Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Feb 1.
Published in final edited form as: Biotechnol Bioeng. 2022 Nov 3;120(2):399–408. doi: 10.1002/bit.28253

Flow Cytometric Evaluation of Yeast-Bacterial Cell-Cell Interactions

Ming Lei 1, Vikas D Trivedi 1, Nikhil U Nair 1, Kyongbum Lee 1,*, James A Van Deventer 1,2,*
PMCID: PMC10072783  NIHMSID: NIHMS1881741  PMID: 36259110

Abstract

Synthetic cell-cell interaction systems can be useful for understanding multicellular communities or for screening binding molecules. We adapt a previously characterized set of synthetic cognate nanobody-antigen pairs to a yeast-bacteria coincubation format and use flow cytometry to evaluate cell-cell interactions mediated by binding between surface-displayed molecules. We further use fluorescence-activated cell sorting (FACS) to enrich for a specific yeast-displayed nanobody within a mixed yeast-display population. Finally, we demonstrate that this system supports characterization of a therapeutically relevant nanobody-antigen interaction: a previously discovered nanobody that binds to the intimin protein expressed on the surface of enterohemorrhagic E. coli. Overall, our findings indicate that the yeast-bacteria format supports efficient evaluation of ligand-target interactions. With further development, this format may facilitate systematic characterization and high throughput discovery of bacterial surface-binding molecules.

Keywords: Yeast display, cell-cell interaction, nanobody, flow cytometry

Introduction

Surface components of bacterial cells such as cell membrane or cell wall proteins are an emerging class of therapeutic targets (Garai & Blanc-Potard, 2020). For example, pathogenic bacteria present virulence factors on their cell surfaces in response to an external physiological stimulus to promote adherence to gut epithelial cells (Pizarro-Cerda & Cossart, 2006). Similarly, in the context of biofilm formation, various cell surface adhesion and colonization factors can participate in adhesion by selectively binding to extracellular matrix proteins on host cells such as fibronectin, collagen, and mucin (Chagnot et al., 2013). Traditional methods for discovery of surface component-binding antibodies and peptides have employed screens utilizing soluble, purified forms of a target protein of interest with cells displaying a library of binding molecules (Islam et al., 2021); or conversely, a soluble library of binding molecules with a surface-anchored target protein (Ruano-Gallego et al., 2019). Recent studies have shown that cell-cell interaction assays can also facilitate the discovery of novel binding molecules that recognize targets presented on cell surfaces. To date, these efforts have utilized microbial display platforms to discover binding proteins that recognize specific components of mammalian cell surfaces (Lown et al., 2021; Stern et al., 2017). Multiple groups have quantitatively analyzed yeast-mammalian cell interaction systems to investigate the roles of binding affinity, surface density of yeast-displayed ligands, surface density of mammalian-displayed ligands, selection conditions, and linker lengths on the detection and enrichment of cognate binding interactions (Bacon et al., 2019; Krohl et al., 2020; Lown et al., 2021; Stern et al., 2017). In addition to their use in ligand discovery, these systems can support the generation and characterization of synthetic interactions mediated by surface-displayed proteins. Previous studies have shown that E. coli cell-cell interactions modulated by orthogonal interactions between surface displayed molecules can form cell aggregates (Glass & Riedel-Kruse, 2018; Kozlowski et al., 2021). Additionally, surface-displayed molecules can enable interactions between different cell types for applications in modeling host-microbe interactions or discovery of antibacterial therapeutics (Chun et al., 2020; Lahav-Mankovski et al., 2020). Further characterization of these synthetic interaction systems in a yeast display format can facilitate discovery of novel binding or antibacterial molecules that may be generalizable towards different bacterial species. In this study, we use flow cytometry assays to evaluate yeast-bacteria cell-cell interactions mediated by surface-displayed nanobody-antigen pairs. The characterizations of yeast-bacteria cell-cell interactions performed in this study demonstrate the utility of this system for evaluating ligand-target interactions on cell surfaces. The present study reports the development and characterization of a novel yeast-bacteria cell-cell interaction system that is able to detect high-affinity nanobody-antigen interactions and motivates future library screening applications using this platform for discovery of bacterial surface-binding molecules.

Results and Discussion

To investigate the potential for yeast and E. coli to form complexes driven by specific interactions, we used a previously described set of nanobody (Nb)-antigen (Ag) pairs known to exhibit mutual specificity to one another. In prior work, both the nanobodies and the antigens were displayed on the surfaces of E. coli (Glass & Riedel-Kruse, 2018). Here, we moved the nanobodies into a standard yeast (S. cerevisiae) display format (see Materials and Methods in supplemental information for details). Nanobodies are covalently anchored on the yeast surface using the Aga1p-Aga2p display system (Cherf & Cochran, 2015). Yeast displaying Nb1, Nb2, or Nb3 were stained with calcofluor white (CFW), a blue fluorescent dye that stains the chitin in yeast cell walls and minimally stains gram-negative bacteria such as E. coli. Antigens were anchored on the E. coli surface using the intimin N-terminus from EHEC O157:H7 (Salema et al., 2013). E. coli displaying Ag1, Ag2, or Ag3 were stained with SYTO 9, a green fluorescent dye that permeates cell membranes and binds to nucleic acids. Since SYTO 9 can stain both yeast and E. coli cells, SYTO 9-stained E. coli cells were washed with phosphate buffered saline (PBS) before coincubation with yeast to remove excess unbound dye and minimize the detection of false positive events. Though the specificity, half-life, and toxicity of such fluorescent dyes should be considered, a major advantage of using fluorescent dyes is that they can facilitate flow cytometry analysis without the need to genetically engineer cells to express fluorescent reporter proteins. Additionally, we showed that the SYTO 9 dye stains E. coli cells uniformly, versus an alternative covalent Alexa Fluor 488 NHS ester dye which leads to nonuniform labeling (Figure S1). The use of a uniformly labeling fluorescent dye can expand the range of cell types that can be tested in this flow cytometric cell-cell interaction format to include genetically intractable organisms.

For reproducible evaluation of cell-cell interaction events, we experimentally determined several important assay parameters, including cell densities, cell ratios, and incubation times, and flow cytometer voltage settings (Supplementary Discussion). We next evaluated the extent to which the two cell types, when co-incubated, associated with one another as a function of different interaction pairs (Figure 1A). Figure 1B depicts two-dimensional flow cytometry dot plots used to evaluate yeast and E. coli cell complexes. High levels of blue and green fluorescence are expected when flow cytometry events include at least one yeast cell and at least one E. coli cell. Upon coincubation of yeast and E. coli displaying Nb2 and Ag2, respectively, the frequency of events exhibiting high levels of both blue and green fluorescence increased (~30-40% of total events) relative to coincubation of cells displaying noncognate nanobody-antigen pairs (~10-15% of total events). A similar trend was observed with yeast displaying Nb3 and E. coli displaying Ag3. These results suggest that cognate Ab-Nb pairs mediate specific cell-cell interactions. We did not observe any increases in double-positive fluorescent events upon coincubation of Nb1-displaying yeast and Ag1-displaying E. coli. A contributing factor may be that E. coli cells induced to express Ag1 also tend to form aggregates with other Ag1 expressing E. coli cells. This was observed in the previous study that characterized these three nanobody-antigen interaction pairs (Glass & Riedel-Kruse, 2018). This tendency to self-aggregate may interfere with potential yeast-E. coli interactions. In any case, our data strongly suggest that the formation of yeast-E. coli complexes can be mediated by specific sets of interactions.

Figure 1.

Figure 1.

Open in a new tab

Flow cytometry and microscopy characterization of yeast-bacteria cell complexes. A) Schematic of orthogonal interactions between nanobodies and antigens displayed on yeast and E. coli, respectively. B) Representative flow cytometry dot plots for all combinations of coincubations for yeast cells displaying Nb1, Nb2, or Nb3 with E. coli cells displaying Ag1, Ag2, or Ag3. Red outlined plots represent expected cognate interactions. C) Percentage of events in quadrant 2 (Q2, high levels of blue and green fluorescence detected) following coincubation for 30 min. Error bars represent the standard deviation of 3 biological replicates. *Indicates significant difference (t-test p-val<0.05) compared to noncognate interactions. D) Representative phase contrast and overlaid fluorescence microscopy images of low, high, and control (noncognate) interactions between calcofluor-stained yeast displaying Nb2 or Nb3 (blue) and SYTO 9-stained E. coli displaying Ag2 or Ag3 (green). Scale bars indicate 5 μm.

To visualize yeast-E. coli complexes for evidence of direct physical interactions between the two cell types, cell suspensions were collected using fluorescence microscopy. Qualitatively, more yeast-E. coli complexes were detected in samples with cognate interaction pairs, consistent with the observations in flow cytometry analysis. Microscopy images for cell complexes representative of both low and high interaction events are shown for the Nb2-Ag2 and Nb3-Ag3 interaction pairs (Figure 1D).

After developing a flow cytometry assay for yeast-bacteria cell-cell interactions, we attempted to enrich for a cognate binding molecule in this interaction format. Previous studies have shown that this type of enrichment is possible using bacteria-mammalian cell (Salema et al., 2016), yeast-mammalian cell (Yang et al., 2019), yeast-yeast (Bacon et al., 2019), and phage-mammalian cell screens (Jones et al., 2016). However, to the best of our knowledge, enrichment of cognate binding has not been demonstrated in a yeast-bacteria cell-cell interaction format. In a proof-of-concept experiment, we performed a single round of sorting for nanobody-antigen interaction events from a coincubation culture of yeast and E. coli. Yeast populations, prepared at 1:1:1, 10:1:10, and 100:1:100 ratios of cells displaying Nb1, Nb2, and Nb3, respectively, were screened against E. coli displaying Ag2. While enrichment conditions are more stringent in traditional antibody discovery using soluble forms of proteins, the approximate 1:200 cognate to nonspecific binder ratio used in this study is similar to ratios used in previously described model sorts with cell-cell interaction systems (Krohl et al., 2020). Cell-cell interactions were evaluated for the pre-sort and post-sort yeast populations with Ag2-displaying E. coli (Figure 2A). The increased percentages of double-positive fluorescent events observed via flow cytometry suggested that there was substantial enrichment of yeast displaying Nb2 from all mixed populations of Nb1:Nb2:Nb3 (Figure 2A). The percentages of double-positive fluorescent events in control yeast-bacteria populations containing individual yeast clones displaying Nb1, Nb2, or Nb3 are depicted alongside the pre-sort and post-sort samples (Figure 2B).

Figure 2.

Figure 2.

Open in a new tab

Cell sorting on mixed yeast display-nanobody populations against an E. coli-displayed target antigen. A) Dot plots of coincubations containing pre- and post- sort yeast populations. B) Percentage of events in Q2 for incubation of Ag2-expressing E. coli with control yeast clones each displaying a single nanobody and with yeast populations comprised of mixed ratios of cells displaying one of three nanobodies, pre- and post-sort. C) Fold change of sequencing reads in pre-sort and post-sort samples. D) Correlation of fold enrichment values determined by flow cytometry analysis versus amplicon sequencing.

The fold enrichment could not be directly calculated using the percentage of interaction events. A calibration curve was fit to the 100:1:100, 10:1:10, 1:1:1, and Nb2-only interaction data to enable estimation of the fraction Nb2-displaying yeast within the population based on its binding to Ag2-displaying E. coli (Figure S2B). There was a strong linear correlation between the fraction of Nb2-displaying yeast versus the detected number of interaction events. The fold enrichment was determined by first calculating the expected fraction of Nb2-displaying yeast in the post-sort samples based on the detected interactions events, then dividing this value by the known initial pre-sort fraction of Nb2-displaying yeast. To corroborate the results of the flow cytometry analysis, we performed amplicon sequencing on yeast plasmid DNA isolated from pre-sort and post-sort samples to quantify the percentage of each of the three nanobody genes detected in each sample. The sequencing results indicated enrichment of Nb2 clones from all three starting ratios (Figure 2C). Additionally, there was a significant (p-value = 0.0004) linear correlation between the enrichment values calculated using flow cytometry analysis and sequencing (Figure 2D). On the other hand, the two assays yielded different fold changes in enrichment. These differences can be attributed to technical differences between both assays. The flow cytometry assay measures the percentage of double positive interaction events for a given population of co-incubated yeast and E. coli cells, which includes nonspecific cell-cell interactions. The DNA sequencing estimates the number of genes present in the full population, including cells in the population that did not exhibit detectable protein display at the time of flow cytometry analysis (Figure S3). To further evaluate the utility of this cell-cell interaction system for enriching rarer cognate interactions, we conducted a series of two-round model enrichments from mixed populations with starting ratios of 1:1:1, 10:1:10, 100:1:100, 1000:1:1000, and 10,000:1:10,000 (Figure S4). After two rounds of enrichment, we observed statistically significant enrichment from a starting Nb1:Nb2:Nb3 ratio of 1000:1:1000 via flow cytometry (Figure S4), and statistically significant round-over-round enrichment from 100:1:100 and 10:1:10 starting ratios. Thus, the cell-cell interaction system enabled single- and multi-round enrichment of a yeast-displayed nanobody to its cognate antigen presented on the E. coli surface.

To test the capability of this cell-cell interaction system for characterizing therapeutically relevant nanobody-antigen interactions beyond the model interaction pairs used above, we expressed a previously discovered nanobody IB10 on the yeast surface. IB10 is a high affinity binder discovered in a screen against the surface intimin protein of enterohemorrhagic E. coli (EHEC) (Ruano-Gallego et al., 2019). Intimin is a protein antigen that mediates bacterial attachment to host tissues through interaction with a translocated intimin receptor (Tir) that is secreted by EHEC and inserted into the host membrane. Given that a previously developed intimin overexpression model was shown to express intimin at biologically relevant levels comparable to that of wild type EHEC during host cell infection (Brady et al., 2011), we aimed to evaluate how yeast displaying the IB10 nanobody would interact with E. coli overexpressing intimin. For the cell-cell interaction events mediated by the IB10-intimin nanobody-antigen interaction pair, we observed a significant increase in the frequency of events exhibiting high levels of both blue and green fluorescence relative to the Ag2 and DH5α controls (Figure 3), suggesting that this system can be applied towards biologically relevant antigen targets using an E. coli surface overexpression model that is frequently used to evaluate intimin targeting (Brady et al., 2011).

Figure 3.

Figure 3.

Open in a new tab

Application of interaction model system for surface-displayed intimin and anti-intimin nanobody. A) Representative dot plots for yeast-E. coli coincubations for yeast displaying IB10 with E. coli displaying intimin or Ag2, or with the wild-type DH5α strain. B) Percentage of events in Q2 for combinations of yeast displaying IB10, Nb2, or the nondisplaying wild-type RJY100 strain with E. coli displaying intimin, Ag2, or the nondisplaying wild-type DH5α strain. Error bars represent the standard deviation of 3 biological replicates for yeast displaying nanobodies.

To differentiate between specific and nonspecific cell-cell interactions, coincubation conditions such as total cell density, yeast to bacteria cell ratio, and incubation time can be optimized for specific use cases. Cell-cell binding interactions can be affected by the affinity of the binding molecule to the target molecule and surface display densities on both yeast and bacterial cell surfaces. In the experiments described above, the number of displayed constructs on the yeast surface using the galactose-inducible Aga1p-Aga2p fusion system is approximately 3 × 104 fusion molecules per cell (Boder & Wittrup, 1997). The number of antigens on the E. coli surface displayed using the scheme employed in this work has not been quantified to our knowledge, but a rough estimate is 3.6 × 104 molecules per cell for aTc induction of intimin-anchored proteins (Wentzel et al., 2001). This study provides proof-of-concept demonstrations for differentiating high affinity nanobody-antigen interactions from nonspecific interactions. Although specific KD values for these nanobody-antigen pairs are not available as they were not reported in the previous cell-cell interaction study (Glass & Riedel-Kruse, 2018), they can be considered representative of high affinity interactions. The Kd of the Nb2-Ag2 interaction, which led to reliable cell-cell interactions in comparison to background levels, has been previously reported to be 190 ± 30 nM at room temperature (Guilliams et al., 2013). The threshold of detection for this cell-cell interaction system is applicable for discovery of nanobodies of similar or higher affinities for therapeutic applications (Muyldermans, 2013). Regarding the effect of surface display density, the expression of full-length nanobodies on the yeast surface is heterogeneous. Consistent with prior studies utilizing the galactose-inducible Aga1p-Aga2p display system, some cells appear to possess higher surface-densities of nanobodies available for interaction (Figure S3) (Chun et al., 2020; Huang & Shusta, 2005). We showed that yeast surface display densities and linker lengths play important roles in cell-cell interactions mediated by specific binding events (Figure 4).

Figure 4.

Figure 4.

Open in a new tab

Effect of yeast surface display density and linker length on detected yeast-E. coli interactions. A) Correlation between normalized median red fluorescence intensity of cells displaying cognate full-length nanobody and the baseline subtracted percentage of interaction events. B) Percentage of interaction events detected for different nanobody-antigen interaction pairs with varying linker lengths used for nanobody display on the yeast surface.

As a step towards investigating a wider range of coincubation conditions, we attempted to modulate the surface density of nanobodies present on the yeast surface to better understand the role of ligand surface density (or avidity) in this system. To control ligand display levels, we treated yeast cells with varying concentrations of dithiothreitol (DTT), which can have the combined effects of 1) reducing the two disulfide bridges linking Aga1p and Aga2p to dissociate the displayed protein from the cell surface (Stern et al., 2017) and 2) reducing the disulfide bridges within the displayed protein to change it to a conformationally inactive state. In independently conducted experiments of yeast and E. coli surface display of three interaction pairs—IB10-intimin, Nb2-Ag2, and Nb3-Ag3—we determined that the nanobody display level (and therefore surface density) was linearly correlated (p-values = 0.0009, 0.0047, and 0.0007, respectively) with the frequency of detected interaction events; higher display levels led to higher percentages of interactions (Figure 4A and Figure S5). The use of longer linkers may also change yeast-bacteria interaction strengths by altering the accessibility of displayed ligands. The linker length used in the main experiments of this study is 40 amino acids. A recent study using a yeast-mammalian cell interaction system investigated the use of an intermediate linker (80 amino acids) and long linker (641 amino acids) for cell-cell interaction experiments. The study reported that the long linker allowed for higher enrichment with yeast-mammalian cell interactions, likely due to improved surface accessibility, although the surface density of displayed ligands and percent of ligand-displaying cells were lower for cells displaying the long linker compared to the intermediate linker (Lown et al., 2021). We investigated cell-cell interactions with the IB10-intimin and Nb2-Ag2 nanobody-antigen interaction pairs using both the 80- and 641-amino acid linkers. In our hands, we observed an increase not only in cell-cell interactions but also in nonspecific cell-cell interactions with the longer linkers. Based on these data, it was not possible to determine if the longer linkers significantly enhanced specific nanobody-antigen interactions under these conditions (Figure 4B). Further exploration of the effects of these parameters on cell-cell interactions are warranted in a future study to fully understand the advantages and limitations of the longer linkers.

In this study, we demonstrate that flow cytometry can be used to identify yeast-bacteria interactions mediated by binding of surface available cognate nanobody-antigen pairs. Furthermore, we validate the use of fluorescent dyes to stain cells in this assay and show representative microscopy images of yeast-E. coli cell-cell complexes. These cell-cell complexes are stable enough to be enriched by FACS, as determined by both flow cytometry and sequencing of sorted populations. For model enrichment, under the conditions with which we performed validations, our data support the feasibility of screening small libraries, such as in the case of an affinity maturation or deep mutational scanning campaign. Finally, to test the capability of this cell-cell interaction system for characterizing therapeutically relevant nanobody-antigen interactions beyond the set of previously characterized synthetic interaction pairs, we test a biologically relevant model of EHEC adhesin (i.e., intimin) overexpression and show that the cell-cell interaction assay can discriminate between the cognate IB10-intimin interaction and non-specific interactions.

We evaluated a yeast-bacteria interaction system that could potentially be used as a platform in protein discovery. A yeast display system offers many advantages for ligand discovery in flow cytometry format (Cherf & Cochran, 2015; Valldorf et al., 2021), including: 1) a eukaryotic translation apparatus that facilitates biosynthesis of complex polypeptides including genetic encoding of different chemical functionalities and other chemical groups amenable to selective modification on the yeast surface (Islam et al., 2021; Lewis et al., 2021); 2) a robust cell wall to facilitate straightforward handling; 3) efficient propagation relative to other eukaryotic organisms; and 4) lack of susceptibility to many antibacterial compounds (Parachin et al., 2012). Bacterial surface-binding molecules have broad applications for biofilm control (Koo et al., 2017), therapeutic inhibition of bacterial adhesion to host cells (Kline et al., 2009), and selective targeting of bacterial species within a microbial community (Ting et al., 2020). With further optimization and understanding of the sensitivity requirements of the cell-cell interaction system, the yeast display-based cell-cell interaction assay described in the present study can provide a flexible platform for characterization of these binding molecules.

Experimental Procedures

Materials.

All DNA fragments and oligonucleotides for molecular cloning and sequencing were purchased from Genewiz. Restriction enzymes were purchased from New England Biolabs. Epoch Life Science miniprep kits were used for E. coli plasmid purification. Materials used for yeast plasmid purification are described in the methods section Preparation for amplicon sequencing and data analysis. Calcofluor white was purchased from Sigma-Aldrich. SYTO 9 and Alexa Fluor™ 488 NHS Ester (Succinimidyl Ester) were purchased from Thermo Fisher Scientific.

Plasmid construction.

Previously characterized orthogonal nanobody-antigen interaction pairs (Nb1-Ag1, Nb2-Ag2, Nb3-Ag3) used in an E. coli-E. coli interaction format were adapted to a yeast-E. coli interaction format by cloning the nanobodies into a pCTcon2 yeast display vector. The set of previously designed and characterized nanobody-antigen interaction pairs adapted to this study are Nb1 (anti-Akt3PH)-Ag1 (Akt3PH), Nb2 (anti-EPEA)-Ag2 (EPEA), and Nb3 (anti-P53TA)-Ag3 (P53TA). Plasmids for E. coli surface display of Nb1, Ag1, Nb2, Ag2, Nb3, and Ag3 were obtained from Addgene: pDSG372 (Addgene plasmid #115602), pDSG358 (Addgene plasmid #115599), pDSG375 (Addgene plasmid #115603), pDSG419 (Addgene plasmid #115600), pDSG398 (Addgene plasmid #115604), pDSG360 (Addgene plasmid #115601). The pCTcon2-nanobody constructs were cloned by PCR amplification of the nanobody genes with the oligonucleotides listed in Table S1, digestion of the pCTcon2 vector with NheI and BamHI (New England Biolabs), and Gibson assembly. The IB10 nanobody (Ruano-Gallego et al., 2019) was synthesized as a double-stranded, linear DNA fragment (FragmentGENE, Genewiz) and cloned by Gibson assembly into pCTcon2. pDSG_Intimin_full_length was provided by the Mougous lab at the University of Washington. pCT80 and pCT641 were provided by the Hackel lab at the University of Minnesota.

Culture conditions.

RJY100 yeast cells were transformed with the pCTcon2 nanobody display vectors. Yeast containing the pCTcon2 nanobody-display vector were grown in 2 ml SD-SCAA −Trp −Ura +pen/strep pH 4.5 in 15 ml polypropylene tubes at 30°C in a shaking incubator overnight to saturation, diluted to OD = 1 in 2 ml of the same medium, grown for an additional 2–3 h at 30°C to approximately OD = 2. Pen/strep was used at a 1× total concentration from a 100× stock solution, containing 10000 IU penicillin and 10000 μg/ml streptomycin (Corning). An aliquot of each culture corresponding to an OD of approximately 1 in a 4 ml volume was spun down. The supernatant was aspirated, and the cell pellet was resuspended in 4 ml SG-SCAA −Trp −Ura +pen/strep. The cultures were induced by incubation in a 20 °C shaking incubator for 16 hours. Three biological replicates were performed for each of Nb1, Nb2, and Nb3-displaying cells by picking three different colonies on the yeast transformation plates and treating them the same throughout the experiment. Wild-type RJY100 was grown in SDSCAA −Ura +pen/strep or SGSCAA −Ura +pen/strep following the corresponding steps above. Full-length surface display was assessed by labeling both the N-terminal hemagglutinin (HA) and C-terminal c-Myc tag flanking the displayed nanobody as described in a previous study (Stieglitz et al., 2018). One E. coli colony transformed with each of Ag1, Ag2, Ag3 were grown in 2 ml LB kan (50 μg/ml kanamycin) at 37 °C in a shaking incubator overnight to saturation, diluted to OD 0.05 in 4 ml LB kan, grown for an additional 2–3 hours to an OD of approximately 0.6, then induced with a final anhydrotetracycline (aTc) concentration of 200 ng/ml. The cultures were induced for approximately 16 hours at 37 °C in a shaking incubator. Wild-type DH5α was grown in LB, otherwise following the same steps above. ATc was prepared by making a 1 mg/ml stock solution in ethanol, then diluting it 1:10 in water to make a 100 μg/ml working stock solution.

Fluorescence staining.

After induction, the OD of each culture was measured using a cuvette with a 1 cm path length by diluting the culture 1:10 in PBS pH 7.4 (Note: Precise OD measurements are critical for these experiments since the frequency of nonspecific interactions are correlated to the ratios of yeast: E. coli. See Figure S6 for details). A corresponding volume of each culture required to prepare a 2 ml volume at OD 1 was pipetted from each tube and pelleted. Each cell pellet was resuspended in 2 ml PBS. All centrifugation steps were performed at 5000 rpm (corresponding to approximately 2000 × g) for 2 minutes on an Eppendorf 5424 tabletop centrifuge, unless noted otherwise. Calculations assume 1 x 109 cells/ml E. coli at OD 1 and 3 x 107 cells/ml yeast at OD 1 (Ausubel, 2002). To stain yeast, 2 μl of a 1 g/L stock solution of calcofluor white (CFW) was added to each of the 2 ml volumes (1 μl per OD1/ml) and incubated for at least 15 minutes. To stain E. coli, 2 μl of a 3.34 mM stock solution of SYTO 9 in DMSO (LIVE/DEAD® BacLight™ Bacterial Viability and Counting Kit) was added to each of the 2 ml volumes (corresponding to 1 μl SYTO9 per 1 ml of cell suspension at OD 1) and incubated in the dark for at least 15 minutes. For microscopy, 4× higher cell densities were used for practical reasons: the density at which cell complexes were optimally detected above background on the flow cytometer is not suitable for microscopy because the cells are too sparsely distributed on the agarose pads. The tubes were briefly mixed, then incubated on a rotary wheel for 15 minutes while covered from light. E. coli cells were washed by pelleting the cells, aspirating the supernatant, and resuspending in PBS, twice. SYTO 9 was previously shown to stain yeast cells, whereas CFW did not significantly stain E. coli at the concentrations used, so no wash step was deemed necessary for CFW-stained yeast prior to coincubation with E. coli.

Flow cytometry.

Flow cytometry analysis was performed on an Attune NxT flow cytometer. For all the figures in the main text and unless otherwise specified, the relevant flow cytometer settings were: 25 μl/min flow rate, 10000 events collected per sample, FSC:1, SSC:200, BL1, RL1, VL1: 250. The minimum FSC voltage was initially used since previous experiments determined that higher FSC voltages caused many events to appear outside the axis limits for coincubated cultures, especially for those with an expected interaction. Though the FSC scatter data captures the same trends as reported in fluorescence data since a higher FSC can indicate formation of large cell aggregates, these data exhibit high coefficients of variation (CVs) associated with low voltage settings. Therefore, FSC scatter data was not used in analyses reported here. Higher FSC voltages led to increased detection of events for the same sample of E. coli (see graph in Figure S8 for voltages used). An FSC voltage of 200 was used for the experiments testing the effect of BSA on cell-cell interactions. Data analysis was performed using FlowJo and Microsoft Excel. Statistical analysis (two-tailed t-test) was performed in Microsoft Excel.

Microscopy.

Microscopy was performed using a DMi8 automated inverted microscope (Leica Microsystems, #11889113) equipped with a CCD camera (Leica Microsystems, #DFC300 G), and LED405 (Ex 375-435 nm, Em 450-490 nm, exposure time – 30 ms, gain 3.3) and YFP (Ex 490–510 nm, Em 520–550 nm, exposure time – 30 ms, gain 3.3) filter cubes. 1 μl of coincubation cultures were spotted on agarose pads (2 % w/v, 1 mm thick).

Fluorescence-activated cell sorting.

Cell sorting was performed on a Bio-Rad S3e Cell Sorter. 1:1:1, 10:1:10, and 100:1:100 mixtures of Nb1:Nb2:Nb3 displaying yeast cells were prepared by normalizing ODs and adding corresponding volumes to achieve the indicated ratios. Sorting gates were drawn on dot plots with FL1 and FL4 on the axes, with gates positioned to capture events exhibiting high levels of fluorescence in the FL1 green emission channel. The FL4 far-red emission channel was chosen since the S3e sorter model used in this work is not equipped with a laser capable of exciting CFW. FL4 has minimal overlap with the emission spectrum of SYTO 9. The same general procedure as in the analytical flow cytometry experiment was followed for quantification of cell-cell interactions. Pre-sort populations (prior to mixing with E. coli) were saved for yeast miniprep and amplicon sequencing. After sorting, cells were recovered in SD-SCAA −Trp −Ura +pen/strep pH 4.5 for 2 days. Post-sort populations were saved and coincubation assays were performed concurrently with single and mixed ratios of nanobody-displaying yeast as controls.

Preparation for amplicon sequencing and data analysis.

Yeast miniprep was performed using the E.Z.N.A.® Plasmid DNA Mini Kit I (OMEGA Bio-tek, #D6943-02) with slight modifications. Approximately 5-10 OD*ml cells were resuspended in solution I (250 μl) supplemented with lyticase (1000 U, Sigma, catalog# SAE0098-20KU) and incubated at 37 °C for 1 h. The remaining steps were performed according to the manufacturer’s protocol. The extracted DNA was quantified using a spectrophotometer and approximately 100 ng was used as template for PCRs. All samples were prepared for amplicon sequencing using unique, barcoded primers flanked by Illumina sequencing adaptors (Table S1). 15–20 cycles of PCR were performed to amplify the region spanning the nanobody using the primers listed in Table 1. The barcoded samples from triplicate experiments were pooled and sent for amplicon sequencing (2×250 bp, Genewiz, New Jersey, USA). For each pooled sample, we received approximately 100,000 sequencing reads. These data were processed according to a previously described bioinformatic workflow using Geneious Prime® 2020.2.4 (Trivedi et al., 2021). Briefly, the reads were paired and merged using the BBMerge package and filtered for poor-quality reads using the BBDuk package. The reads were then mapped onto the reference genes (Nb1, Nb2, and Nb3) using BowTie2 (Langmead & Salzberg, 2012). The number of reads mapped to Nb1, Nb2 and Nb3 were used to estimate the fold enrichment. Statistical analysis (Pearson correlation) was performed in MATLAB.

Supplementary Material

Supplementary material

Acknowledgments

This work was supported in part by a grant from the National Institute of General Medical Sciences of the National Institutes of Health (1R35GM133471 to J.A.V.) and the Karol Family Professorship (to K.L.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or Tufts University. Plasmids pCT80 and pCT641 were a gift from the laboratory of Benjamin Hackel. Plasmid pDSG_Intimin_full_length was a gift from the laboratory of Joseph Mougous.

Footnotes

Conflicts of interest

There are no conflicts of interest to declare.

References

  1. Ausubel FM, Brent R, Kingston RE, Moore DD, Seldman JG, Smith JA, Struhl K (2002). Short Protocols in Molecular Biology (5th ed.). [Google Scholar]
  2. Bacon K, Burroughs M, Blain A, Menegatti S, & Rao BM (2019). Screening Yeast Display Libraries against Magnetized Yeast Cell Targets Enables Efficient Isolation of Membrane Protein Binders. ACS Comb Sci, 21(12), 817–832. 10.1021/acscombsci.9b00147 [DOI] [PubMed] [Google Scholar]
  3. Boder ET, & Wittrup KD (1997). Yeast surface display for screening combinatorial polypeptide libraries. Nat Biotechnol, 15(6), 553–557. 10.1038/nbt0697-553 [DOI] [PubMed] [Google Scholar]
  4. Brady MJ, Radhakrishnan P, Liu H, Magoun L, Murphy KC, Mukherjee J, … Leong JM (2011). Enhanced Actin Pedestal Formation by Enterohemorrhagic Escherichia coli O157:H7 Adapted to the Mammalian Host. Front Microbiol, 2, 226. 10.3389/fmicb.2011.00226 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chagnot C, Zorgani MA, Astruc T, & Desvaux M (2013). Proteinaceous determinants of surface colonization in bacteria: bacterial adhesion and biofilm formation from a protein secretion perspective. Front Microbiol, 4, 303. 10.3389/fmicb.2013.00303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cherf GM, & Cochran JR (2015). Applications of Yeast Surface Display for Protein Engineering. Methods Mol Biol, 1319, 155–175. 10.1007/978-1-4939-2748-7_8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chun J, Bai J, & Ryu S (2020). Yeast Surface Display System for Facilitated Production and Application of Phage Endolysin. ACS Synth Biol, 9(3), 508–516. 10.1021/acssynbio.9b00360 [DOI] [PubMed] [Google Scholar]
  8. Garai P, & Blanc-Potard A (2020). Uncovering small membrane proteins in pathogenic bacteria: Regulatory functions and therapeutic potential. Mol Microbiol, 114(5), 710–720. 10.1111/mmi.14564 [DOI] [PubMed] [Google Scholar]
  9. Glass DS, & Riedel-Kruse IH (2018). A Synthetic Bacterial Cell-Cell Adhesion Toolbox for Programming Multicellular Morphologies and Patterns. Cell, 174(3), 649–658 e616. 10.1016/j.cell.2018.06.041 [DOI] [PubMed] [Google Scholar]
  10. Guilliams T, El-Turk F, Buell AK, O'Day EM, Aprile FA, Esbjorner EK, … De Genst E (2013). Nanobodies raised against monomeric alpha-synuclein distinguish between fibrils at different maturation stages. J Mol Biol, 425(14), 2397–2411. 10.1016/j.jmb.2013.01.040 [DOI] [PubMed] [Google Scholar]
  11. Huang D, & Shusta EV (2005). Secretion and surface display of green fluorescent protein using the yeast Saccharomyces cerevisiae. Biotechnol Prog, 21(2), 349–357. 10.1021/bp0497482 [DOI] [PubMed] [Google Scholar]
  12. Islam M, Kehoe HP, Lissoos JB, Huang M, Ghadban CE, Berumen Sanchez G, … Van Deventer JA (2021). Chemical Diversification of Simple Synthetic Antibodies. ACS Chem Biol, 16(2), 344–359. 10.1021/acschembio.0c00865 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jones ML, Alfaleh MA, Kumble S, Zhang S, Osborne GW, Yeh M, … Mahler SM (2016). Targeting membrane proteins for antibody discovery using phage display. Sci Rep, 6, 26240. 10.1038/srep26240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kline KA, Falker S, Dahlberg S, Normark S, & Henriques-Normark B (2009). Bacterial adhesins in host-microbe interactions. Cell Host Microbe, 5(6), 580–592. 10.1016/j.chom.2009.05.011 [DOI] [PubMed] [Google Scholar]
  15. Koo H, Allan RN, Howlin RP, Stoodley P, & Hall-Stoodley L (2017). Targeting microbial biofilms: current and prospective therapeutic strategies. Nat Rev Microbiol, 15(12), 740–755. 10.1038/nrmicro.2017.99 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kozlowski MT, Silverman BR, Johnstone CP, & Tirrell DA (2021). Genetically Programmable Microbial Assembly. ACS Synth Biol, 10(6), 1351–1359. 10.1021/acssynbio.0c00616 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Krohl PJ, Kim KB, Lew L, VanDyke D, Ludwig SD, & and Spangler JB (2020). A suspension cell-based interaction platform for interrogation of membrane proteins. AIChE Journal, 66(12). https://doi.org/ 10.1002/aic.16995 [DOI] [Google Scholar]
  18. Lahav-Mankovski N, Prasad PK, Oppenheimer-Low N, Raviv G, Dadosh T, Unger T, … Margulies D (2020). Decorating bacteria with self-assembled synthetic receptors. Nat Commun, 11(1), 1299. 10.1038/s41467-020-14336-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Langmead B, & Salzberg SL (2012). Fast gapped-read alignment with Bowtie 2. Nat Methods, 9(4), 357–359. 10.1038/nmeth.1923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lewis AK, Harthorn A, Johnson SM, Lobb RR, & Hackel BJ (2021). Engineered protein-small molecule conjugates empower selective enzyme inhibition. Cell Chem Biol. 10.1016/j.chembiol.2021.07.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lown PS, Cai JJ, Ritter SC, Otolski JJ, Wong R, & Hackel BJ (2021). Extended yeast surface display linkers enhance the enrichment of ligands in direct mammalian cell selections. Protein Eng Des Sel, 34. 10.1093/protein/gzab004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Muyldermans S (2013). Nanobodies: natural single-domain antibodies. Annu Rev Biochem, 82, 775–797. 10.1146/annurev-biochem-063011-092449 [DOI] [PubMed] [Google Scholar]
  23. Parachin NS, Mulder KC, Viana AA, Dias SC, & Franco OL (2012). Expression systems for heterologous production of antimicrobial peptides. Peptides, 38(2), 446–456. 10.1016/j.peptides.2012.09.020 [DOI] [PubMed] [Google Scholar]
  24. Pizarro-Cerda J, & Cossart P (2006). Bacterial adhesion and entry into host cells. Cell, 124(4), 715–727. 10.1016/j.cell.2006.02.012 [DOI] [PubMed] [Google Scholar]
  25. Ruano-Gallego D, Fraile S, Gutierrez C, & Fernandez LA (2019). Screening and purification of nanobodies from E. coli culture supernatants using the hemolysin secretion system. Microb Cell Fact, 18(1), 47. 10.1186/s12934-019-1094-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Salema V, Manas C, Cerdan L, Pinero-Lambea C, Marin E, Roovers RC, … Fernandez LA (2016). High affinity nanobodies against human epidermal growth factor receptor selected on cells by E. coli display. MAbs, 8(7), 1286–1301. 10.1080/19420862.2016.1216742 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Salema V, Marin E, Martinez-Arteaga R, Ruano-Gallego D, Fraile S, Margolles Y, … Fernandez LA (2013). Selection of single domain antibodies from immune libraries displayed on the surface of E. coli cells with two beta-domains of opposite topologies. PLoS One, 8(9), e75126. 10.1371/journal.pone.0075126 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Stern LA, Csizmar CM, Woldring DR, Wagner CR, & Hackel BJ (2017). Titratable Avidity Reduction Enhances Affinity Discrimination in Mammalian Cellular Selections of Yeast-Displayed Ligands. ACS Comb Sci, 19(5), 315–323. 10.1021/acscombsci.6b00191 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Stieglitz JT, Kehoe HP, Lei M, & Van Deventer JA (2018). A Robust and Quantitative Reporter System To Evaluate Noncanonical Amino Acid Incorporation in Yeast. ACS Synth Biol, 7(9), 2256–2269. 10.1021/acssynbio.8b00260 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ting SY, Martinez-Garcia E, Huang S, Bertolli SK, Kelly KA, Cutler KJ, … Mougous JD (2020). Targeted Depletion of Bacteria from Mixed Populations by Programmable Adhesion with Antagonistic Competitor Cells. Cell Host Microbe, 28(2), 313–321 e316. 10.1016/j.chom.2020.05.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Trivedi VD, Chappell TC, Krishna NB, Shetty A, Sigamani GG, Mohan K, … Nair NU (2021). In-depth sequence-function characterization reveals multiple paths to enhance phenylalanine ammonia-lyase (PAL) activity. ACS Catal., 12(4), 2381–2396. https://doi.org/ 10.1021/acscatal.1c05508 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Valldorf B, Hinz SC, Russo G, Pekar L, Mohr L, Klemm J, … Zielonka S (2021). Antibody display technologies: selecting the cream of the crop. Biol Chem. 10.1515/hsz-2020-0377 [DOI] [PubMed] [Google Scholar]
  33. Wentzel A, Christmann A, Adams T, & Kolmar H (2001). Display of passenger proteins on the surface of Escherichia coli K-12 by the enterohemorrhagic E. coli intimin EaeA. J Bacteriol, 183(24), 7273–7284. 10.1128/JB.183.24.7273-7284.2001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Yang Z, Wan Y, Tao P, Qiang M, Dong X, Lin CW, … Lerner RA (2019). A cell-cell interaction format for selection of high-affinity antibodies to membrane proteins. Proc Natl Acad Sci U S A, 116(30), 14971–14978. 10.1073/pnas.1908571116 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material
NIHMS1881741-supplement-Supplementary_material.pdf (2.4MB, pdf)

RESOURCES