Differential context-specific impact of individual core promoter elements on transcriptional dynamics

Abstract

Eukaryotic transcription occurs in bursts that vary in size and frequency, but the contribution of individual core promoter elements to transcriptional bursting is not known. Here we analyze the relative contributions to bursting of the individual core promoter elements-CCAAT, TATAA-like, Sp1BS, and Inr-of an MHC class I gene in primary B-cells during both basal and activated transcription. The TATAA-like, Sp1BS, and Inr elements all function as negative regulators of transcription, and each was found to contribute differentially to the overall bursting pattern of the promoter during basal transcription. Whereas the Sp1BS element regulates burst size, the Inr element regulates burst frequency. The TATAA-like element contributes to both. Surprisingly, each element has a distinct role in bursting during transcriptional activation by γ-interferon. The CCAAT element does not contribute significantly to the constitutive transcriptional dynamics of primary B-cells, but modulates both burst size and frequency in response to γ-interferon activation. The ability of core promoter elements to modulate transcriptional bursting individually allows combinatorial fine-tuning of the level of MHC class I gene expression in response to intrinsic and extrinsic signals.

FIGURE 1:

Single mRNA transcripts detected by FISH in primary B-lymphocytes. (A) Sequence of the PD1 proximal promoter. Promoter elements are boxed and the mutations in each element indicated in red. (B) Detection and counting of single PD1-Cy3-labeled and CD19-Cy5-labeled mRNA transcripts in primary B-cells stimulated ex vivo. Left panel: Maximum-intensity projection of smFISH staining of B-cells; red and green spots indicate single PD1 and CD19 mRNAs, respectively. White arrow indicates CD19 transcription site with multiple nascent mRNAs. Right panel: Green squares identify individual mRNA molecules and circles identify transcription sites. (C) PD1/CD19 smFISH in a transgene-nonexpressing mouse. (D) smFISH with anti PD1-Cy3 and H2K^b-Cy5 probes in primary B-cells show nonoverlapping detection of endogenous and transgenic MHC class I mRNA. (E) Colocalization of smFISH spots for CD19 with two probe sets targeted to the 5′ and 3′ regions of the mRNA. (F) CD19 smFISH in B-cell stimulated culture. Image shows a CD19-expressing lymphocyte and a CD19-negative lymphocyte. (G) Bimodal distribution of CD19 transcripts indicates non–B-cells remaining in the population after stimulation. Black line indicates threshold for CD19 positive B-cells gated for downstream analysis. Scale bars represent 5 µm.

FIGURE 2:

Core promoter mutations alter PD1 mRNA distribution and nascent transcript dynamics. (A) Representative smFISH images of B-lymphocytes probed with Cy3-labeled PD1 oligonucleotides, derived from either the WT or the indicated PD1 core promoter element mutation. (B) Binned and normalized PD1 mRNA distributions of the WT transgene in B-cells (blue, top panel; gray, remaining panels), and of the core promoter mutant transgenes (red).

FIGURE 3:

Bursting dynamics of PD1 core promoter mutants. (A) Fano factor vs. mean mRNA/cell of PD1 core promoter element mutants in B-lymphocytes stimulated in the absence or presence of γ-interferon. (B) A best fit curve generated from the two-state bursting telegraph model to a PD1 mRNA distribution. The two-state model can be represented as combined rates of a, a gene turning on; b, a gene turning off; and c, the rate of transcript firing with a burst. (C) Burst parameters that that satisfy the χ² null hypothesis test (p > 0.1) for the WT distribution and for the core promoter mutants. Nonoverlapping distributions indicate samples with significantly different bursting kinetics. Burst size is mRNA/cell; burst frequency is unitless.

FIGURE 4:

γ-Interferon treatment differentially affects PD1 transcription in core promoter mutants. (A) Normalized histograms of PD1 RNA expression in B-lymphocytes from WT and core promoter mutant lines stimulated in the absence or presence of γ-interferon. Gray, no γ-interferon; red, + γ-interferon. (B) Expression of core promoter mutants relative to WT in the presence or absence of γ-interferon, as determined from mean expression levels in A. (C) Differential bursting kinetics upon γ-interferon treatment in WT and mutant PD1 genes. Triangles indicate γ-interferon–treated samples in respective panels.

FIGURE 5:

Tissue-specific expression dynamics of H2K^b. (A) smFISH of H2K^b RNA in primary T- (left) and B-lymphocytes (right). (B) Pooled and normalized H2K^b mRNA distributions in primary T- and B-lymphocytes. (C) The differential bursting dynamics of the two distributions.

FIGURE 6:

Frequency and relative intensity of active transcription sites in PD1 core promoter mutants. (A) Frequency of active PD1 transcription sites over all alleles present in B-lymphocytes from WT and core promoter mutants in the presence and absence of interferon (IFN). Error bars represent the Poisson standard error. (B) Average number of nascent transcripts per PD1 transcription site in the presence and absence of interferon. Number of transcripts per given transcription site calculated by normalizing intensity to the mean intensity of a single mRNA in that field. (C) Comparison of the fraction of the inferred burst frequency with the measured fraction of alleles transcribing, in the absence of interferon. Error bars of the burst frequency represent the χ² range of solutions. (D) Comparison of the inferred burst size with measured nascent RNA in the absence of interferon. Error bars of the burst size represent the χ² range of solutions.