Differential prolyl hydroxylation by six Physcomitrella prolyl-4 hydroxylases

The hydroxylation of proline residues to 4-trans-hydroxyproline (Hyp) is a common post-translational protein modification in plants, mediated by prolyl 4-hydroxylases (P4Hs). Hyps predominantly occur in a group of cell wall proteins, the Hydroxyproline-rich glycoproteins (HRGPs), where they are frequently O-glycosylated. While prolyl-hydroxylation and O-glycosylation are important, e.g. for cell wall stability, they are undesired on plant-made pharmaceuticals. Sequence motifs recognized for prolyl-hydroxylation derived from vascular plants were proposed but did not include data from mosses, such as Physcomitrella. Here, a phylogenetic reconstruction of plant P4Hs identified six P4Hs in four subfamilies in mosses. We analysed the amino acid sequences and structural environments around Hyps in Physcomitrella utilizing 73 Hyp sites in 24 secretory proteins from multiple MS/MS datasets, and found that prolines in close proximity to other prolines, alanine, serine, threonine and valine were preferentially hydroxylated. About 95% of the Hyp sites were predictable with a combination of previously defined motifs and methods. In our data, AOV (Ala-Hyp-Val) was the most frequent prolyl-hydroxylation pattern. Additionally, short arabinose chains were attached to Hyps in two cell-wall pectinesterases. A combination of 443 AlphaFold structure models and our MS data of peptides with nearly 3000 proline sites found Hyps predominantly on protein surfaces in disordered regions. Moss-produced human erythropoietin (EPO) exhibited plant-specific O-glycosylation with arabinose chains on two Hyps. This modification was significantly reduced in a P4H1 single knock-out (KO) Physcomitrella mutant. Quantitative proteomics after isotope labelling with different P4H-KO moss mutants revealed specific changes in the amount of proteins, including HRGPs, and modified prolyl-hydroxylation pattern from the mutants, suggesting a differential function of the six Physcomitrella P4Hs. Quantitative RT-PCR proved a differential effect of single P4H KOs on the expression of the other five p4h genes, suggesting a partial compensation of the mutation. AlphaFold-Multimer models for Physcomitrella P4H1 and its target EPO peptide superposed with the crystal structure of Chlamydomonas P4H1 and a peptide substrate suggested significant amino acids in the active centre of the enzyme that form H-bonds with the peptide substrate, and revealed differences between P4H1 and the other five Physcomitrella P4Hs. HighlightsMosses, such as Physcomitrella, encode six prolyl 4-hydroxylases in four subfamilies Computational analysis reveals specific environment for proline hydroxylation AlphaFold structure models and mass spectrometry reveal features of modification Quantitative proteomics and mutant analysis identify relevant modifying enzyme Modelling suggests relevant amino acids in active centre of the modifying enzyme