Single-molecule investigations of single-chain cellulose biosynthesis

Abstract

Cellulose biosynthesis in sessile bacterial colonies originates in the membrane-integrated bacterial cellulose synthase (Bcs) AB complex. We utilize optical tweezers to measure single-strand cellulose biosynthesis by BcsAB from Rhodobacter sphaeroides. Synthesis depends on uridine diphosphate glucose, Mg²⁺, and cyclic diguanosine monophosphate, with the last displaying a retention time of ∼80 min. Below a stall force of 12.7 pN, biosynthesis is relatively insensitive to force and proceeds at a rate of one glucose addition every 2.5 s at room temperature, increasing to two additions per second at 37°. At low forces, conformational hopping is observed. Single-strand cellulose stretching unveiled a persistence length of 6.2 nm, an axial stiffness of 40.7 pN, and an ability for complexes to maintain a tight grip, with forces nearing 100 pN. Stretching experiments exhibited hysteresis, suggesting that cellulose microstructure underpinning robust biofilms begins to form during synthesis. Cellohexaose spontaneously binds to nascent single cellulose strands, impacting polymer mechanical properties and increasing BcsAB activity.

Keywords: biosynthesis; cellulose; cellulose synthase; optical tweezers; single-molecule studies.

Fig. 1.

BcsAB cellulose synthesis. (A) Schematic of the BcsAB synthesis assay in which a single BcsAB complex (Protein Data Bank accession no. 4P00) is enveloped in a surface-bound nanodisc. A cellulose-binding DNA aptamer-coated bead binds the cellulose product strand, and the position and applied force are measured with nanometer and piconewton resolution using optical tweezers. (B) Cellulose synthesis traces. The black dotted line indicates the average velocity of 0.22 ± 0.01 nm s⁻¹ (SEM, n = 176).

Fig. 2.

Biochemical controls. (A) All BcsAB complexes sampled in the absence of UDP-glc (n = 10) showed no activity, while synthases sampled before removal (n = 18) and after replenishment (n = 13) displayed clear motility. Green and red points indicate synthesis and no synthesis of individual BcsAB, respectively. Distributions are shown as box plots. (B) An example trace of UDP-control buffer exchanges while monitoring the same tether shows that synthesis halts without available monomer. (C) Sampled synthases before introduction (n = 14), in the presence (n = 16), and after removal (n = 6) of an EDTA control buffer reveal that velocity slows to 0.05 ± 0.01 nm s⁻¹ (SEM). Distributions are shown as box plots. (D) Mg²⁺/EDTA control buffer exchange on a single tether indicates synthesis is severely hindered after Mg²⁺ chelation. Red and blue arrows point to the moments Mg²⁺ was chelated and replenished, respectively. All activity was recovered when both control buffers were washed out. Insets in B and D are the raw traces, including large perturbations where buffers were exchanged midexperiment. Blue indicates a complete synthesis solution, yellow indicates flow, and red indicates control buffer. We note the force range for each example trace, with the typical range spanning 2 to 3 pN. The break between control regions in D at ∼450 s is from preparing the next flow step, but no buffer exchange occurred.

Fig. 3.

C-d-GMP controls and fluorescence. (A) Different complexes were sampled for synthesis before removal, at varying time points after removal, and after replenishment of c-d-GMP. Motility persists in most cases. (B) We performed a null experiment starting without c-d-GMP, introducing it to the system, and removing again. C-d-GMP is necessary for synthesis but remains bound for long periods of time. Distributions in A and B are shown as box plots. (C) Single-tether c-d-GMP controls show synthesis is relatively unaffected when the activator is removed from solution. The applied force range is 5 to 8 pN for this example. (D) Schematic of fluorescence assay in which BcsAB is bound to the surface in the same design as the synthesis experiments, except the coverslip includes a nonstick PEG brush layer between the coverslip and the complex. If c-d-GMP–labeled DY-547 (c-d-GMPf) is bound to BcsAB, we detect fluorescence. The signal disappears in a single step when the molecule dissociates or photobleaches. (E) The number of bound and fluorescing c-d-GMPf decreases over time (n = 247). We record incredibly long bond lifetimes of c-d-GMPf, as 70% of molecules remain associated past 30 min, with 46% persisting until the 60-min acquisition time limit. An exponential fit (red) reveals a time constant of 82.5 min corresponding to an off-rate of 2.0 × 10⁻⁴ s⁻¹. Due to potential photobleaching, our results show the lower bound of the time constant.

Fig. 4.

Kinetics analysis. (A) Recorded polymerization velocities from motility traces are binned and averaged every 2 pN (blue circles) and fit to a general Boltzmann relationship (n = 176), revealing a distance to the mechanical transition state of 4.0 nm. The correlation indicates synthesis is biochemically limited, and synthesis begins to stall with an assisting load of 12.7 pN (black dashed line). Unloaded velocity (purple square) was recorded from a change in contour length over time (SI Appendix, Fig. S9). Error bars denote SEM. (B) Probabilty density function (PDF) vs. extension and retraction. (B, Top) We detected a large range in extensions (2 to 100 nm, positive distance change) and retractions (2 to 20 nm, negative distance change). The average extension during motility was 10.6 ± 1.9 nm (SEM, n = 73), and the average retraction was −4.6 ± 0.7 nm (SEM, n = 26). Exponential fits (dotted black lines) generated length scales of 5.6 and −4.0 nm, respectively, while gamma distribution fits reveal peaks at 4.1 and −2.6 nm (dashed vertical black lines). (B, Bottom) In the presence of cellohexaose, the extension and retraction profiles were best fit to gamma distributions, with peak locations appearing larger than for single cellulose at 6.8 and −6.0 nm (dashed vertical red lines). Transition magnitudes below 3 nm were not observed with cellohexaose present. Outlier extensions greater than 40 nm were excluded from diagrams but included in mean calculations. (C) Rapid extensions and retractions of 3 to 10 nm during cellulose synthesis at ∼2 pN. Force reference markers note the slight decrease in applied force as cellulose is synthesized. (D) A histogram of distances from the mean trajectory for the Inset in C is fit to the sum of two Gaussian distributions separated by a displacement of 4.8 nm. The mean distance between states is 5.0 ± 0.1 nm (SEM, n = 201 from 51 molecules). The ratio of amplitudes (a₂/a₁) is equal to the ratio of the equilibrium force (1.8 ± 0.2 pN, SEM, n = 201 segments from 51 molecules) to the acquisition force. Very few segments (<1%) displayed multimodal behavior and were excluded from this analysis.

Fig. 5.

Single cellulose strand stretching and effects of cellohexaose hybridization. (A) At the beginning of consecutive stretches, 26% of cellulose strands exhibit hysteresis until microstructures are unfurled and cellulose reaches a fundamental state. (B) During stretching, cellulose undergoes sudden elongations of random distances between regions of stability. Within larger events (yellow bars) exist smaller jumps (Inset, red bars). The change in extension is likely due to the unfolding of microstructures. (C) Cellulose product likely forms hairpins or other secondary microstructures upon or after extrusion from the complex. BcsA is shown in cyan, BcsB is shown in yellow, and the residues comprising the complex’s exit pore are shown in red. Formation of the secondary structure through hydrogen bonding (Inset, black dotted lines) between strand segments may also assist with translocating cellulose through the synthase. (D) Single cellulose chain follows the eWLC model after being fully extended (gray points and black fit). Experiments with cellohexaose (pink points and red fit) show a larger persistence length and axial stiffness. The dashed lines are theoretical fits using the same contour length of each respective data curve but the persistence length and axial stiffness of the opposite condition, with or without cellohexaose. The juxtaposition highlights the change in axial stiffness caused by cellohexaose hybridization. Bar graphs display the increase in (E) persistence length, (F) axial stiffness, (G) single-molecule synthesis velocity, and (H) bulk BcsAB overnight activity, represented as disintegrations per minute (DPM), in the presence of cellohexaose. Error bars in E-H denote SEM.