Protein tyrosine sulfation: A critical posttranslation modification in plants and animals

Protein tyrosine sulfation is a posttranslational modification restricted to proteins that transit the secretory pathway that was first described by Bettelheim in bovine fibrinopeptide B in 1954 (1). Subsequent pioneering work by Wieland Huttner's group (2, 3) and others characterized the enzyme activity responsible for the reaction, called tyrosylprotein sulfotransferase (TPST), that catalyzes the transfer of sulfate from 3′-phosphoadenosine 5′-phosphosulfate to the hydroxyl group of peptidyl tyrosine residues to form a tyrosine O⁴-sulfate ester. The enzymes' subcellular localization in the trans-Golgi network and its widespread tissue and cellular distribution have been well documented in animals, and several dozen tyrosine-sulfated proteins, mostly of animal origin, have been described, many of which play important roles in inflammation, hemostasis, immunity, and other processes (2–4). Finally, the general importance of protein tyrosine sulfation in protein–protein interactions has become widely accepted.

In 1998, TPST-1 was purified from rat liver, and mouse and human TPST-1 cDNAs were identified (5). A second isoenzyme called TPST-2 was reported later that year (6, 7). Importantly, TPST orthologs from numerous vertebrate and invertebrate species are readily identifiable based on sequence homology with the mouse or human enzymes. However, despite the presence of tyrosine-sulfated peptides in plants first reported in 1996 (8) and the sequencing of several plant genomes, no plant TPST orthologs have been identified. These observations strongly indicated that TPSTs in plants have very limited sequence homology with those in the animal kingdom and identification of plant TPSTs would require some heavy lifting.

The heavy lifting has been provided by Komori et al. and is documented in a article published in this issue ofPNAS (9). They performed some elegant, old-fashioned protein purification work and accomplished a ≈ 9,100-fold enrichment of TPST activity by affinity chromatography on a peptide substrate column followed by chromatography on hydroxyapatite. The putative TPST was then identified by photoaffinity cross-linking as a 62-kDa polypeptide, and internal peptide sequence was obtained by in-gel tryptic digestion and mass spectrometry from which a full-length cDNA was isolated that encodes a 500-residue polypeptide precursor (Fig. 1).

Fig. 1.

Domain structure of TPSTs. Human TPST-1 and TPST-2 are type II transmembrane proteins of similar size (370–377 residues). Each has a short N-terminal cytoplasmic domain, a single transmembrane domain (TM), and a putative ≈ 40-residue stem region, followed by a luminal catalytic domain. Arabodopsis TPST is a 500-residue type I transmembrane proteins with a 24-residue N-terminal signal peptide (SP), a 446-residue luminal catalytic domain, followed by a single TM and a short cytoplasmic domain. Potential N-glycosylation sites are depicted as lancets. HS6ST2, heparan sulfate 6-O-sulfotransferase 2.

The work by Komori et al. (9) identifies several significant differences between the TPST enzyme system in Arabidopsis (and likely other plant species) and animals. The first notable difference is that the Arabidopsis TPST is a type I transmembrane protein, in contrast to all TPSTs of animal origin that have type II transmembrane topology (4). The precursor includes an N-terminal 24-residue signal peptide and the mature polypeptide is predicted to have a 446-residue luminal catalytic domain with six potential N-glycosylation sites followed by a single transmembrane helix and a short cytoplasmic domain. The second difference is that the Arabidopsis TPST has no significant sequence homology with TPSTs of animal origin, which includes the lack of recognizable 5′-PSB and 3′PB motifs. These two motifs are highly conserved in all known cytosolic and membrane sulfotransferases and are involved in binding of the 5′ and 3′ phosphate groups of the reaction product 3′, 5′-ADP, respectively (10). The only significant homology of note is that residues 371–447 of Arabidopsis TPST align with a region near the C terminus of heparan sulfate 6-O-sulfotransferase 2. In light of the major difference in the primary structure of TPSTs in Arabidopsis and animal species, it will be interesting to assess whether the tertiary structures and catalytic mechanisms of the enzyme in plants and animals are similar. A third difference of note is that only a single TPST gene is apparent in the Arabidopsis genome and other plant species. In contrast, animal genomes have two TPST genes, with the exception of Drosophila melanogaster. One can only speculate on the reason for this. Perhaps animal species have far more TPST substrates than plants and thus evolved more than one enzyme to effectively sulfate all of them. However, it is still formally possible that Arabidopsis has a second distantly related Tpst gene. It would be of interest to directly address this question by determining whether the loss-of-function mutant of the Arabidopsis TPST lacks the ability to synthesize protein-tyrosine sulfate.

Komori et al. (9) also show that the Arabidopsis TPST enzyme system shares several general features with animals. As in the mouse and human systems, the Arabidopsis enzyme is widely expressed in plant tissues and is localized in the Golgi. Likewise, the sites of tyrosine sulfation in plant proteins are highly acidic in nature like most tyrosine-sulfation sites in animal proteins. However, at present, PSK and PSY1 are the only tyrosine-sulfated proteins that have been described in Arabidopsis. In addition, a loss-of-function mutant of the Arabidopsis TPST displayed a markedly abnormal phenotype that includes severely stunted growth, early senescence, and other abnormalities. These findings indicate that protein-tyrosine sulfation is of crucial importance to normal growth and development in plants. The overall severity of the phenotype is similar to that observed in Tpst1;Tpst2 double knockout mice that have severely impaired postnatal survival (11). Most importantly, the phenotype of the Arabidopsis TPST loss-of-function mutant cannot be fully explained based on the current understanding of the TPST substrate repertoire, as is the case for the phenotypes observed in mice with deficiencies of Tpst1 and/or Tpst2. Thus, it is almost certain that many TPST substrates await description in both plants and animals. Komori et al. deserve congratulations for their important contribution to the field, but much work remains to be done.