Covalent modification of histones has an important role in regulating chromatin dynamics and transcription. Whereas most covalent histone modifications are reversible, until recently it was unknown whether methyl groups could be actively removed from histones. Using a biochemical assay coupled with chromatography, we have purified a novel JmjC domain-containing protein, JHDM1 (JmjC domain-containing histone demethylase 1), that specifically demethylates histone H3 at lysine 36 (H3-K36). In the presence of Fe(ii) and alpha-ketoglutarate, JHDM1 demethylates H3-methyl-K36 and generates formaldehyde and succinate. Overexpression of JHDM1 reduced the level of dimethyl-H3-K36 (H3K36me2) in vivo. The demethylase activity of the JmjC domain-containing proteins is conserved, as a JHDM1 homologue in Saccharomyces cerevisiae also has H3-K36 demethylase activity. Thus, we identify the JmjC domain as a novel demethylase signature motif and uncover a protein demethylation mechanism that is conserved from yeast to human.

We have previously shown functional differences in fibrinogen from hyperhomocysteinemic rabbits compared to that in control rabbits. This acquired dysfibrinogenemia is characterized by fibrin clots that are composed of abnormally thin, tightly packed fibers with increased resistance to fibrinolysis. Homocysteine thiolactone is a metabolite of homocysteine (Hcys) that can react with primary amines. Recent evidence suggests that Hcys thiolactone-lysine adducts form in vivo. We now demonstrate that the reaction of Hcys thiolactone with purified fibrinogen in vitro produces fibrinogen (Hcys fibrinogen) with functional properties that are strikingly similar to those we have observed in homocysteinemic rabbits. Fibrinogen purified from homocysteinemic rabbits and Hcys fibrinogen are similar in that (1) they both form clots composed of thinner, more tightly packed fibers than their respective control rabbit and human fibrinogens; (2) the clot structure could be made to be more like the control fibrinogens by increased calcium; and (3) they both form clots that are more resistant to fibrinolysis than those formed by the control fibrinogens. Further characterization of human fibrinogens showed that Hcys fibrin had similar plasminogen binding to that of the control and an increased capacity for binding tPA. However, tPA activation of plasminogen on Hcys fibrin was slower than that of the control. Mass spectrometric analysis of Hcys fibrinogen revealed twelve lysines that were homocysteinylated. Several of these are close to tPA and plasminogen binding sites. Lysines are major binding sites for fibrinolytic enzymes and are also sites of plasmin cleavage. Thus, modification of lysines in fibrinogen could plausibly lead to impaired fibrinolysis. We hypothesize that the modification of lysine by Hcys thiolactone might occur in vivo, lead to abnormal resistance of clots to lysis, and thereby contribute to the prothrombotic state associated with homocysteinemia.