Utilization of a photoactivatable antigen system to examine B-cell probing termination and the B-cell receptor sorting mechanisms during B-cell activation

Significance

B-cell receptor (BCR) and antigen engagement induces several responses resulting in B-cell activation. However, it has been difficult to study these responses due to their dynamic nature. To solve this problem, a photoactivatable antigen, caged 4-hydroxy-3-nitrophenyl acetyl (caged-NP), was developed. B cells contacting caged-NP exhibited probing behaviors that are cell intrinsic with strict dependence on F-actin remodeling. B-cell probing behaviors were terminated within 4 s after the photoactivation of caged-NP. The termination of B-cell probing was concomitant with the accumulation response of the BCRs into the BCR microclusters. The analysis of temporally segregated single molecule images demonstrated that antigen binding induced trapping of BCRs into the BCR microclusters is a fundamental mechanism for B cells to acquire antigens.

Abstract

Antigen binding to the B-cell receptor (BCR) induces several responses, resulting in B-cell activation, proliferation, and differentiation. However, it has been difficult to study these responses due to their dynamic, fast, and transient nature. Here, we attempted to solve this problem by developing a controllable trigger point for BCR and antigen recognition through the construction of a photoactivatable antigen, caged 4-hydroxy-3-nitrophenyl acetyl (caged-NP). This photoactivatable antigen system in combination with live cell and single molecule imaging techniques enabled us to illuminate the previously unidentified B-cell probing termination behaviors and the precise BCR sorting mechanisms during B-cell activation. B cells in contact with caged-NP exhibited probing behaviors as defined by the unceasing extension of membrane pseudopods in random directions. Further analyses showed that such probing behaviors are cell intrinsic with strict dependence on F-actin remodeling but not on tonic BCR signaling. B-cell probing behaviors were terminated within 4 s after photoactivation, suggesting that this response was sensitive and specific to BCR engagement. The termination of B-cell probing was concomitant with the accumulation response of the BCRs into the BCR microclusters. We also determined the Brownian diffusion coefficient of BCRs from the same B cells before and after BCR engagement. The analysis of temporally segregated single molecule images of both BCR and major histocompatibility complex class I (MHC-I) demonstrated that antigen binding induced trapping of BCRs into the BCR microclusters is a fundamental mechanism for B cells to acquire antigens.

The immune system uses immune receptors to sense and acquire antigens. Antigen binding induces a series of dynamic changes in the biophysical behaviors and biochemical features of the immune receptors, and these changes determine the fate of a cell (1–3). However, it has been difficult to accurately capture and thus comprehensively investigate these changes because they usually occur very quickly after immune recognition (1, 4). For example, recent live cell imaging studies showed that the B lymphocytes swiftly accumulate the surface-expressed B-cell receptors (BCRs) into the contact interface between the B cells and the antigen-presenting substrates to form a specialized membrane structure, the B-cell immunological synapse (IS) (1, 4). Moreover, both our studies and those of others showed that these accumulation events are sensitive to the biochemical and biophysical features of the antigens that B cells likely encounter in vivo (4, 5). These features include but are not limited to antigen density (6, 7), antigen affinity (6, 7), antigen valency (8–13), the mobility of the antigen (14–17), the stiffness of the substrates presenting the antigen (18, 19) and the mechanical forces delivered to the BCRs by the antigens (20, 21). These facts highlight a long-standing question in immunology: how can the initiation of B-cell activation process the information of antigen specificity, density, affinity, valency, mobility, substrate stiffness, and mechanical forces in such an efficient way? To attempt to address this intriguing question, a detailed understanding of the precise BCR sorting mechanisms within the B-cell IS during the initiation of B-cell activation is required. However, it is technically difficult to accurately capture these events due to their fast and dynamic nature. It is challenging to capture an entire molecular event from the same B-cell before and immediately after antigen recognition, and it is more difficult to capture multiple events in parallel from multiple cells in a synchronized manner. An attractive solution for this dilemma is to develop a precisely controllable trigger point for BCR and antigen recognition by using photoactivatable antigens, which are initially inactive but become immediately active on the illumination of UV light. Indeed, photoactivatable systems have been used to investigate the kinetics of second-messenger activity through caged calcium (22) and caged cAMP (23). Additionally, the T-cell receptor was studied using major histocompatibility complex (MHC) presenting photoactivatable peptide (24, 25).

In this report, we dissected the dynamic responses during the initiation of B-cell activation by using a photoactivatable antigen based experimental system in combination with high-resolution high speed total internal reflection fluorescence microscopy (TIRFM) imaging techniques. We caged the widely used model antigen 4-hydroxy-3-nitrophenyl acetyl (caged-NP) that is only converted to its antigenic form after exposure to UV photons. The photoactivation of caged-NP in contact with NP-specific B1-8-BCR–expressing B cells provides a precisely controllable trigger point to perform high resolution temporal analyses of the formation of BCR microclusters and the B-cell IS in response to antigen stimulation. By combining the unique strengths of the caged-NP–based photoactivatable antigen system with TIRFM-based live cell and single molecule imaging techniques, we examined the basal response of a quiescent B cell exposed to coverslips presenting the caged-NP for 360 s and then examined the changes in the responses of the same B cell immediately after the recognition of the photoactivated-NP antigen for another 360 s. To our knowledge, this system represents the first temporally seamless imaging experimental procedure for the study of the molecular events during the initiation of B-cell activation. We illuminated the probing behaviors in quiescent B cells as defined by the unceasing extension of membrane pseudopods in random directions. We found that BCR and antigen recognition promptly terminated the probing responses. We also dissected the sophisticated BCR sorting mechanism within the B-cell IS during the initiation of B-cell activation.

Results

The Development of a Caged-NP–Based Photoactivatable Antigen System.

The NP hapten is specifically recognized by B1-8 antibody or B1-8-BCR–expressing B cells mainly due to the hydroxyl (-OH) group in the phenol ring (26). To mask the antigenicity of NP, we conjugated a UV-sensitive moiety, 4,5-dimethoxy-2-nitrobenzyl (DMNB) to its -OH group to generate DMNB-NP (Fig. 1A and Fig. S1). Analyses by ¹H- and ¹³C-NMR verified the correct conjugation of DMNB to NP (Fig. S2 A and B). To facilitate the downstream imaging experiments, a carrier peptide with His₆-tag, ASTGKTASACTSGASSTGS-His₆, was further conjugated to either WT NP (WT-NP) or DMNB-NP (caged-NP) (Fig. 1A) at the N terminus through a 6-aminocaproic acid linker as characterized by reverse-phase HPLC (RP-HPLC) (Fig. 1B) and electrospray ionization-MS (ESI-MS) (Fig. 1C). To examine the photo-cleavage efficiency, the photolysis kinetics of the caged-NP was examined on exposure to UV photons (365 nm, 18 mW/cm²). We used RP-HPLC to quantify the photolysis kinetics of the caged-NP by determining the percentage of photolyzed peptides in a population of total peptides including the caged-NP peptides and the photolyzed-NP peptides (Fig. 1D). We found that the caged-NP had rapid photolysis kinetics with greater than 40% photolyzed peptides detected in 10 s. We also evaluated the antigenicity of WT-NP, caged-NP, and caged-NP after photoactivation (referred to as photoactivated-NP hereafter) using NP-specific B1-8 antibodies in an antigen dose-dependent ELISA (Fig. 1E). As a system control, anti–His₆-tag antibodies (the carrier peptide contains a His₆ tag) confirmed that all these NP-conjugated peptides were similarly coated onto the ELISA plate (Fig. S3A). WT-NP showed robust binding capability to NP-specific antibodies, whereas caged-NP lost the binding abilities (Fig. 1E). On photoactivation, we observed the recovered binding of the photoactivated-NP to NP-specific antibodies, and this recovery efficiency was positively correlated to the coating concentration of the caged-NP peptide (Fig. 1E). Thus, a caged-NP–based photoactivatable antigen system was developed.

Fig. 1.

The development of a caged-NP–based photoactivatable antigen system. (A) The schematic presentation of the conjugation of a UV-sensitive moiety, DMNB (caging group), to the -OH group of the NP hapten antigen to generate DMNB-NP. (B) The RP-HPLC analysis of WT-NP or caged-NP. The retention time is shown in minutes. (C) The ESI-MS analysis of WT-NP or caged-NP. The mass-to-charge ratio is shown in m/z. (D) The photolysis kinetics of the caged-NP peptide (1 mg/mL) on photoactivation with UV photons (365 nm, 18 mW/cm²) for different exposure times (0, 5, 10, 20, 30, and 60 s) as examined by RP-HPLC to quantify the percentage of photolyzed peptides in the total peptide population. Bars represent mean ± SD of three independent experiments. (E) The ELISA evaluation of the binding capacity of anti-NP antibodies to WT-NP- or caged-NP-peptide before (0 s) or after photoactivation with UV photons (365 nm, 18 mW/cm²) for different exposure time (30, 120, and 300 s). Each peptide (20 µg/mL) was photoactivated for the indicated amount of time and was then diluted to 10 or 5 µg/mL in the binding detection.

Fig. S1.

Synthetic route for DMNB-NP-OH. Synthetic route for DMNB-NP-OH: 5-dimethoxy-2-nitrobenzaldehyde (shown as 2) (1.06 g, 5 mmol) was dissolved in anhydrous THF (30 mL) and cooled to 0 °C. NaBH₄ (188 mg, 5 mmol) was added and the mixture was stirred for 6 h at 0–5 °C. Then the reaction was quenched by addition of cold water (50 mL), and the solution was extracted with CH₂Cl₂. The organic layer was washed with saturated NaCl twice and dried over anhydrous Na₂SO₄, followed by filtration and evaporation to give 4,5-dimethoxy-2-nitrobenzyl alcohol (shown as 3) as a yellow solid (90% yield, 1.92 g) without further purification. PBr₃ (150 μL, 1.6 mmol) was added to a solution of 4,5-dimethoxy-2-nitrobenzyl alcohol (shown as 3) (213 mg, 1 mmol) in dry THF (10 mL) at 0 °C, and the reaction was stirred for 2 h. Then the reaction mixture was quenched with cold water (20 mL) and extracted with EtOAc. The organic layer was washed with saturated NaCl twice, dried over anhydrous Na₂SO₄, filtered, and concentrated to quantitatively affordable level with the corresponding 1-(bromomethyl)-4,5-dimethoxy-2-nitrobenzene (shown as 4) (266 mg, 0.96 mmol). NP-OH (4-hydroxy-3-nitrophenyl acetic acid) (shown as 5) (197 mg, 1 mmol) was dissolved in CH₃OH (5 mL) and cooled to 0 °C. SOCl₂ (145 μL, 2 mmol) was added dropwise, and the mixture was stirred for 2 h. The reaction was quenched by addition of cold water (10 mL), and the solution was extracted with EtOAc and washed with saturated NaHCO₃ and saturated NaCl twice for each. The organic layer was dried over anhydrous Na₂SO₄, and concentrated under reduced pressure to give methyl 2-(4-hydroxy-3-nitrophenyl) acetate (NP-OMe) (shown as 6) as a yellow solid (95% yield, 200 mg) without further purification. A solution of NP-OMe (shown as 6) (200 mg, 0.95 mmol) in DMF (10 mL) and anhydrous K₂CO₃ (278 mg, 2 mmol) was stirred at room temperature for 10 min. After that, 1-(bromomethyl)-4,5-dimethoxy-2-nitrobenzene (shown as 4) (266 mg, 0.96 mmol) was added to the reaction mixture and stirred for another 1 h. The reaction was quenched by addition of 1 M cold HCl solution (20 mL), and extracted with CH₂Cl₂. The organic layer was washed with saturated NaCl twice, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure, followed by flash chromatography (CH₂Cl₂:PE = 2:1) to afford the desired product methyl 2-(4-(4,5-dimethoxy-2-nitrobenzyloxy)phenyl)acetate (DMNB-NP-OMe) (shown as 7) as a yellow solid (47% yield, 161 mg). The reaction should be quenched no longer than 2 h to avoid side reactions. Light was avoided at the beginning of this step to prevent the decomposition of products. DMNB-NP-OMe (shown as 7) (162 mg) was dissolved in a solution of THF/1 M NaOH (vol/vol) (20 mL) and stirred vigorously for 1 h. The reaction was kept away from light to avoid decomposition; 1 M HCl (20 mL) was added to the mixture to quench the reaction. The mixture was extracted with CH₂Cl₂ and washed with 1 M HCl and saturated NaCl twice for each. The organic layer was dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure to afford the final product 2-(4-(4,5-dimethoxy-2-nitrobenzyloxy)phenyl)acetic acid (DMNB-NP-OH) (shown as 1) (87% yield, 136 mg, 0.39 mmol) without further purification. The overall yield was 39%.

Fig. S2.

¹H-NMR and ¹³C-NMR for DMNB-NP-OH. ¹H-NMR (A) and ¹³C-NMR (B) analyses for the quality control of DMNB-NP-OH. TLC was performed on silica 60F-254 plates. The spots were visualized by UV light. Flash column chromatography was performed on silica gel 60 (300–400 mesh). ¹H and ¹³C NMR spectra were recorded on a JOEL 300-MHz instrument at room temperature in (CD₃)₂SO. Chemical shifts (δ) were reported relative to TMS (¹H 0 ppm) or (CD₃)₂SO (¹H 2.50 ppm) for ¹H-NMR and (CD₃)₂SO (39.52 ppm) for ¹³C-NMR spectra. J was given in Hertz, and the splitting patterns were designed as follows: s, singlet; d, doublet. ¹H NMR [300 MHz, (CD₃)₂SO] δ 3.66 (s, 2H), 3.87 (s, 3H), 3.91 (s, 3H), 5.56 (s, 2H), 7.39 (d, J = 8.94, 1H), 7.51 (s, 1H), 7.59 (dd, J = 8.58 and 2.07, 1H), 7.71 (s, 1H), 7.90 (d, J = 2.43, 1H), 10.18 (s, 1H). ¹³C NMR (300 MHz, (CD₃)₂SO) δ 38.79, 56.05, 67.71, 108.08, 110.32, 115.11, 26.28, 126.98, 128.32, 136.27, 138.51, 138.85, 147.74, 149.61, 153.46, 172.34.

Fig. S3.

ELISA evaluation of the binding capacity of anti–His-tag antibodies to WT-NP or caged-NP peptides before or after the photoactivation; B-cell probing behaviors are cell intrinsic without dependence on caged-NP. (A) The binding capacity of the His₆ tag on either WT-NP or caged-NP peptides was detected and compared by anti–His-tag antibodies in ELISA. Both WT-NP and caged-NP peptides were exposed to UV photons (365 nm, 18 mW/cm²) for the time duration as indicated (0, 30, 120, and 300 s). The group with no NP peptides coated was used as negative control. Bars represent mean ± SD from three repeats. (B–E) Representative TIRFM images at the indicated time points of the B1-8 primary B cells prelabeled with DyLight 649-conjugated Fab fragment anti-IgM that were placed on coverslips alone (B) or coverslips presenting PLBs (C), ICAM-I (D), or anti–MHC-I antibodies (E). In each condition, 1% casein was used as blocking reagent. The basal responses of quiescent B cells on each of these coverslips were first examined for 420 s by TIRFM imaging. See Movies S3–S6 for complete time-lapse TIRFM images: Movie S3, coverslips alone; Movie S4, PLBs; Movie S5, ICAM-I; Movie S6, anti–MHC-I antibodies. The white arrowheads indicate the membrane protrusions of the probing B1-8 primary B cells. (Scale bar, 1.5 µm.) (F) Statistical comparison with quantity the number of membrane pseudopods of each B cell in different conditions. Bars represent mean ± SD of 20 cells from one representative of three independent experiments. Two-tailed t tests were performed for statistical comparisons.

Photoactivation Promptly Terminates the Probing Behaviors of Quiescent B Cells.

We combined the unique strengths of the photoactivatable NP antigen system with the TIRFM-based live cell imaging system to examine the precise behavior changes of NP-specific B cells before and after photoactivation. We first imaged the basal behaviors of a single B cell in its quiescent state on coverslips presenting the caged-NP for a sufficient amount of time (e.g., 360 s) and then examined the behavior changes of the very same B cell immediately on photoactivation and thereafter for another 360 s. To the best of our knowledge, this represents the first design of a seamless imaging experimental approach to capture the changes in the molecular events of the same B cell in a sufficient temporal domain (e.g., an examination of 360 s in the quiescent status immediately followed by an examination of 360 s in the activated status) in response to antigen recognition.

In all of the following photoactivation-based seamless imaging experiments, NP-specific B cells prelabeled with Dylight 647-conjugated Fab fragment anti-mouse IgM antibodies were first placed on coverslips presenting the caged-NP antigen for 10 min to blunt any potential behavior changes of the B cells that were introduced into the system by the acute landing and adhesion responses of the B cells. Thus, imaging experiments were only performed in the condition that the B cells formed steady-state contact with the coverslips after the 10-min incubation time. We found that the NP-specific B cells in contact with caged-NP exhibited the unceasing extension of membrane pseudopods in random directions, for which we termed as the probing behavior hereafter in this report. This probing behavior of quiescent B cells can be readily captured in both J558L-B1-8-IgM cells (Fig. 2A, as shown by the white arrowheads; see Movie S1 for the best visional effects with full frame information) and B1-8 primary B cells from IgH ^B1-8/B1-8 Igκ^−/− transgenic mice (26, 27) (Fig. 2B and Movie S2 for the best visional effects). Further experiments showed that these probing behaviors were not induced by caged-NP as similar results were captured from B1-8 primary B cells that were placed on control coverslips without caged-NP (Fig. S3B and Movie S3). These probing behaviors were not induced by nonspecific stimulation from the glass to the cells as the B1-8 primary B cells that were placed on coverslips presenting fluid planar lipid bilayers (PLBs), which were used to insulate the direct contact of the cell membrane to the glass, similarly exhibited the probing behaviors (Fig. S3C and Movie S4). To further exclude the possibility that these probing behaviors might reflect the membrane projections that are transiently entering into the TIRFM imaging plane, we imaged B1-8 primary B cells that were placed on coverslips presenting either ICAM-1 or anti–MHC-I antibodies, both of which have been used to pretether and preadhere B cells to the surface of coverslips in the literature (28, 29). The probing behaviors were readily observed in both cases (Fig. S3 D and E and Movies S5 and S6). Furthermore, a series of pharmaceutical inhibitor experiments showed that the probing behaviors were terminated in B cells pretreated with cytochalasin D to disrupt F-actin or with jasplakinolide to stabilize F-actin, suggesting that B-cell probing behaviors were dependent on the remodeling of F-actin (Fig. S4 A and B). In contrast, these probing behaviors were independent of tonic BCR signaling as B cells pretreated with the Src family kinase inhibitor PP2 or the Syk kinase inhibitor piceatannol still maintained these behaviors (Fig. S4C and D). Thus, B-cell probing behaviors are cell intrinsic with dependence on F-actin remodeling but not on tonic BCR signaling.

Fig. 2.

Photoactivation promptly terminates the probing behavior of quiescent B cells. (A and B) Shown are the representative TIRFM images at the indicated time points of the same J558L-B1-8-IgM cells (A) or B1-8 primary B cells (B) prelabeled with DyLight 649-conjugated Fab fragment anti-IgM that were placed on coverslips coated with caged-NP before and after photoactivation. The basal responses of quiescent B cells on coverslips presenting caged-NP were first examined for 360 s by TIRFM imaging. Immediately starting with photoactivation (time 0 s as indicated), the behaviors of the same B cells were continuously examined for another 360 s. See Movies S1 and S2 for complete time-lapse TIRFM images. The white arrowheads indicate the membrane protrusions of the probing J558L-B1-8-IgM cells (A) or B1-8 primary B cells (B) before the photoactivation. (Scale bar, 1.5 µm.) (C and D) Statistical comparison with quantification for the number of membrane pseudopods per TIRFM image in each J558L-B1-8-IgM B cell (C) or B1-8 primary B cell (D) before and after photoactivation. Bars represent mean ± SD of 20 cells from one representative of three independent experiments. Two-tailed t tests were performed for statistical comparisons in C and D.

Fig. S4.

B-cell probing behaviors are cell intrinsic with strict dependence on F-actin remodeling but not tonic BCR signaling molecules. (A–D) Representative TIRFM images at the indicated time points of J558L-B1-8-IgM B cells that were pretreated with cytochalasin D (A), jasplakinolide (B), PP2 (C), or piceatannol (D). The basal responses of the quiescent B cells on coverslips presenting caged-NP were examined for 360 s by TIRFM imaging. The white arrowheads indicate the membrane protrusions of the B cells with the probing behaviors. (Scale bar, 1.5 µm.) (E and F) Representative TIRFM images at the indicated time points of J558L-B1-8-IgM B cells prelabeled with DyLight 649-conjugated Fab fragment anti-IgM that were placed on PLBs presenting caged-NP. The basal responses of the B cells in their quiescent state in contact with caged-NP were first examined for 360 s by TIRFM imaging. Immediately starting photoactivation (time 0 s as indicated), the behavior changes of the same B cells were continuously examined for another 360 s. (Scale bar, 1.5 µm.) The white arrowheads indicate the membrane protrusions of the B-cell probing behaviors. In the still TIRFM images at the time point of 200 s, the dynamics of the BCR microclusters as indicated by either yellow (Upper) or blue (Lower) colored boxes were shown at 500% magnification for better resolution. The yellow colored arrowheads in the upper panel of the magnified images indicate the increase of mFI of BCR microclusters that were stationary, whereas the blue colored arrowheads in the lower panel of the magnified images indicate the increase of mFI of BCR microclusters that were performing retrograde mobility to the center of the B-cell IS. (G and H) Trajectories of three representative BCR microclusters by means of x vs. y footprints accumulated in the time course of −300 to 0 (G) and 0–300 s (H) as depicted in E and F. (I) Statistical quantification of the normalized mFI to show the synaptic accumulation of BCR in the same B cells before and after photoactivation of caged-NP as depicted in E and F. The curve from −360 to 0 s showed the dynamic changes of the J558L-B1-8-IgM cells in contact with mobile caged-NP, whereas the curve from 0 to 360 s showed the changes of the responses of the same cells in contact with photoactivated-NP. Bars represent mean ± SEM of 10 cells from one representative of three independent experiments.

On conversion of caged-NP to photoactivated-NP, the behavior of the same NP-specific B cells changed drastically. TIRFM imaging demonstrated that in as short as 4 s after photoactivation (see the 0- vs. 4-s TIRFM images in Fig. 2 or examine Movies S1 and S2), the same B cells that were originally probing promptly terminated this behavior as quantified by the significantly reduced number of membrane pseudopods per cell per TIRFM image (Fig. 2 C and D and Fig. S3F). In contrast, the photoactivation triggered B-cell probing termination events were not observed in B1-8 primary B cells that were placed on coverslips alone (Fig. S5A). Similarly, photoactivation did not induce the probing termination event of primary B cells from C57BL/6-WT mice that were placed on coverslips presenting caged-NP (Fig. S5B). These experiments suggested that the probing termination behavior was induced by the recognition of BCR and antigen. Furthermore, NP-specific B cells that have been exposed to bioactive WT-NP antigen for 10 min before TIRFM imaging did not exhibit the probing behaviors in the 360-s time courses before and after photoactivation (Fig. S5 C and D). All these additional experiments consistently suggested that photoactivation-induced phototoxicity does not contribute to the probing cessation event.

Fig. S5.

Photoactivation-induced phototoxicity cannot induce probing termination event. (A) Representative TIRFM images at the indicated time points of the same B1-8 primary B cells prelabeled with DyLight 649-conjugated Fab fragment anti-IgM that were placed on coverslips without antigen. (B) Representative TIRFM images at the indicated time points of the same primary B cells from C57BL/6-WT mice prelabeled with DyLight 649-conjugated Fab fragment anti-IgM that were placed on coverslips coated with caged-NP. (C and D) Representative TIRFM images at the indicated time points of the same B1-8 primary B cells or J558L-B1-8-IgM cells prelabeled with Dylight 649-conjugated Fab fragment goat anti-mouse IgM that were placed on coverslips coated with WT-NP. In A–D, B cells were first placed on coverslips presenting the indicated type of antigens for 10 min to blunt any responses of the B cells that were introduced into the system by the acute landing and adhesion of B cells. In the imaging experiments, the basal responses of quiescent B cells on coverslips presenting caged-NP were first examined for 360 s by TIRFM imaging. Immediately starting photoactivation (time 0 s as indicated), the behaviors of the same B cells were continuously examined for another 360 s. The white arrowheads indicate the membrane protrusions of the B cells exhibiting the probing behaviors. (Scale bar, 1.5 µm.) (E–H) Statistical comparison with quantify the normalized mFI of the synaptically accumulated BCRs in the same B cells before and after photoactivation as depicted in A–D, respectively. The curve from −360 to 0 s showed the dynamic changes of the normalized mFI of BCRs in B cells before photoactivation, whereas the curve from 0 to 360 s showed the changes of the BCR mFI of the same cells after photoactivation. Bars represent mean ± SEM of at least 18 cells from one representative of three independent experiments.

The Termination of B-Cell Probing Is Concomitant with the Synaptic Accumulation of the BCRs.

Next, we precisely quantified the changes in the size of the contact area of the B cells on the coverslips and the mean fluorescence intensity (mFI) of the BCR molecules within the contact area using the abovementioned seamless imaging experimental approaches. We calculated the mFI instead of the total FI of BCRs within the B-cell contact area because the former can better reflect the changes in the density (or concentration) of the BCRs at the contact interface, whereas the latter can increase if the B cells simply spread over the coverslips. Thus, the mFI value can better represent the efficiency of the accumulation of the BCRs within the B-cell IS. The results showed that both the size of the B-cell contact area and the mFI of the BCRs did not change in quiescent B cells in contact with the caged-NP in the whole time course of 360 s (Fig. 3 A–E and Movie S1). Strikingly, both values rapidly and drastically increased in the very same J558L-B1-8-IgM B cells immediately after photoactivation (Fig. 3 B–E). The same conclusion was acquired in B1-8 primary B cells (Fig. 3 F and G and Movie S2). In contrast, photoactivation did not drive the synaptic accumulation of BCR molecules in the experiments where B1-8 primary B cells were placed on coverslips alone (Fig. S5E) or the experiments where primary B cells from C57BL/6-WT mice were placed on coverslips presenting caged-NP (Fig. S5F). These results and the additional results from NP-specific B cells that have been exposed to bioactive WT-NP antigen for 10 min before the photoactivation (Fig. S5 G and H) suggested that the BCR accumulation event was not induced by the potential phototoxicity after photoactivation. To further exclude the possibility that the increase in the BCR mFI resulted from the distance changes in the z dimension of B cells with coverslips after photoactivation, we also similarly quantified these changes from B cells that were placed on coverslips presenting both caged-NP and anti–MHC-I antibodies. Anti–MHC-I antibodies have been used in the literature to uniformly pretether B cells to the surface of coverslips (29). In this system, we similarly captured the drastic BCR accumulation event on photoactivation (Fig. S6A).

Fig. 3.

The termination of the B-cell probing behavior is concomitant with the synaptic accumulation of the BCRs. (A) Shown are the TIRFM images at the indicated time points of the same J558L-B1-8-IgM cells that were placed on coverslips coated with the caged-NP before and after photoactivation. The basal responses of the B cell in its quiescent state in contact with the caged-NP were first examined for 360 s by TIRFM imaging. Immediately starting with photoactivation (time 0 s as indicated), the behavioral changes of the same B cells were continuously examined for another 360 s. See Movie S1 for the complete time-lapse TIRFM images. (Scale bar, 1.5 µm.) (B–E) Statistical comparison with quantification for the size of the contact area (B) and the mFI of the BCRs (D) from the same B cells before and after photoactivation as depicted in A. The curve from −360 to 0 s showed the dynamic changes of the J558L-B1-8-IgM cells in contact with the caged-NP, whereas the curve from 0 to 360 s showed the changes in the responses of the same cells in contact with the photoactivated-NP. The corresponding curves following a normalization step to the value at time 0 s are also shown in C (for B) and E (for D). Bars represent mean ± SEM of 15 cells from one representative of three independent experiments. (F and G) A similar experiment was performed as described for D and E. Shown are the mFI (F) and the normalized mFI (G) of the BCR molecules within the contact area of the same B1-8 primary B cells before and after photoactivation of the caged-NP. Bars represent mean ± SEM of 10 cells from one representative of three independent experiments.

Fig. S6.

Termination of B-cell probing is concomitant with the synaptic trapping of BCRs. (A) Shown are the statistical comparisons to quantify the normalized mFI of the synaptically accumulated BCRs of J558L-B1-8-IgM cells before and after photoactivation on coverslips presenting caged-NP and anti–MHC-I antibodies. The curve from −360 to 0 s showed the dynamic changes of the J558L-B1-8-IgM cells before photoactivation, whereas the curve from 0 to 360 s showed the changes of the responses of the same cells after photoactivation. Bars represent mean ± SEM of at least 10 cells from one representative of three independent experiments. (B) Shown are the representative TIRFM images at indicated time points showing the landing events into TIRFM imaging plane of a fluorescent bead landing. See Movie S7 for complete time-lapse TIRFM images. (C) Statistical comparisons to quantify the dynamics of the normalized mFI of the fluorescent bead during its landing event to the coverslips as depicted in B.

Our recent studies and those of others showed that the BCR microcluster is the most basic unit during the initiation of B-cell activation (15, 29–34). However, it has been difficult to accurately capture and then investigate the dynamic changes of the BCR microclusters in the same B cells before and after the recognition of antigens. In all those published studies, a strategy was used to directly monitor the responses after loading B cells onto coverslips presenting bioactive antigens. However, that strategy will inevitably lead to the activation of the BCRs and the formation of BCR microclusters immediately on contact of B-cell membrane pseudopods with active antigen-presenting surfaces. Thus such an experimental strategy makes it difficult to examine the true basal responses in quiescent B cells and the conversion from the quiescent to the activated state in the same B cells. It is also difficult to separate the acute landing and adhesion events of B cells from the BCR engagement induced B-cell spreading response. Thus, these conventional imaging procedures will likely overestimate the extent of B-cell responses that are truly induced by antigen-BCR recognition. For example, TIRFM imaging showed that the normalized mFI of a fluorescent bead, which should be a fixed value, drastically increased during its stochastic landing event on the surface of the coverslips (Fig. S6 B and C and Movie S7). We propose that the photoactivated antigen-based seamless imaging approaches can overcome these obstacles.

We showed that B cells in contact with coverslips presenting caged-NP did not form stable and persistent BCR microclusters, although we frequently observed the formation of dynamic and transient puncta structures of BCR molecules that were concomitant with the probing behaviors of the B cells (Figs. 2 and 3 and Movies S1–S6). Immediately after photoactivation, the same B cells terminated the probing behavior and began to form truly prominent BCR microclusters (Fig. 4A and Movie S8). To accurately quantify the spatial-temporal changes of the biophysical features of the BCR microclusters, we placed control beads in the same imaging field of the B cells to precisely calibrate the vibration of the whole TIRFM imaging system (Fig. 4A and Movie S8). Then we analyzed these time-lapse images following our published protocol (6), using a 2D Gaussian function based mathematical fitting method to accurately quantify the mFI (integrated FI/size) and the position in the x and y coordinates of both the BCR microclusters and the calibration bead (Fig. 4 B–E). First, we examined the lateral motility of both the calibration bead and the BCR puncta structure by generating the x and y coordinates from the entire TIRFM imaging time course, which was presented as a typical trajectory plot. It was clear that the calibration bead did not move beyond one pixel (150 nm) in all of the time course (Fig. 4 C and D and Movie S8). In marked contrast, the BCR puncta structures in the same TIRFM imaging field showed highly motile behavior in quiescent B cells contacting caged-NP (Fig. 4C). After photoactivation, the BCR puncta structures became stationary and exhibited a trajectory range of less than one pixel (150 nm) in 300 s, comparable to a control bead (Fig. 4D and Movie S8). Next, we quantified the dynamic changes in the mFI of the BCR puncta structures in quiescent B cells in contact with caged-NP and in the same B cells upon photoactivation. The mFI of the BCR puncta structures in contact with the caged-NP did not increase in the whole 300-s time course (Fig. 4 A and B), suggesting the lack of enrichment of BCR molecules within these puncta structures in quiescent B cells. However, immediately after photoactivation, we readily observed the formation of prominent and stable BCR microclusters that drastically increased the mFI over time in the same B cell (Fig. 4 A–D).

Fig. 4.

The antigen engaged BCRs are accumulated into stationary BCR microclusters. (A) Representative TIRFM images show the response of the same BCR puncta before (−300 to 0 s) and after (0–300 s) photoactivation of caged-NP (time 0 s as indicated). (Scale bar, 1.5 µm.) Indicated by the white boxes in the −240 s TIRFM image are one representative typical BCR punctum (Left) and one representative calibration bead (Right). A fraction of the entire time-lapse TIRFM images from −12 to 36 s is shown in a refined temporal manner. See Movie S8 for complete time-lapse TIRFM images. (B–D) The normalized mFI (B) of the BCR microclusters from the whole time course (−300 to 300 s) are shown along with the trajectories by means of the x vs. y footprints accumulated in the time course of −300 to 0 s (C) and of 0–300 s (D) as depicted in A. In B, bars represent mean ± SEM of 435 BCR microclusters analyzed from 15 cells in one representative of three independent experiments. (E) The pseudocolor 2.5D Gaussian images of the same BCR microcluster at the indicated times from B cells in contact with caged-NP (Upper) or photoactivated-NP (Lower) are shown. (Scale bars, 1.5 μm.)

To make sure that the above observations were not induced by the immobile nature of the antigens that were tethered to coverslips, we examined the response of the NP-specific B cells that were placed on PLBs presenting caged-NP antigens. B cells exhibited the typical probing behaviors in contacting with caged-NP antigens on PLBs and BCR puncta structures from these B cells similarly showed highly motile behaviors (Fig. S4 E–G and Movie S9). Photoactivation of the antigens on PLBs also triggered B-cell probing termination response that was concomitant with the formation of prominent BCR microclusters (Fig. S4 E–H and Movie S9). The newly formed BCR microclusters were still lack of obvious mobility immediately after photoactivation, although at the later stage some but not all BCR microclusters exhibited the obvious centripetal movement toward the central region of the B-cell IS (Fig. S4 E–H, as shown by the yellow arrowheads for stationary BCR microclusters and blue arrowheads for the ones showing centripetal movements; see Movie S9 for the best visional effects). These results are largely consistent with our published studies and those of others showing that there was a lag time to allow B cells to reach the maximal spreading before the BCR microcluster exhibited an obvious centripetal movement toward the center of B-cell IS (6, 7). In either the stationary BCR microclusters (Fig. S4 E and F, as shown by the yellow arrowheads) or the BCR microclusters retrograding to the center of the B-cell IS (Fig. S4 E and F, as shown by the blue arrowheads; see Movie S9 for the best visional effects), it was evident that the BCR microclusters increased their mFI over time (Fig. S4I, or see the representative BCR microclusters by arrowheads in Fig. S4 E and F and Movie S9). These data suggested that photoactivation promptly induced a rapid and drastic accumulation of the BCRs into the BCR microclusters. Because the BCR microclusters could be stationary in these events, we propose that the BCR molecules could be accumulated into the BCR microclusters in a passive trapping manner.

The Termination of B-Cell Probing Is Sensitive to the Affinity Between the BCR and the Antigen.

B cells show excellent capability to discriminate the affinity between the BCR and the antigen during the initiation of B-cell activation. We thus investigated if the photoactivation triggered termination of probing would be sensitive to the BCR and antigen affinity. The J558L-B1-8-IgM cells that were used in the above experiments in this report expressed a B1-8-IgM-High BCR with high affinity to NP (K_a = 5 × 10⁶ M⁻¹); in contrast, B1-8-IgM-Low BCRs showed a much low affinity (K_a = 1.25 × 10⁵ M⁻¹) to NP as described in detail in our previous studies and those of others (6, 35). Thus, we continued to image the responses of J558L-B1-8-IgM-Low cells using the photoactivatable antigen-based seamless imaging system. J558L-B1-8-IgM-Low cells in contact with the caged-NP also showed the typical B-cell probing response. In marked contrast to the observation that high-affinity J558L-B1-8-IgM cells terminated the probing behavior to accumulate the BCRs into the B-cell IS (Fig. 3 A–E and Movie S1), probing termination event was only mild in the case of J558L-B1-8-IgM-Low cells, where attenuated but constant probing behaviors always occurred throughout the imaging time course after photoactivation (Fig. S7 A and B and Movie S10). Concomitant with the termination of the probing behaviors, high-affinity J558L-B1-8-IgM cells formed a more stable cell body, which overtly facilitated the BCR microclusters to become prominently stationary. All these early events resulted in an increased size of the B-cell contact area and the synaptic accumulation of the BCRs in high affinity J558L-B1-8-IgM cells (Fig. 3 A–E and Movie S1). However, J558L-B1-8-IgM-Low cells did not show a drastic increase in the size of the B-cell contact area nor was there an obvious synaptic accumulation of the BCRs (Fig. S7 A and B and Movie S10). Thus, the termination of the probing behavior of B cells on BCR engagement is an event that is sensitive to BCR and antigen affinity.

Fig. S7.

Termination of B-cell probing is sensitive to the affinity between BCR and antigen. (A) Shown are the TIRFM images at indicated time points of the same J558L-B1-8-IgM-Low cells that were placed on coverslips coated with caged-NP before and after photoactivation. The basal responses of the B cells in their quiescent state in contact with caged-NP were first examined for 360 s by TIRFM imaging. Immediately starting photoactivation (time 0 s as indicated), the behavior changes of the same B cells were continuously examined for another 360 s. See Movie S10 for complete time-lapse TIRFM images. (Scale bar, 1.5 µm.) (B) Statistical comparisons to quantify the normalized mFI of the synaptically accumulated BCRs in J558L-B1-8-IgM-Low cells before and after photoactivation as depicted in A. The curve from −360 to 0 s showed the dynamic changes of the J558L-B1-8-IgM-Low cells in contact with caged-NP, whereas the curve from 0 to 360 s showed the changes of the responses of the same cells in contact with photoactivated-NP. Bars represent mean ± SEM of at least 12 cells from one representative of three independent experiments.

The B-Cell IS Is a Dynamically Open yet Highly Selective Membrane Structure.

The observation that the accumulation of BCR molecules into the B-cell IS was driven by BCR engagement raised several intriguing questions. Is the B-cell IS a dynamic or steady structure? Is the B-cell IS an open or closed structure? Is the B-cell IS a selective or promiscuous structure? To answer these questions, we performed a two-color TIRFM imaging experiment in combination with the photoactivation system to simultaneously examine the spatial and temporal changes of the BCRs and control molecules (MHC-I or lipid molecules) within the B-cell IS. We observed that both BCR and MHC-I molecules were transiently expressed in the TIRFM images in a highly dynamic manner in quiescent B cells contacting caged-NP (Fig. 5A and Movie S11). Furthermore, Pearson’s correlation index (PCI) analysis suggested that the BCR and MHC-I molecules only weakly codistributed in a quiescent B cell (Fig. S8 A and B). We continued to capture the dynamic changes of these two molecules in the same B cell on photoactivation and readily observed robust accumulation of BCRs with very mild but reproducible accumulation of MHC-I molecules within the B-cell IS (Fig. 5 A–C and Movie S11). We further validated this observation using Dil-stained membrane lipid molecules (Fig. S8 C–E and Movie S12). These data suggested that the B-cell IS is a dynamically open and selective membrane structure. To further confirm these observations, we performed a long-term (30 min) two-color TIRFM imaging experiment to examine the accumulation and spatial and temporal dynamics of both the BCRs and the MHC-I molecules within the B-cell IS (Fig. 5 D–F and Movie S13). As expected, we observed a robust and steady accumulation of the BCRs into the B-cell IS with an initial exponential growth pattern followed by a linear growth pattern (Fig. 5 D and E). We also observed a synaptic accumulation of MHC-I in response to the photoactivation of caged-NP in the long-term experiments, although the accumulation was obviously much weaker compared with the case of the BCRs (Fig. 5 D and F). PCI analysis of the long-term TIRFM images examining the synaptic codistribution of the BCR and MHC-I molecules showed that the PCI value slightly increased in the first 5 min after photoactivation and then decreased over time (Fig. S8 A and B). Interestingly, the synaptic accumulation of the MHC-I molecules mainly occurred at the central but not the peripheral region of the B-cell IS (Fig. S8B). These results defined the unexpected heterogeneity in terms of “molecule crowding” within the B-cell IS (Discussion).

Fig. 5.

The B-cell IS is a dynamically open yet highly selective membrane structure. Representative two-color TIRFM images at the indicated time points show the spatial-temporal changes in the dynamic behaviors of either the BCR (green) or MHC-I (red) molecules in J558L-B1-8-IgM cells that were placed on coverslips coated with caged-NP before and after photoactivation. The basal behaviors of both the BCR and MHC-I molecules in quiescent B cells in contact with caged-NP were first examined for 240 s by TIRFM imaging. Immediately starting with photoactivation (time 0 s as indicated), the behaviors of these two molecules on the same B cells were continuously examined for another 360 s. See Movie S11 for complete time-lapse TIRFM images. (Scale bar, 1.5 µm.) (B and C) Statistical quantification of the normalized mFI to show the synaptic accumulation of the BCR (B) or MHC-I (C) molecules in the same B cells before and after photoactivation of caged-NP as depicted in A. The curve from −360 to 0 s showed the dynamic changes of the J558L-B1-8-IgM cells in contact with caged-NP, whereas the curve from 0 to 360 s showed the changes in the responses of the same cells in contact with photoactivated-NP. Bars represent mean ± SEM of 10 cells from one representative of three independent experiments. (D–F) A similar experiment was performed as described for A–C except that a long TIRFM imaging time course of 30 min was performed after photoactivation. See Movie S13 for complete time-lapse TIRFM images. (Scale bar, 1.5 µm.)

Fig. S8.

The B-cell IS is a dynamically open yet highly selective membrane structure. (A) Given is the PCI analysis to show the dynamics of the colocalization of BCR and MHC-I. The PCI value from −600 to 0 s showed the dynamics of the colocalization of these two molecules in quiescent B cells in contact with caged-NP, whereas the value from 0 to 1,800 s showed the dynamics of the colocalization of these two molecules in the same B cells on photoactivation of the caged-NP. Bars represent mean ± SEM of at least 10 cells from one representative of three independent experiments. (B) Representative TIRFM images of BCR (green) and MHC-I (red) within the IS of the same J558L-B1-8-IgM cells that were placed on coverslips coated with caged-NP before (−350 s) and after photoactivation at the indicated time points (210, 910, and 1,750 s). Relative FI along the white line was also given on the right of each paired TIRFM images. (Scale bar, 1.5 µm.) (C) Representative two-color TIRFM images at the indicated time points were given to show the synaptic accumulation of either BCR (green) or membrane lipid molecules (red) in the same J558L-B1-8-IgM cells that were placed on coverslips coated with caged-NP before and after photoactivation. The basal behaviors of both BCR and membrane lipid molecules in quiescent B cells in contact with caged-NP were first examined for 360 s by TIRFM imaging. Immediately starting photoactivation (time 0 s as indicated), the behaviors of these two molecules on the same B cells were continuously examined for another 360 s. See Movie S12 for complete time-lapse TIRFM images. (Scale bar, 1.5 µm.) (D and E) Statistical quantification of the normalized mFI to show the synaptic accumulation of BCR (D) or membrane lipid molecules (E) in the same B cells before and after photoactivation of the caged-NP as depicted in C. The curve from −360 to 0 s showed the dynamic changes of the J558L-B1-8- IgM cells in contact with caged-NP, whereas the curve from 0 to 360 s showed the changes of the responses of the same cells in contact with photoactivated-NP. Bars represent mean ± SEM of at least 10 cells from one representative of three independent experiments.

The Decreased Lateral Mobility of Antigen-Engaged BCRs Accounts for Their Synaptic Trapping.

A unique advantage of the photoactivatable NP system is that we can accurately generate temporal segments at an unprecedented resolution immediately after BCR engagement for dozens of B cells with a highly synchronized trigger point. The schematic (Fig. 6A) depicts our experimental strategy for the single molecule imaging (SMI) experiment for the BCR and MHC-I molecules. Each B cell was subjected to SMI of 600 frames in 18 s (30 ms/frame) for the first SMI reading. The second SMI reading was taken after an interval of 10 s followed by the third SMI reading and another 10-s interval until the sixth reading. Photoactivation by 405-nm laser was only executed during the the first and second reading as depicted in Fig. 6A. Taking advantage of the photoactivatable antigen system, we can for the first time, to our knowledge, accurately quantify the Brownian motility of single BCR molecules in the same B cells before and immediately after BCR engagement in a temporally seamless manner (Fig. 6A and Movie S14).

Fig. 6.

The decreased lateral mobility of antigen-engaged BCRs accounts for their synaptic trapping events. (A) A schematic representation showing the experimental strategy during the single molecule imaging (SMI) of the BCRs. Each B cell was subjected to single BCR molecule imaging of 600 frames in 18 s (30 ms/frame) for the first SMI reading. The second SMI reading was taken after an interval of 10 s followed by the third SMI reading and another 10-s interval until the sixth reading. Photoactivation was only applied during the first and second TIRFM imaging cycles. See Movies S14 (BCR) and S15 (MHC-I) for complete time-lapse TIRFM images. (B and C) The accumulated trajectory footprints of the BCR (B) or MHC-I (C) molecules from the representative B cells in three segmented TIRFM imaging time courses of 0–18, 56–74, and 112–130 s. The instant Brownian diffusion coefficients (D₀) were shown as pseudocolored trajectories. The display range of the pseudocolor is based on the D₀ value of 1e⁻⁴ to 1e⁰ μm²/s. (D–I) The D₀ values from the BCR (D–F) or MHC-I (G–I) single molecules that were observed from the indicated TIRFM imaging temporal segments (0–18, 28–46, 56–74, 84–102, 112–130, and 140–158 s) from SMI imaging experiment as depicted in A. All of the D₀ values were displayed as MSD plots (D and G), CDP plots (E and H), or mean ± SD scattered plots (F and I). Data represent the single BCR or MHC-I molecules of indicated numbers (E and H) for each condition from three independent experiments. Two-tailed t tests were performed for statistical comparisons in F and I.

Analyses of the single BCR trajectory footprints suggested that single BCR molecules were highly mobile in quiescent B cells contacting caged-NP. However, on photoactivation, the BCRs from the same B cells became gradually less mobile over time (Fig. 6B and Movie S14). Tracking hundreds of single BCR molecules showed that their short-range mean-square displacements (MSDs) were linearly dependent on time, indicating free diffusion movement (Fig. 6D). A simple comparison of the six temporally segregated MSD plots from the same batch of B cells suggested that, on photoactivation, the Brownian motility of the BCR molecules became increasingly confined over time (Fig. 6D). Moreover, the short-range diffusion coefficients of each individual BCR molecule were calculated and their distribution was analyzed and displayed as a cumulative distribution probability (CDP) plot (Fig. 6E). It was quite clear that the Brownian diffusion coefficients of the BCRs in the same batch of B cells gradually decreased over time (Fig. 6E). The mean diffusion coefficient of the BCRs in the same batch of B cells also decreased (Fig. 6F). Moreover, we performed subpopulation analysis to mathematically sort thousands of BCRs into either fast or slow populations by a maximum likelihood estimation approach as detailed in Materials and Methods. The results confirmed that the slow BCR population in the same batch of B cells increased, whereas the fast population decreased over time on photoactivation (Fig. 7 A–F, Fig. S9A, and Movie S14). In contrast, we did not observe these dynamic changes for the MHC-I molecules (Fig. 6 C and G–I, Figs. S9B and S10 A–F, and Movie S15). Because the overall trend of the Brownian motion of the BCRs (but not the control MHC-I molecules) was to slow down after the photoactivation of caged-NP (Figs. 6 and 7, Figs. S9 and S10, and Movies S14 and S15), we propose that the decreased lateral mobility of antigen engaged BCRs shall account for their trapping and accumulation within the B-cell IS. We also propose that antigen engagement induced freezing of the lateral mobility of the BCRs is the fundamental mechanism for the growth feature of the BCR microclusters (Discussion).

Fig. 7.

The probability density function (PDF) plot and subpopulation analysis of the instant diffusion coefficient of BCR molecules. The D₀ values from each temporal segment of (A) 0–18, (B) 28–46, (C) 56–74, (D) 84–102, (E) 112–130, and (F) 140–158 s were fitted by a mixture model of two components as described in Materials and Methods. The proportion and absolute values of both fast and slow populations in each temporal segment are also shown on the top left corner within each PDF plot.

Fig. S9.

The PDF plot and subpopulation analysis of the instant diffusion coefficient of both BCR and MHC-I molecules. (A and B) Proportions of the slow mobile population of either BCR (A) or MHC-I (B) molecules in each of the temporal segments. Data in each individual temporal segment were acquired in Fig. 7 for BCR and Fig. S10 for MHC-I. (C) MSD plot of the puncta-structured membrane protrusions in quiescent B cells before photoactivation. Due to the high mobile nature of these membrane pseudopods, we increased the TIRFM imaging speed from one frame per 4 s to five frames per 1 s. We calculated the x and y coordinates for accurate spatial position of the puncta-structured membrane protrusions and determined their trajectories using method as detailed in Materials and Methods.

Fig. S10.

PDF plot and subpopulation analysis of the instant diffusion coefficient of MHC-I molecules. PDF plots of instant diffusion coefficient (D₀) and the subpopulation analysis of MHC-I molecules. D₀ values from each temporal segment of (A) 0–18, (B) 28–46, (C) 56–74, (D) 84–102, (E) 112–130, and (F) 140–158 s were fitted by mixture model of two components as described in Materials and Methods. Proportion and absolute values of both fast and slow population in each temporal segment were also given on the top left corner within each PDF plot.

Discussion

In this report, by combining the photoactivatable antigen system with TIRFM-based live cell and single molecule imaging techniques, we examined the basal responses of a quiescent B cell in contact with coverslips presenting the caged-NP for 360 s and then examined the response of the very same B cell immediately after the immune recognition of the photoactivated-NP antigen for another 360 s. We define our system as the first, to our knowledge, temporally seamless imaging experimental procedure for the study of the molecular events that occur during B-cell activation. Our approach is quite different from the conventional experimental system of capturing the response of a B cell that is loaded onto the coverslips presenting bioactive antigens directly. The conventional experimental systems could not accurately separate the landing and adhesion responses of a B cell from the B-cell activation responses that were truly induced by BCR and antigen recognition. With the aid of this unique experimental system, a series of unprecedented observations were captured to help us better understand B-cell and BCR behavioral changes during the initiation of B-cell activation.

First of all, we readily recorded the probing behaviors of a quiescent B cell in contact with biologically inert caged-NP, which is defined by the periodical formation of membrane pseudopods in random directions. The probing behaviors require the remodeling of F-actin as B cells pretreated with latrunculin B to disrupt F-actin or with jasplakinolide to stabilize F-actin lose the probing behaviors. These probing behaviors seem to be independent of tonic BCR signaling as B cells pretreated with the Src family kinase inhibitor PP2 or the Syk kinase inhibitor piceatannol maintain the probing behaviors. Previous studies by Brodovitch et al. used TIRFM imaging to examine the acute adhesion response of T cells when landing to nonactivating surfaces. They showed that T cells initially touched the surfaces through highly dynamic membrane pseudopods (protrusions) of submicrometer diameter, and substantial antigen-independent spreading was only observed after multicontacts formed with a lag time of 1 min (36). Importantly, they found that these T-cell membrane pseudopods exhibited autonomous movements with a typical duration of a few seconds or less and an axial amplitude of 60–70 nm (36). In our experimental system, B cells have been placed on coverslips presenting inert caged-NP for 10 min before the TIRFM imaging; thus, we cannot examine the axial displacement of B-cell membrane pseudopods that shall be only obvious during the acute landing and adhesion responses of lymphocytes. Instead we calculated the x and y coordinates for accurate spatial position of the puncta-structured membrane protrusions in quiescent B cells before photoactivation and determined their trajectories using single particle tracking method as reported in our published studies (14, 18, 30). Strikingly, the results suggested that B cells seem to survey the microenvironments at a mean square based lateral displacement of 0.05 µm²/s (equal to 500 nm ²/s) (Fig. S9C). If only calculating those trajectories showing directed Brownian mobility and long tracking steps (>30 steps), a linear lateral displacement of 76.65 ± 15.375 nm/s can be acquired. This value is well consistent with the observation in T-cell studies showing that the T-cell membrane protrusions only displayed limited lateral movement with a value below 100-nm amplitude in most of the cases (36). Moreover, the relatively large SD (76.65 ± 15.375 nm/s) indicated the diverse range of the lateral mobility of B-cell membrane protrusions, largely consistent with the T cells studies showing that some membrane protrusions with rapid lateral displacement of hundreds of nanometers can be captured among a pool of protrusions with slow displacement of less than 100-nm amplitude (36). We propose that it shall be of essential interest to investigate the mechanism accounting for the high heterogeneous nature of the lateral displacement of membrane protrusions and their diverse functions for both B and T cells to survey the microenvironment.

Second, we systematically compared the probing behaviors of the same B cell before and after photoactivation. We found that photoactivation promptly terminated the probing behaviors. Interestingly, the termination of the probing behavior seems to be dependent on the interactions between the BCR and antigen molecules, as neither the interaction between MHC-I and anti–MHC-I antibodies, nor that between integrin and the adhesion molecule ICAM-1, terminated the B-cell probing behaviors. This highly selective and strict dependence on the interactions between the BCR and antigen molecules may be linked to the well-documented actin remodeling events in response to BCR and antigen recognition (15, 16, 28, 29). All these in vitro data support the early intravital imaging studies, which showed that B cells terminated the migratory behavior on recognition of cognate antigens on the surface of antigen-presenting cells (APCs) and subsequently developed an enlarged contact interface for the acquisition of the antigenic information (37, 38). Thus, the seamless imaging system in this report may be a better representation of what occurs in vivo because the probing behavior of B cells allows them to survey the membrane surface of the APCs for the presence of cognate antigen. We speculate that if there is no cognate antigen, the B cells will continue with the probing behavior. In marked contrast, the presence of BCR specific antigens will efficiently terminate the B-cell probing behaviors and lead to the synaptic accumulation of the BCRs and an increase in the size of the B-cell contact area. Furthermore, the findings in this report demonstrated that the termination of the probing behavior of B cells on BCR engagement is very sensitive to the affinity between the BCR and the antigen, suggesting the potential possibility that B cells could use probing termination event to discriminate antigen affinity.

Third, we captured the growth feature of the BCR microclusters in response to the immune recognition of the BCR and cognate antigens, consistent with our previous studies (6). We also showed that the BCRs could be accumulated into the stationary BCR microclusters in a passive trapping mechanism. Using a series of two-color TIRFM imaging experiments in combination with the photoactivatable system, we simultaneously examined the spatial and temporal changes of both the BCRs and control molecules (MHC-I or lipid) within the B-cell IS. We observed that only BCRs but not MHC-I, nor lipid molecules could be stably trapped within the B-cell IS. Based on all these data, we speculate that both MHC-I and general lipid molecules had the freedom to diffuse into, and out of the B-cell IS. Additionally, an interesting finding was that the freedom to diffuse out of the BCR microclusters is not without limits. When the B-cell IS matures and forms the central and peripheral structures (as examined in a 30-min TIRFM imaging experiment of live B cells in Fig. 5D and Fig. S8B), a significant accumulation of MHC-I within the central area of the B-cell IS could be observed. These unexpected results indicated the molecular crowding effects that were induced by the F-actin rich cytoskeleton structures corralling the central area of a B-cell IS. Because recent studies by Dustin and colleagues demonstrated the polarized release of TCR-enriched vesicles at the T-cell IS (39), it shall be of interest to examine the biological functions of the heterogeneous molecular crowding effects within the B-cell IS.

Last, this report suggested that the decreased lateral mobility of antigen engaged BCRs accounted for the synaptic accumulation of the BCRs into the BCR microclusters. This supporting data were obtained by accurately quantifying the Brownian diffusion coefficients of thousands of BCR molecules on the same B cells before and immediately after BCR engagement in a temporally seamless manner. All these seamless single molecule experiments were accomplished with our photoactivatable NP system, which is capable of generating subminute scaled temporal segments immediately after BCR engagement for dozens of B cells with a synchronized trigger point. Indeed, on photoactivation, we found that the Brownian diffusion coefficient of the BCRs in the same batch of B cells gradually decreased over time, consistent with the early end point studies showing that the mobility of BCRs decreased a few minutes after the activation by mobile antigens on PLBs (6, 24, 28, 31). It is worth noting that even though the overall trend of the Brownian motion feature of the BCRs was to slow down, this change was not absolutely monotonic as we observed that the slow-moving BCR molecules regained motile behavior. These fluctuations fit well with the early observation by Batista and colleagues examining the mobility of BCRs upon the recognition of membrane bound antigens on PLBs (28), which depicted the nonmonotonic mode of regulation of the single molecule motile behavior of the BCRs by actin cytoskeleton system. The contemporary “picket and fence” model of the plasma membrane (40–42), where the actin cytoskeleton represents the fence and the transmembrane proteins represent the picket, predicts that the membrane receptors will undergo oligomerization-induced trapping upon ligand binding, thus showing the decreased Brownian diffusion coefficient. Interestingly, in our experimental system, this oligomerization-induced trapping phenotype was only observed with BCR molecules but not with MHC-I molecules, which suggested that these events were induced by the recognition of BCR and antigen. Overall, these results suggested that the decreased lateral mobility of antigen engaged BCRs accounts for their synaptic accumulation.

Thus, by using this photoactivatable antigen-supported seamless imaging system, we can for the first time to our knowledge, readily capture the behavioral changes of the same B cell before and after BCR antigen recognition. All of these results improved our understanding of the B-cell probing termination behaviors and the sophisticated BCR sorting mechanisms within the B-cell IS during B-cell activation.

Materials and Methods

Mice, Cells, Antibodies, and Plasmids.

High-affinity J558L-B1-8-IgM-High and low-affinity J558L-B1-8-IgM-Low cells were constructed and maintained as reported (6). For brevity, we used J558L-B1-8-IgM cells to indicate high-affinity J558L-B1-8-IgM-High cells in most parts of this report except in the places in comparison with J558L-B1-8-IgM-Low cells. B1-8 primary B cells were isolated from the spleen of IgH ^B1-8/B1-8 Igκ^−/− transgenic mice as previously reported (6). DyLight 649-conjugated Fab anti-mouse IgM constant region antibodies were purchased from Jackson ImmunoResearch Laboratories. Goat whole IgG anti–MHC-I antibodies were purchased from BioLegend. Dil cell-labeling was purchased from Invitrogen. ICAM-I antibodies were purchased from R&D.

Chemical Reagents and Solvents.

Fmoc amino acids, HCTU [2-(6-chloro-1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate], DIEA (N,N-diisopropylethylamine), HOBt (hydroxybenzotriazole), DIC (N,N'-diisopropylcarbodiimide), Rink amide-AM resin, TFA (trifluoroacetic acid), TIPS (triisopropylsilane), and phenol were acquired from GL Biochem. Piperidine, DMF (dimethylformamide), and CH₂Cl₂ were purchased from Sinopharm Chemical Reagent Co. All reagents and solvents used for small organic compound synthesis were bought from Sinopharm Chemical Reagent Co. or Alfa Aesar with the highest commercial quality and were used without further purification. Anhydrous solvents (THF, CH₂Cl₂) were obtained from a dry solvent system (passed through a column of alumina).

Peptide Synthesis.

Peptide synthesis vessels were acquired from Synthware Glass Co. All of the peptides were synthesized manually according to the standard protocol. For the regular coupling reaction, Rink amide-AM resin was coupled with 4 eq. amino acid, 3.6 eq. HCTU in 0.4 M DIEA (8 eq.)/DMF solution; 20% (vol/vol) piperidine in DMF was used for deprotection. The following Fmoc amino acids were used: Fmoc-His(Trt)-OH, Fmoc-Ser(^tBu)-OH, Fmoc-Gly-OH, Fmoc-Thr(^tBu)-OH, Fmoc-Ala-OH, Fmoc-Cys(Trt)-OH, Fmoc-Lys(Boc)-OH, and Fmoc-ε-Acp-OH. After the N-terminal Fmoc-ε-Acp-OH was coupled to the resin, the final deprotection was performed, followed by addition of 2 eq. 4-hydroxy-3-nitrophenyl acetic acid (NP-OH), 5 eq. HOBt in 0.2 M DIC (5 eq.)/DMF solution to afford the NP-conjugate peptide. For the synthesis of DMNB-NP-conjugate peptide, the corresponding 4,5-dimethoxy-2-nitrobenzoyl-NP-acetic acid (DMNB-NP-OH) was added. After the assembly was completed, peptides were cleaved from the resin by treatment with a TFA mixture [88% (vol/vol) TFA, 2% (vol/vol) TIPS, 5% (vol/vol) H₂O, 5% (vol/vol) phenol]. The crude peptide was then analyzed and purified by RP-HPLC, and the molecular weight of each peptide was confirmed by ESI-MS.

RP-HPLC and MS.

Analytical and semipreparative RP-HPLC was performed with a Prominence LC-20AT with SPD-20A UV/Vis detector. All separations involved a mobile phase of 0.1% (vol/vol) TFA in water (solvent A) and 0.1% (vol/vol) TFA in acetonitrile (solvent B). Analytical separations were performed using a linear gradient (5–95%) of buffer B in buffer A over 30 min after an initial isocratic phase of 5% buffer B in buffer A for 2 min with a 1 mL/min flow rate on a Vydac C18 column (5 μm, 4.6 × 250 mm). A Vydac C18 column (10 μm, 10 × 250 mm) with a 4 mL/min flow rate was used for semipreparative separations. Data were recorded and analyzed by LC-Solution software. Product containing fractions were identified by ESI-MS. ESI-MS was performed on an Agilent 1200/6340 mass spectrometer in the Center of Biomedical Analysis. The buffers for MS analysis were the same but using formic acid instead of TFA to increase MS ionization of peptides.

Photochemistry.

Caged peptides in PBS buffer (pH 7.4, containing 10 mM DTT) (1 mg/mL, 20–50 μL) were placed in a 1.5-mL Eppendorf tube on ice. Irradiation was performed at 365 nm using the Omnicure S1500 (EXFO Photonic Solutions Inc.) with a light intensity of 18 mW/cm². RP-HPLC analysis was used to determine the photolysis kinetics. Peptides were quantified based on the HPLC standard curves of the caged and noncaged derivatives. The amount of each peptide was normalized to the amount of the internal standard (Benzamide).

ELISA.

We measured the binding capacity of anti-NP IgG to different NP peptides by ELISA. Each peptide (20 µg/mL) was photoactivated for the indicated amount of time and was then diluted to 10 or 5 µg/mL Theses NP peptides were then coated on maxisorb plates (Nunc) by an overnight incubation at 4 °C. The plates were further blocked with 0.3% (wt/vol) gelatin in PBS buffer (2 h at 37 °C), followed by the addition of anti-NP antibodies. After an incubation at 37 °C for 1 h, 1:10,000 diluted HRP-conjugated goat anti-mouse IgG was used to detect NP antibody. After washing, the substrate solution, composed of 0.325% orthophenylenediamine dihydrochloride (OPD; Sigma) and 0.085% H₂O₂ in 0.3 M Tris-citrate buffer, pH 6.0, was added. After incubation at room temperature for 10 min in the dark, the reaction was stopped with 2.5 M H₂SO₄ and the plates were measured using the ELISA plate reader (Bio-Rad) at 490 nm.

Photoactivation of Caged-NP and Molecule Imaging by TIRFM.

The behaviors of NP-specific B cells before and after the photoactivation of caged-NP were examined by TIRFM imaging. Caged-NP peptides were first tethered to acid-treated coverslips for 30 min at 37 °C. Caged-NP peptides can also be captured to the coverslips presenting PLBs. In this case, we first prepared Ni-NTA–containing PLBs following our published protocol (6, 18, 30, 43). Briefly, we mixed 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC; Avanti Polar Lipids) and 1,2-dioleoyl-sn-glycero-3-[N(5-amino-1-carboxypentyl) iminodiacetic acid]-succinyl (nickel salt; DOGS–Ni-NTA; Avanti Polar Lipids) in a 9:1 DOPC/DOGS–Ni-NTA ratio. Small unilamellar vesicles (SUVs) were formed by sonication of the mixed lipids and clarified by ultracentrifugation and filtering. Acid treated coverslips were first coated with Ni-NTA–containing SUVs (0.1 mM) to form PLBs, and then caged-NP peptides containing a His₆-Tag were tethered to PLBs. The coverslips were washed with PBS before the blocking step with 200 µL of 0.1% (wt/vol) casein at 37 °C for 1 h. After further washing, NP-specific B cells were placed into the chamber and incubated for 10 min at 37 °C. Next, the photoactivation experiments by the 405-nm laser can be performed. TIRFM images were acquired using an Olympus IX-81 microscope equipped with a TIRF port, ANDOR iXon+ DU-897D electron-multiplying EMCCD camera, Olympus 100 × 1.45 NA objective TIRF lens, and a 405-, 568-, and 640-nm laser (Coherent). The acquisition was controlled by the Metamorph system (MDS Analytical Technologies). All of the photoactivation and TIRFM imaging were performed at 37 °C unless otherwise indicated. The total photoactivation time of 30 s was achieved by 10 cycles of the following imaging acquisition itinerary: (i): 405 nm for caged-NP photoactivation (exposure time, 3 s); (ii) 647 nm for BCR imaging (exposure time, 100 ms); and (iii) 568 nm for MHC-I or lipid imaging if necessary (exposure time, 100 ms). Starting from the 11th cycle, the 405-nm photoactivation step was removed from the itinerary. The interval time between each cycle was adjusted to 4 s automatically. DyLight 649-conjugated Fab anti-mouse IgM constant region antibodies were used to mark the IgM BCRs. The mFI of BCRs within the B cell’s contact interface with the coverslips tethering photoactivated-NP were processed and analyzed through Image J (National Institutes of Health) or Matlab (MathWorks) software following our published protocol (6, 18, 30, 43).

Two-Color Time-Lapse Live-Cell Imaging by TIRFM.

By using two-color TIRFM, lipid or MHC-I molecules were imaged in parallel with the BCRs for 420 s on J558L-B1-8-IgM cells in contact with coverslips coated with caged-NP antigen. Without stopping the TIRFM imaging procedure, we photoactivated caged-NP by 405-nm total reflection photons and continued to capture the behavior changes of both BCRs and the other molecules in photoactivation and thereafter. During the acquisition, two-color TIRFM images were taken every 4 s through multiple dimensional acquisition modes controlled by the Metamorph system. TIRFM images were processed to acquire the mean fluorescence intensity using Image Pro Plus (Media Cybernetics) and Image J (National Institutes of Health) software as mentioned above.

Single-Molecular Tracking and Analysis.

Single BCR molecule imaging was performed following our published protocol (6). Briefly, prelabeled J558L-B1-8-IgM cells were imaged by TIRFM with a 640-nm laser at an output power of 10 mW at the objective lens in the epi-fluorescence mode. A subregion of 100 × 100 pixels of the available area of the electron-multiplying CCD chip was used to achieve an exposure time of 30 ms/frame, the time resolution of which was found to be sufficient to reliably track the single molecule BCRs as reported (6, 31). Single molecule tracking of BCR molecules was analyzed as described in our previous study (6, 31). MSD and short-range diffusion coefficients for each BCR molecule trajectories were calculated from positional coordinates and plotted as CDP.

Subpopulation Analysis of Molecule Motion.

BCR or MHC-I molecules in each temporal segment are grouped into a fast population and a slow population, achieved by mathematically fitting the distribution of the instant diffusion coefficient ($D_0$) by a mixture model of two components. The fitting begins by transforming $D_0$ to its logarithm, $x=\text{log}\left(D_0\right)$. For the slow population, the PDF, noted by $p_s$, is assumed to be Gaussian

$$p_s\left(\mu ,\sigma ,x\right)=\frac{1}{\sqrt{2\pi }\sigma }\text{exp}\left[-\frac{{\left(x-\mu \right)}^2}{2{\sigma }^2}\right].$$

For a molecule exhibiting 2D Brownian motion of average diffusion coefficient D, $D_0$ follows χ² distribution of 2°. Hence the PDF $p_B$ is

$$p_B\left(D,x\right)=\frac{\text{exp}\left(x\right)}{D}\exp \left[-\frac{\text{exp}\left(x\right)}{D}\right].$$

The fast population is modeled by a mixture of groups of Brownian molecules. Considering the complexity of cell conditions, Brownian molecules will follow χ² distributions of parameter D slightly different to each other. To formulate this, the PDF of the fast population $p_f$ equals to $p_B$ blurred by a Gaussian function

$$p_f\left(D,{\sigma }_s,x\right)=\underset{-\infty }{\overset{\infty }{{\displaystyle \int }}}{p}_B\left(D,x-t\right)\frac{1}{\sqrt{2\pi }{\sigma }_s}\text{exp}\left(-\frac{t^2}{2{\sigma }_S^2}\right)dt.$$

The log likelihood function $L\left(\mu ,\sigma ,D,{\sigma }_s,q\right)$ on molecule set $\{x_1,\hspace{0.17em}{x}_2,\dots ,x_n\}$ is

$$L\left(\mu ,\sigma ,D,{\sigma }_s,q\right)={\displaystyle \sum }_{i=1}^{\infty }\log \left[q{p}_f\left(D,{\sigma }_s,x_i\right)+\left(1-q\right)p_s\left(\mu ,\sigma ,x_i\right)\right],$$

Where q is the proportion of the fast population. Parameters are estimated by maximizing the log likelihood function, solved by the matlab function fminsearch.

Quantification of BCR Microclusters and B-Cell Probing Behaviors.

The mathematical quantification of BCR microclusters for precise FI and position information was performed following our published protocol by a MatLab-based 2D Gaussian analysis algorithm (6, 43). This function was used to quantify each of the 2D FI profiles for some critical parameters of each microcluster, such as position (xc, yc) and integrated FI (I). B-cell probing behaviors were quantified by counting the number of membrane pseudopods/protrusions per cell per TIRFM image. In this report, the membrane protrusions were defined as the puncta-structured membrane protrusions showing distinct distances to the boundaries of the centralized cell body of B cells in TIRFM images.

Pretreatment of B Cells with Pharmaceutical Inhibitors.

B cells were pretreated with 0.2 µM cytochalasin D for 10 min at 37 °C to block actin polymerization or were pretreated with 1 µM jasplakinolide for 45 min at 37 °C to block actin depolymerization (6). B cells were pretreated with 50 µM of the Src family tyrosine kinase inhibitor PP2 for 10 min at room temperature or with 50 µM Syk inhibitor piceatannol for 10 min at room temperature (6).