A growing body of evidence suggests that only a few amino acids (“hot-spots”) at the interface contribute most of the binding energy in transient protein-protein interactions. However, experimental protocols to identify these hot-spots are highly labor-intensive and expensive. Computational methods, including evolutionary couplings, have been proposed to predict the hot-spots, but they generally fail to provide details of the interacting amino acids. Here we showed that unbiased evolutionary methods followed by biased molecular dynamic simulations could achieve this goal and reveal critical elements of protein complexes. We applied the methodology to selected G-protein coupled receptors (GPCRs), known for their therapeutic properties. We used the structure-prior-assisted direct coupling analysis (SP-DCA) to predict the binding interfaces of A2aR/D2R, CB1R/D2R, A2aR/CB1R, 5HT2AR/D2R, and 5-HT2AR/mGluR2 receptor heterodimers, which all agreed with published data. In order to highlight details of the interactions, we performed molecular dynamic (MD) simulations using the newly developed AWSEM energy model. We found that these receptors interact primarily through critical residues at the C and N terminal domains and the third intracellular loop (ICL3). The MD simulations showed that these residues are energetically necessary for dimerization and revealed their native conformational state. We subsequently applied the methodology to the 5-HT2AR/5-HTR4R heterodimer, given its implication in drug addiction and neurodegenerative pathologies such as Alzheimer’s disease (AD). Further, the SP-DCA analysis showed that 5-HT2AR and 5-HTR4R heterodimerize through the C-terminal domain of 5-HT2AR and ICL3 of 5-HT4R. However, elucidating the details of GPCR interactions would accelerate the discovery of druggable sites and improve our knowledge of the etiology of common diseases, including AD.

DNA double-strand breaks (DSBs) can be detected by label-based sequencing or pulsed-field gel electrophoresis (PFGE). Sequencing yields population-average DSB frequencies genome-wide, while PFGE reveals percentages of broken chromosomes. We constructed a mathematical framework to combine advantages of both: high-resolution DSB locations and their population distribution. We also use sequencing read patterns to identify replication-induced DSBs and active replication origins. We describe changes in spatiotemporal replication program upon hydroxyurea-induced replication stress. We found that one-ended DSBs, resulting from collapsed replication forks, are population-representative, while majority of two-ended DSBs (79-100%) are not. To study replication fork collapse, we used strains lacking the checkpoint protein Mec1 and the endonuclease Mus81 and quantified that 19% and 13% of hydroxyurea-induced one-ended DSBs are Mec1-and Mus81-dependent, respectively. We also clarified that Mus81-induced one-ended DSBs are Mec1-dependent.

Intrinsically disordered regions (IDRs) and intrinsically disordered proteins (IDPs) play pivotal roles in cellular signaling, molecular recognition, and the regulation of various biological processes. These flexible and conformationally dynamic protein segments are difficult to study using structural analysis methods and computational approaches including AlphaFold. Therefore a critical challenge arises when attempting to understand the structural basis of protein-protein interactions involving IDRs. Here we demonstrate that the poorly characterized C16orf87 protein, which we rename as MHAP1, forms a stable complex with HDAC2 and MIER1. These three proteins all contain IDRs whose structure is unknown. We implemented an integrative approach combining experimental crosslinking data with computational modeling techniques (I-TASSER, HADDOCK, AlphaFold) to probe the IDR-driven assembly of the HDAC1:MIER2:MHAP1 complex and build an integrative structural model of this complex. The C-terminal domain of HDAC2, a poorly characterized IDR, promotes interactions between the ELM2 domain of MIER1 as well as the N- and C-termini of MHAP1. These results contrast with most current literature, including the results from AlphaFold alone that are missing structural information on HDAC C-domain. The approach herein can be generalized to study other complexes, emphasizing the need for integrative approaches in determining the 3D structures of IDR/IDP-driven complexes.

Multiple phenotypic protein expressions arise from one genome represent variations in the protein relative abundance and their stoichiometry. A lack of definite compositional parts challenges the modeling of protein megacomplexes and cellular architectures. Despite the advance in protein structural predictions with Artificial Intelligence (AI), the mechanism of protein interactions and the emergence of megacomplexes they assemble remains unclear. Here, we present a statistical-physics framework of grand canonical ensemble to explore the protein interactions that drive the emergent assembly of a megacomplex using the observational mass spectrometry datasets including protein relative abundance and the cross-linked connections. Using chromatin remodeler megacomplex, INO80, as an example, we discovered a class of “divergent” protein that plays a critical role in orchestrating the assembly beyond nearest neighbors, dependent on the excluded volumes exerted by others. With the constraints of the excluded volumes by varying the volume fraction of INO80 subunits, these divergent subunits orchestrate and form clusters with selective components growing into configurationally distinct architectures. We propose a machinery view for the INO80 chromatin remodeler complex where each loosely associated subunits can be occasionally recruited for parts as attachment into a core assembly driven by excluded volumes. Our computational framework provides a mechanistic insight into taking the volume fractions as necessary physicochemical variables representing cell states to remodel the configurations of protein megacomplexes with structurally loose modules.

Summary: Histone Deacetylase (HDAC) 1 and 2 are key enzymatic components in multiple large chromatin remodeling complexes including NuRD, SIN3, and CoREST. In addition, both HDAC 1 and 2 contain a large intrinsically disordered region (IDR) within their C-terminal domain (CTD). How HDAC1/2 assemble into these complexes and the structure of the CTD IDR remains poorly understood. Here, we used HDAC1/2 to isolate their protein interaction networks from cells and used crosslinking mass spectrometry (XL-MS) coupled with the Integrative Modeling Platform to build structural models of the NuRD, SIN3A, and CoREST complexes. Next, we implemented an AlphaFold-enabled XL-MS constrained modeling approach to investigate how HDAC1 could assemble into these complexes. We show that the CTD IDR of HDAC1 folds into alpha helices in these complexes. Finally, we built a complete integrative structural model of a NuRD subcomplex including the abundant HDAC1:MBD3:MTA1:GATAD2B:RBBP4 subunits, which included 6 IDRs. The approaches used herein are broadly applicable for the study of protein complexes and protein interaction networks that can provide important insights into IDRs.