An Inexpensive Way to Record and Quantify Bacterial Swarming


 Bacterial swarming refers to a rapid spread, with coordinated motion, of flagellated bacteria on a semi-solid surface1. There has been extensive study on this particular mode of motility among microbiologists and biophysicists because of its interesting biological and physical properties (e.g. enhanced antibiotic resistance2, turbulent collective motion3). The existing equipment for recording swarm expansion rate can easily go beyond tens of thousands dollars4, yet the conditions are not accurately controlled, resulting in large variations across the assays. Here, we report a reliable protocol to perform reproducible bacterial swarming assays and an inexpensive way to record and quantify the swarming activity by time-lapse photography. This novel protocol consists of three main parts: 1) building a “homemade”, environment-controlled photographing incubator; 2) performing bacterial swarming assay; 3) taking serial photos over time and calculating the swarming rate. The homemade incubator is economical, easy to operate, and has wide applications. In fact, this system can be applied for any slow evolving biological process that needs to be monitored by camera under a controlled environment.


Introduction
Coordinated multicellular migration across a moist surface known as bacterial swarming is an important phenotype that has been studied for decades in laboratories, including those inclinical settings. The colony expansion speed and morphological patterns are key parameters to describe bacterial swarming activity. We built an environment-controlled incubator to perform swarming assay.
Inside the incubator, a digital camera was mounted on the top to take time-lapse images of the swarming activity. After the recording, images were transferred to a laptop for quantification of the swarming rate. The swarming area can be calculated manually using ImageJ plugins or identified automatically using a python script we developed. The system was tested by over 1,000 swarming events to assure the stability and reproducibility.
At first glance, one may think swarming assay is simple: just inoculate bacteria on an agar petri dish, then taking a picture of the plate after a period of incubation. However, it can be challenging to perform swarming assays with reproducible results and record high quality images for further analysis. Here we list a few technical challenges one may encounter and provide corresponding solutions: 1. A typical swarming event may take about 10 to 20 hours for the bacteria to cover a 9-cm petri dish.
Besides the full coverage time, sometimes we need to know how long the lag phase lasts, by what time branches form at the colony edge, and at the microscopic level, when cell elongation happens.
Thus, the researcher needs to check at the plate regularly. Suppose one starts the assay during daytime, he/she may need to scan the plates every 30 minutes over night. Otherwise, he/she may miss the details. In our protocol, time-lapse photography helps to capture the key frames so that the researcher does not have to stay up late or regularly check in order not to miss key events.
2. Bacterial swarming is highly sensitive to the environment. Fluctuations in temperature and humidity will cause large differences in swarming rate and colony pattern. In this case, a stable humidity and temperature-controlled environment is critical for the assay. In our system, we utilized a thermo-insulated tent, a humidity control unit, and a temperature control unit to minimize the environmental fluctuation during the assay. One can readily set different humidity and temperature for swarming strain screening.
3. Since agar gel contains nearly 99% water, condensation readily forms on the plate lid, which obscures photo taking. We designed the incubator in such a way that when the plates are invertedly placed on the platform, the temperature of the lid is slightly higher than that of the agar, which prevents condensation. 4. Taking a clean and clear photo of swarming plates is tricky. The swarming plate has three optical surfaces: the plate lid, the agar surface and the plate bottom. When the camera flashlight or auxiliary front light is used, some light is reflected by the lid and bottom of the plate to the camera, forming unexpected light spots in the photos. In our design, for one plate photo shooting, we used circularly fluorescent backlight plus an adjustable light shield. For multi-petri dish (up to 9) event, LED light strip was used for sidelight illumination. Image quality will be better using the light shield because the light field is well calibrated, but the efficiency is higher when using LED light since it shines on a 4 larger area to fit 9 petri dishes at a time.
5. Quantification of the swarming rate takes much effort. Usually the researchers scan the plates and measure the radius of the swarm colony using a ruler. For an irregular shape of colony, they usually make a rough estimation of "effective radius". When multiple plates are used in the assay and one wants to plot the swarming area vs. time instead of just calculating the average swarming rate, it is a lot of work. In our case, developed a python script helped calculate the swarming colony area of each frame automatically. The accuracy was confirmed by ImageJ free selection tool.
Compared with other protocols in this field, our protocol has its distinctive advantages.
(1) Affordability. In 2015, Shrout team from University of Notre Dame developed a protocol on preparation, imaging and quantification of swarming assay 4 . However, in their protocol, they used a commercial equipment "Bruker in vivo imaging station". This product is no longer available from the company and a used one costs over $58,000 (dotmed.com). A lot of microbiology labs may not need many complex functions of the equipment, such as x-ray imaging or in vivo fluorescence imaging.
These labs may not afford or be willing to invest such a big amount of money on a bacterial swarming chamber. In contrast, the cost of our system is around $1,000 with everything included.
(2) Accuracy. For Bruker imaging station, to maintain humidity, they proposed to place a plate of water on each of the swarming plates. In this way, it is hard to control the humidity within specific range because different amount of water will result in different humidity. Nevertheless, in our design, the chamber humidity is well controlled. One can set the parameters before the assay starts and the environment is controlled dynamically using the digital sensor and controllers.
(3) Efficiency. In 2018, independent researcher Maria Cobo developed a time lapse imaging chamber for bacterial colony morphology observation 5 which allows for photography one petri dish a time while the temperature is not controlled. In our case, up to nine 9-cm petri dishes can fit inside the incubator, increasing efficiency. It is not merely having a larger box since adjusting the size of the box will alter the optical path in a subtle way. The geometry of our incubator and light field for photography were well coupled to ensure best lighting. In Dr. Cobo's protocol, the time-lapse video flickers a lot, and fog forms on the lid. In our protocol, these problems have been completely fixed.
This protocol includes three parts: (1) assembly of the incubator; (2) swarming assay preparation; and (3) image taking and data processing. By strictly following the procedure, we guarantee that one can get a homemade bacteria incubator with stable swarming results and high-quality images as we had in our lab. The overall cost is around $1,000 dollars depending on what camera you use in the system. Indeed, any digital camera that has a "Manual Mode" is adequate for the purpose. Swarming plate preparation is not difficult, but one needs to be very careful on certain details. For instance, bacterial swarming system is very sensitive to small environment perturbation as well as surface conditions such as the roughness of the agar surface. In the photo shooting part, we will show in detail how to tune the camera settings. Finally, we'll explain how to manually process the data using ImageJ, and also automatically quantify the swarming area using our Python script. 14. Set the humidifier outside the tent and connect the power cord to the humidity controller outlet.

Materials (i) For building the photography incubator
Extend the extractable plastic mist tube through the hole on the tent wall into the tent beneath the The temperature is set to 37 °C for SM3 and shaking frequency 200 rpm (revolutions per minute).
Prepare the overnight suspension around 5 pm so that it will be ready for use around 9 am the next morning. 28. Once the autoclave is done, put the bottle of medium back to the magnetic stirrer with heating function off but stirring function on. In this step, we want the medium to cool down to 40~50 °C and the constant stirring is to avoid non-uniformity in the agar.

ii) Prepare the swarm plates
29. Once the target temperature is reached, set a flame and use a pipet aid to transfer LB agar medium to 6 petri dishes, with 15 ml on each petri dish.
30. Turn off the flame and wait for the agar to solidify. This will take about 2 hours.

(Optional)
Use the freshly made plate or store the swarming plates in 4 °C cold room with them inverted for up to 2 days. Whenever you plan to use the swarming plates, you need to dry the plates in the hood first. This is to remove water on the agar surface. If there's water on the agar surface, the bacterial motility may be swimming rather than swarming. Thus, the next step is crucial.
32. Remove the lid of swarming plates in the hood. When the room humidity is above 50%RH, dry the plates for 20 minutes. When the room humidity is below 30%RH, the drying time is about 10 minutes.
When the humidity is between 30-50%, you can adjust the drying time accordingly to around 15 minutes. Do not over dry the plates; otherwise, bacterial cells may not be able to swarm, due to either surface friction or dryness.
33. Use micropipette to inoculate 2 μL overnight bacteria suspension on the center of a swarm plate.
Transfer the swarm plates into the incubator after the inoculation drop dried (3D hemisphere turns to a 2D circle). 35. In the preview of the camera, you should see the plate sitting in the center of the screen.

Part 3, time-lapse photo taken and swarming rate quantification
Otherwise move around the light shield to align the camera with the sample.
36. Rotate the nuts on the threaded rod to adjust the position of the acrylic sheets. The distance between the sample platform and Level I is about 1.5 inches. The distance between Level I and Level II is slightly under 1 inch (about 2 cm) while Level II and Level III are separated by 3.5 inches. If you see a round light spot on the petri dish, slightly lift Level III. If the petri dish is too dark, slightly lower Level III. If you cannot get a good image by adjusting Level III, then adjust Level II slightly up or down.
Caution: Calibration of the light shield takes practice. There is a subtle distance relationship between each sheet to achieve the best image quality depending on the camera setting. We want the light to shine through the transparent agar and be reflected from the swarm colony. Once you find the right position, tighten all the nuts and the position is locked for later imaging.
37. Set the camera focal length to 35 mm -65 mm. Adjust the zoom ring to have the sample occupy the full screen but not exceed the border. Use "M" manual mode to focus on the bacteria colony. The aperture is set to F5.6 -F7.1 and adjust the shutter speed until the resultant exposure value is 0 or -⅓. This is to make image processing easier because overexposed images will lose information details.
38. For multi-plate assay, remove the light shield, turn on the LED light strip, and place the uncut acrylic sheet on the sample platform. Place the swarm plates inverted on the acrylic sheet so that water will not condensate on the lid. Check the camera preview to make sure all the plates are within the range of the screen.

Reason & solution:
Sometimes certain model of products will be out of stock. In this case, one can try to cut the 36'' into two halves with a hacksaw. To cut the circle, if one plans to use a hole saw or circle cutter, make sure to clamp the acrylic sheet first. One alternative is to go to an engineering workshop or carpenter's shop. They have the saber saw or jigsaw to cut the circle.

Problem 2:
Swarming strain does not swarm. They just form a dense colony spot by growth.

Reason & solution:
The reason for this is due to roughness of the agar surface or the plate is too dry. Always use fresh plates or two days old at most. Beyond two days, the agar plate will become significantly drier than before and may inhibit swarming. Drying time in the hood may be too long.
When the lab humidity is below 30% RH, 10 min of drying is enough. Also, pour the plates when the agar solution is not too cold. Cold agar will result in rough surface since solidification has already taken place in the bottle, forming small clusters. Finally, double check the tryptone or yeast.
Occasionally, for certain Lot numbers, the chemicals do not dissolve thoroughly, which may change the texture of the agar surface and increase surface friction.

Reason & solution:
The agar concentration may be too low, or the drying time too short so the cells are swimming not swarming on the plates. Different swarming species may have different agar concentration tolerance. For SM strains, 0.5% agar is the concentration to distinguish swarmers and non-swarmers while for B. Subtilis 3610, 0.7% agar is the concentration to tell non-swarmer B. Subtilis DS215 from the swarming wild type. To fix the problem, one can try to increase the drying time, pour the plate when agar solution is colder, or raise the agar concentration.

Problem 4:
The swarming rate of the same strain varies a lot among different trials.

Reason & solution:
The swarming rate may vary under different optical density (OD 600 ). For SM bacteria strains, when the cells are at the first half of the growth exponential phase, or the death phase, the swarming lag is shortened and variations in swarming rate and morphology was noticed.
However, in stationary phase (OD 600 : 1.5~1.8), the swarming rate hardly fluctuates. For, different bacteria species, situation may differ, but our suggestion is to use stationary phase cells every time one performs the assay.   Taking photos using the light shield or LED light strip. a, Taking photos using the light shield. The circular fluorescent bulb is placed between Level I and the sample platform. The distance between Level II and Level I is about 2 cm while the distance between Level II and Level III is 3.5''. b, Taking photos using the LED light strip. The LED light is fixed 1 inch from the sample platform on the horizontal zinc-plated slotted angles. If the light is too strong, one can insert a white paper belt to diffuse and reflect the light. The black cloth near the camera is necessary for blocking reflections from the plates.
20 Figure 4 Quantification of Enterobacter sp. SM1 and SM3 swarming motility. 5 μL overnight bacterial culture of SM1 and SM3 were inoculated on 0.5% LB agar plates and incubated in the swarming incubator (repeated 10 times). Time lapse images were taken every half an hour, and the swarm area was measured for each image. Data are represented as means with 95% confidence interval. When not shown, the errors are smaller than the size of the symbols.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.