Genetic studies conducted over the last 20 years have been hugely successful in identifying
variation across the genome that is associated with disease risk. The inflammatory
bowel diseases (IBD) Crohnʼs disease and ulcerative colitis have been exemplars of
this success, with now > 250 individual genomic risk coordinates identified. This
success sometimes overshadows the more humbling fact that mechanistic understanding
is very scarce for how these genetic risk factors conspire with environmental triggers
to cause disease. For some risk loci, we completely lack any understanding of the
very basic function of their gene products.
An open reading frame on chromosome 13 (C13orf31) was one of these intriguing, entirely
blank, loci. A coding variant, which leads to a valine-for-isoleucine substitution
at amino acid position 254 (I254V) increases risk for Crohnʼs disease as well as for
leprosy, a chronic infection with Mycobacterium leprae. Very rare deleterious loss-of-function
variants detected in consanguineous families, such as an arginine-for-cysteine substitution
at amino acid 284 (C284R), cause autosomal recessive (i. e. monogenic) forms of very-early
onset Crohnʼs disease, or, indeed the same variant, Stillʼs disease. Stillʼs disease
(or systemic juvenile idiopathic arthritis, sJIA), where this gene is the sole known
cause of its monogenic form, is the paradigm of autoinflammation-cum-autoimmunity. It starts in toddlers as an IL-1-mediated periodic fever syndrome,
which morphs over weeks into a destructive, debilitating arthritis, which is thought
to be T cell driven. The C-terminal half
of the C13orf31 gene product is indeed evolutionarily highly conserved, and the
X-ray crystallographic structures of several of its bacterial orthologues had been
solved in the mid-2000s. These structures suggested that the protein was an enzyme.
However, not even a sub-domain of this protein exhibited any homology to domains or
proteins of known function, hence its function remained entirely enigmatic.
We discovered that this protein, which we named FAMIN [1], is an unprecedented, single pocket multi-functional purine nucleoside-metabolising
enzyme, with its activities conserved from bacteria to man. Specifically, FAMIN irreversibly
deaminates adenosine to inosine, and reversibly phosphorolytically cleaves inosine,
guanosine and methyl-thio-adenosine into hypoxanthine, guanine, and adenine and their
respective ribose-1-phosphates. These activities had been thought to be the domain
of namesake enzymes ADA, PNP and MTAP in any form of life. Additionally, FAMIN directly
cleaved adenosine into adenine, an activity that had been considered outrightly absent
from eukaryotic life [2]. Since purine nucleotide de novo synthesis directly yields inosine monophosphate
(IMP), from which adenosine and guanosine monophosphates (AMP, GMP) are generated,
hitherto it had been thought that ADA, PNP and MTAP are the sole source of
the nucleobases adenine, guanine and hypoxanthine. We solved the structure of
FAMINʼs bacterial orthologue YlmD in complex with its substrate inosine [2]. This demonstrated that inosine indeed binds to the predicted single pocket of the
enzyme, with the ribose coordinated by a Cys-His-His triad via a zinc. Intriguingly,
the orthologous cysteine in FAMIN is mutated in the loss-of-function variant (C284R)
that causes monogenic Crohnʼs and Stillʼs disease. We further demonstrated that the
I254V variant is not only hypomorphic, but also leads to a qualitative catalytic change.
Whilst FAMIN-254I prefers adenosine phosphorolysis to adenine, FAMIN-254V predominantly
deaminates adenosine to inosine with onwards phosphorolytic cleavage to hypoxanthine.
Adenosine, adenine and ribose are primordial metabolites from which life has emerged
from prebiotic chemistry. The discovery of such multifunctional purine enzyme was
particularly surprising, since
central purine metabolism had been considered settled for the last half century.
This posited the question how such multifunctional purine enzyme would affect metabolic
and immune function. We discovered that FAMIN enables a purine nucleotide cycle (PNC)
in macrophages, in which the enzyme is particularly highly expressed. A PNC had originally
been described by Lowenstein and Tornheim in extracts of skeletal muscle [3]. It consists of the cyclical deamination of AMP to IMP, and re-amination back to
AMP via the intermediate succinyl-AMP. The re-amination is fuelled by GTP hydrolysis
and consumes aspartate, whose carbons are then released as fumarate, which can be
reversibly hydrated to malate. As such, the FAMIN-enables PNC controlled electron
(e
−) and proton (H+) transfer from cytoplasm, primarily from glycolysis, to mitochondria. Indeed, reduced
or absent FAMIN activity results in gradual cytoplasmic acidification and reductive
stress due to an increased NADH/NAD+ ratio in the
cytoplasm. This is linked to a decrease in cellsʼ extracellular acidification
rate (ECAR), a measure of glycolytic activity as protons are co-exported from a cell
with lactate, and a decrease in oxygen consumption rate (OCR), reflective of mitochondrial
oxidative phosphorylation. Despite these profound metabolic perturbations, reduced
or absent FAMIN activity does not have a transcriptional footprint, further arguing
for a very ancient mechanism.
Hence the in-depth investigation of a single disease risk gene has revealed a fundamental
biochemical function that is conserved from bacteria to man, with ramifications that
go far beyond the associated diseases.
