View all hidden authors and organizations

Edited by Caroline S. Harwood, University of Seattle, approved on September 17, 2020 (reviewed on May 20, 2020)

Form obeys function is true for organisms, including bacteria, because the shape and form they adopt contribute to their biological properties. However, it is unclear whether bacteria can actively change their shape to adapt to the environment. Vibrio cholerae is a pathogen that causes diarrheal cholera and adopts a characteristic "comma"-shaped cell morphology. Here, we show that the intracellular signal molecule cycle di-GMP drives the curved Vibrio cholerae to adopt a straight cell morphology, which is conducive to a fixed biofilm lifestyle, while the curved Vibrio cholerae is more suitable for swimming. Our research provides a clear example of how bacteria actively change their form to affect function.

The cell morphology of rod-shaped bacteria is determined by the rigid network of peptidoglycans that form the cell wall. The change in the shape of the rod, such as a curved rod, occurs by manipulating the process of cell wall synthesis. The human pathogen Vibrio cholerae usually exists in the form of curved rods, but straight rods are observed under certain conditions. Although this seems to be a regulated process, the regulatory pathways that control the transformation of Vibrio cholerae cells and the benefits of switching between rod-shaped and curved shapes have not yet been determined. We proved that the cell shape of Vibrio cholerae is regulated by the bacterial second messenger cyclic dimeric guanosine monophosphate (c-di-GMP) by post-transcriptionally inhibiting the expression of crvA, which encodes an intermediate filamentous protein The gene is necessary for the bending of Vibrio cholerae. cholera. This regulation is mediated by the transcription cascade, which also induces the production of biofilm matrix components, indicating that the cell shape is co-regulated with the infertility induction of Vibrio cholerae. In the process of microcolony formation, wild-type Vibrio cholerae cells tend to exist in a straight rod shape, while maintaining high curvature through genetically engineered cells will reduce microcolony formation and biofilm density. Conversely, direct Vibrio cholerae mutants slowed their swimming speed when they used flagella to move in liquid. Our results indicate that the regulation of bacterial cell shape is a mechanism that increases the lifestyle adaptability of plankton and biofilms.

Morphology is an important biological feature because it determines how organisms interact with their physical world. All aspects of form, such as shape, length, and presence of appendages, are all affected by selection pressure and help to adapt to a specific niche (1). Not surprisingly, bacteria take on different shapes, from simple rods and cocci to spirals and curves. However, the fitness benefits of each shape are not always clear (2).

The shape of the bacteria largely depends on the structure of the peptidoglycan (PG) cell wall, because bacteria without PG adopt a polymorphic or l-shaped shape (3, 4). The main form of gram-negative bacteria is rod-shaped. During the growth of rod-shaped cells, new PG subunits are added to the growing PG layer through successive rounds of PG lysis and subunit insertion. These subunits are along the side of the rod, the middle cell or in some cases The extreme point of (3, 5). The curved rod shape is common in many aquatic organisms, such as the freshwater bacterium Caulobacter crescentus and the opportunistic human pathogen Vibrio cholerae. In his 1928 book entitled "The Morphological Variation of Vibrio cholerae", Arthur Henrici measured the changes in the shape of Vibrio cholerae over time during the batch culture growth curve, and observed the change from spherical to straight and curved Various shapes of rods (6). The curvature of the Caulobacter crescentus bacterium is mediated by the protein crescentin (7). Similarly, the study of Bartlett et al. (8) Determine that the gene crvA encoding the periplasmic intermediate filament protein CrvA introduces curvature by changing the insertion rate of PG. Mutants lacking crvA grow straight, have reduced migration in soft agar, and are less toxic in animal models of infection (8). The ORF adjacent to crvA, annotated as crvB, works with CrvA, and CrvAB together is sufficient to produce curvature in normal straight cells (9).

Interestingly, in Vibrio cholerae, the change of cell shape is controlled by quorum sensing (QS), which is a phenomenon in which microorganisms communicate and coordinate behaviors based on population density (8). Specifically, Henrici and Bartlett et al. Highly curved rods are observed in the high cell density (HCD) population, and straighter rods are observed in the low cell density (LCD) population (6, 8). In addition to changing the shape of cells, QS also regulates many population density-dependent phenotypes through complex regulatory pathways, such as biofilm formation. One of these is the bacterial second messenger cyclic dimeric guanosine monophosphate (c-di-GMP). ) (Figure 1) (10, 11).

QS and circulating di-GMP signaling and related phenotypes in Vibrio cholerae. Vibrio cholerae showed an inverse relationship between cell density and intracellular c-di-GMP. Under HCD conditions with high concentrations of autoinducers, the intracellular concentration of c-di-GMP is lower (11, 12). Under HCD conditions, the curvature of cells increases and their tendency to adhere to surfaces and form biofilms decreases. In contrast, cells under LCD conditions have low concentrations of autoinducers and high c-di-GMP (see Reference 13). Under LCD conditions, cells are more likely to form biofilms (14). In addition, Vibrio cholerae has lost its typical curved rod shape and appears as a straight rod (6).

In many bacteria, flagellar movement and biofilm formation are regulated by c-di-GMP signaling (15). C-di-GMP is produced by an enzyme called diguanylate cyclase (DGC) and is degraded by an enzyme called phosphodiesterase (PDE). In Vibrio cholerae, c-di-GMP binds and activates the transcription factors VpsR and VpsT (16, 17). Then, these transcription factors induce the expression of operons, which encode enzymes necessary for the synthesis of Vibrio polysaccharides (VPS), which are the main extracellular polysaccharide components of the biofilm matrix, and matrix-related proteins that lead to the formation of mature biofilms (18 ⇓ ⇓ –21). In addition, c-di-GMP inhibits motility by inhibiting the transcription of flagellar biosynthesis genes, inhibiting the expression of transcription factor TfoY by binding to c-di-GMP-dependent riboswitches, and inducing MshA fimbriae extension to mediate surface attachment ( 12, 22). The intracellular c-di-GMP concentration is controlled by a variety of signals, including QS, where cells in the LCD state have a higher c-di-GMP concentration than cells in HCD (11).

Although it has been demonstrated that the cell shape of Vibrio cholerae changes between a curved shape and a straight shape, the regulatory pathways that control this transition and the potential ecological benefits of this change are not fully understood (6, 8). In this study, we reported that increased intracellular c-di-GMP concentration straightened Vibrio cholerae by reducing CrvA expression. Locking Vibrio cholerae in a curved form reduces the formation of microcolony and biofilm formation. Conversely, compared to curved wild-type (WT) cells, Vibrio cholerae cells locked in a straight morphology will reduce swimming speed. Therefore, curved and straight cell morphologies are the best morphologies for movement and biofilm formation, respectively. Our results demonstrate how bacteria actively control cell shapes to adapt to different lifestyles.

When exploring the effect of c-di-GMP on the stress response pathway of Vibrio cholerae (23, 24), we noticed that strains with high intracellular c-di-GMP concentrations are more often straight rod-shaped. We explored this by using strains that ectopic express isopropyl-β-D-thiogalactoside (IPTG)-inducible DGC (QrgB) to quantify cell shape parameters at different intracellular c-di-GMP concentrations. It was found that the DGC synthesizes c-di-GMP or catalyzes inactive DGC (QrgB*). We have previously demonstrated that this expression of QrgB in Vibrio cholerae produces physiologically relevant concentrations of c-di-GMP (12). We observed that when IPTG was not added, the curvature of the cell axis was reduced in strains containing active DGC due to the leaky expression of the inducible promoter (Figure 2, P value <1e-3). When IPTG was added to increase QrgB expression, a further decrease in curvature was observed (P value<6e-4), supporting our hypothesis that high c-di-GMP concentration would cause the cells to straighten (Figure 2). When the QrgB* mutant protein was induced, this curvature change was not observed, indicating that the morphological change is unique to the synthesis of c-di-GMP (Figure 2). Increasing IPTG in QrgB*-expressing cells resulted in a 1.16-fold increase in cell length, while QrgB-expressing cells, regardless of IPTG, resulted in a 1.02-fold decrease in width (SI Appendix 1, Figure S1). Although these subtle changes are statistically significant based on our data, they may reflect experimental noise and have nothing to do with biology. Therefore, c-di-GMP mainly changes the cell morphology of Vibrio cholerae by reducing the curvature.

C-di-GMP reduces cell curvature in a dose-dependent manner. (A) Representative phase contrast photomicrographs of cells in the early quiescent phase. The cells contain plasmids encoding IPTG inducible inactive DGC (QrgB*) with 1 mM IPTG and IPTG inducible active DGC (QrgB) with 0, 0.1, 0.5, and 1 mM IPTG (OD600 = 1.3). (Scale bar on the last image of the panel, 5 µm.) (B) In a population expressing inactive DGC (QrgB*, light) or active DGC (QrgB, dark), the distribution of cell curvature as a function of IPTG concentration. The boxes represent the first, second, and third quartiles. The dots represent the average value. Each distribution represents 1,000 to 1,200 cells, analyzed and aggregated from four to five separate experiments. (Inset) Estimate the contribution of each experimental factor to the change in the median (point = mean, thick line = 90% confidence interval (CI), thin line = 98% CI). The P value for the 90% CI effect that does not include zero is ≤0.05. The P value for 98% CI effects that do not include zero is ≤0.01.

In Vibrio cholerae, three c-di-GMP-dependent transcription factors (VpsR, VpsT, and FlrA) and two c-di-GMP-dependent riboswitches (Vc1 and Vc2) are known to trigger multiple phenotypes. Genes such as biofilm formation, exercise, DNA repair and catalase production (16, 19, 22⇓ ⇓ ⇓ ⇓ ⇓ –28). We hypothesize that VpsT or VpsR are involved in the c-di-GMP-dependent regulation of cell curvature because of their role in regulating multiple c-di-GMP-dependent phenotypes (23, 24, 29). To test this hypothesis, we measured the cell shape parameters of the ∆vpsR and ∆vpsT mutants under high (QrgB) and unchanged (QrgB*) concentrations of c-di-GMP. Although the width changed little under all experimental conditions, in the absence of vpsT or vpsR, QrgB expression reduced cell length (P value <6e-3) (SI Appendix 1, Figure S2). In addition, regardless of c-di-GMP and genetic background, overexpression of VpsT will cause a slight increase in length (P value <6e-3) (SI Appendix 1, Figure S2). These data indicate that c-di-GMP and VpsT can play a role in other aspects of cell morphology.

The c-di-GMP-mediated decrease in curvature observed in the paternal background is lost in the ΔvpsR mutant (Figure 3). The VpsR expression of the multicopy plasmid in the ΔvpsR mutant restored the c-di-GMP-dependent decrease in cell curvature (P value <1e-2). In the ∆vpsT background, QrgB expression leads to a decrease in cell curvature (P value<4e-2), but the degree is different from the maternal background (Figure 3). Regardless of the concentration of c-di-GMP (P value<5e-4), the supplementation of VpsT in the multi-copy plasmid resulted in a decrease in cell curvature (P value<5e-4), although more was observed when the concentration of c-di-GMP increased. Large curvature reduction (P value<1e-2). VpsT is a c-di-GMP-dependent transcription factor, but this result shows that high levels of VpsT expression are maintained at the same concentration of c-di-GMP Enough to reduce the curvature (16). To determine whether VpsT requires c-di-GMP to reduce curvature, we overexpressed VpsT in a Vibrio cholerae mutant lacking 12 DGC coding genes (Δ12 DGC), which had previously been shown to have undetectable c-di- GMP level (30). In this context, VpsT overexpression does not reduce curvature, indicating that c-di-GMP is required for this regulation (SI Appendix 1, Figure S3).

The transcription factors VpsR and VpsT are necessary for the C-di-GMP-dependent reduction of curvature. The cell curvature distribution of different mutant strains in the early stable population supplemented with 100 µM IPTG expressing inactive DGC (QrgB*, light color) or active DGC (QrgB, dark color) (OD600 = 1.3). The ΔvpsR strain is complementary to an empty plasmid or a plasmid expressing VpsR or VpsT. The ΔvpsT strain is complementary to an empty plasmid or a plasmid expressing VpsR or VpsT. The boxes represent the first, second, and third quartiles. The dots represent the average value. Each distribution represents 1,000 to 1,200 cells, analyzed and aggregated from four to five separate experiments. (Inset) Estimate the influence of QrgB, VpsT and VpsR expression on the median curvature of the population. (Point = average, thick line = 90% CI, thin line = 98% CI).

Since VpsR directly activates vpsT transcription, we hypothesized that the role of VpsR in controlling cell shape is to increase VpsT expression, thereby reducing curvature (17). We tested this hypothesis by overexpressing VpsT in the ∆vpsR mutant, and found that regardless of the intracellular c-di-GMP concentration, the cells were linear (Figure 3). The overexpression of VpsR in the vpsT mutant only partially restored the inhibition of curvature by c-di-GMP, indicating that vpsT is completely dependent on c-di-GMP to reduce cell curvature. In general, this analysis supports the view that under high c-di-GMP conditions, VpsR mainly inhibits curvature indirectly by inducing the transcription of vpsT, while VpsT and c-di-GMP are sufficient to reduce curvature.

In Vibrio cholerae, cell bending is caused by the intermediate filamentous protein CrvA by reducing net growth on the short axis relative to the long axis (8). We hypothesized that the c-di-GMP-dependent inhibition of cell curvature is due to the negative regulation of crvA transcription. To verify this hypothesis, we generated a transcription reporter gene with a crvA promoter (PcrvA, 358 bp upstream of the crvA transcription start site) fused with luciferase. When integrated with crvA ORF at a heterologous site on the chromosome, this promoter region is sufficient to complement the ΔcrvA mutant (SI Appendix 1, Figure S4D).

Contrary to our hypothesis, PcrvA transcriptional activity is not affected when the c-di-GMP concentration is increased by QrgB expression at any culture density examined (Figure 4A) (slope difference, 95% confidence interval [CI] [- 0.11, 0.10]). Consistent with previous studies that observed increased curvature at high cell densities (6, 8), PcrvA transcriptional activity increased with increasing cell density (Figure 4A) (P value <2.5e-4).

The regulation of CrvA expression occurs at the post-transcriptional level. (A) Normalized luciferase reporter gene activity of PcrvA-luciferase transcription fusion as a function of cell density (OD600) under high (QrgB, dark) and unchanged (QrgB*, light) c-di-GMP conditions (Relative Light Unit [RLU]) The conditions in the population that grow to the early stable period and supplement the 100 µm IPTG. Dots represent aggregated data from four biological replicates. (B) qRT-PCR analysis of relative crvA transcription levels between high (QrgB) and unchanged (QrgB*) c-di-GMP conditions. Each dot represents an independent repetition. (C) CrvA-HIS grows to the early stationary phase under high (QrgB, dark) or unchanged (QrgB*, light) c-di-GMP conditions (left) and ∆vpsT (right) strains, supplemented by 100 µM IPTG. The shape represents independent repetitions, and the y-axis is log10. (Inset) Estimate the effect of QrgB and VpsT expression on CrvA accumulation. (Point = average, thick line = 90% CI, thin line = 98% CI).

Next, we quantified the effect of c-di-GMP on the abundance of crvA mRNA to test whether c-di-GMP affects CrvA accumulation through a post-transcriptional mechanism. The abundance of crvA mRNA under high c-di-GMP conditions is approximately 1.5 times lower (Figure 4B) (95% CI = [-2.4, -0.7]). To determine whether CrvA protein levels have changed, we generated a C-terminal 6-histidine tag (CrvA-HIS) fusion. The curvature of cells in the CrvA-HIS background is approximately 82% of that of WT (95% CI = [78 to 84%]), indicating that CrvA-HIS is functional (SI Appendix 1, Figure S5). Western blot analysis showed that at high c-di-GMP concentrations, the reduction of CrvA protein in the wild-type background was greater (95% CI = [-10, -1.2] than in the ∆vpsT background (95% CI = [- )] 2.4 to 0.51]) (Figure 4C). These data indicate that high c-di-GMP concentration reduces crvA mRNA and CrvA protein levels in a VpsT-dependent manner without changing the activity of crvA promoter, indicating that c-di-GMP has a negative effect on cells. Reduce the curvature of CrvA abundance through a post-transcriptional mechanism.

The ability of Vibrio cholerae to attach to the surface and initiate biofilm formation depends on the intracellular c-di-GMP and the transcription factors VpsR and VpsT (11, 19, 25). Recent reports indicate that the yield of c-di-GMP and VPS substrates increases within hours of surface attachment (31, 32). This indicates that c-di-GMP may be dependent on changes in cell shape, because Vibrio cholerae is initiating biofilm formation, and VpsT induces the production of VPS and related biofilm matrix proteins (16, 32, 33). Since the form fits the function, we hypothesized that the straight cells of Vibrio cholerae are better at forming biofilms. To test this hypothesis, we first cultivated wild-type Vibrio cholerae under static conditions and measured the morphology of attached cells that changed over time from the initial contact to the early stages of microcolony formation. We chose these time points because in the later stages of biofilm development, Vibrio cholerae grows in dense clusters, and some cells in the biofilm become perpendicular to the surface, making it difficult to segment and reproduce from two-dimensional (2D) images. Quantify curvature (34, 35). We observed that during the experiment (7 hours), the cells attached to the glass surface became straighter than the cells initially seeded (Figure 5) (P value<1.2e-2). To assess whether only the cells associated with the surface became straight, we collected the cells from the planktonic stage at the 8 hour time point and measured the curvature. It is important that 8 hours after seeding, adherent and non-adherent cells are still at a low cell density due to seeding conditions (for experimental details, please refer to Materials and Methods). The planktonic cells also showed a straight cell morphology similar to the morphology of the cells attached to the surface, indicating that the changes in the cell shape during the experiment may reflect the QS control of the cell shape (SI Appendix 1, Figure S6). These data indicate that the relationship between cell shape, population status, and c-di-GMP is similar in adherent and non-adherent Vibrio cholerae cells (6, 8). Therefore, under conditions that promote biofilm formation, Vibrio cholerae cells exhibit a linear morphology.

When microcolonies are formed, Vibrio cholerae becomes straighter. (Top) Representative images from the inoculum and WT Vibrio cholerae attached to the glass coverslip after 2, 4, 5, 6, and 7 hours. The cells were stained with FM4-64 before imaging. (Scale bar, 5 µm.) (Bottom) Cell curvature distribution at different time points after attachment. The boxes represent the first, second, and third quartiles. The dots represent the average value. Each distribution represents 500 to 1,000 cells analyzed and pooled from two independent experiments. (Inset) The estimated effect of time after surface attachment on the curvature of the median population. (Point = average, thick line = 90% CI, thin line = 98% CI).

A single Vibrio cholerae cell inoculated on the surface can produce large three-dimensional biofilms through surface adhesion along the horizontal axis of the cell, surface replication, and conversion from horizontal to vertical cell positioning (34, 35). This process largely depends on the VPS matrix and related matrix proteins, as well as cell-cell and cell-surface contact (34⇓ –36). In addition, single-cell studies of Vibrio cholerae biofilm formation images show that the cells are relatively straight, including strains with elevated c-di-GMP concentrations, although the curvature has never been directly measured (32, 34, 35). We found that the adherent Vibrio cholerae grows into a straight rod under the conditions of biofilm formation, which indicates that this morphology may be suitable for biofilm formation. If this is the case, we hypothesize that locking the cells in a curved form, regardless of the intracellular c-di-GMP concentration, will have a negative impact on the development of biofilms. We tested this hypothesis by using the PBAD promoter (pCrvA) to express CrvA from a multicopy plasmid in the context of ΔcrvA. We found that the basic expression of this construct (that is, without the addition of arabinose) was sufficient to restore the curvature to the same level as that of wild-type planktonic cells (SI Appendix 1, Figure S4A). Compared with QrgB*, cells maintain their curvature when CrvA and QrgB are co-expressed, indicating that c-di-GMP does not negatively regulate the expression of ectopic CrvA (SI Appendix 1, Figure S4A). A slight increase in curvature was observed in cells expressing CrvA ectopic at high c-di-GMP concentrations, although it is unclear whether this difference is biologically significant (SI Appendix 1, Figure S4A). Therefore, we generated three different populations: 1) cells capable of transitioning between curved and straight rods (WT carrying the control vector pHERD20C), 2) constitutively straight cells (ΔcrvA-pHERD20C), and 3) Regardless of the concentration of c-di-GMP, the composition is curved (ΔcrvA-pCrvA).

Through microcolony analysis 8 hours after inoculation, we found that the area of ​​single microcolony of WT and linear populations averaged 10 µm2 and 12 µm2, indicating that the linear mutant was not damaged in the formation of microcolony (Figure 6 A and B). Curved The microcolonies formed by the mutants are smaller than the size of WT (5 µm2, P value <2.5e-4) and contain only two to three cells (the area of ​​one cell = 2.6 ± 1.4 µm2) (Figure 6 A and B). By analyzing the total surface fluorescence of the image in Figure 6, we found that the WT and straight populations have similar fluorescence levels. The curved population has a higher level of fluorescence; however, this difference is not statistically significant (SI Appendix 1, Figure S7). In any case, the data shows that the reduction in the area of ​​the microcolonies that lock the curved cells is not due to the inability to adhere to the surface under the test conditions.

Curvature affects the development of microcolony and the formation of biofilm at the population level. (A) Representative images of microcolonies of WT Vibrio cholerae carrying the control vector pHERD20C (left), ∆crvA carrying the control vector pHERD20C (middle), and the CrvA expression vector pCrvA (right). Stain microcolonies with FM4-64 before imaging. (The scale bar in the lower right corner of each image, 20 µm.) (B) The distribution of microcolonies pooled from two independent experiments, totaling 700 to 2,000 microcolonies. The boxes represent the first, second, and third quartiles. The dots represent the average value. Each distribution represents 500 to 1,000 cells analyzed and pooled from two independent experiments. (C) Biofilm biomass formed by different strains. The dot shapes represent independent repetitions. (Inset) Estimate the effect of genomic crvA or ectopic CrvA expression on the median microcolony area and biofilm biomass (dot = average, thick line = 90% CI, thin line = 98% CI).

Based on the changes in microcolony formation, we hypothesized that locking the cells in a curved rod would result in a reduction in the biomass of mature biofilms. Using crystal violet to quantify the total biomass of attached cells cultured statically in a glass tube, we found that the WT and straight populations have indistinguishable biofilm biomass (Figure 6C) (biomass difference, 95% CI = [-0.40, 0.22] ). Compared with other strains, the biofilm biomass of the bent mutant is reduced, supporting the view that the reduction of microcolony formation leads to a decrease in the overall biofilm biomass (Figure 6C) (biomass difference, 95% CI = [-0.91,- 0.28], P value = 1.8e-3). In view of the indistinguishability of the surface adhesion and growth rate of these strains (SI Appendix 1, Figure S8), we concluded that the observed differences in microcolony and biofilm formation are due to the cell shape. Our analysis of cell curvature in the process of biofilm formation shows that cell shape is a regulated process, and the decoupling of curvature from c-di-GMP changes will have a negative impact on Vibrio cholerae biofilm formation.

Because wild-type Vibrio cholerae bends under growth conditions that promote a planktonic lifestyle, such as low intracellular c-di-GMP concentration, we hypothesized that the bend provides advantages for cells that use flagellar movement to swim in liquid. To test this hypothesis, we cultured wild-type and ΔcrvA mutants in minimal medium supplemented with pyruvate to the late exponential stage (most cells are highly motile under these conditions) and tracked individual cells to quantify their swimming speed. We determined that curved rods are on average 5.5% faster than straight rod swimming (Figure 7A) (relative speed increases, 95% CI = [5.5%, 5.9%], P value <1e-5). The difference in reversal frequency between the trajectories of each strain is not significant (relative difference in reversal frequency, 95% CI = [-0.87%, 2.8%]) (Figure 7B), indicating that cell curvature does not significantly affect cell behavior. In addition, compared with WT cells, the cell length of ΔcrvA cells is increased by 1.08 times, and the main feature is a 4-fold difference in curvature (SI appendix, Figure S9). Therefore, curvature is beneficial for moving cells because it can significantly increase the speed of bacteria that are already swimming relatively fast. Overall, our results support that Vibrio cholerae regulates CrvA expression through c-di-GMP signaling to regulate its cell shape, thereby increasing its adaptability during biofilm formation and free swimming.

The curvature increases the swimming speed of the flagellar cells. The distribution of (A) swimming speed and (B) recovery frequency of WT (light) and ΔcrvA mutants (dark) was determined by four independent repeated single-cell tracking that grew to the early stable phase. The dots represent the average value. Each distribution represents more than 1,000 cell trajectories. (Inset) Estimate the effect of missing crvA on swimming speed and reversal frequency (dot = average, thick line = 90% CI, thin line = 98% CI).

Rod-shaped derivatives, such as helical or arc-shaped, are adapted to genes with specific shape-changing functions (2). Although significant progress has been made in understanding how bacteria construct and manipulate their cell walls to change cell shape, little is known about whether and how these processes are affected by environmental conditions. Through a variety of c-di-GMP synthesis and degradation enzymes to sense and respond to specific environmental cues, the c-di-GMP signaling network is one of the main mechanisms by which bacteria integrate environmental information to regulate their lifestyles (such as motility and biofilm formation) In this study, we show that c-di-GMP is a regulator of cell shape in the aquatic organism Vibrio cholerae, where high c-di-GMP concentration reduces the cell curvature to generate straight rods. This response depends on the genes crvA and the biofilm-promoting transcription factors VpsR and VpsT, prompting us to evaluate the influence of shape on biofilm formation. We found that the adherent cells in the developing microcolonies adopt a straight morphology, and the retention of curvature leads to defects in the formation of microcolonies and the production of mature biofilms. Conversely, when pushed by a single flagella in a liquid environment, straight rods swim slower than curved cells. This work emphasizes the importance of controlling the shape of cells when changing from a single-cell plankton lifestyle to the development of multicellular communities such as biofilms. In addition, the regulation of crvA provides an example of bacterial second messengers controlling cell shape.

In 1928, Arthur Henrici, a pioneer in microbiology, discovered that the cell shape of Vibrio cholerae changes with the culture density. He pointed out: "Embryonic cells [those cells in the early growth curve] are therefore large, plump and straight [...] Therefore, mature cells Slender and curved, it is a typical Vibrio form" (6). In addition, Bartlett et al. When sampling from the quiescent population, the cells were found to be more curved, and it was determined that the result was caused by QS (8). A study comparing the transcripts of Vibrio cholerae from LCD cells and HCD cells found that crvA is a QS regulatory gene, which is more highly expressed in the HCD state (37). Vibrio cholerae uses QS as a mechanism to change intracellular c-di-GMP, where cells in LCD culture have a higher concentration of c-di-GMP than cells in HCD culture (Figure 1) (11, 38). In addition, if the production of c-di-GMP is stimulated by another mechanism (for example, ectopic expression of DGC), cells in HCD can form biofilms, indicating that c-di-GMP is epistatic to QS (11). In our study, we also observed a similar phenomenon, in which the curvature of the HCD cells in the early quiescent phase decreased with the expression of QrgB (Figure 2). Therefore, the discovery of cells with high c-di-GMP concentration in LCD directly supports our conclusion that c-di-GMP controls cell shape in Vibrio cholerae (6, 8, 11).

The c-di-GMP-dependent transcription factor VpsT is sufficient to suppress the curvature of planktonic cells. In addition, our data suggests that VpsR has a potentially direct role in regulating curvature, because VpsR and c-di-GMP partially reduce curvature in the context of ∆vpsT (Figure 3). These data indicate that the VpsR/VpsT regulatory node is responsible for straightening cells during microcolony formation. Although we cannot test this directly because vpsR and vpsT are necessary for microcolony formation, some studies strongly support this model (16, 25, 26, 33). Specifically, adherent cells produce VPS within 15 minutes, and produce structural matrix proteins within 1 hour after surface attachment, both of which are positively regulated by the VpsR/VpsT regulatory node and high c-di-GMP concentration ( 32). In addition, the production of c-di-GMP in Vibrio cholerae attached to the surface increased over time, reaching a peak around 6 hours after attachment (31). For other bacteria, such as Pseudomonas aeruginosa, similar increases in c-di-GMP have been observed after surface attachment (39, 40). These studies support a model that after attaching to the surface, the c-di-GMP concentration increases and activates the VpsR/VspT regulatory node, leading to surface adhesion. As cells divide on the surface, the negative regulation of crvA leads to changes in the curvature of progeny cells during biofilm development (Figure 8).

c-di-GMP is a model for controlling the cell shape and biofilm formation of Vibrio cholerae. (A) Under low c-di-GMP conditions (eg, HCD, the presence of bicarbonate) (11, 41), VpsR cannot induce VpsT and VPS operons. Under these conditions, due to the production of CrvA, Vibrio cholerae is flagellated and exists as a curved rod (8). (B) Under high c-di-GMP conditions (eg, low temperature, LCD) (11, 42), c-di-GMP-binding VpsR is active and induces the expression of VpsT, which in turn interacts with c-di- Incorporate good manufacturing practices. VpsR and VpsT together form a node in the c-di-GMP signaling network, inducing the expression of the VPS operon, leading to the formation of mature biofilms (16, 28). In addition, the VpsR/VpsT node activated by c-di-GMP negatively inhibits the expression of crvA at the post-transcriptional level through an unknown mechanism (dashed arrow indicates inhibition), resulting in a decrease in the level of CrvA protein and producing straight rods during cell division.

When examined under a microscope, pandemic isolates of Vibrio cholerae usually retain their Vibrio-like morphology (43, 44). However, Wucher et al. It was found that nutrient deficiencies in pandemic isolates induce cell filamentation, which contributes to chitin colonization and VPS-independent biofilm formation under flow conditions (44). Although the cells are longer than the reference strain, they seem to retain their curvature, indicating that this mechanism is not related to crvA and c-di-GMP (8, 44). However, in multiple reports, individual Vibrio cholerae cells in the VPS-dependent microcolonies appear straight rather than curved under underwater flow cell conditions; however, the degree of change in cell shape has never been quantified (32 , 34, 35). Therefore, this study quantified the curvature of the 2D developing microcolonies (Figure 6).

Although the curvature of Vibrio cholerae has been shown to promote exercise in plate-based exercise assays, it has not been determined how shape affects specific parameters of exercise, such as chemotaxis at the single-cell level or swimming speed (8). Our results show that the influence of cell shape on movement is not limited to promoting movement in confined spaces, such as the matrix formed in low agar movement assays. By tracking individual cells swimming in a liquid medium, we found that constitutive straight cells swim slower than curved cells (Figure 7). Our research provides experimental evidence to support the theoretical results, that the predicted intermediate curvature increases the cell swimming efficiency of the elongated rod in solution by increasing rotational resistance rather than translational resistance, resulting in more energy transfer to the flagella and more Fast swimming speed (45, 46).

Examples of Vibrio or spiral forms include Campylobacter crescentis, Helicobacter pylori, and Vibrio cholerae, as well as many bacteria found in reservoirs (47). Curvature may affect how each organism adapts to its own ecological niche. For example, the curvature of C. crescentus is important for positioning the cell poles near the surface during biofilm growth under flow conditions. The straight mutant cannot form a biofilm under these conditions, but has no defects under static conditions, which is contrary to the conclusion drawn here for Vibrio cholerae (48). The spiral shape of the human pathogen Helicobacter pylori is considered to be an adaptive feature that promotes the health of the human host; however, direct clinical isolates of H. pylori have been identified as having obvious advantages and disadvantages in certain physiological environments during infection (49 ). These results emphasize that the shape of cells can have different effects on bacterial species according to their own specific physiological functions.

Rod-shaped cells have been observed in all three areas of life, indicating that this shape has an evolutionary advantage compared to other shapes (5). For example, mutations that change the morphology of the rod-shaped bacteria Escherichia coli or Rhodobacter sphaeroides can also negatively affect biofilm formation, which may affect adaptability in certain environments (50). In addition, the analysis of cell packaging in dense clusters emphasizes the importance of rod-shaped microbial communities in cell positioning and self-organization (51, 52). Although spiral or curved bacteria like H. pylori and C. crescentus can obtain the adaptive benefits of curved or straight cells through mutagenesis and loss of function, Vibrio cholerae has evolved to use environmental sensing and c-di-GMP Signal to achieve this. 48, 49). Our work shows that the cell shape in bacteria is a plasticity feature that is regulated by environmental signals to maximize swimming speed in the plankton state and cell surface interactions during the development of biofilms.

Except for the SI appendix, Figure S3 (SI appendix S1, using Vibrio cholerae E7946 and Δ12DGC mutants), all experiments used strains derived from Vibrio cholerae C6706 Str2. (For a complete list of strains used in this study, see SI Appendix 2, Table S1.) As shown in Figure 2-4, a biofilm mutant derivative (ΔvpsL) was used as the parent strain to prevent high c- Aggregates are formed under di-GMP, allowing individual bacteria to be imaged in solution. The expression vectors for ΔvpsT and ΔvpsR mutants and VpsT and VpsR were previously generated (17). pKAS32 was used to construct deletion and knock-in strains (53). Unless otherwise stated, all cloning was done by Gibson Assembly (NEB). For the deletion construct pKAS32_ΔcrvA, pKAS32 was digested with XbaI and SacI and purified by gel extraction (Promega). The primers (Integrated DNA Technologies) are designed with NEBuilder (www.nebuilder.com, NEB) to contain the appropriate 5'end and 3'gene-specific end required for Gibson assembly. Specifically, using Vibrio cholerae gDNA as a template, 700 bp upstream and downstream of crvA (VCA1075) were amplified by PCR (Q5 polymerase, NEB). For the knock-in construct pKAS32_crvA-HIS, the replacement allele with a 6-histidine tag before the stop codon was amplified by PCR, where the HIS tag was incorporated 5'of the oligonucleotide primer. For the knock-in construct pKAS32_pcrvA-CrvA, PCR was used to amplify the natural promoter and open reading frame (ORF) of crvA (relative to the ATG start codon -358 bp, until the translation stop codon) and inserted into the locus VC1807, one This pseudogene encodes a true frameshift, providing a neutral site for insertion into the genome (54). For pBBRLux_pcrvA, the crvA upstream sequence from -358 to -1 bp relative to the ATG start codon was amplified, and Gibson was cloned into the BamHI and SpeI restriction sites of pBBRlux. For pHERD20C_CrvA, the ORF of crvA was amplified and cloned into the KpnI and SacI sites of pHERD20C. Successful clones were screened by colony PCR using GoTaq polymerase (Promega), and sequenced by Sanger sequencing (Genewiz Inc.) to ensure that no mutations were incorporated during the cloning process. The construction of knockout and knock-in strains was carried out according to the protocol of Skorupski et al. (53). The plasmid was transferred from S17-λpir E.coli to Vibrio cholerae, and reverse selection was performed with Polymixin B (10 U/mL). All plasmids and oligonucleotides can be found in the SI appendix, tables S2 and S3 (SI appendix S2) (55, 56). Unless otherwise specified, both E. coli and Vibrio cholerae use ampicillin (100 µg/mL), kanamycin (100 µg/mL) and/or chloramphenicol (10 µg/mL) in LB when needed Reproduction. Unless otherwise specified, the inducer IPTG is usually 100 µM.

The overnight culture of Vibrio cholerae was diluted 1:100 into 2 mL LB, and the appropriate concentrations of ampicillin and IPTG were added. The culture grows to an OD600 of 1.3 to 1.5, at which time the cells are diluted to an OD600 of 0.5 in a microcentrifuge tube. Cut the 1% agarose pad in deionized water into ∼20 × 20 mm squares and place them on a microscope slide (75 × 25 × 1.0 mm [length {L} × height {H} × width {W}] , Alkali Scientific Inc.) Spot a total of 2 µL of diluted culture on a glass cover slip (22 × 22 mm, #1.0 thickness, Alkali Scientific), and then gently place the cover slip on the agarose pad. Use a Nikon Eclipse Ti-E inverted microscope equipped with a 100x phase contrast oil immersion objective lens (1.4 NA), Nikon perfect focus system, Prior H117 ProScan motorized stage, Lumencor PEKA white light LED translucent lens for phase contrast microscope inspection light source and Andor Zyla 4.2 sCMOS camera. The microscope and camera are controlled by a computer workstation with MATLAB (Mathworks Inc.) and Micromanager version 1.4 (micromanager.org). Use a flat-field image of a clean glass slide to correct for uneven illumination and image artifacts. Use the central axis method in the Fiji plug-in MicrobeJ to detect and segment cells. The thresholds are set as follows: area (0 to 4.5, μm2), length (maximum 1.5, μm) and width (0 to 3, μm) (57). Use R to analyze the data of the segmented image to plot the curvature (defined as the average value of the reciprocal of the radius of curvature along the given cell axis) (μm-1), and the width (defined as the average value of the width measurements taken) along a given The central axis of the cell (μm) and the length (defined as the pole-to-pole distance) (μm) (57). Crop representative images and use Fiji software (58) to add scale. Calculate the median cell length, width, and curvature of each biological replicate. Use different linear mixed-effect models with lognormal distribution link functions to calculate the median posterior probability distribution and test the P-values ​​of different factors (QrgB, VpsR, VpsT, and CrvA induction). Figure 1 and ED1: (curvature, length, width) ∼QrgB* IPTG + (QrgB* IPTG|replicate), Figure 2: (curvature, length, width) ∼complement+QrgB: complement+(complement+QrgB:complement|copy ), Figure 4: (curvature, length, width) ∼ time + (time|copy).

The overnight culture was diluted 1:1,000 in 1 mL 1×PBS (Sigma) in a 10-fold serial dilution. Place the UV-sterilized #1 coverslip (22 × 22 cm) in a six-well plate (Costar) immersed in 1 mL of LB. By further diluting 5 times, inoculate the cultures of the six groups of biological replicates into the wells with glass slides, the final dilution is 1:5,000, and then gently rotate. Allow microcolonies to develop during static incubation at 21°C. At a given time point, aspirate the medium of two biological replicates, wash the wells with 1 mL 1× PBS, and use 200 µL of membrane stain N-(3-triethylammonium propyl) to adhere to the coverslip Bacteria stained)-4-(6-(4-(diethylamino)phenyl)hexatrienyl)pyridinium dibromide (FM4-64) (Sigma), final concentration 20 µg/mL, lasting 5 minute. Remove stains by additional washing with 1 mL of 1× PBS. Arrange a small portion of agarose pads (~5 × 5 mm) in a 20 × 20 cm square pattern on a glass microscope slide. The glass coverslip containing the stained microcolonies was then inverted and placed on top of the agarose pad. Biofilm imaging was performed using a Leica DM5000b epi-fluorescence microscope equipped with a Spot Pursuit CCD camera and X-cite 120 illumination system with a 100x brightfield objective lens (1.4 NA). Use the dsRed filter set (excitation 560/40 nm, emission 620/60 nm) to acquire images. Each slide has at least 20 fields of view for each biological replicate at each time point. Use the manual interface option to manually outline the cells in the microcolonies in MicrobeJ, with at least 500 to 1,000 cells in each repeat. The data from MicrobeJ analysis is exported to R for analysis. The posterior probability distribution of the influence of time on the morphology of adherent cells is calculated using a linear mixed effects model, which has a log-normal distribution link function [(curvature, length, width) ~ time + (time | repetition)].

For imaging of non-adherent cells, the overnight culture was diluted 1:100 in 1 mL 1×PBS (Sigma) in a 10-fold serial dilution. The 1:100 diluted overnight culture was diluted by adding 20 μL to three separate wells of a 6-well plate (Costar) containing 4 mL of LB for additional dilution, then gently swirling and incubating at 21 °C for 7 to 8 Hour. For each biological replicate, collect 3 of the 4 mL cultures and perform membrane filtration using a stainless steel vacuum filter (Millipore 13 mm) without touching the bottom of the well. Collect the cells on a pre-soaked 0.45 µm polyvinylidene fluoride membrane filter (Millipore), transfer to a 1.5 mL microcentrifuge tube, and resuspend in 20 µL membrane stain FM4-64 (20 µg/mL) 5 minutes. A total of 5 µL of stained cells were spotted on the glass coverslip, which was gently placed on the agarose pad. Cell imaging is performed by the same method as microcolony imaging.

For single-time point biofilm analysis, an eight-well micro-chamber slide (μ-Slide, eight-well glass bottom, ibidi) was used. A 1:1,000 diluted overnight culture was added to 200 µL of LB with an appropriate concentration of chloramphenicol in a ratio of 1:5 to a single well of a micro-chamber slide. Each slide has three strains that are biologically replicated. Microcolonies are developed by statically incubating microchamber slides at 21°C for 8 hours, resulting in WT microcolony sizes between 10 and 20 µm2. Remove the medium from the slide by aspiration, wash each well with 200 µL 1× PBS, and stain the microcolonies with FM4-64 (150 µL) at a final concentration of 20 µg/mL for 5 minutes. The remaining stain was washed with 200 µL 1× PBS, and the biofilm was imaged by a fluorescent microscope using a Leica DM5000b epi-fluorescence microscope inverted microchamber (place the glass bottom up), as described above. At least 20 fields of view were captured for each repetition of each strain, and the resulting image was processed with Fiji (enhanced contrast, saturated pixel percentage, 0.3) (58). Then use MicrobeJ to analyze the processed image and set as follows: background type (dark), detection mode (basic), area (4.5 to 100, μm2) and circularity (0 to 1). The data is exported to R for analysis. For all fluorescence microscopes, crop representative images and add scale bars using Fiji Software (58). A linear mixed effects model with a lognormal distribution link function is used to calculate the posterior probability distribution of mutations and complementary effects in the microcolony region [area∼strain+complementary+(strain+complementary|copy)]. In order to measure the total surface fluorescence, the above image was processed by Fiji (subtracting the background, rolling ball radius: 50 pixels) (58) and the Fiji measurement function was used to analyze the total integrated density. The data is exported to R for analysis.

Four biologically replicated overnight cultures of Vibrio cholerae containing PcrvA and PcrvA transcriptional fusion of luciferase in pBBRlux were diluted 1:100 in 1 mL LB, and supplemented with ampicillin, ampicillin in a 1.5 mL microcentrifuge tube Chloramphenicol and IPTG. Aliquot a total of 150 µL of cell solution into the wells of a black 96-well plate (Costar), and repeat the technique. Incubate the plate at 35°C with shaking at 220 RPM. Hourly, use Envision microplate reader (Perkin-Elmer) for luciferase (relative light unit [RLU]) and OD595 measurement. A linear mixed effect model with a normal distribution link function [RLU ∼ OD600:QrgB + (OD600:QrgB|replicate)] is used to calculate the posterior probability distribution of the relationship between light emission and light density.

RNA isolation and qRT-PCR analysis were performed according to the previously described protocol (24). In short, the overnight culture was diluted to an initial OD600 of 0.040 in 2 mL LB supplemented with ampicillin and IPTG and grown at 35°C and 220 RPM until the OD600 was ~1.3. 1 mL was co-precipitated for each replicate sample, and RNA was extracted using TRIzol reagent according to the manufacturer's instructions (Thermo Fisher Scientific). The purified RNA was quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific). A total of 5 µg of purified RNA was treated with DNase (Turbo DNase, Thermo Fisher Scientific). Use GoScript Reverse Transcription Kit (Promega) for cDNA synthesis. Dilute cDNA 1:30 into molecular biology grade water, and use SYBR Green (Applied Biosystems) to quantify the amplification. The reaction included 5 µL 2.5 µM primer 1, 5 µL 2.5 µM primer 2, 5 µL diluted cDNA template and 15 µL 2× SYBR Green, which consisted of dNTP and AmpliTaq Gold DNA polymerase. Each plate has technical replicates and biological replicates, and there is no reverse transcriptase reaction to detect genomic DNA contamination. The StepOnePlus Real Time PCR system is used for qRT-PCR with the following thermal cycling conditions: 95 °C for 20 s, then 95 °C for 2 s and 60 °C for 30 s for 40 cycles. The melting curve is included to ensure that the PCR product has a single amplicon. The data is analyzed by the ∆∆Ct method using gyrA as the reference target. The experiment was repeated for 10 biological replicates on different days, and data from each experiment was pooled.

Overnight cultures of parents or ∆vpsT backgrounds that integrate 6-histidine-tagged crvA on the crvA locus (CrvA-HIS) with QrgB* or QrgB in a 50 mL baffled flask in 10 mL LB Dilute 1:100 in the medium, supplemented with ampicillin and IPTG. The culture was grown to an OD600 of ~1.3, collected by centrifugation (2.5 minutes at 7,000 × g), resuspended in 200 µL lysis buffer (20 mM Tris·HCl, 1% SDS, pH 6.8), and transferred to 1.5 mL Microcentrifuge tube and boil at 95 °C for 10 minutes. Centrifuge the boiled lysate (1 minute, 12,000 × g) to precipitate insoluble materials, and transfer the supernatant to a new 1.5 mL microcentrifuge tube. The sample was diluted 1:20 in 1x PBS, and the protein concentration was quantified by a detergent compatible protein assay using bovine serum albumin as the standard (BioRad). The lysate was normalized to 5 µg/µL in 2× loading buffer (lysis buffer supplemented with 5 mM β-mercaptoethanol and Coomassie Blue). A total of 20 µL of standardized lysate and size standards are loaded into a prefabricated 4 to 20% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel (4 to 20% Mini-PROTEAN) TGX precast protein gel, Bio-Rad (Precision Protein Plus, Bio-Rad). The gel was run at 90 V for 90 minutes in 1×Tris/glycine/SDS running buffer (Bio-Rad) at room temperature. The blot was transferred to a nitrocellulose membrane using a Mini Trans-Blot system (Bio-Rad) with TBST (Tris buffered saline containing Tween 20, pH 8.0)-methanol (20% vol/vol). The transfer was performed at 4 °C at 250 mAmps for 2 hours. After transfer, remove the blot and place it in blocking buffer (TBST supplemented with 5% skimmed milk), and incubate for 1.5 hours with stirring at room temperature. Remove the blocking buffer and replace it with 20 mL of blocking buffer, supplemented with 4 µL of anti-HIS primary antibody (mAb, mouse, GenScript), and incubate the blot overnight at 4 °C with stirring. The next day, wash the blot 3 times by removing the used blocking buffer, adding 20 mL of blocking buffer, and stirring at room temperature for 2 to 3 minutes. After the last wash, add 20 mL of blocking buffer, supplemented with 5 µL of horseradish peroxidase (HRP) conjugated rabbit anti-mouse secondary antibody (Bio-Rad), and incubate at room temperature for 2 hours with stirring. After incubating with the secondary antibody, wash the blot with 5 cycles of 10 mL blocking buffer and stir for 1 minute. Mix reagents 1 and 2 from the Pierce ECL kit according to the manufacturer's instructions, and use the Amersham Imager 600 with the "Chemiluminescence" setting to detect chemiluminescence. The image is removed from the imager, uploaded to Fiji, cropped to remove protein standards, and processed by contrast enhancement (contrast enhancement, 0.3% saturated pixels). Use the gel analyzer function to measure the band strength. Two independent experiments with two biological replicates were repeated on different days (n = 4). The posterior probability distribution of the influence of VpsT and QrgB on CrvA-His accumulation is calculated using a linear mixed-effects model, which has a log-normal distribution link function [strength∼ QrgB*VpsT + (QrgB*VpsT|replicate)].

Dilute the overnight culture of Vibrio cholerae 1:100 into 1 mL LB and supplement with chloramphenicol, put it into a new sterile 18 × 150 mm borosilicate test tube, and let it stand at 21 °C Incubate for 8 hours. After 8 hours, remove the culture medium and unattached bacteria, and wash the biofilm twice with 1 mL of 1× PBS (remove the PBS washing solution by suction). Add 1 mL of Crystal Violet (CV) (0.4%) solution to each tube and stain the biofilm for 10 minutes. The stain was removed by suction, and the stained biofilm was washed twice with 1 mL of 1× PBS. Elute the CV with ethanol and measure the absorbance at 570 nm (OD570). The posterior probability distribution of the mutation and complementary effects of biofilm biomass is calculated using a linear mixed effects model, which has a log-normal distribution link function [biomass∼strain+complement+(strain+complement|copy)].

Wild-type and crvA mutant cells were grown in the lowest salt of M9 (52 mM Na2HPO4, 18 mM K2HPO4, 18.69 mM NH4Cl, 2 mM MgSO4, pH 7) and supplemented with 10 µM FeSO4, 20 µM C6H9Na3O9, and 36 mM sodium pyruvate ( shake). RPM) in liquid culture at 37 °C. The culture was sampled during the early stabilization period (1.9 × 109 colony forming units [cfu]/mL), diluted to 107 cfu/mL in fresh medium, and incubated for 15 minutes before tracking. Follow the previously described protocol (59) to track swimming cells in the liquid. In short, polyvinylpyrrolidone (PVP) was added to the sample at 0.05% wt/vol to prevent adhesion to the glass slide. A total of 6 µL of each sample was dropped on the glass slide and trapped under a 22 × 22 mm, #1.5 cover glass sealed with wax and paraffin to form a thin water film (10 ± 2 µm) for video microscopy . The sample is kept at 37 °C during the tracking process. A 40× objective lens (Plan Fluor 40×, Nikon Instruments, Inc.) mounted on an inverted microscope (Eclipse Ti-E) was used to record the swimming cell’s performance at 20 frames per second using an sCMOS camera (Andor Zyla 4.2, Oxford Instruments) Image, Nikon Instruments). Use phase contrast to irradiate the cells. Use a custom script (59) to analyze images to detect and locate cells, and use the µ-track package (60) to reconstruct cell trajectories. The analysis of cell trajectories was done in MATLAB (The Mathworks, Inc.), as described previously (59). When the cells are not reversed, the average swimming speed of each cell is calculated by the average instantaneous speed along the trajectory. Calculate the posterior probability distribution and P value of the difference between average swimming speed and reversal frequency for each strain using a linear mixed effects model with a log-normal distribution link function [(average velocity, reversal frequency) ∼ strain+ (1| copy)].

The overnight culture was diluted 1:100 into LB supplemented with appropriate antibiotics, grown until the culture reached the exponential phase, and further diluted to an OD590 of 0.002. Transfer a total of 200 μL of the diluted culture to a 96-well microplate (Costar). The culture was grown in a Sunrise (Tecan) plate reader at 37°C for 24 hours, and OD590 was measured every 10 minutes. In order to calculate the growth rate, by calculating the first and second derivatives of OD590 with respect to time, the OD measurement value is trimmed to the exponential stage of growth, and then only the first and second derivative measurements are retained. The growth rate is extracted by fitting an exponential function to the pruned data using a nonlinear least squares algorithm.

All statistical analysis is done using Bayesian sampling of generalized linear models with their respective mixed effects (see each section), using RSTAN (61) and BRMS packages (62) in R(63) and four chains, each Each chain has 1,000 warm-up iterations and at least 2,500 sampling iterations. The non-information prior is set as the default value generated by BRMS. These plots are generated using ggplot2 and tidybayes package (64, 65). The posterior probability distribution of the effect size of each experimental treatment is used to calculate the 90% and 98% confidence intervals reported in the illustration. Given the model and data, use Bayes' rule to calculate the posterior probability to update the prior distribution of each parameter. The confidence interval is the Bayesian analogy of the confidence interval. It is defined as the interval in which the parameter value decreases with a certain probability. Given the model and sample data, the probability that the population parameter is within the 90% confidence interval is 90%. The 90% confidence interval is defined as an interval. If different samples from the same population are used to construct multiple times, the population parameters will be included 90% of the time. The confidence interval provides a more intuitive and direct representation of the possible values ​​of the overall parameter. A detailed explanation can be found in the references. 66, 67.

All research data is included in the article and supporting information.

This work was supported by CMW NIH grants GM109259, GM110444 and AI143098; NSF Grant 1714612 to YSD and CMW; and Michigan State University funding to YSD

↵1 Current address: Department of Molecular, Cellular and Developmental Biology, University of Michigan, Ann Arbor, Michigan 48103.

Author contributions: NLF, BYH, NTQN, JLF, YSD, and CMW design research; NLF, BYH, and NTQN conducted research; NLF, BYH, NTQN, JLF, YSD, and CMW analysis data; NLF, YSD, and CMW wrote this paper .

The author declares no competing interests.

This article is directly contributed by PNAS.

View relevant content online, such as comments.

This article contains online support information at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2010199117/-/DCSupplemental.

Released under PNAS license.

Thank you for your interest in advertising on PNAS.

Note: We only ask you to provide your email address so that the people you recommend the page to know that you want them to see it and that it is not spam. We do not capture any email addresses.

Feedback privacy/legal

Copyright © 2021 National Academy of Sciences. Online ISSN 1091-6490. PNAS is a partner of CHORUS, COPE, CrossRef, ORCID and Research4Life.