News & Views | Published:


Unexpected role for vitamin B2

Nature volume 522, pages 427428 (25 June 2015) | Download Citation

An enzyme has been found that alters the molecular structure of vitamin B2, adding a fourth ring to its existing three-ring system. The product catalyses new types of chemistry in concert with certain other enzymes. See Letters p.497 & p.502

Many vitamins work by expanding the biochemical repertoire of enzyme-catalysed reactions. For example, riboflavin (vitamin B2) does this for a wide variety of flavoproteins, which catalyse many of the redox reactions in central metabolic pathways and act in the electron-transport chains of cells. Two papers in this issue, by White et al.1 (page 502) and Payne et al.2 (page 497), report a previously unknown cofactor derived from riboflavin. The findings solve the long-standing mystery of how a pair of bacterial enzymes — and their counterparts in yeast — catalyse crucial reactions known as decarboxylations, and further expand the repertoire of enzymatic reactions.

White et al. relate that the molecular structure of riboflavin, which contains three rings, can be modified by the addition of a prenyl group (a hydrocarbon group containing five carbon atoms, also known as an isoprenyl group) to form a fourth ring (Fig. 1). The authors call the resulting compound prenylated flavin mononucleotide (prenylated FMNH2). They used high-resolution crystal structures, biochemical studies, spectroscopy and computational calculations to characterize prenylated FMNH2 and to determine how it is employed as a cofactor in decarboxylation reactions — in which a carboxylate group (CO2) is removed in the form of carbon dioxide from a substrate.

Figure 1: Prenylated riboflavin.
Figure 1

Flavin mononucleotide (FMNH2) is a common enzyme cofactor that is derived from riboflavin (vitamin B2). White et al.1 show that the bacterial UbiX enzyme and the fungal Pad1 enzyme modify FMNH2 in the presence of dimethylallyl monophosphate (DMAP), a precursor of prenyl groups, to generate a previously unknown cofactor, prenylated FMNH2. Payne et al.2 show that prenylated FMNH2 is an essential cofactor for the decarboxylase enzymes UbiD and Fdc1; apo-forms of UbiD and Fdc1 lack a bound cofactor and so are inactive. Me, methyl group.

Surprisingly, the enzyme that forms prenylated FMNH2 is UbiX, a protein first found in the bacterium Escherichia coli. The authors show that UbiX works in tandem with another enzyme, called UbiD, to mediate the decarboxylation of an intermediate in the biosynthesis of coenzyme Q (see Fig. 1a of the paper1). Coenzyme Q is an essential lipid in the electron-transport chain and a potent cellular antioxidant3.

How exactly UbiX and UbiD cooperate to perform the decarboxylation reaction has been quite puzzling. Both enzymes are present in a wide array of prokaryotes (bacterial and archaeal microorganisms), and many other microbes, including fungi, rely on homologous enzymes — proteins that share ancestry with UbiD and UbiX — to decarboxylate various acids. White et al. show that prenylated FMNH2 is first oxidized by Fdc1, a fungal homologue of UbiD, and then used to decarboxylate cinnamic acid (see Fig. 1a of the paper1).

So how does prenylated FMNH2 form? White et al. reveal that when UbiX is supplied with dimethylallyl monophosphate (a precursor of prenyl groups), it prenylates riboflavin's three-ring system. This results in the formation of a fourth ring to generate prenylated FMNH2 (Fig. 1). The authors show that this previously unknown cofactor enables Fdc1 to decarboxylate substrates in vitro, in the absence of UbiX. These results explain the findings of other studies4,5 that showed that an unknown small molecule — but not UbiX itself — is required to activate the ability of UbiD (and Fdc1) to decarboxylate substrates.

Payne et al.2 show that apo-Fdc1 (an apoenzyme is one without its bound cofactor) catalyses unusual biochemistry only when in the presence of prenylated FMNH2. They find that Fdc1 first oxidizes prenylated FMNH2 to generate an 'iminium' form of the cofactor. This activates and prepares Fdc1 to decarboxylate a wide assortment of substrates known as α,β-unsaturated aromatic carboxylic acids. Many of these derive from the microbial breakdown of lignin, the structural component of the secondary cell walls of plants6. The authors propose a mechanism for these decarboxylation reactions known as 1,3-dipolar cycloaddition (see Fig. 4d of the paper2). Although this mechanism is well known to organic chemists, its use by enzymes has until now been speculative7.

The findings raise several questions about the biological use of prenylated riboflavin cofactors. First, there is no obvious amino-acid sequence for binding prenylated FMNH2: UbiX and UbiD (or Fdc1) do not share similar amino-acid sequences or structures, and the interactions of these proteins with the prenyl group seem to be restricted to providing an appropriately shaped binding site. This complicates the identification of other enzymes that might make use of this cofactor.

It is also surprising that UbiX uses dimethylallyl monophosphate as a source of prenyl groups, rather than the commonly used source dimethylallyl diphosphate. It will be important to determine the metabolic origin of dimethylallyl monophosphate and to explore whether it is used in other enzyme reactions. Dimethylallyl diphosphate is generated either through a series of biochemical reactions known as the mevalonate pathway, or through the methylerythritol phosphate pathway. In both pathways, dimethylallyl diphosphate is reversibly converted to another compound, isopentenyl diphosphate. A third, related pathway has recently been characterized8 in the archaeon Thermoplasma acidophilum. This alternative route directly produces isopentenyl monophosphate — could it be that this compound interconverts with dimethylallyl monophosphate, in the same way that dimethylallyl diphosphate and isopentenyl diphosphate interconvert?

Can Pad1 (the fungal homologue of UbiX) and Fdc1 also mediate decarboxylation steps in the biosynthesis of coenzyme Q? UbiX from E. coli and Pad1 from yeast have been shown to perform the same function: a mutant form of E. coli in which the ubiX gene is deleted cannot synthesize coenzyme Q, but the synthesis is restored9 if the mutant is engineered to express yeast Pad1. The two current papers now identify both Pad1 and UbiX as FMN prenyltransferases, enzymes that synthesize prenylated FMNH2 as a diffusible small molecule.

But what about Fdc1 and UbiD? Do these enzymes both recognize 3-polyprenyl-4-hydroxybenzoic acid (the substrate that is decarboxylated in the biosynthesis of coenzyme Q)? Unfortunately, 3-polyprenyl-4-hydroxybenzoic acid is not commercially available, and so the authors of the current papers could not perform direct assays of Fdc1 or UbiD with this substrate. But the synthesis of coenzyme Q is not impaired when the genes that express Pad1 and Fdc1 are both deleted from two species of yeast9,10. The enzyme responsible for the decarboxylation step of coenzyme Q biosynthesis in eukaryotes (organisms that include fungi, plants and animals) therefore remains an open question.



  1. 1.

    et al. Nature 522, 502–506 (2015).

  2. 2.

    et al. Nature 522, 497–501 (2015).

  3. 3.

    Mitochondrion 7, S2–S7 (2007).

  4. 4.

    , & Biochim. Biophys. Acta 436, 800–810 (1976).

  5. 5.

    , , , & ACS Chem. Biol. 10, 1137–1144 (2015).

  6. 6.

    & New Phytol. 187, 273–285 (2010).

  7. 7.

    , & J. Am. Chem. Soc. 135, 17638–17642 (2013).

  8. 8.

    , , & Biochemistry 53, 4161–4168 (2015).

  9. 9.

    , , , & Arch. Biochem. Biophys. 467, 144–153 (2007).

  10. 10.

    , , , & J. Biosci. Bioeng. 109, 564–569 (2010).

Download references

Author information


  1. Catherine F. Clarke and Christopher M. Allan are in the Department of Chemistry and Biochemistry, University of California, Los Angeles, Los Angeles, California 90095, USA.

    • Catherine F. Clarke
    •  & Christopher M. Allan


  1. Search for Catherine F. Clarke in:

  2. Search for Christopher M. Allan in:

Corresponding author

Correspondence to Catherine F. Clarke.

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing