Imaging Flow Cytometry to Study Microbial Autoaggregation

Summary

This protocol describes a quantitative approach to measure microbial autoaggregation using imaging flow cytometry.

Abstract

Beneficial and probiotic bacteria play essential roles in their hosts, providing various health benefits, including immunity to infectious diseases. The Lactobacillaceae family consists of Gram-positive bacteria with confirmed probiotic properties. This study utilizes Lactobacillaceae species as a model to demonstrate the effectiveness of single-cell high throughput analysis in studying cellular aggregation. The focus is on analyzing the response of these beneficial species to simple carbohydrates from the diet.

The study showcases how Imaging Flow Cytometry (IFC) can overcome the fundamental differences in the assembly of probiotic bacteria in the presence and absence of carbohydrates. IFC combines the power and speed of conventional flow cytometry with the spatial resolution of microscopy, enabling high-rate complex morphometric measurements in a phenotypically defined manner across a library of beneficial bacterial strains and conditions. This protocol provides insights into the autoaggregation of Lactobacillaceae species and sheds light on their response to dietary carbohydrates, contributing to understanding the mechanisms behind the beneficial effects of these probiotic bacteria.

Introduction

Bacterial autoaggregation is considered a primary step in biofilm formation. In this process (sometimes also called autoagglutination or flocculation), bacteria of the same type form multicellular clumps that eventually settle at the bottom of culture tubes or attach to their target tissue or surface¹.

Autoaggregation is a widely observed phenomenon and has been shown so far in Gram-negative pathogens such as the opportunistic pathogen Acinetobacter baumannii², the dental pathogen Aggregobacter actinomycetemcomitans³, and the emerging pathogen Burkholderia pseudomallei⁴. Autoaggregation has also been described in several probiotic Gram-positive strains^5,6,7,8. In Lactobacillus (L.) acidophilus, autoaggregation was partially mediated by S-layer proteins and correlated with an adhesion to xylan⁷. Similarly, we found a correlation between glucose-dependent autoaggregation and induction of the adhesive properties (as judged by adhesion to mucin) of the probiotic species Lacticaseibacillus rhamnosus GG, Lacticaseibacillus casei, L. acidophilus, Lacticaseibacillus paracasei, and Lactiplantibacillus plantarum⁵. The increased autoaggregation most likely reflected changes in the expression of cellular adhesins in response to glucose and its catabolites⁹. While the molecular mechanisms of autoaggregation remain to be determined, it has been shown that this process alters the phenotype of the aggregated bacteria and grants them enhanced tolerance to environmental stressors¹, as well as increased sensitivity to quorum sensing molecules¹⁰.

Several approaches have been used to measure autoaggregation; one experimental approach is to let cultures stand statically in narrow culture tubes for a given time. Control cultures remain turbid, whereas autoaggregation cultures will settle at the bottom of the tube. A more quantitative approach measures autoaggregation by sedimentation or settling assay¹¹.

Flow cytometry has also been increasingly employed in recent years to investigate bacterial autoaggregation. This method is appropriate for analyzing particles between approximately 0.5 and 1000 µm in size. The single bacterium or formed aggregates are suspended in fluid, fed into a stream, and can be detected one by one¹¹. Recording forward scattered light allows measuring the relative size of the cell or aggregate. It is relatively fast and straightforward but cannot detect several parameters, such as aggregate size or the average number of cells in aggregates. Therefore, this approach can be complemented microscopically, allowing more parameters to be checked¹². However, traditional microscopy is time-consuming and thereby limits the number of tested samples and the statistical power of the analysis. In general, imaging flow cytometry provides several features compared to traditional flow cytometry, such as simultaneous analysis of cell morphology and phenotype, conducting image-based analysis, detecting rare events, and validating flow cytometry data¹³. These advantages enhance the capabilities of flow cytometry and facilitate a more detailed examination of cell populations.

This study provides a valuable protocol for imaging flow cytometry to monitor autoaggregation in lactic acid bacteria (LAB). These Gram-positive rods are facultative anaerobes and belong to the LAB group. These efficient glucose fermenters generate lactic acid as their main end-product of carbohydrate metabolism¹⁴. These bacteria are beneficial core members of the microbiome and are naturally found in the gastrointestinal tract (GIT) of humans and animals, as well as in the urogenital tract of females¹⁵. Therefore, the exact characterization of their autoaggregation properties is of high biotechnological and clinical interest.

Our previous findings indicated that the basal level of autoaggregation differs between different probiotic strains. This heterogeneity is affected by the different carbohydrates used as a carbon source⁵. To overcome this fundamental property of probiotic bacteria, the effects of carbohydrates from the diet were monitored on autoaggregation on the single-cell level using IFC. This IFC-based approach combines the power and speed of traditional flow cytometers with the microscope's resolution. Therefore, it allows high-rate complex morphometric measurements in a phenotypically defined way^16,17. This approach can be extended to other probiotic and pathogenic bacteria, combined with fluorescent reporters to monitor gene expression, and fluorescently labeled strains to monitor the presence and abundance of specific bacterial species in heterogeneous aggregates.

Protocol

The .ast file, with the template for Lacticaseibacillus rhamnosus GG (LGG) as an example, is provided in Supplementary Coding File 1.

1. Media preparation

1. Prepare Lactobacilli MRS broth following the manufacturer's instructions (55 g in 1 L of deionized water) and Lactobacilli MRS agar plates with 1.5% (w/v) agar (see Table of Materials). After autoclave sterilization, the medium is ready to use directly or can be stored at 4 °C.
2. Prepare 50% (15 g in 1 L of deionized water) Tryptic soy broth (TSB) (see Table of Materials), TSB supplemented with 1% (w/v) D-(+)-glucose and 1% (w/v) D-(+)-raffinose, then autoclave to sterilize the medium. It is preferred to store it at 4 °C to avoid contamination.
3. For a buffered medium with glucose, prepare TSB supplemented with 1% (w/v) D-(+)-glucose, 5 mM of potassium phosphate monobasic together with potassium dibasic, and 100 mM MOPS (3-(N-morpholino) propane sulfonic acid, see Table of Materials) pH 7. Then autoclave to sterilize the medium.

2. Preparation of samples

NOTE: This step involves the evaluation of cellular aggregation in response to fermentable or non-fermentable sugar from our diet. A schematic representation of the sample preparation process is depicted in Figure 1.

1. Streak the Lactobacillaceae strain from a frozen glycerol stock (50% glycerol) on a Lactobacilli MRS broth plate (1.5% agar).
   NOTE: In the example experiment presented here, Lactobacillus acidophilus (ATCC 4356), Lacticaseibacillus casei (ATCC 393), Lacticaseibacillus casei subsp. paracasei (ATCC BAA-52), Lactiplantibacillus plantarum (ATCC 8014), and Lacticaseibacillus rhamnosus GG (ATCC 53103) (LGG) probiotic strains were used. To expand this methodology to additional probiotic members of the firmicutes phylum, Bacillus coagulans (ATCC 10545) was selected.
2. Grow the cells overnight at 37 °C under static conditions.
3. Inoculate 5 mL of Lactobacilli MRS broth with a single colony and grow it at 37 °C under static conditions overnight.
4. Dilute the cell culture from the previous step (1:100) in 3 mL of 50% Tryptic soy broth (TSB), TSB supplemented with 1% (w/v) D-(+)-glucose, or TSB supplemented with 1% (w/v) D-(+)-raffinose.
5. Incubate overnight at 37 °C under static conditions.
6. Prepare the sample by mixing the cells uniformly in the medium, and gently vortex the cell culture tube from step 2.5. Gently invert the test tube until complete mixing takes place. Transfer 200-300 µL into a 1.5 mL microcentrifuge tube.
   ​NOTE: It is essential not to sonicate the sample to avoid breaking the aggregates. During the data acquisition from steps 3.4 to 3.10, there may be instances where the samples are too concentrated. If it is observed that the sample is running very slowly or not running at all, one can dilute the sample with the appropriate fresh medium.

3. Data acquisition

1. Set up the imaging flow cytometer according to the instructions provided by the manufacturer (see Table of Materials).
2. Set the 785 nm laser to 5 mW for darkfield measurement (equivalent to side scattering in conventional flow cytometry, abbreviated as SSC). Use a 60x lens with a numerical aperture (N.A) of 0.9, select high sensitivity, and set the speed to low.
3. Before collecting any data, ensure the stability of the calibration beads flow in the SSC channel. Click the play button in the "fluidics" box and examine the quality of the bead images. The bead images should appear sharp and crisp.
4. Press the Load button and place a tube with a sample (from step 2.6) into the holder.
5. Create a new scatterplot in the "workspace" box. Click New scatterplot and select the area in the brightfield channel corresponding to the SSC channel intensity. Exclude the calibration beads that run in the instrument along with the sample.
   NOTE: Beads can be excluded by loading a sample with buffer only and identifying the beads on the Area vs. SSC plot. Then draw a gate around the bead population using the Create Polygon Region button.
6. In the "Acquisition Settings" box, enter the sample name, specify the location for the data storage, choose the population (all cells without beads) to record from, and set the number of events to be collected. Typically, 10,000-20,000 events are sufficient.
7. Click the Record button located in the "Acquisition" box, and the acquisition process will automatically end once the specified number of events is reached.
8. Click the Return button to unload the tube, and remove the tube from the holder when prompted in a small window.
9. Repeat steps 3.5-3.9 for each sample.
10. Save the template to maintain the uniformity of data acquisition for subsequent repeats.

4. Data analysis

1. Measure the percentage (%) of autoaggregation from the population.
   1. Analyze the data acquired on the imaging flow cytometer using IDEAS software (see Table of Materials).
   2. Create a data analysis file (.daf) from the raw file (.rif) by loading the (.rif) file into the analysis software. Click the File button and open the required (.rif) file. In the opened window, click on use acquisition analysis and then click OK.
   3. Create a histogram of gradient RMS (a measurement of image contrast and focus) in the brightfield channel to identify focused events. In the analysis area, click the New Histogram button. In the "New Histogram" window, choose the population from the acquisition to analyze, and in the "x axis feature," select Gradient RMS of the brightfield channel.
      1. Place a gate by clicking the Create Line Region button in the analysis area to exclude non-focused events (Figure 2A).
         NOTE: Typically, the "Focused" line region should include 80%-90% of the events with the highest Gradient RMS score, but specific thresholds for "in-focus" cells should be determined for each experimental setup.
   4. Create a scatterplot of the area (µm²) versus the aspect ratio (width divided by the length of a best-fit ellipse) of the brightfield channel. Press the New Scatterplot button in the analysis area. In the "New Scatterplot" window, choose the Focused population from step 4.1.3.
      1. Select Area of the brightfield channel for the x-axis feature and Aspect Ratio of the brightfield channel for the y-axis feature. This scatterplot allows gating on the focused population to distinguish singlets and small aggregate events from larger microbial aggregates and chains.
   5. On this scatterplot, draw a gate using the Create Rectangle Region button (or Create Polygon Region) based on the area value for the aggregation events population and another gate for the singlets and small aggregates population (Figure 2B).
      NOTE: In certain backgrounds, separating small aggregates from singlets may be challenging, so they can be combined into the same gate. The same applies to larger aggregates and chains (Figure 3).
   6. Manually review and evaluate images of events within each gate to verify the reliability of the gating strategy. If needed, adjust the area value for the gate. Perform this by selecting the reviewed population from the "Populations" drop-down menu.
      NOTE: The gating has no specific area value, as it can vary depending on the characteristics of the bacteria being analyzed.
   7. Save the data analysis as a template. Click the File tab, select Save as Template.ast, and open the next sample (.rif) file under the same template to create the data analysis file (.daf) (Supporting .ast file as an example for LGG).
   8. Create a statistics table to enumerate key events for further analysis. Click on the Reports tab, then click on Define Statistics Report.
   9. In the new window, click Add Columns. Add the singlets/aggregates count and %gated statistics. Under "statistics," choose %gated/count, and under the selected population, choose singlets/aggregates. Click Add Statistics to add the statistic to the list and save it as a template.
   10. Click on Generate statistic report and choose the statistic template and the (.daf) files for analysis.
2. Measure the size distribution of the aggregates.
   1. To analyze the mean size of the aggregation events, plot a histogram of the area (µm²) of the aggregation events population from step 4.1.5 (Figure 2C).
   2. Save the data analysis as a template. Click on the File tab, select Save as Template, and open the next sample (.rif) file under the same template for the data analysis file (.daf).
   3. Repeat steps 4.1.7-4.19. For the aggregates' size, choose mean under statistics and select aggregates under the selected population. Under features, select Area of the brightfield channel. Click on Add Statistics to add the statistic to the list and save it as a template.
   4. Repeat step 4.1.9 to generate a statistic report.

Representative Results

The results demonstrate that this method can easily measure the differences in autoaggregation in response to dietary sugars in LAB bacteria. By separating individuals from aggregates, the method allows calculating the percentage of the population of the aggregation events out of all events in response to fermentable or non-fermentable sugars from the diet. Additionally, it was possible to measure if there are differences in the mean size of the aggregate's population between treatments.

The representative images in Figure 3 demonstrate the gating strategy for each LAB bacterium. In LGG and L. paracasei, single cells, small aggregates, large aggregates, and chains were detected (Figure 3). However, chains and larger aggregates could not be separated.

The representative images in Figure 4 revealed substantial variations in the aggregation characteristics among different LAB strains presented in response to fermentable or non-fermentable sugars, which were measured by the population percentage and the average size of aggregation events⁵.

Figure 1: Workflow of the sample preparation process. Please click here to view a larger version of this figure.

Figure 2: Gating strategy for evaluating aggregation events population and size. (A) Gating of focused events from all events based on Gradient RMS using the brightfield channel. (B) Separation of singles and small aggregates from larger aggregates by plotting the area (µm²) versus the aspect ratio (width divided by the length of a best-fit ellipse) of the brightfield channel. (C) The area of the brightfield channel among aggregation events provides information about the size of the particles. Please click here to view a larger version of this figure.

Figure 3: Cell images of gating strategy. (A) LGG. (B) L. paracasei in TSB (control). Please click here to view a larger version of this figure.

Figure 4: Representative brightfield images of aggregation events population in different conditions. (A) LGG. (B) L. paracasei. (C) L. plantarum. (D) L. casei. The conditions include TSB medium (control), TSB medium supplemented with glucose and raffinose (1% w/v), and TSB medium supplemented with glucose (1% w/v) and buffer. Please click here to view a larger version of this figure.

Supplementary Coding File 1: The .ast file with the template for LGG as an example. Please click here to download this File.

Discussion

Flow cytometry is a widely used method for quantifying fluorescence intensities in eukaryotic cells, but it may not provide accurate measurements for bacterial cells due to their larger size or small aggregates. These factors can significantly impact the precise quantification of autoaggregation and the basal level of aggregate formation in different conditions. To address this, imaging flow cytometry (IFC) was employed to gain a better resolution of how carbohydrates affect the aggregation of probiotic bacteria⁵. IFC combines the statistical significance of large sample sizes in traditional flow cytometry with the per-cell information content of standard microscopy by collecting numerous digital images per sample and extracting numerical image-based features¹⁶.

Previously, IFC has been used to study various bacterial behaviors, such as antibiotic production in Bacillus subtilis and the composition of dual-species biofilms^18,19. While sporadic use of Imagestream flow cytometry for measuring autoaggregation has been reported²⁰, it is believed to be the first demonstration of its usefulness in studying probiotic bacteria⁵. The findings indicate that IFC is efficient and reliable, serving as a replacement for both traditional flow cytometry and microscopy in studying microbial aggregation of LAB strains, especially considering the significant differences in basal levels of autoaggregation among strains⁵. This approach is particularly valuable for accurate high-throughput screening of multiple strains and growth conditions.

However, the IFC-based method has certain limitations and considerations. It requires a higher number of cells to generate data sets compared to traditional microscopy methods. Additionally, the optical and digital resolution of IFC images is relatively lower than that of conventional microscopy due to the inherent properties of the method. While traditional flow cytometry lacks the reliability and quality assurance of optical scans, Imagestream flow cytometry is commonly used as an alternative to conventional flow cytometry rather than fluorescent microscopy. Moreover, determining the area value that distinguishes single cells from aggregates is a critical step, and its accurate identification is crucial to avoid losing vital data. As mentioned in the protocol, this value varies across bacterial species and should be carefully determined.

Further developments can be made to expand the method’s application, such as investigating aggregation under additional environmental factors. Combining image analysis with flow cytometry data allows for more comprehensive characterization of microbial aggregates, including size, shape, and composition. This method can also be applied to enhance the high-throughput characterization of microbial aggregates in naturally occurring aquatic habitats²¹. Furthermore, it can be easily adapted to examine aggregation in complex microbiome communities by labeling target species and cell types.

Overall, the findings highlight that this approach is straightforward, informative, and highly useful for evaluating the aggregation properties of bacteria in both mono-species and complex communities. It enables a better understanding of the molecular mechanisms underlying autoaggregation by comparing the number of cells and the size of aggregates among different deletion, overexpression strains, and various chemical and physical manipulations.

Materials

14 mL culture tubes	Falcon	352051
15 mL centrifuge tube	Falcon	352096
Bacto Agar	Baeton,Dickinson and Company	214010
Bacto Typtic Soy Broth	Baeton,Dickinson and Company	211825
D-(+)-Glucose	Sigma	G7021-1KG
D-(+)-Raffinose pentahydrate	Sigma	83400-25G
Difco Lactobacilli MRS broth	Baeton,Dickinson and Company	288130
EASY-LOCK MICROPR. 1.5 mL (Eppendorf)	FL medical	23053
IDEAS Software	Amnis/EMD Millipore	N/A	Details available at: https://www.merckmillipore.com/INTL/en/20150212_144049?ReferrerURL=https%3A%2F%2Fwww.google.com%2F&bd=1
ImageStream X Mark II	Amnis/EMD Millipore	N/A	Details available at: https://www.merckmillipore.com/INTL/en/20150121_205948?ReferrerURL=https%3A%2F%2Fwww.google.com%2F
MOPS, 3-(N-morpholino)propanesulfonic acid	Fisher bioreagents	BP308-500
Potassium phosphate dibasic	Fisher Scientific, 174.18 g/mol	BP363-1
Potassium phosphate monobasic	Sigma, 136.09 g/mol	P0662-500G