The iron chaperone hypothesis
The first functional hypothesis on frataxin was suggested by Grazia Isaya and co-workers alongside her discovery of iron binding and aggregate formation. Frataxin was put forward as a ferritin-like scavenger that keeps iron in a bio-available form (Adamec et al. 2000; Cavadini et al. 2002). This hypothesis was supported by the evidence that ferritin can partially complement absence of frataxin (Zanella et al. 2008). However, independent evidence suggests that this is not the major function of frataxin. First, the effect of other ions on oligomerization is such that physiological concentrations of magnesium or calcium salts within the mitochondria would stabilize CyaY and Yfh1 in an iron bound monomeric state and destabilize oligomerization (Adinolfi et al. 2002). Second, it was independently shown that oligomerization is unessential when the protein functions as an iron chaperone during heme and Fe-S cluster assembly (Aloria et al. 2004). In higher organisms, mitochondria contain a specialized ferritin (Levi et al. 2001), that, when expressed in yeast, can prevent iron accumulation caused by Yfh1 deletion (Campanella et al. 2004), suggesting that iron storage is a redundant function, at least in mammals.
In search for functional interactions
Clues on the function of proteins often come from identification of interacting partners. This approach was attempted also for frataxin. The first indication of functional interactions with other proteins came from a bioinformatics study which investigated whether the frataxin gene co-occurred with other genes (Huynen et al. 2001). It was found that, over the 56 genomes available at the time, two genes, hscA and hscB/JAC1, had identical phylogenetic distribution with the frataxin/CyaY gene. The hit was particularly clear as, in the authors' words, ‘these genes have not only emerged in the same evolutionary lineage as the frataxin gene, they have also been lost at least twice with it, and they have been horizontally transferred with it in the evolution of the mitochondria'. The corresponding gene products encode hscA and hscB, two proteins involved in the synthesis of Fe-S clusters in proteobacteria suggesting a role of frataxin in this machinery (Fig. 2a).
Figure 2. A link between frataxin and the proteins from the Isc machinery. (a) Schematic representation of the E. coli Isc operon. It is worth notice that frataxin does not belong to the operon. (b) Human frataxin interactome as provided by the STRING database (http://string-db.org/). Frataxin is shown as a red node (FXN). It has been linked to the following proteins: LYRM4, LYR motif (or Isd11); IscU, scaffold protein; INS, insulin; GAA, alpha-glucosidase; IscS, Nfs1 desulfurase; LEP, leptin; ACO2, mitochondrial aconitase; TTPA, alpha-tocopherol transfer protein; EPAS1, endothelian PAS domain protein 1; PIP5K1B, phosphatidylinositol-4-phosphate-5-kinase. The links in the network are color coded as follows: linked in databases (green connectors), from text mining (black) and from experimental evidence (pink). Notice that there is no node for ferrochelatase that is found instead when starting from yeast Yfh1 (not shown). (c) Model of the triple complex between CyaY (in yellow), IscS (blue) and IscU (red) as described in Prischi et al. (2010). Top: CPK representation to show the surface complementarity; bottom: ribbon representation to show some of the details of the molecular recognition.
Download figure to PowerPoint
This elegant study initiated a long search for experimental links between frataxin with this metabolic pathway. Several different hits now define a rather complex interactome (Fig. 2b). Independent experimental evidence was published proving interactions between frataxin and other components of the same pathway. Early pathophysiological studies in a conditional mouse model reproducing the heart cardiomyopathy of FRDA also pointed to a primary involvement of frataxin in Fe-S cluster biogenesis (Puccio et al. 2001).
It was shown that, in the absence of Yfh1, activity of Fe-S-containing enzymes (aconitase, succinate dehydrogenase) and cluster incorporation into apo-ferredoxin decrease specifically (Chen et al. 2002; Duby et al. 2002; Mühlenhoff et al. 2002). Finally, it was found that yeast and human frataxin bind to the central Fe-S cluster assembly complex, which is composed of the Nfs1 enzyme and the scaffold protein Isu (Gerber et al. 2003; Ramazzotti et al. 2004). These two proteins are, respectively, the cysteine desulfurase that converts cysteine to alanine and a highly reactive persulfide used in the synthesis of sulfur bio-organic derivatives and the transient scaffold protein on which the cluster assembles (Bandyopadhyay et al. 2008; Py et al. 2011) (Fig. 2a). They are highly conserved in most organisms. The constitutive Nfs1 homodimer and two copies of monomeric Isu (or their bacterial orthologs IscS and IscU) form a binary complex.
The direct frataxin partner however remained controversial together with the state under which the proteins interact. Early genetic and biochemical studies were careful to generically indicate interaction between frataxin and the Nfs1/Isu complex (Gerber et al. 2003; Ramazzotti et al. 2004). Later on, it was reported that both Yfh1 and human frataxin interact with Isu (Cook et al. 2006; Bencze et al. 2007). It was suggested that human frataxin binds Isu with high affinity and in an iron-dependent way. It is however rather difficult to reconcile these reports with the interaction surface that was mapped on the frataxin β-sheet, since this region is not involved in iron binding. This interaction could also not be reproduced in the bacterial system where CyaY primarily pulls down endogenous IscS while IscU is pulled-down indirectly through its interaction with IscS (Layer et al. 2006; Adinolfi et al. 2009). The surface of interaction between IscS and CyaY was suggested by mutagenesis in an elegant study also presenting the X-ray structure of the IscS/IscU complex (Shi et al. 2010). A more accurate, although low-resolution, description of the interaction was obtained by a combination of SAXS, NMR, and modeling techniques (Prischi et al. 2010) (Fig. 2c). This study showed that two CyaY monomers bind IscS in a pocket that forms between the active site and the IscS dimer interface. The interaction between CyaY and IscS is iron independent as it was mapped in regions of the two proteins with complementary charges. CyaY packs mainly against IscS, which acts as a mold. A direct interaction with IscU is possible but only in the context of the ternary complex. CyaY binding increases the affinity of the IscS/IscU complex. Independent work using mammalian recombinant proteins confirmed these results by showing that human frataxin too interacts with a preformed complex composed of Nfs1, Isu and Isd11 in an iron-independent fashion (Tsai and Barondeau 2010; and Schmucker et al. 2011). Furthermore, mutagenesis experiments in the mammalian system are consistent with the surface of interaction suggested in the yeast and bacterial systems (Leidgens et al. 2010).
Another functional link was suggested by an interaction between frataxin and ferrochelatase, the enzyme that catalyzes the final step of heme biosynthesis by inserting the ferrous ion into porphyrin (Foury and Cazzalini 1997; Lesuisse et al. 2003; Taketani 2005). A direct high-affinity interaction between frataxin and ferrochelatase was reported both for Yfh1 and for human frataxin (Yoon and Cowan 2004; Bencze et al. 2007). These observations could explain why frataxin deficiency leads to kinetic inhibition of the heme pathway (Schoenfeld et al. 2005).
Finally, independent reports have linked frataxin to mitochondrial aconitase (Bulteau et al. 2004), subunits of complex II of the mitochondrial respiratory chain (Gonzalez-Cabo et al. 2005) and several chaperones (GRP75, Ssc1) (Shan et al. 2007). The relevance of these interactions has, however, recently been questioned, as different approaches were unable to identify or reproduce some of the previously described interactions (Schmucker et al. 2011; AP unpublished results). The potential involvement of frataxin in other mitochondrial pathways thus needs to be further validated through complementary in vivo and in vitro approaches. The only pathway that seems to be consolidated by increasing experimental evidence is that of Fe-S cluster biogenesis.
The role of frataxin in iron-sulfur cluster formation
Identification of interactions between frataxin and iron–sulfur cluster components led to different hypotheses on the level of involvement of the frataxin protein in this metabolic pathway. It was first suggested that frataxin is the iron donor which helps to solubilize and transport iron to the cluster biogenesis machine (Yoon and Cowan 2003). While cysteine is the accepted preferential donor of sulfur, nothing is known about how the largely insoluble iron gets to the complex. It was shown that cluster formation on the scaffold Isu is appreciably more efficient in the presence of frataxin leading to the conclusion that frataxin is an iron chaperone. This study was however limited by the absence of IscS/Nfs1, a component which may be crucial for understanding the biological functioning of the complex.
Later on, it was shown that the role of frataxin goes well beyond that of an iron transporter (Adinolfi et al. 2009). These studies greatly benefitted from the establishment of an enzymatic assay followed by different spectroscopies which, under strict anaerobic conditions, follow the kinetics of iron-sulfur cluster formation on the IscU/Isu. When the experiment was repeated comparing the rates of enzymatic cluster formation on IscU, it became clear that frataxin has a role in regulating the kinetics. Extensive work at this level was carried out using a prokaryotic model (Prischi et al. 2010; Iannuzzi et al. 2011). This work showed unequivocally that CyaY acts as an enzyme inhibitor. The structural information suggests at least two non-mutually exclusive mechanisms to explain these data. The presence of frataxin increases the affinity of the interaction between the scaffold IscU and the desulfurase IscS. This implies that detachment of IscU, which is needed to continue the reaction, is less favored. The effect could also be the consequence of a slower movement of the loop that brings the persulfide group from the active site to IscU, because of the steric impediment caused by the presence of CyaY.
Further studies however showed that also this interpretation was probably too simple: the effects of eukaryotic frataxin on the same enzymatic process seem to be opposite of those observed in bacteria: frataxin increases the rates of the enzymatic cluster formation thus acting as an activator (Gakh et al. 2010; Tsai and Barondeau 2010). Although confirming a role of frataxin as a regulator, it proposes the question of whether this is an unusual example of an enzyme that inverts or reverts its mode of control through evolution or if there is a different factor that is responsible for the observed behavior. To resolve the discrepancy, the enzyme kinetics were measured in parallel for the human and E. coli systems individually interchanging all the components (Bridwell-Rabb et al. 2012). This study unambiguously revealed that the effect of frataxin as an activator or an inhibitor is solely determined by which cysteine desulfurase is present, and not by the identity of the frataxin ortholog. A possible explanation is that the two systems are not entirely equivalent since the recombinant Nfs1 used in these studies was produced in E. coli where it is soluble only when co-expressed with Isd11, a protein absent in prokaryotes. Isd11 is poorly characterized: little remains known about the stoichiometry and the binding site on Nfs1. Native mass spectrometry studies on the mammalian ternary and quaternary complex suggest a stoichiometry of two ISD11 per NFS1 (Schmucker et al. 2011). This stoichiometry is however not easily compatible with the symmetry of the NFS1 dimer. Isd11 was proposed to stabilize Nfs1 through a direct interaction and to activate the enzyme through a conformational change (Adam et al. 2006; Wiedemann et al. 2006; Pandey et al. 2011). It is possible that Isd11 could interfere with frataxin or alter the Nfs1 conformation. New data are necessary to clarify this crucial issue. More recently, it was shown that the presence of human frataxin helps the efficiency of formation of [Fe4S4] clusters that can be transferred to mammalian aconitase (ACO2) to generate an active enzyme (Colin et al. 2013).
Conclusions and future perspectives
The identification of the FRDA gene in 1996 has finally opened new avenues for the comprehension of the molecular causes of FRDA. Yet, we still seem to be a long way away from the moment when we can be sure of the precise role of frataxin. This small and yet highly conserved essential protein has so far eluded most efforts along this direction thanks to its novel fold and its unusual properties. Two important milestones, however, have been the determination of its iron-binding properties and the link to a precise, although still understudied metabolic pathway. At this point, the most promising perspective seems to be that of dissecting the machine responsible for Fe-S cluster biogenesis and understanding the precise role of frataxin in it. This task will require, among other things, the development of new tools to study not just an isolated molecule or a protein complex but a complex network of competing multiple interactions. While making the project much more demanding, this constitutes one of the main challenges for the next several decades. In the words of Bruce Alberts (2012), we can conclude that ‘we now know that the most interesting properties of cells are “emergent” properties, resulting from elaborate networks of interactions between many different molecules that include the protein machines. Scientists currently lack the ability to decipher this complexity, and gaining this ability will require great ingenuity and many new developments whose exact nature is unpredictable. Much of the work will need to be done through small-science research in relatively simple systems, such as the E. coli bacterium, with the hope that what is learned will lead to new approaches and principles that can be transferred to the more complicated cells of mammals.'