A Fifth of the Protein World: Rossmann-like Proteins as an Evolutionarily Successful Structural unit

Abstract

The Rossmann-like fold is the most prevalent and diversified doubly-wound superfold of ancient evolutionary origin. Rossmann-like domains are present in a variety of metabolic enzymes and are capable of binding diverse ligands. Discerning evolutionary relationships among these domains is challenging because of their diverse functions and ancient origin. We defined a minimal Rossmann-like structural motif (RLM), identified RLM-containing domains among known 3D structures (20%) and classified them according to their homologous relationships. New classifications were incorporated into our Evolutionary Classification of protein Domains (ECOD) database. We defined 156 homology groups (H-groups), which were further clustered into 123 possible homology groups (X-groups). Our analysis revealed that RLM-containing proteins constitute approximately 15% of the human proteome. We found that disease-causing mutations are more frequent within RLM domains than within non-RLM domains of these proteins, highlighting the importance of RLM-containing proteins for human health.

Keywords: Rossmann-fold; domains classification; minimal Rossmann-like motif; protein evolution.

Fig 1.. Minimal Rossmann-like motif (RLM) definition.

(A) RLM SSEs adapted from 5-formly-3-hydroxy-2-methylpyridine 4-carboxylic acid (FHMPC) 5-dehydrogenase (PDB: 4om8) are numbered and colored in rainbow, with a magenta loop (usually catalytic) between the first β-strand - element I (β1) and the first α-helix - element II (α1). The second α-helix - element IV (α2) forms crossover between the second β-strand - element III (β2) and the third β-strand - element V (β3). The crossover loop is a loop at the N-terminal part of α2. Element IV can be α-helix, β-strand, or loop. The unlabeled SSEs (colored in slate) are considered as an insertion to the RLM, which can occur between element III (β2) and element IV (α2) or in any of the loops connecting the RLM SSEs. (B) An interaction matrix defines RLM search strategy using ProSMoS program [34]. Interaction type “T” considers the angle between vectors corresponding to particular RLM elements. (C) RLM 2D topology diagram. All colors correspond to the panel (A).

Fig 2.

Distributions of (A) RLM X-groups sizes, for all X-groups which contain more than two F-groups; (B) relative contact order (CO) values for RLM-containing domains (red), Ig-like domains (green) and OB-like domains (blue).

Fig 3.

Distributions of average score for representative domains (comparison inside and outside of all homology groups) show a need for functional consideration and manual curation.

Fig 4.. Example of distant homology that can be detected by active site similarity.

RLM SSEs are colored in rainbow with active site residues shown in magenta. (A) DNA-adenine methyltransferase (PDB: 1yf3) binds SAH. (B) 12-hydroxydehydrogenase /15-Oxo-prostaglandin 13-reductase (PDB: 1v3v) binds NADP. (C) Caffeoyl-CoA O-methyltransferase (PDB: 1sus) binds SAH. Sequence homology probability was calculated using HHpred [49], structure similarity score was calculated using DALI [59]. (D) HHpred sequence alignment of caffeoyl-CoA O-methyltransferase (PDB: 1sus) and 12-hydroxydehydrogenase / 15-Oxo-prostaglandin 13-reductase (PDB: 1v3v). Gly residues from Gly-rich loop are colored by magenta. (E) HHpred sequence alignment of caffeoyl-CoA O-methyltransferase (PDB: 1sus) and DNA-adenine methyltransferase (PDB: 1yf3). Gly residues from Gly-rich loop are colored by magenta.

Fig 5.. POR family domains.

(A) Domain III of pyruvate:ferredoxin oxidoreductase (PDB: 6ciq) with CoA bound. Gly residues from Gly-rich loop colored by magenta. (B) Binding site superposition of pyruvate:ferredoxin oxidoreductase’s domain III (PDB: 6ciq, binds CoA), shown in beige, and UDP-N-acetylmuramoyl-L-alanine-D-glutamate ligase (PDB: 2jfg, binds ATP), shown in slate green. Adenine rings of ATP and CoA are colored in salmon, ribose rings in olive. Walker A motif is shown by magenta. (C) Sequence alignment of pyruvate:ferredoxin oxidoreductase’s domain III (PDB: 6ciq) and UDP-N-acetylmuramoyl-L-alanine-D-glutamate ligase (PDB: 2jfg). Active site residues shown in magenta. (D) Oxalate oxidoreductase (OOR) subunit delta (PDB: 5c4i). “Plug loop” is colored by magenta. (A-C) Presumed RLM elements are colored by rainbow, and the antiparallel β-strands inserted between β1 and β2 of presumed RLM are colored by yellow.

Fig 6.. Homologous lipase domains may lack the RLM.

The catalytic triad is shown in magenta sticks. The RLM is colored in rainbow (blue, cyan, green, orange, red). The lid is colored yellow. Diundecyl phosphatidylcholine (PDB: PLC) is shown in sticks and colored by elements. (A) Lipase_3 family (PF01764) member Thermomyces lanuginose lipase (PDB: 1einA), which lacks an RLM, is colored light orange. β-strands that correspond to (B) Lipase family (PF00151) human lipase (PDB: 1lpaB) are colored in corresponding colors. Active site of (C) T.lanuginose lipase (PDB: 1einA) and (D) human lipase (PDB: 1lpaB). (E) Dali structure-based sequence alignment of T.lanuginose lipase (PDB: 1einA) and the human lipase (PDB: 1lpaB). Catalytic triad shown by magenta. Secondary structure elements are shown by the same colors as in (A-B).

Fig 7.. Transition between 3(10)- and α-helices results in bending of α-helical part.

(A) Hydrogenase maturating endopeptidase from E.coli (PDB: 1cfz). RLM is shown in rainbow. 3(10)-helix is colored in yellow. Metal-binding residues are shown as sticks. Cd²⁺ is shown as sphere. (B) Main chain sticks representation of RLM elements β1 (C-atoms are colored in blue) and α1 (3(10)-helix part is colored in yellow, α-helix (alpha) part – in cyan) of HYBD. Hydrogen bonds are shown as orange dashed lines.

Fig 8.. Probable evolutionary path between RLM-containing domains from Peptidyl-tRNA hydrolase-like homology group

(A) Hydrogenase maturating endopeptidase from E.coli (PDB: 1cfz); (B) archaeal proteasome activator from Pyrococcus furiosus (PDB: 3vr0); (C) PH0006 protein from Pyrococcus horikoshii (PDB: 2gfq); (D) purine nucleoside phosphorylase from E.coli (PDB: 4ts3), formycin A is shown by sticks and colored by elements. RLM is shown in rainbow. 3(10)-helix is colored in yellow in A and B. Metal-binding residues are shown by sticks in A, B. Residues near metal-binding site are shown by sticks in C.

Fig 9.. Trimer tandem duplication of RLM-containing domains in distant homologs

Individual RLMs from (A) B.thetaiotaomicron glycoside hydrolase N-terminal domains (PDB: 3sgg) and (B) C.difficile Cwp8 RLM trimer (PDB: 5j6q) are colored blue, cyan and orange from the N- to the C-terminus. Conserved residues are shown in red stick. Insertions in Cwp8 are colored magenta.

Fig 10.. Taxonomic distribution of RLM protein across family and homology groups.

A - Archaea, B - Bacteria, E – Eukaryotes, V – Viruses. Percentage shows the ratio of particular taxonomic groups combination from total number of family or homologous groups respectively. The thickness of lines represents the number of RLM F-groups.

Fig 11.. Distribution of DCM fraction within RLM and non-RLM domains.

DCM fraction is defined as ratio of number of residues linked with DCM to the total number of residues in the chain.

Fig 12.. Disease-causing mutations within RLM-containing domains.

(A) Aromatic-L-amino-acid decarboxylase (PDB: 3rbf). Positions of the mutations associated with aromatic L-amino-acid decarboxylase deficiency are colored in magenta. (B) BRCT domains of Breast cancer type 1 (BRCA1) susceptibility protein (PDB: 4y2g). Positions of mutations associated with ovarian and breast cancer are colored in magenta. (C) Calcium-sensing receptor (CaSR) domains of G-protein-coupled receptor (GPCR) (PDB: 5k5s). Positions of mutations associated with hypocalciuric hypercalcemia, hyperparathyroidism, hypocalcemia and epilepsy are colored by magenta. (D) Nucleotide-binding domain of Ras protein (PDB: 4dsn). Positions of mutations associated with Noonan syndrome are colored by magenta. (A-D) RLMs are colored by rainbow.