Clustered Bottlenecks in mRNA Translation and Protein Synthesis

Using a model based on the totally asymmetric exclusion process, we investigate the effects of slow codons along messenger RNA. Ribosome density profiles near neighboring clusters of slow codons interact, enhancing suppression of ribosome throughput when such bottlenecks are closely spaced. Increasing the slow codon cluster size beyond $\sim 3–4$ codons does not significantly reduce the ribosome current. Our results are verified by both extensive Monte Carlo simulations and numerical calculation, and provide a biologically motivated explanation for the experimentally observed clustering of low-usage codons.

Figure 1

(a) The mRNA translation/protein synthesis process. Ribosomes move unidirectionally along mRNA as tRNA (not shown) deliver the appropriate amino acid to the growing chain. Codons with low concentrations of corresponding tRNA result in bottlenecks that locally suppress ribosome motion across it. (b) A simple totally asymmetric exclusion process on $N$ lattice sites used to model mRNA translation in the presence of slow codons bottlenecks or defects.

Figure 2

Placements of slow defects, or rare-usage codons (thick red segments). (a) Two defects near the initiation site straddled by an $n=5$ lattice segment. (b) Two defects in the chain interior, away from the boundaries, $n=6$. (c) A single slow defect near the termination end of the chain, $n=4$.

Figure 3

Matrix-generating algorithm for a three site model. Each possible occupancy of the lattice is associated with a bit pattern, and the state is enumerated with the corresponding decimal value; i.e., since 011 is the binary representation of 3, we label this state 3. Next, divide the states into groups where the first lattice site is occupied (1-states), and where the first lattice site is empty (0-states). Regardless of the number of lattice sites in the TASEP, the transitions between the two classes of states always occur between the first half of the 1-states and the second half of the 0-states (dashed red arrows). To determine the remaining transitions we call the algorithm recursively on both the 1-states and the 0-states, making sure to ignore the highest order bit (i.e., the leftmost lattice site). Finally, we add the transitions between each 0-state and each 1-state resulting from injection at the left edge of the lattice (blue arrows). With the connectivity of the states fully enumerated, one readily assigns the appropriate rate $p_i$ to each of the transitions.

Figure 4

(a) Comparison of steady-state currents $J_1(q)$ for a single defect derived from MC simulations with those obtained from $n$-segment MFT. For $n\ge 4$, the FSMFT results are within 2% of those from MC simulations. The boundary injection or extraction rates were not rate limiting $(\alpha =\beta =10)$. (b) Further reduction of steady-state current as successive, identical defects is added. The first few defects cause most of the reduction in current.

Figure 5

(a) Steady-state currents across a chain with two identical interior defects spaced $k$ sites apart. The current is suppressed most when the defects are closely spaced. (b) The density profiles near a pair of defects ($q=0.15$) of various spacings $k$. The thick vertical bars denote the defect positions for $k=6$. For larger $k$, density boundary layers heal, allowing particle buildup behind the second barrier, enhancing the current. (c) The dependence of the normalized steady-state current $J_2(q;k)/J_1(q)$ as a function of the defect separation. (d) The upper bound for the mean current of a lattice with $m$ defects randomly distributed within the central $N=300$ sites. A lower bound as $m\to N-1$ is $⟨J(m\to N-1){⟩}_{\mathrm{r}\mathrm{a}\mathrm{n}}\to q/4$.