**1. Introduction**

Mutations in the genes that encode complex I are responsible for a large fraction of all mitochondrial diseases. For example, 20–30% of cases in childhood mitochondrial disease (MD) are related to complex I dysfunction [1,2]. There seem to be many reasons for this. Complex I genes make up a large component of the mitochondrial and nuclear genome. Complex I is encoded by seven mitochondrial genes (out of 13 total) and 37 nuclear genes. There are also at least 15 complex I assembly factors [3]. Mitochondrial DNA (mtDNA) is especially prone to mutation due to insufficient DNA repair systems. All of the various functions of complex I can be impacted by mutation. Complex I forms supercomplexes with other members of the respiratory chain, such as complex III (cytochrome *bc1*) and complex IV (cytochrome *c* oxidase). It is metabolically linked to the citric acid cycle by NADH and to ATP synthesis by proton translocation. Furthermore, it is a major site of superoxide generation in mitochondria. Therefore, mutations that alter or degrade complex I function will typically have wider effects on mitochondrial function.

Complex I is a boot-shaped multi-subunit enzyme embedded in the inner mitochondrial membrane (see Figure 1). Its primary role is to oxidize NADH while reducing ubiquinone and translocating protons across the membrane. One arm extends into the matrix space, and it contains the flavin mononucleotide (FMN) and all of the iron–sulfur (FeS) clusters necessary for electron transfer to ubiquinone. The membrane arm contains subunits that translocate the protons from the matrix space to the intermembrane space (IMS). These core functions are carried out by the fourteen "core" subunits that appear in all known examples of complex I, including bacteria. Seven of the core subunits are membrane-embedded, and these seven are all encoded in mtDNA: ND1, ND2, ND3, ND4, ND4L, ND5, and ND6. The other seven core subunits are found in the matrix arm and contain the FMN and all of the FeS clusters. NDUFV1 contains the flavin and one tetranuclear FeS cluster, N3. NDUFV2 contains one binuclear FeS cluster N1a, which is not on the main electron transfer path. NDUFS1 contains one binuclear FeS cluster, N1b, and two tetranuclear clusters, N4 and N5. NDUFS8 contains two ferredoxin-like clusters, N6a and N6b. NDUFS7 contains the tetranuclear cluster N2, which is proximal to the ubiquinone binding site. For recent reviews of complex I, see [4–7]; for supernumerary subunits, see [8]. ‐ ‐

**Figure 1.** Structure of complex I with supernumerary subunits highlighted. Core subunits in the matrix arm are colored light blue. Core subunits in the membrane arm are colored beige. Supernumerary subunits that are described in this review are colored and labeled. Other supernumerary subunits are white. The two views are rotated 180◦ relative to each other. The structure is from the Protein Data Bank file 5xtd (PDB id = 5xtd) [9]. All structural images were generated using Jmol (http://www.jmol.org).

‐ ‐ The remaining thirty-one subunits (one is found in two copies) are supernumerary (or accessory) subunits, and much less is known about their functions. They are typically much smaller than the core subunits, and they are distributed on all surfaces of complex I. Some cross the membrane, while others are localized to the matrix face or the IMS. The naming of these subunits has generally followed their co-purification with various fractions of complex I: FV for the flavoprotein fraction, FS for the FeS protein fraction, FA for the alpha fraction associated with the matrix arm subunits, and FB for the beta fraction associated with the membrane proteins. The exception is NDUFAB1, the acyl carrier protein, which resembles an enzyme in lipid biosynthesis. This subunit appears to have an essential role apart from complex I, and it is the only protein that appears in two copies.

‐ ‐ ‐ ‐ ‐ From an evolutionary point of view, the core subunits can be organized into three modules. The N-module contains NDUFV1, NDUFV2, and NDUFS1, and it is defined by the source of electrons to the complex, the substrate NADH. This module is related to various NAD-linked dehydrogenases. The Q-module contains the remaining peripheral subunits NDUFS2, NDUFS3, NDUFS7, and NDUFS8, as well as two membrane subunits, ND1 and ND3, that contain the remaining FeS clusters and contribute to the ubiquinone binding site. This module is related to various membrane-bound hydrogenases. Finally, the remaining membrane subunits, ND2, ND4, ND4L, ND5, and ND6 compose the P-module (for proton translocating) and are related to subunits of an Na+/H<sup>+</sup> antiporter, the Mrp complex [10]. This grouping of subunits also corresponds to the assembly pathway. Q-module subunits

appear to assemble first, followed by the stepwise addition of the P-module, associated with the three major subunits ND2, ND4, and ND5. ND5 and the N-module enter the complex last.

In this review, 20 supernumerary subunits for which clinical mutations have been identified are described (see Table 1). Some of the mutations are interpreted in light of the structure of the human complex I (Protein Data Bank file 5xtd) or, in some cases, from other species. Some of the mutations are likely to be null mutations in which no protein is made, but the evidence is not always clear for that. Currently, the understanding of the effects of mutations is limited by a lack of knowledge. For example, how do mutations in one subunit affect the stability or import of that subunit? How does the absence of one subunit affect the expression, import, or stability of other subunits? How do mutations affect the assembly of that subunit or of other subunits? Do supernumerary subunits have roles that can be affected by mutation while assembly remains normal?


**Table 1.** Clinical missense mutations in supernumerary subunits and their surface contacts. IMS: intermembrane space.


**Table 1.** *Cont.*

\* Other mutations include nonsense mutations, splicing mutations, start codon mutations, insertions, and deletions. Mutations are defined by the change in the protein (p.) and the amino acid substitution, Lys45Thr, lysine at position 45 changed to threonine. See text for more details.
