controller (0 – 2 bar pressure range), two rotary valves (Mux Distributor), flow sensors and a microfluidic chip, ensures a constant environment for long time experiment and allows to quickly change between growth medium or temperature. Additionally, the PID feedback loop permits an effective control over the flow rate while keeping the stability and responsiveness of pressure-driven flows.

The following colour code was given to ease the protocole follow-up:

Figure 4: Picture of the microfluidic setup employed for the study of bacteria adaptation to stress and environmental changes comprising an OB1 pressure-driven flow controller, flow sensor, Mux Distribution and microfluidic chip.  Courtesy of the authors.

HARDWARE:

• Pressure & Flow controller: Impose a given pressure in order to create a stable and pulseless flow.
• Reservoirs: Contain your medium, buffer, bacteria suspension or samples. Various cap sizes are available, from Eppendorf to bottles.
• Rotation valve (MUX-Distributor): Select the injected liquid.
• Flow Sensor: Monitor and control the flow rate in real time.
• microfluidic chip: Contains your bacteria. Compatible with microscopy.
• Computer: Control all the parameters with our software and automate your experiment by creating injection sequences.

REAGENTS:

• Bacteria: Escherichia coli BW25113 and mutant strains
• Medium: minimal M9 medium with Glucose /  minimal M9 medium with Glucose supplemented with Casaminoacids filtered (0.22µm). We also add antibiotics, in particular kanamycin and chloramphenicol.
• Wash buffer: Isopropyl alcohol + ultrapure water

Microfluidic experimental protocol

1. Control all the instruments through a single interface

Figure 5: Picture of the ESI microfluidic software interface employed for the study of bacteria adaptation to stress and environmental changes comprising an OB1 pressure-driven flow controller, flow sensor, Mux Distribution and microfluidic chip.  Courtesy of the authors.

Window 1: [ORANGE] Main window 

• You have the list of all the connected devices. If any device is missed, you can press the “recharge” button on the top/left
• Press on the play button to open the windows 2 and 3
• Press the 3 dots symbol on top/right to open window 4

Window 2: [RED] MUX top and bottom window

• Press on the channel relative to the growth medium you want to flow in your circuit

Window 3: [GREEN] OB1 window

• Press the power bottom to turn on the relative channel
• If you select Regulator mode you can set a pressure to be maintained constant (flow rate will change based on the circuit’s resistance )
• If you select Sensor mode you can set a flow rate to maintain constant (pressure will change based on the circuit’s resistance)

Window 4: [BLUE] ESI Sequence Scheduler window

• Press the folder symbol with a plus to load an already existing file.
• For example, for an upshift experiments open the upshift.sq file

 Window 5: [PURPLE] Microscope window (Ti-Eclipse Nikon inverted microscope)

2. Bacteria preparation

a) Grow the bacteria overnight in filtered growth medium. NOTE: Always use a growth medium filtered at 0.2 µm in order to avoid clogging the microfluidic device.

b) In the morning, dilute the bacteria (ex 20 µl in 1.5ml). Put them at 30°C for around 3 hours (we want them to be in exponential phase when they go into the device). 

3. Microfluidic Setup preparation

1. Using a 20µl pipette put around 150 µl of BSA in the inlets (top and bottom) until you see BSA going out from the outlets. This reduces attachment of bacteria to the walls.
   
2. Put the device at 30°C and wait 30 mins
3. Fill the reservoirs with the growth medium
4. Fill the tubes with a growth medium.
5. In Sensor mode with the flow rate at 80 µl/min let the growth mediums flow inside the tubes and fill them completely by switching the MUX distributors between different falcons to avoid bubbles in the entire circuit.
6. Set the flow rate ch1_bottom to 20µl/min and ch2_top to 10µl/min (dust will be washed out from the channels).
7. Let the medium flow through the device for 30 mins
8. Set the flow rate ch1_bottom to 10µl/min and ch2_top to 20µl/min (the bubbles left will go out from the channels).
9. In the meantime, start an acquisition in brigth field by selecting around 5 positions to check if the device is stable and there are no problems with the autofocus.
10. Once there are no more bubbles in the channels we can inject bacteria.
11. Put 0.5 ml of pr-eculture in a 1 ml syringe.  Stop pumping.  Unplug the 2 input tubes from the device, set them in a tube and start pumping again (to avoid bubbles formation).
12. Inject the bacteria by hand (by slowly moving the syringe plunger, not by pushing on it) and check each 0.1 ml if you have enough bacteria in the channels (without unplugging the syringe).
13. Set flow rate ch1_bottom to 7µl/min and ch2_top to 15µl/min (to help bacteria to go inside the channels).

Figure 6 : Image analysis allows for the detection of the bacteria in each time frame in order to carry out segmentation and tracking using a script created by the author’s lab and their collaborators. Courtesy of the authors.

4. Run your microfluidic experiment

a) Start the flow of the medium

• Open window 4
• Select the desired pipeline (see below)
• Select first step and press play

b) Start the acquisition

5. Automate your microfluidic experiment

ESI software

Use the ESI sequence scheduler to automate your microfluidic experiment. In this protocole, the following sequence was created and performed:

STEP 1: Start the acquisition of pressure and or/ flow rate measurements.

STEP 2 – 3: Set both mux distributors (DIST) to valve 1 (slow growth medium)

STEP 4:  Start pumping (OB1) with a constant flow rate in both channels but with different flow rates (we can save our OB1 personalized settings in a text file to be loaded for each kind of experiment)

STEP 5: Maintain these conditions for 2h 10ms

STEP 6 – 7: Switch both the mux distributors to valve 2 (fast growth medium)

STEP 8:  Maintain these conditions for 2h 10ms

Figure 7: Picture of the ESI software interface focusing on the Sequence Scheduler employed for the study of bacteria adaptation to stress and environmental changes comprising an OB1 pressure-driven flow controller, flow sensor, Mux Distribution and microfluidic chip.  Courtesy of the authors.

/!\The flow sensors are fragile. A drop of liquid can cause malfunctioning for several hours. Keep them dry /!\

Results

Image Analysis

Image acquisition takes place on over 20 fields of view, each one with 10 channels, and each channel with about 10 bacteria. This can result in up to 10 000 cells in a 7 hour experiment. Automated image segmentation and tracking are therefore important steps in the analysis process. 

Movies obtained from the microscope are in the .nd2 format and can be opened with the Fiji software. Background subtraction is performed using a 50-pixel rolling ball technique and different positions are stored separately as a set of tiff image files. Channels are selected manually and stored in different folders. In order to perform segmentation and tracking on mother machine dataset, we use the codes developed by Mia Panlilio, Cambridge University [3]. Necessary modifications were made based on our experimental setup.

Figure 8:  Microfluidic study of the growth medium evolution over 30 minutes illustrated by the evolution of the fluorescence for Luria-Bertani growth medium (blue curve) and pure water (red curve). Courtesy of the authors.

To measure how quickly the growth medium changes within the device, the fluorescence of the LB (Luria-Bertani) growth medium was measured.

Here,  we switched from water to LB. Time 0 is the time at which the growth medium enters the device. This data was acquired in the last channels before the exit.

The time necessary to completely fill the device with the new growth medium is negligible in the timescale of our experiments.

Images of bacteria in the microfluidic device

Figure 9 : Image analysis allows for the detection of the bacteria in each time frame in order to carry out segmentation and tracking using a script created by the author’s lab and their collaborators. Courtesy of the authors.

The figure x shows the average values for all the bacteria in the fields of view of the experiments.

There are changes in division time, cell volume and fluorescence upon an adaptation to a switch from a growth medium containing glucose to one containing both glucose and amino acids (GluCaa).

In the first growth medium, the cells spend a lot of energy synthetizing their own amino acids and therefore grow more slowly.

Once the amino acids are added, the glucose is used mostly for energy production and the growth rate can increase.

The fluorescence comes from the expression of the green fluorescent protein (GFP) under control of a promoter chosen by us to report on the activity of a specific regulatory pathway. One can see that the cell division time becomes shorter and cell volume increases as the bacteria adapt to a faster growth rate. In addition, in this case the fluorescence also increases soon after the change in growth medium, indicating that gene expression has been induced as part of the cellular response.

Figure 10:  Population average change in division time (min), volume and fluorescence as the bacteria adapt from a poor to a rich growth medium over an 8 hour experiment obtained with the microfluidic setup. Courtesy of the authors.

Hints & tips

• Automate the process with the ESI scheduler
• Quickly switch between 8 different growth media with the Mux Distributor.
• Easily implement a FOR cycle to repeat these steps multiple times

CONTACT US FOR MORE 😉

References