The Evolution of Chromatin Proteins and Prediction of Novel Factors in Chromatin Dynamics National Center for Biotechnology Information National Institutes of Health L. Aravind

Extraordinary diversity of eukaryotes “ crown group” Most studied “ microbial eukaryotes” Most diverse and prevalent animals fungi Slime molds plants Chlorophytes rhodophytes diatoms Heteroloboseans parbasalids Diplomonads Euglenozoa ciliates Apicomplexans

A lot of action around the conserved histones High conservation of core chromatin components: histones Non-linear expansion of transcription factors Increasing number and complexity of chromatin proteins

Extracting information from proteins at different levels ASH1: Histone methylase Application of bulk property analysis: e.g. Seg and Coils Globular domains are identified Insects Vertebrates Nematodes Inferred presence in the ancestral animal The phylogenetic tree suggests that though the protein was present in the last common ancestor of these animal-lineages, it underwent independent proliferation in the insect and vertebrate lineages Phyletic and phylogenetic analysis Sequence profile analysis of globular domains Structural relationship analysis Identification of active sites, binding pockets, catalytic mechanisms and other key features Understanding the evolutionary origins of domains

The phylogenetic tree suggests that though the protein was present in the last common ancestor of these animal-lineages, it underwent independent proliferation in the insect and vertebrate lineages

Identification of active sites, binding pockets, catalytic mechanisms and other key features

Understanding the evolutionary origins of domains

Early domain universe Peering back in time Evolutionary branching order Functional modifications Predicting novel functions The largest assemblage of homologous domains that can unified by sequence features is formally a superfamily Several superfamilies may share a common folding pattern and arrangement of secondary structure elements. An assemblage of structurally related homologous domains is termed a fold

Evolutionary branching order

Functional modifications

Predicting novel functions

Overview The provenance of the MORC ATPases and prediction of a specific role in chromatin dynamics Architectural evolution of chromatin proteins in eukaryotes: Organismal complexity and chromatin protein complexity Unusual functional linkages in non-model systems Architectural analysis and prediction of novel functions The NAD metabolite network in eukaryotic chromatin A conserved enzyme predicted to catalyze two modifications of chromatin proteins

Morc1- mutated in Microrchidia Required for completion of prophase I and localization of endonuclease Spo11 What is its role? Eukaryotic Morcs have emerged from bacterial Restriction- Modification components The MORC ATPases Contains a catalytically active GHKL ATPase domain- presence of lysine finger Specific homologs are found in bacterial type III restriction-modification systems Belongs to a vast radiation of bacterial GHKL ATPases associated with bacterial restriction modification systems and the MutL family of DNA repair proteins GHKL ATPase domain Active site

Morc1- mutated in Microrchidia

Required for completion of prophase I and localization of endonuclease Spo11

What is its role?

Contains a catalytically active GHKL ATPase domain- presence of lysine finger

Specific homologs are found in bacterial type III restriction-modification systems

Belongs to a vast radiation of bacterial GHKL ATPases associated with bacterial restriction modification systems and the MutL family of DNA repair proteins

Prediction of a specific role for Morcs in chromatin dynamics Eukaryotic Morcs fused to methylated and acetylated histone- or methylated DNA-binding domains Prokaryotic homologs mainly fused to diverse nuclease domains A model CH3 CH3 CH3 MORC Type-III RM system Epi Epi Epi Eukaryotic MORCs

Eukaryotic Morcs fused to methylated and acetylated histone- or methylated DNA-binding domains

Prokaryotic homologs mainly fused to diverse nuclease domains

Overview The provenance of the MORC ATPases and prediction of a specific role in chromatin dynamics Architectural evolution of chromatin proteins in eukaryotes: Organismal complexity and chromatin protein complexity Unusual functional linkages in non-model systems Architectural analysis and prediction of novel functions The NAD metabolite network in eukaryotic chromatin A conserved enzyme predicted to catalyze two modifications of chromatin proteins

Domain architectures of chromatin proteins One can represent protein domain architectures as an ordered graph Reveals several details about the behavior of particular domains Not all enzymatic domains operating on chromatin proteins are equally connected Most highly connected are the SNF2/SWI2 ATPases – recognize all kinds of histone modifications, DNA features and also potentially certain soluble NAD metabolites Both histone methylases and histone acetylases are highly connected but the corresponding demethylases are much more connected than the deacetylases – deacetylation is much less dependent on direct interactions with other modifications than methylation

One can represent protein domain architectures as an ordered graph

Reveals several details about the behavior of particular domains

Not all enzymatic domains operating on chromatin proteins are equally connected

Most highly connected are the SNF2/SWI2 ATPases – recognize all kinds of histone modifications, DNA features and also potentially certain soluble NAD metabolites

Both histone methylases and histone acetylases are highly connected but the corresponding demethylases are much more connected than the deacetylases – deacetylation is much less dependent on direct interactions with other modifications than methylation

Domain architectures of chromatin proteins and organizational complexity Several major histone modifying enzyme emerged early but show low diversity of fused modified peptide recognition domains -- little cross-talk between modifications The eukaryotic crown group (plants, amoebozoa, fungi, animals) have greater complexity in domain architectures Integration of numerous signals by chromatin proteins The breeding ground for multicellularity Evolution of domain architecture networks in the histone methylation-demethylation systems of eukaryotes

Several major histone modifying enzyme emerged early but show low diversity of fused modified peptide recognition domains -- little cross-talk between modifications

The eukaryotic crown group (plants, amoebozoa, fungi, animals) have greater complexity in domain architectures

Integration of numerous signals by chromatin proteins

The breeding ground for multicellularity

Analysis of architectures reveals novel lineage-specific connections Potential functional connections between methylation-based histone modification and ubiquitination is elaborated in the chromalveolate lineage Chromo Tudor Methylated histone- binding domains OTU-type DUB Ubiquitin C-terminal hydrolase JAB-type DUB RING finger Ubiquitin E3 ligase Ubiquitin-like domain Deubiquitinating enzymes BMB/PWWP Ciliates and oomycetes oomycetes Histone deubiquitination based on specific methylation marks? Ch Jab Otu UbCH Tu R Ubl Ch Ch Otu UbCH Ch Ubl Ch R Ch UbCH Tu Jab BM BM

Lineage-specific diversification of the NAD metabolite network in eukaryotes Macro PARP SIR2 SWI/SNF ATPase H Poly-ADP ribose polymerase SWI/SNF ATPase NAD-dependent Histone deacetylase ADP-ribose/derivatives Binding domain Histone domain NAD PARP SIR2 H Ac H N ADPr T Ac-ADPr T ADPr H Poly-ADPr NAD ? ? Potential MACRO domain target MACRO domain Active site H Macro PARP Macro SWI/SNF ATPase Macro SIR2 Macro macroH2A Some animals Filamentous fungi, Entamoeba , oomycetes Several eukaryotes Animals, ciliates, Naegleria Macro Plasmodium, ciliates

Prediction of a novel NAD metabolite binding domain in the DBC1 family DBC1 gene homozygously deleted in breast cancer and some other tumors CARP1(CCAR1) pro-apoptotic gene; mediator complex component; upregulates p21 (cyclin inhibitor) expression DBC1 is a negative regulator of SIRT1 and promotes hyperacetylation of targets like p53 Inactive version of the NUDIX domain that is likely to bind O-AcADPR or ADPR, i.e. SIRTuin/PARP metabolites Comparable to the calcium channel TRPM2, which contains a catalytically compromised NUDIX domain DBC1 family might serve as a nexus for integrating regulatory inputs from NAD metabolites, RNAs and perhaps calcium

DBC1 gene homozygously deleted in breast cancer and some other tumors

CARP1(CCAR1) pro-apoptotic gene; mediator complex component; upregulates p21 (cyclin inhibitor) expression

DBC1 is a negative regulator of SIRT1 and promotes hyperacetylation of targets like p53

Inactive version of the NUDIX domain that is likely to bind O-AcADPR or ADPR, i.e. SIRTuin/PARP metabolites

Comparable to the calcium channel TRPM2, which contains a catalytically compromised NUDIX domain

DBC1 family might serve as a nexus for integrating regulatory inputs from NAD metabolites, RNAs and perhaps calcium

Two-headed chromatin protein modifying enzymes? Chromatin modifying enzymes typically contain a single modifying catalytic domain, even if they might be fused to several peptide-recognition or DNA-binding domains An exception is the fusion of the SET methyltransferase domain with an amino-acid ligase domain predicted to be a polyglutamylase Ancestral eukaryote Kinetoplastids (e.g. Trypanosoma ) Protein-lysine Protein-lysine-CH 3 SET domain Methyltransferase Protein-glutamate A single enzyme modifying both lysines via methylation and glutamate via polyglutamylation appears to have been an ancestral feature of eukaryotes A second version independently emerged in kinetoplastids Protein-poly-glutamate Amino acid ligase X SET aa-ligase aa-ligase SET

Chromatin modifying enzymes typically contain a single modifying catalytic domain, even if they might be fused to several peptide-recognition or DNA-binding domains

An exception is the fusion of the SET methyltransferase domain with an amino-acid ligase domain predicted to be a polyglutamylase

A single enzyme modifying both lysines via methylation and glutamate via polyglutamylation appears to have been an ancestral feature of eukaryotes

A second version independently emerged in kinetoplastids

Concluding remarks Computational analysis suggests that are several interesting chromatin associated biochemistries still lying open for experimental investigation in model organisms … But Non-model eukaryotes, including some major pathogens, have several novel connections suggesting biochemistries beyond what we have learned from the models …

Acknowledgements Vivek Anantharaman S. Balaji Max Burroughs (RIKEN, Japan) Lakshminarayan Iyer M. Madan Babu (MRC-LMB, UK) Abhiman Saraswati Thiago Venancio Present group members at NCBI Former group members