Replication-dependent size reduction precedes differentiation in Chlamydia trachomatis

Abstract

Chlamydia trachomatis is the most common cause of bacterial sexually transmitted infection. It produces an unusual intracellular infection in which a vegetative form, called the reticulate body (RB), replicates and then converts into an elementary body (EB), which is the infectious form. Here we use quantitative three-dimensional electron microscopy (3D EM) to show that C. trachomatis RBs divide by binary fission and undergo a sixfold reduction in size as the population expands. Conversion only occurs after at least six rounds of replication, and correlates with smaller RB size. These results suggest that RBs only convert into EBs below a size threshold, reached by repeatedly dividing before doubling in size. A stochastic mathematical model shows how replication-dependent RB size reduction produces delayed and asynchronous conversion, which are hallmarks of the Chlamydia developmental cycle. Our findings support a model in which RB size controls the timing of RB-to-EB conversion without the need for an external signal.

Fig. 1

Temporal analysis of chlamydial developmental forms using a three-dimensional electron microscopy approach. a Serial block-face scanning electron microscopy analysis (SBEM) was used to generate a three-dimensional computational reconstruction of the chlamydial inclusion in a C. trachomatis-infected HeLa cell at 28 h post infection (h.p.i.). Micrographs (middle) are shown for sections 154 and 132 (3/4 and halfway up from equator, respectively) and section 88 (equator), with segmentation markings for inclusion membrane (green), RBs (dark blue), dividing RBs (light blue), IBs (orange) and EBs (red). Scale bar: 1000 nm. b Entire chlamydial inclusions from representative infected cells at 16, 28, and 36 h.p.i. Scale bar: 1000 nm. Pie charts showing mean numbers of each chlamydial form per inclusion are grouped into three developmental phases: RB replication only (no IBs or EBs), onset of RB-to-EB conversion (IBs + EBs ≤50% of chlamydiae), and EB accumulation (IBs + EBs >50% of chlamydiae). All four chlamydial forms inside each inclusion were identified and counted: 12 h.p.i. (n = 50 inclusions), 16 h.p.i. (n = 31), 20 h.p.i. (n = 22), 24 h.p.i. (n = 10), 28 h.p.i. (n = 13), 32 h.p.i. (n = 10), 36 h.p.i. (n = 9), 40 h.p.i. (n = 10)

Fig. 2

Volume analysis of the chlamydial inclusion and chlamydial forms during the developmental cycle. The data presented in this figure is compiled from a total of 155 inclusions: 12 h.p.i. (n = 50), 16 h.p.i. (n = 31), 20 h.p.i. (n = 22), 24 h.p.i. (n = 10), 28 h.p.i. (n = 13), 32 h.p.i. (n = 10), 36 h.p.i. (n = 9), 40 h.p.i. (n = 10). Error bars represent standard deviation. a Temporal change in inclusion volume and number of chlamydiae/inclusion. The data is presented in log scale. b Linear relationship between inclusion volume and total number of chlamydiae within that inclusion. Each dot represents a single inclusion, color-coded by its developmental phase, as described in Fig. 1b. c Temporal change in inclusion volume and total volume of chlamydiae within the inclusion. d % inclusion volume occupied by chlamydiae was calculated for each inclusion as total chlamydial volume divided by inclusion volume

Fig. 3

RB size decreases and becomes heterogeneous as the developmental cycle progresses. The data for a and b were compiled from a total of 140 inclusions: 12 h.p.i. (n = 50 inclusions), 16 h.p.i. (n = 31), 20 h.p.i. (n = 22), 24 h.p.i. (n = 9), 28 h.p.i. (n = 7), 32 h.p.i. (n = 8), 36 h.p.i. (n = 5), 40 h.p.i. (n = 8). Error bars represent standard deviation from the mean. a Temporal change in volume of RBs and dividing RBs. Average volume of all RBs or dividing RBs in each inclusion was first determined, and then reported as the mean RB or dividing RB volume for all inclusions at each time point. Mean values are reported and error bars indicate standard deviation. The decrease in RB size was statistically significant between 12 h.p.i. and all later time points (highest p-values were between 12 and 16 h.p.i.: p = 0.00025, t-value = 3.9, df = 45, and between 12 and 24 h.p.i.: p < 0.0001, t-value = 4.6, df = 28, unpaired t-test). b Ratio of dividing RB volume to RB volume during the developmental cycle. For each time point, the ratio was first determined for each inclusion, and then reported as the mean of ratios for all inclusions at that time point. c Size histograms for all RBs within a single inclusion at 24 h.p.i. (n = 40) and 40 h.p.i. (n = 240), distributed into 0.1 μm³ bins. Insets show the smallest bin subdivided into five 0.02 μm³ bins with same y-axis scale

Fig. 4

RB replication by binary fission. Size histogram of nascent daughter cells at 24 h.p.i. The daughter/parent ratio was calculated as the volume ratio of each daughter cell to its dividing RB parent. Volumes were measured for 228 daughter cells from 114 dividing RBs in two 24 h.p.i. inclusions

Fig. 5

Analysis of size-dependent control of RB-to-EB conversion using a stochastic mathematical model. a Proposed model in which the size of an RB determines whether it can convert or continues to replicate. RBs become progressively smaller because they divide, on average, at less than twice their starting size. They can only convert into an EB below a permissive size. These two elements of RB size control ensure that the RB population expands before conversion occurs. The figure demonstrates how weak control of RB size at replication can produce size heterogeneity and lead to asynchronous conversion by varying the number of replication cycles required to reach the conversion size threshold. b Wiring diagram to show the four different variables in the system and the four possible transformations. Details of the mathematical model provided in Supplementary Notes. c Mean volume of the RB population within an individual inclusion, measured experimentally (16 h.p.i. n = 8 RBs, 24 h.p.i. n = 40, 32 h.p.i. n = 245, 40 h.p.i. n = 240), or produced by the size control model, at selected time points. Error bars indicate standard deviation. d Histograms of RB size obtained with the mathematical model for single inclusions at 24 and 40 h.p.i. recapitulate the experimental data in Fig. 3c. e Two sample time courses from the model illustrating how different RB lineages culminate in different times of RB-to-EB conversion. Each time course consists of successive rounds of RB replication (blue line) followed by conversion to an IB (orange line) and then EB (red line). Each newly produced RB shown by an open circle. f Histogram showing time of RB-to-EB conversion predicted by the mathematical model for all EBs produced in a single inclusion by 40 h.p.i. g Growth curves showing the mean number of each chlamydial form/inclusion. The graph on the left was produced by the stochastic size control model, while the graph on the right shows growth curves from the 3D EM analysis of Chlamydia-infected cells (12 h.p.i. n = 50 inclusions, 16 h.p.i. n = 31, 20 h.p.i. n = 22, 24 h.p.i. n = 10, 28 h.p.i. n = 13, 32 h.p.i. n = 10, 36 h.p.i. n = 9, 40 h.p.i. n = 10)