Cooperative action in eukaryotic gene regulation: Physical properties of a viral example

The Epstein-Barr virus (EBV) infects more than $90\%$ of the human population, and causes glandular fever as well as several more serious diseases. It is a tumor virus, and has been widely studied as a model system for cell transformation in humans. A central feature of the EBV life cycle is its ability to persist in human B cells in different latency states, denoted latency I, II, and III. In latency III the host cell is driven to cell proliferation and hence expansion of the viral population without entering the lytic pathway, while the latency I state is almost completely dormant. We here study the effective cooperativity of the viral C promoter, active in latency III EBV cell lines. We show that the unusually large number of binding sites of two competing transcription factors, one viral and one from the host, serves to make the switch sharper (higher Hill coefficient), either by cooperative binding between molecules of the same species when they bind, or by competition between the two species if there is sufficient steric hindrance.

Figure 1

(Color online) Illustration of the two blocking scenarios. EBNA-1 ($E$) binds black sites while Oct-2 ($O$) binds blue sites. (a) The single-side blocking model where a bound $E$ blocks binding of $O$ to the closest site to the right, and a bound $O$ blocks $E$ binding to the closest site to the left. (b) The double-side blocking model where one bound $E$ or $O$ blocks the opposite molecular species binding on both neighboring sites.

Figure 2

Hill coefficient curves for varying numbers of binding sites $N$ in the promoter. The promoter is defined as active when more than eight EBNA-1 are bound. Hence, the Hill coefficient at the limit of low EBNA-1 concentrations (low $P$) is 8, independent of $N$. At high $P$, the Hill coefficient approaches $N-7$, yielding higher effective cooperativity for longer promoters.

Figure 3

(Color online) Hill coefficient curves for both the single- and double-sided blocking models, for various concentrations of Oct-2 (legends show Oct-2 concentration). Without Oct-2, both models have the same coefficient curve (solid line). For the single-side blocking, an increase in Oct-2 does not lead to any effect on the Hill coefficient (cyan circles). On the other hand, in the double-sided blocking model, an increase in Oct-2 concentration dramatically shifts the curve (five black marker lines). For saturating levels of Oct-2, the effective Hill coefficient approaches 10.5.

Figure 4

Hill coefficient curves with no Oct-2 and increasing strength of cooperative binding between EBNA-1. The cooperative binding is varied from $0\%-40\%$ of the DNA binding strength of EBNA-1 $\left(0-6.2\text{kcal}∕\text{mol}\right)$. With no competition and only added cooperative interaction, the effective Hill coefficient changes dramatically from 3.5 up to 16 for the largest added cooperative binding energy.

Figure 5

Hill coefficient curves for the single-sided blocking model at a high concentration of Oct-2, and different strength of cooperative binding between EBNA-1. The cooperative binding is varied from $0\%-40\%$ of the DNA binding strength of EBNA-1. The effect of adding cooperative binding energy for EBNA-1 doubles the effective Hill coefficient from 3.5 to 7.

Figure 6

Hill coefficient curves for the double-sided blocking model at high concentration of Oct-2, and different strength of cooperative binding between EBNA-1. The cooperative binding is varied from $0\%-40\%$ of the DNA binding strength of EBNA-1. The double blocking in itself gives a relatively high effective Hill coefficient, and the extra cooperative interactions almost double this coefficient from 10.5 up to 18.