Cross-links of proteins without annotated polynucleotide binding

The third group of proteins found cross-linked to RNA contains proteins without annotated RNA (or DNA) binding properties. Interestingly, this group contains exclusively metabolic enzymes (sum-marized in Table 3.9). Three of those enzymes, adenosylhomocysteinase, alcohol dehydrogenase, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) contain a Rossmann fold domain. Since this domain has been proposed to have RNA-binding properties (discussed in 4.1.5), the structure and function of these three proteins will be described in more detail.

(1) Adenosylhomocysteinase (Sah1p) peptide ECINIKPQVDR (position E320–R330) was found to cross-link via C321 (spectrum shown in Figure B.71). Sah1p hydrolyzes S-adenosyl-homocystein (AdoHcy) into homocystein and adenosine. AdoHcy is a competitive inhibitor of S-adenosyl-L-methionine-dependent methyl transferase reactions. As the latter are important for formation of the cap structure on viral mRNAs, Sah1p inhibition leads to antiviral activity. A structure of the human homolog S-adenosylhomocystein hydrolase is available and shown in Figure 3.27. Yeast and human protein share 70% sequence identity, and the cross-linked peptide is almost identical to the human sequence EKVNIKPQVDR (E320–R330). The only exception is the cross-linked C321, the human protein has a lysine residue at this position. The peptide lies within the NAD binding Rossmann fold domain but does not contact the cofactor directly (see Figure 3.27b)^[118].

(2) A cross-linked peptide of the sequence YSGVCHTDLHAWHGDWPLPVK was identified (see Figure B.72). This sequence can be found in two highly homologous alcohol dehydrogenases (around

3.4 Application of the automated data analysis workflow to complex systems 111

Table 3.9: Cross-linked proteins without annotated RNA-binding function

recommended name gene synonym protein feature

1 adenosylhomocysteinase SAH1 Sah1p Rossmann fold

2 alcohol dehydrogenase 1/3 ADH1/ADH3 ADH1/ADH3 Rossmann fold 3 glyceraldehyde-3-phosphate TDH2/TDH3 GAPDH 2/ Rossmann fold

dehydrogenase 2/3 GAPDH 3

4 enolase 1/2 ENO1/ENO2 Eno1p/Eno2p

5 inorganic pyrophosphatase IPP1 IPP1

6 peroxiredoxin TSA1 TSA1 Tsa1p

7 phosphoglycerate kinase PGK1 PGK

8 pyruvate kinase 1 CDC19 PK

Numbers are indicated at the corresponding paragraphs in the text.

(a) Overview of the tetrameric structure. (b) Zoom on the cross-linked peptide.

Figure 3.27: Structure of human S-adenosylhomocysteine hydrolyase in complex with nicotin-amide-adenine-dinucleotide (NAD, light green) and the inhibitor fluoroneplanocin A (dark green). In the tetrameric structure, each subunit binds one molecule NAD and inhibitor.

pdb 3NJ4^[118]

80% amino acid identity), namely, alcohol dehydrogenases 1 and 3 (Adh1p: Y40–K60, Adh3p: Y67–

K86). Alcohol dehydrogenases catalyze the oxidation of an alcohol to its corresponding aldehyde or ketone by NADH. In yeast fermentation, they additionally catalyze the reverse reaction, e.g. the reduction of acetaldehyde to ethanol. Adh1p is the main ADH expressed and active during anaerobic fermentation; Adh3p is a mitochondrial protein. A crystal structure for Adh1p is available (B.V.

Plapp, B.R. Savarimuthu, S. Ramswamy; PDB ID 2HCY; no related publication). In the tetramer (see Figure 3.28a), each subunit binds two zinc ions although only one is catalytically active. The cross-linked cysteine residue C44 (Adh3p: C71) lies in a catalytic pocket and coordinates zinc together with H67 and C154 (see Figure 3.28b). All these residues are conserved between Adh1p and Adh3p^[119]. The cross-linked peptide does not lie within the C-terminal Rossmann fold.

(3) Another pair of highly homologous metabolic enzymes, glyceraldehyde-3-phosphate dehydroge-nases (GAPDH) 2 or 3, are identified by the common peptide ETTYDEIKK (E250–K258)

cross-(a) Overview of the homotetrameric structure. (b) Zoom on one catalytic pocket.

Figure 3.28: Structure of yeast alcohol dehydrogenase 1 in complex with nicotinamide-8-iodo-adenine-dinucleotide (light green) and trifluoroethanol (dark green). The cross-linked peptide is shown in orange, and the cross-linked cysteine residue in red. Additional Zn-coordinating residues are in dark blue, with residues involved in NAD-binding in light blue.

(pdb 2HCY, B.V. Plapp, B.R. Savarimuthu, S. Ramaswamy, no related publication.)

linked to [AU –HPO₃] (spectrum in Figure B.74). GAPDH is involved in the sixth reaction of glyco-lysis, namely, the oxidation and phosphorylation of glyceraldehyde-3-phosphate to 1,3-bisphospho-glycerate by NAD⁺ and orthophosphate. It was shown that the 43 N-terminal amino acids of the human protein are sufficient to preserve RNA-binding activity of the respective GST fusion protein, but do not preserve the protein’s preference to AU-rich elements^[120]. A structure is available for GAPDH 3 (I. Garcia-Saez, F. Kozielski, D. Job, C. Boscheron; PDB ID 3PMY; no related publica-tion). The structure shows a homodimer, with each subunit containing a NAD binding Rossmann fold (see Figure 3.29a); however only tetramers are catalytically active. The cross-linked peptide can be found on the surface of the protein, with the peptide N-terminus in a loop region containing amino acids E250, T251, and T252 (see Figure 3.29b). It is proximal to the Rossmann fold of the respective other subunit. The N-terminal protein region homologous to the human protein fragment sufficient for RNA-binding is highlighted in yellow in Figure 3.29a. The cross-link identified could point to the protein region responsible for RNA binding selectivity.

In addition to GAPDH, three other glycolytic enzymes were found to cross-link to RNA, namely enolase, phosphoglycerate kinase, and pyruvate kinase. While these proteins are best known for their role in glycolysis, several other functions and interactions have been described (reviewed in

[121]).

(4) Both yeast enolases Eno1p and Eno2p are highly homologous (96% sequence identity). There-fore, the peptide IGLDCASSEFFK (I244–K255) found cross-linked to [U +152 –H₂O] via C248 (see spectrum in Figure B.73) could be of either or both enolases. Enolase partakes in the ninth step of the glycolytic pathway, catalyzing the dehydration of 2-phosphoglycerate (2PG) to phos-phoenolpyruvate (PEP).

3.4 Application of the automated data analysis workflow to complex systems 113

(a) Overview of the dimeric structure. (b) Zoom on the cross-linked peptide.

Figure 3.29: Structure of yeast glyceraldehyde-3-phosphate dehydrogenases (GAPDH) 3 in com-plex with meso-erythritol (dark green), sodium and NAD (light green). The 41 N-terminal amino acids connected with RNA-binding are shown in yellow. The cross-linked peptide is shown in orange.

(pdb 3PYM, I. Garcia-Saez, F. Kozielski, D. Job, C. Boscheron, no related publica-tion.)

(5) Yeast inorganic pyrophosphatase is a cytoplasmic protein and hydrolyzes phosphoanhydride. Its peptide NCFPHHGYIHNYGAFPQTWEDPNVSHPETK (N83–K112) was found cross-linked via residue C84 (see spectrum in B.75).

(6) Peroxiredoxin TSA1 is a cytoplasmic protein and belongs to a family of thiol-specific peroxidases.

Interestingly, C171, the cross-linked residue in the peptide NGTVLPCNWTPGAATIKPTVEDSK (N165–K188; spectrum shown in B.76), is the resolving cysteine in the catalytic reaction. After the peroxidatic cysteine (C47 in TSA1) is oxidized, it reacts with the resolving cysteine to form a disulfide bond which, in turn, is reduced by thioredoxin to complete the catalytic cycle^[122]. (7) Phosphoglycerate kinase (PGK) catalyzes the seventh reaction of glycolysis, transferring a phos-phate group from 1,3-bisphosphoglycerate (1,3-BPG) to ADP, yielding 3-phosphoglycerate (3PG) and ATP. PGK peptide YVLEHHPR (Y49–R56) was found cross-linked to RNA; the cross-linked amino acid residue could not be determined (spectrum shown in Figure B.77).

(8) Pyruvate kinase (PK) catalyzes the tenth and final step of glycolysis, conversion of phospho-enolpyruvate (PEP) to pyruvate, thereby transferring phosphate from PEP to ADP to yield ATP.

Two regions of pyruvate kinase were identified to interact with RNA: Peptide NCTPKPTSTTET-VAASAVAAVFEQK (N370–K394) was found cross-linked via C371 and peptide YRPNCPIILVTR (Y414–R425) via C418 (see spectra in Figures B.78 and B.79).