Begitt et al., (2014). STAT1-cooperative DNA binding distinguishes type 1 from type 2 interferon signaling.
Interferons (IFNs) are a class of glycoproteins known as cytokines, which are important in cell signaling, and are produced by a range of cells. They are known to trigger a variety of immune responses, such as the activation of macrophages and the modulation of cellular environments into those less permissive of viral replication. Interferons are generally divided into classes based upon the type of receptor through which they signal; Interferon type 1 binds to a cell surface receptor known as IFNAR (Interferon Alpha/Beta Receptor), while Interferon type 2 binds to INFGR (Interferon Gamma Receptor). There are many type 1 interferons in humans, including interferon alpha (IFN-α), and one type II interferon, interferon gamma (IFN-γ).
Figure 1.JAK-STAT pathway activation by type I and II interferons. Interferons bind to receptors at the plasma membrane. The type I IFNreceptor is composed of two subunits (IFNAR1 and IFNAR2) which are associated with the Janus activated kinases (JAKs). JAKs associated with type I IFNs, once activated, phosphorylate STAT2 or STAT1, which associates with RF9 to form ISGF3 complexes. In the nucleus, these bind ISRE DNA sequences to initiate gene transcription. The type II IFN receptor is composed of two subunits (IFNGR1 and IFNGR2) which are associated with JAK1 and JAK2. The JAKs associated with type II IFNs act upon STAT1, which forms a dimer (GAF), and binds GAS DNA sequences to initiate gene transcription. The consensus GAS element and ISRE sequences are shown, where N represents any nucleotide. Figure taken from .
Upon binding to their respective receptors IFNs (Figure 1), through Janus kinase (JAK), trigger a cascade of tyrosine phosphorylations leading to the activation of the signal transducer and activator of transcription (STAT) complexes. These transcription factors regulate a variety of genes, many of which are relevant for immune response. Different interferons activate different STAT pathways, which activate different genes:
- IFN-α (Type I IFN) leads to the production of STAT1-STAT2 heterodimers, which additionally require IRF9 (interferon-regulatory factor 9) for transcriptional activity. This heterotrimer, ISGF3 (interferon-stimulated gene factor 3) initiates transcription at so-called ISREs (INF-stimulated response elements) sites.
- IFN-γ (Type II IFN) binding results in the formation of STAT1 homodimers known as GAF (Gamma-activated factor), which bind palindromic DNA sequences known as GAS (Gamma-activated sequence) sites.
It has also been established that both GAF and ISGF3, through cooperative DNA binding, are able to recruit additional complexes to adjacent GAS and ISRE sites, respectively. This polymerisation is dependent upon the N-terminal domain of STATs, a site which is conserved within that family of proteins. This conserved domain, however, is not necessary for the formation of the individual dimers STAT1-STAT2 (1 above) nor STAT1-STAT1 (2 above). Hence, substitution of alanine for phenylalanine in the N-terminal domain (position 77) disrupts the region, prevents the polymerisation of transcription factors, and of cooperative DNA binding
Cooperative DNA binding can act as an important switching mechanism, as exemplified by phage lambda where is acts to determine whether the phage is in a lytic or lysogenetic state. In human cells, the sparsity of IFN-regulated genes in proximity to GAS or ISRE sites suggests an accessory role for STAT1 cooperativity in INF signaling. Begitt et al.  investigated the role of cooperative DNA binding in type 1 and type 2 INF-controlled gene expression, using in vivo, in vitro and in silico methods, the latter of which are represented by models BIOMD0000000500 and BIOMD0000000501.
The basis for much of this work was the generation of a mouse mutant (STAT1F77A) incapable of polymerising additional STAT dimers due to an amino acid substitution as described above. This strain demonstrably expressed STAT1 at levels comparable to those seen in the wild type. Following incubation of cells with IFN-α or IFN-γ, electrophoretic mobility shift assays demonstrated that both GAF and ISGF3 complexes could not bind DNA cooperatively in STAT1F77A mice.
Figure 2. Bacterial load in bone marrow macrophages. Wild type and STAT1F77A cells preincubated with IFN-γ and untreated controls were infected with Listeria monocytogenes. Bacterial survival was assessed over a 24 hour period by measuring colony forming unit capacity atvarious sample times. Wild type cells preincubated with IFN-γ (blue line) responded best, while wild type and STAT1F77A cells that were not preincubated with IFN-γ (black and grey line, respectively) responded comparably. The increased listeriosis in STAT1F77A cells (red line) is due to a reduction of Listeria-induced apoptosis. Figure taken from .
Additionally, wild type and STAT1 modified mice infected with the VSV (Vesicular stomatitis virus), which is known to induce IFN-α mediated response, both responded similarly. This indicated that STAT1 cooperativity was not required in the mounting of a type 1 IFN immune response. Conversely, type 2 IFN responses in STAT1F77A mice was found to be severely compromised as evidenced by the antibacterial capabilities of macrophages preincubated with IFN-γ from wild type and STAT1F77A mice (Figure 2). Collectively, it was demonstrated that defective STAT1 interfered specifically with IFN type 2 signaling, and did not affect type 1.
The authors then used microarray and RT-PCR techniques to examine induction of gene transcription following IFN treatment. IFN-α induction gave comparable results between wild type and STAT1F77A cells, while in IFN-γ treated cells, 404 of the 456 IFN-γ induced genes were found to be expressed at levels less than a half of those seen in the wild type.
Promoter recruitment was also examined using chromatin immunoprecipitation (ChIP). It was found that where tandem GAS sequences were present in a target promoter, STAT1F77A derived cells were not able to recruit STAT1 when induced with IFN-γ; it was found that only single GAS sites were able to bind STAT1 in STAT1F77A cells. However, it was also established that STAT1 associated normally in response to IFN-α, and confirmed that GAF has a negligible role in type 1 IFN signaling.
One would anticipate cooperative DNA binding to occur with multiple adjacent binding sites, however with GAS sites only a single site was required. Indeed, sequence analysis indicated that there was only one discernible GAS site in 99% of IFN-γ upregulated genes. Using low stringency sequence constraints to determine the presence of tandem GAS sites, it was reasoned that there could be a maximum of around 15% of such GAS sites. This scenario was modeled mathematically to determine the effects of single (BIOMD0000000500) and double (BIOMD0000000501) GAS sites. In the model, GAS sites are surrounded by non-GAS sites which are defined as having a 50-fold lower binding affinity. Essentially, for the single GAS model, this comprises 3 sites with a central GAS site flanked by non-GAS sites, while the double GAS model has tandem GAS sites flanked by non-GAS sites.
Figure 3. Schematic reaction examples. Schematic examples are given of reversible STAT/DNA binding under different conditions:a) binding of a STAT dimer to a GAS site occupied by 2 STAT dimersb) binding of a STAT dimer at a Gas site, where adjacent non-GAS sites are already occupied by STAT dimersc) reversible polymerisation where 2 bound STAT dimers bind cooperatively to form a tetramer. Figure taken from .
For simulation, STAT1 dimers were allowed to bind to any sites in the model freely, aside from the binding affinities assigned. Once bound to a site, the dimer was allowed to polymerise to adjacent STAT1 dimers, with a specified polymerisation rate. Since only dimers were allowed to unbind from GAS sites, depolymerisation was first required in the case of cooperatively bound STAT1 dimers. The single GAS site model consisted of 12 reversible STAT/DNA binding reactions (for example, Figure 3a,b), and 6 polymerisation reactions (for example, Figure 3c). The double GAS site model comprised 32 STAT/DNA reactions and 25 polymerisation reactions.
Using the COPASI parameter scan function, different initial free STAT1 concentrations in the range 10-5 to 10-15 were used and run to steady state. Results were presented as fractional GAS site occupancy: the proportion of DNA fragments where a STAT1 molecule is bound. It was found that the double GAS site/cooperative binding model was able to achieve 99% GAS site occupancy using 13-fold lower transcription factor concentration than the non-cooperative binding model (Figure 4). Even where the binding affinities for GAS and non-GAS sites were modified to differ 200-fold, cooperativity still markedly increased promoter occupancy. To test whether increasing STAT1 concentration alone could rescue IFN-γ induced promoter recruitment in STAT1F77A mice, an immune response was mimicked by exposing cells for 24 hrs to IFN-γ under culture conditions, a process known as IFN-γ priming. This was found to increase STAT1 concentration around 5-fold, but did not rescue gene expression function in STAT1F77A cells. Thus single-site cooperative DNA binding enhances promoter recruitment, an observation verified by both mathematical and experimental methods.
Figure 4. Mathematical modeling of STAT1 promoter binding. Binding of STAT1 at promoter sites containing single (1xGAS) or tandem (2xGAS)GAS sites was performed. STAT1 affinity was defined as being 50-fold greater for GAS sites over non-GAS sites, while dissociation (off) rates were 60-fold higher for high cooperativity over low cooperativity. It can be seen that 99% GAS site occupancy (indicated by arrows for single GAS site) is achieved at more than 10-fold lower STAT1 concentration with high cooperativity. Figure taken from .
This work elegantly demonstrates, through the use of a wide range of techniques, that activated STAT1 cooperative DNA binding and polymerisation at GAS sites plays a major role in type II interferon responses and antibacterial activity through interferon gamma.
- Leonidas C. Platanias. Mechanisms of type-I- and type-II-interferon-mediated signalling Nat Rev Immunol. 2005 5:375-386
- Begitt, A.et al. STAT1-cooperative DNA binding distinguishes type 1 from type 2 interferon signaling. Nat Immunol. 2014 15:168-176