DNA is a highly reactive molecule and is susceptible to damage caused by both endogenous and exogenous factors, including hydrolytic and oxidative agents, UV and ionizing radiation, as well as exposure to alkylating and crosslinking agents. Additionally, mutations may occur during the DNA replication process. While mutagenesis plays a role in DNA maintenance and evolution, it can also lead to many diseases, including different types of cancers. To address this process, cells possess sophisticated and complex systems, including cell cycle checkpoints, damage tolerance mechanisms, various DNA repair processes, and the ability to undergo cell death, for example through apoptosis [1], [2].

Five DNA repair pathways are considered the major ones: base excision repair (BER), nucleotide excision repair (NER), mismatch repair (MMR), homologous recombination (HR), and non-homologous end joining (NHEJ) [1], as further discussed in Section 2. In addition to these, other DNA repair mechanisms include interstrand crosslink (ICL) repair and direct chemical reversal. ICL repair employs components from the major pathways to counteract exogenous agents, such as nitrogen mustards, platinums, psoralens, and mitomycin C, as well as endogenous products like oxygen radicals and aldehydes [2], [3]. Conversely, direct chemical reversal relies on a single protein with high substrate specificity and does not involve sugar-phosphate backbone incision or base excision. Examples include alkyltransferases and dioxygenases, which repair damage caused by alkylating agents, and lyases and photolyases, which reverse ultraviolet-induced lesions [4].

In numerous cancer types, DNA repair mechanisms are disrupted or dysregulated, resulting in elevated mutagenesis and genomic instability, which in turn promotes cancer progression [5], [6], [7]. Likewise, aging is linked to the degradation of chromosomal ends and the diminished efficacy of a combination of DNA damage response pathways. Moreover, several diseases, including neurodegenerative disorders, stem from the concurrent breakdown of multiple DNA repair processes [1].

In order to manage these complex processes, a set of proteins must operate within different pathways. Despite their variability, all these proteins must undergo a fundamental step before participating in DNA repair pathways: they need to be imported into the cell nucleus. Therefore, understanding the transport of nuclear proteins through the nuclear envelope and the regulation of import processes is crucial for comprehending their function.

Nuclear transport pathways involve proteins that are members of the β-karyopherin family, which can bind cargo directly (e.g., importin-β, transportin-1, transportin-3, and importin-13) or through adaptor proteins (e.g., importin-α, snurportin-1, and symportin-1). Other transport factors, involved in the transport of heat-shock proteins (e.g., Hikeshi) or RanGDP (e.g., NTF2) have also been described [8].

The first and best-characterized nuclear targeting signal is the classical nuclear localization sequence (cNLS) recognized by the protein importin-α (Impα) [9], [10]. Impα is an adaptor protein that can simultaneously binds to a cargo protein containing a nuclear localization sequence (NLS) and the carrier protein, importin-β (Impβ) [11]. This trimeric complex (Impα + NLS/cargo protein + Impβ) is then translocated through the nuclear pore complex (NPC) via transient interactions between Impβ and nucleoporins in the NPC. This process, known as the classical nuclear import pathway, is thought to be the most extensively used nuclear import mechanism in the cell [12], [13], [14]. cNLSs are composed of one or two basic clusters of amino acids [11], [12], [15], [16]. In agreement, Impα contains two NLS-binding sites, S1 (major site) and S2 (minor site). NLSs can be classified as monopartite, when they bind to a single site, and bipartite, when they bind to both sites (minor and major sites) simultaneously [10], [15], [17].

Monopartite NLSs have a consensus sequence defined by K(K/R)X(K/R) (corresponding to positions P2-P5), while bipartite NLSs contain two basic clusters separated by a linker region, with a consensus sequence of KRX10–12K(K/R)X(K/R) (corresponding to positions P1’-P2’ binding to the S2 (minor) site, and P2-P5, binding to the S1 (major) site) [10], [15], [17], [18], [19]. For monopartite NLS, the major binding site is the main binding region; however, in many crystal structures, binding to the minor site has also been observed. This is attributed to the high concentration of high-affinity NLS peptides used in the crystallization experiments [10]. More recent studies have also demonstrated that the presence of specific residues in positions adjacent to NLSs (N- and C-terminally), including P-1, P0, P1, P6, P0’, P3’ and P4’, can also contribute favorably to binding and increase the affinity of NLSs for Impα [19], [20], [21], [22], [23].

Several Impα variants have arisen due to duplication events in some organisms. The variants are divided into three families (α1, α2, and α3) expressed in metazoan species, and a fourth is classified as α1-like, including Impα from Viridiplantae and Fungi [24], [25], [26]. Mus musculus has six Impα variants (or isoforms) (cf. [27]), while seven Impα variants are found in humans (cf. [28]). Some of these variants have preferences for specific NLSs and may be associated with specific roles [24].

Most of the structural studies of NLSs from DNA repair proteins have employed the M. musculus Impα variant α2 [29] (labelled MmImpα in this review), also referred to as variant α1 (e.g., [30], [31]); this protein was the first mammalian Impα characterized structurally [10]. Some structural studies of NLS from DNA repair proteins have employed the human Impα (HsImpα) variants α1 and α3 (e.g., [32], [33], [34], [35]). Because the core of both NLS binding regions (major and minor sites) is strictly conserved in all mammalian Impα proteins, the structures are all directly comparable.

Experimental structural studies on complexes formed by NLS peptides and Impα have proven to be fundamental in uncovering key binding characteristics, such as monopartite or bipartite binding modes, preferential binding sites (major or minor), and even specific ligand-receptor interactions or reasons why a peptide does not bind to Impα [8]. Bioinformatic NLS predictors (e.g., [36]) often provide inconsistent data [32], [37], reinforcing the importance of structural coordinates available in databases (e.g., Protein Data Bank – PDB) for comprehensive analysis. Several studies using functional methods have proposed putative NLS sequences in DNA repair proteins, including the endonuclease III homolog (NTH1) from the BER pathway, replication protein A (RPA) from the NER pathway, and exonuclease 1 (EXO1) from the MMR pathway (cf. [38]). However, the lack of structural data prevents a detailed examination of the binding properties of these complexes.

In this review, we summarize the experimental structural and affinity studies on Impα and NLSs from DNA repair proteins, to uncover the general features of nuclear transport of DNA repair proteins. The article is divided into four sections: (i) a brief overview of DNA repair mechanisms; (ii) a review of the structures of complexes between Impα and NLSs from DNA repair proteins; (iii) a comparison of the affinities of these complexes; and (iv) a concluding discussion integrating all the information.