Many enzymes require a non-protein cofactor for activity, and for several such proteins the cofactor-bound (holo) three-dimensional structure is known while the cofactor-free (apo) conformational ensemble is not, leaving open both what the apo ensemble looks like at atomic resolution and by what kinetic pathway cofactor binding converts it into the holo state. The open objective concerns two systems — the prFMN-dependent decarboxylase Fdc1 from Aspergillus niger (studied by mass spectrometry) and human D-amino acid oxidase (studied by flavin-binding kinetics) — whose atomistic apo conformational ensembles and apo-to-holo cofactor-binding kinetic pathways remain undetermined. The broader question of general principles by which unbound ensembles encode cofactor responsiveness is outside the scope of this question.

Background

A large class of enzymes is active only when bound to a non-protein cofactor: a small organic molecule or metal ion that the surrounding protein cannot synthesize itself. The cofactor-bound protein is called the holoenzyme (holo form) and the cofactor-free protein the apoenzyme (apo form). Crystallography has delivered holo structures for many cofactor-dependent enzymes, but apo structures are far scarcer — a 2024 census of the Protein Data Bank (PDB, the international archive of solved biomolecular structures) called AHoJ-DB found that fewer than half of all catalogued binding sites have any apo form deposited — because apo proteins are often conformationally heterogeneous and resist crystallization. Understanding what the apo protein looks like, and how it responds to its cofactor, is the physical basis of enzyme regulation by cofactor availability.

The general mechanistic question embedded here is the classical distinction between conformational selection and induced fit: either the apo ensemble already transiently populates a cofactor-competent conformation that binding captures (selection), or binding itself drives the protein into the bound conformation (induced fit); real systems commonly mix both, and deciding between them requires kinetic measurements of the binding pathway, not just endpoint structures. The three source systems illustrate the problem.

Ferulic acid decarboxylase Fdc1 is the founding member of the UbiD family of (de)carboxylases and requires the prenylated flavin mononucleotide (prFMN) cofactor, a modified flavin mononucleotide carrying an additional prenyl-derived ring system that is synthesized by the prenyltransferase UbiX. The holo structure of the Aspergillus niger enzyme has been known crystallographically since 2015, but no structure of the cofactor-free A. niger enzyme exists: every deposited A. niger Fdc1 entry in the PDB contains prFMN or a related flavin, and Beveridge and co-workers (2016) report that no crystals were attainable from recombinant A. niger Fdc1 produced without UbiX co-expression, which they attribute to the increased conformational heterogeneity of the apo protein. They therefore compared apo and holo Fdc1 by mass-spectrometric methods: ion mobility–mass spectrometry (IM–MS, which separates gas-phase protein ions by their rotationally averaged collision cross section and thus reports compactness) and hydrogen–deuterium exchange MS (HDX–MS, which maps local solvent accessibility and backbone dynamics by measuring how fast backbone amide hydrogens exchange with deuterated solvent). They found the apo form more open, more dynamic, and less compact than the holo dimer — an ensemble-level, not atomic-resolution, characterization that left both the atomistic apo structure and the kinetic pathway of prFMN binding undetermined. Two cofactor-free crystal structures of Fdc1 from a different species, Saccharomyces cerevisiae, do exist — a wild-type structure with the reaction product 4-vinylphenol bound, and a 2025 open-state structure of an engineered quadruple mutant — but neither characterizes the apo ensemble's dynamics or the binding pathway, and neither is the A. niger enzyme of the source study.

Human D-amino acid oxidase (hDAAO) is a flavoprotein depending on FAD (flavin adenine dinucleotide) that degrades D-serine, a co-agonist of NMDA-type glutamate receptors in the brain, and is a drug target in schizophrenia and D-serine-related disorders. Its holo crystal structure has been known since 2006, but hDAAO binds FAD unusually weakly for a flavoprotein (dissociation constant around 8 micromolar), and Caldinelli and co-workers (2009) noted that no crystals of the apoprotein form had been obtained; using fluorescence, circular dichroism, and limited proteolysis they found the apo form more open and more exposed than the holoenzyme, and showed that reconstitution is sequential: FAD first binds rapidly to a catalytically active intermediate (termed holo*), which then slowly converts — on a timescale of more than an hour at 15 °C — into the final holoenzyme conformation. The structural identity of the apo conformations and of this slow isomerization step remains unknown; later fluorescence-titration evidence indicates that the apoprotein itself exists in two conformations with different FAD affinity.

The bacterial DNA cytosine methyltransferase M.HhaI, which transfers a methyl group from S-adenosylmethionine (AdoMet) to a target cytosine it flips out of the DNA helix, provided the third, grounding source: Zhou and co-workers (2007) used transverse relaxation-optimized spectroscopy NMR (TROSY, a multidimensional NMR technique that sharpens spectra of large proteins) to show that cofactor binding causes structural and dynamical changes extending more than 30 angstroms from the cofactor site — an allosteric response present in the cofactor-bound form. An apo crystal structure of M.HhaI was subsequently determined in 2017.

Across these and related systems, the shared scientific objective is to move from endpoint holo structures and low-resolution ensemble fingerprints to (i) atomic-resolution structural and dynamical descriptions of apo ensembles and (ii) resolved kinetic pathways — intermediates, rate constants, and the selection-versus-induced-fit decomposition — of the apo-to-holo transition, and ultimately to general principles of how an unbound ensemble encodes responsiveness to its cofactor. The field approaches the general question as an active research program rather than a single open problem: PDB-wide apo/holo databases catalogue the coverage gap, AlphaFold2 (a deep-learning method for predicting protein structures) has been adapted to predict apo and holo conformational ensembles, and on the cofactor-loading side a prFMNH2-binding chaperone that assists activation of apo UbiD-family enzymes in vitro was reported in 2024 — but no general answer exists, and for the two systems named above not even the system-level apo ensembles and binding pathways are known.

Problem Statement

For cofactor-dependent enzymes whose holo structure is known but whose cofactor-free state is undetermined at atomic resolution — concretely, the prFMN-dependent decarboxylase Fdc1 from Aspergillus niger and human D-amino acid oxidase — determine, for each system, the structural and dynamical properties of the apo conformational ensemble at atomic resolution and, for human D-amino acid oxidase in addition, the kinetic pathway by which FAD binding converts the apo ensemble into the holo state, specifying the structural identity of the rapidly formed catalytically active holo* intermediate and of the slow holo*-to-holo isomerization, and the extent to which binding proceeds by conformational selection (apo pre-population of a cofactor-competent conformation) versus induced fit (binding-driven refolding). The two systems are independent open problems: a characterization of either system stands on its own. For A. niger Fdc1, the kinetic pathway of prFMN binding is likewise experimentally uncharacterized and is a related open problem, though it is not part of this question. The broader form of the question asks what general features of unbound conformational ensembles encode responsiveness to a protein's cofactor; that general form is the motivating question but is not part of this question. The third source system, the DNA methyltransferase M.HhaI, has had its apo crystal structure determined since 2017 (PDB 5LOD), so it contributes no open instance.

References

  1. Beveridge R, Migas LG, Payne KAP, Scrutton NS, Leys D, Barran PE. 'Mass spectrometry locates local and allosteric conformational changes that occur on cofactor binding.' Nature Communications 7 (2016) 12163. DOI: 10.1038/ncomms12163. https://doi.org/10.1038/ncomms12163
  2. Caldinelli L, Molla G, Sacchi S, Pilone MS, Pollegioni L. 'Relevance of weak flavin binding in human D-amino acid oxidase.' Protein Science 18 (2009) 801-810. DOI: 10.1002/pro.86. https://doi.org/10.1002/pro.86
  3. Zhou H, Shatz W, Purdy MM, Fera N, Dahlquist FW, Reich NO. 'Long-Range Structural and Dynamical Changes Induced by Cofactor Binding in DNA Methyltransferase M.HhaI.' Biochemistry 46 (2007) 7261-7268. DOI: 10.1021/bi602662e. https://doi.org/10.1021/bi602662e
  4. Payne KAP, White MD, Fisher K, Khara B, Bailey SS, Parker D, Rattray NJW, Trivedi DK, Goodacre R, Beveridge R, Barran P, Rigby SEJ, Scrutton NS, Hay S, Leys D. 'New cofactor supports α,β-unsaturated acid decarboxylation via 1,3-dipolar cycloaddition.' Nature 522 (2015) 497-501. DOI: 10.1038/nature14560. https://doi.org/10.1038/nature14560 (first crystal structures of holo A. niger Fdc1 with prFMN)
  5. Bhuiya MW, Lee SG, Jez JM, Yu O. 'Structure and Mechanism of Ferulic Acid Decarboxylase (FDC1) from Saccharomyces cerevisiae.' Applied and Environmental Microbiology 81 (2015) 4216-4223. DOI: 10.1128/AEM.00762-15. https://doi.org/10.1128/AEM.00762-15 (cofactor-free S. cerevisiae Fdc1 structure with 4-vinylphenol bound, PDB 4S13)
  6. Kawazoe T, Tsuge H, Pilone MS, Fukui K. 'Crystal structure of human D-amino acid oxidase: context-dependent variability of the backbone conformation of the VAAGL hydrophobic stretch located at the si-face of the flavin ring.' Protein Science 15 (2006) 2708-2717. DOI: 10.1110/ps.062421606. https://doi.org/10.1110/ps.062421606
  7. Marshall SA, Payne KAP, Fisher K, Titchiner GR, Levy C, Hay S, Leys D. 'UbiD domain dynamics underpins aromatic decarboxylation.' Nature Communications 12 (2021) 5065. DOI: 10.1038/s41467-021-25278-z. https://doi.org/10.1038/s41467-021-25278-z
  8. Pollegioni L, Sacchi S, Murtas G. 'Human D-Amino Acid Oxidase: Structure, Function, and Regulation.' Frontiers in Molecular Biosciences 5 (2018) 107. DOI: 10.3389/fmolb.2018.00107. https://doi.org/10.3389/fmolb.2018.00107
  9. Rondelet G, Fleury L, Faux C, Masson V, Dubois J, Arimondo PB, Willems L, Wouters J. 'Inhibition studies of DNA methyltransferases by maleimide derivatives of RG108 as non-nucleoside inhibitors.' Future Medicinal Chemistry 9 (2017) 1465-1481. DOI: 10.4155/fmc-2017-0074. https://doi.org/10.4155/fmc-2017-0074 (reports PDB 5LOD, apo M.HhaI)
  10. Datar PM, Joshi SY, Deshmukh SA, Marsh ENG. 'Probing the role of protein conformational changes in the mechanism of prenylated-FMN-dependent phenazine-1-carboxylic acid decarboxylase.' Journal of Biological Chemistry 300 (2024) 105621. DOI: 10.1016/j.jbc.2023.105621. https://doi.org/10.1016/j.jbc.2023.105621
  11. Gahloth D, Fisher K, Marshall S, Leys D. 'The prFMNH2-binding chaperone LpdD assists UbiD decarboxylase activation.' Journal of Biological Chemistry 300 (2024) 105653. DOI: 10.1016/j.jbc.2024.105653. https://doi.org/10.1016/j.jbc.2024.105653
  12. Feidakis CP, Krivak R, Hoksza D, Novotny M. 'AHoJ-DB: A PDB-wide Assignment of apo & holo Relationships Based on Individual Protein–Ligand Interactions.' Journal of Molecular Biology 436 (2024) 168545. DOI: 10.1016/j.jmb.2024.168545. https://doi.org/10.1016/j.jmb.2024.168545
  13. Raisinghani N, Parikh V, Foley B, Verkhivker G. 'AlphaFold2-Based Characterization of Apo and Holo Protein Structures and Conformational Ensembles Using Randomized Alanine Sequence Scanning Adaptation: Capturing Shared Signature Dynamics and Ligand-Induced Conformational Changes.' International Journal of Molecular Sciences 25 (2024) 12968. DOI: 10.3390/ijms252312968. https://doi.org/10.3390/ijms252312968
  14. Zhu X, Feng Y, Wang M, Li S, Li H, Liu L, Xue S. 'Integrated strategies for engineering ferulic acid decarboxylase with tunnel conformation and substrate pocket for adapting non-natural substrates.' Biochemical Engineering Journal 213 (2025) 109566. DOI: 10.1016/j.bej.2024.109566. https://doi.org/10.1016/j.bej.2024.109566 (reports PDB 8Y64, cofactor-free open-state S. cerevisiae Fdc1 mutant)