Structural determinants of nucleobase modification recognition in the AlkB family of dioxygenases

Highlights

• Review on the AlkB family of iron- and α-ketoglutarate dependent dioxygenases.

• Structural information on AlkB family base modification recognition.

• Focus on key residues and structural features governing the substrate specificity of the AlkB family proteins.

Abstract

Iron-dependent dioxygenases of the AlkB protein family found in most organisms throughout the tree of life play a major role in oxidative dealkylation processes. Many of these enzymes have attracted the attention of researchers across different fields and have been subjected to thorough biochemical characterization because of their link to human health and disease. For example, several mammalian AlkB homologues are involved in the direct reversal of alkylation damage in DNA, while others have been shown to play a regulatory role in epigenetic or epitranscriptomic nucleic acid methylation or in post-translational modifications such as acetylation of actin filaments. These studies show that that divergence in amino acid sequence and structure leads to different characteristics and substrate specificities. In this review, we aim to summarize current insights in the structural features involved in the substrate selection of AlkB homologues, with focus on nucleic acid interactions.

Introduction

The AlkB family of dioxygenases has been studied extensively over the recent decades. As the name suggests, all of the members of this family share homology with the E. coli DNA repair protein AlkB (EcAlkB). Being part of the adaptive (Ada) response in E. coli, EcAlkB expression is strongly upregulated by DNA damage caused by alkylation. Upon sensing of alkylated adducts, a cysteine residue in the alkyltransferase Ada is irreversibly methylated. The methyl transfer to the protein reduces the negative charge of the zinc–thiolate center, thereby lowering its repulsive interaction with DNA. Consequently, Ada is turned into a strong transcription activator of the Ada operon, consisting of four DNA repair proteins (AlkA, Ada, AlkB and AidB) [1]. Similar adaptive response mechanisms have been found in many bacterial species, although the response genes are sometimes grouped into different operons [2]. Also in higher eukaryotes including yeast, plants and mammals, members of the AlkB family play pivotal roles [[3], [4], [5], [6]]. In humans, this family draws particular interest because of its involvement in DNA repair and progression of cancer, as they are frequently overexpressed in tumors and counteract the effects of alkylating drugs such as temozolamide used in chemotherapy [7,8]. Therefore, homologs such as hALKBH8 and hALKBH3 are targeted in anti-cancer therapies [9,10]. Besides, the AlkB homologue FTO is related to the onset of obesity [11] whereas ALKBH5 activity is regulating fertility in mice [12].

Like the other family members, the prototypical EcAlkB is an iron- and α-ketoglutarate-dependent dioxygenase [13] catalyzing the direct reversal of a wide variety of modified nucleobases. Oxygen is used to convert adducts into an unstable aldehyde, which is spontaneously released from the base, causing complete reversal of the alkylation. During the repair reaction, the required co-factor α-ketoglutarate is converted into succinate and CO₂ [14,15]. Through several biophysical experiments, Ergel et al. [16] showed sequential binding of all (co-) substrates in order to avoid premature turnover of the α-ketoglutarate. Furthermore, Bleijlevens et al. [17] showed that release of unmethylated DNA is facilitated by increased flexibility of the protein upon formation of succinate.

All of the investigated family members contain a 200–300 amino acid-sized AlkB domain adopting a so-called “jelly-roll” topology. This fold, also known as a double-stranded β-helix (DSBH), typically consists of a core of eight arranged β-strands (Fig. 1A and B) [13,18,19]. Strongly conserved in this family is a histidine-x-aspartate (HXD) motif located at the ‘open end’ of the DSBH core, followed by a second histidine further downstream. These three residues are involved in coordinating the catalytic Fe atom. Further complementation of the octahedral coordination of iron is provided by two oxygen atoms of the α-ketoglutarate cofactor. Often, binding of α-ketoglutarate is further stabilized by two additional, conserved arginine residues [[20], [21], [22]]. A feature shared amongst the nucleobase-binding members of the AlkB family is the occurrence of residues that aromatically stack against the damaged bases, although no strict conservation of aromatic residue type is seen between different family members [23]. A third conserved property of this family is the presence of a so-called nucleotide recognition lid consisting of a small β-sheet, contributing to substrate specificity [20].

Nucleic acid-interacting AlkB homologues need to be able to discern damaged bases from canonical bases. For AlkB and homologues repairing 1-methyl adenine (1mA), the mechanism of damage detection is thought to rely on the adoption of a Hoogsteen base pair between 1mA and thymine in duplex DNA. Other lesions, such as 3-methyl cytosine (3mC) show other features in duplex DNA, yet they all have in common that the thermal stability of the base-pairing is weakened [24]. The process of searching and interrogating base pair stability is thought to occur through multiple mechanisms such as sliding, hopping or a combination of both [25]. To our knowledge, no kinetic studies on the AlkB family have been done so far to elucidate their way of scanning. For hALKBH2 and hALKBH3, associated proteins PCNA and ASCC3, respectively, have been identified to assist in sliding over the DNA [26,27]. EcAlkB on the other hand has been reported to interact with single-strand binding protein (Ssb) and RecA, which is possibly promoting the oxidative repair of modified ssDNA [[28], [29], [30]]. Although the sliding/hopping mechanism of EcAlkB is largely unknown, the low affinity for canonical bases and the dynamic state of EcAlkB when bound to the end product succinate is thought to result in a more transient binding. This transiency then promotes one-dimensional scanning for nucleobase adducts [17].

As noted by Huffman et al. [31], upon detection of modified bases, most DNA repair enzymes flip out the damaged base by intercalating in the double helix. hALKBH2 extensively contacts the complementary strand, thereby forcing out the damaged base and filling the vacant space in the double helix with the aromatic ring of a phenylalanine residue (F102) from a DNA-interacting loop. Conversely, EcAlkB forces the nucleosides flanking the methylated base to stack on one another, thereby squeezing out the modified nucleobase. Stacking of these nucleobases requires one of the two sugars to be inverted, by forcing the sugar moiety to rotate 180°. This inversion then further requires stabilization by EcAlkB through a network of hydrogen bonds.

Through bioinformatical sequence analysis of bacterial AlkB homologues, Van den Born et al. [32] proposed four different subfamilies of AlkB homologues, termed 1A, 1B, 2A and 2B, each characterized by the presence of specific amino-acid residues in the nucleotide recognition domain. Human homologues were classified in the same subfamilies. For example, hALKBH2 and hALKBH3 were assigned to subfamily 1B. Together with subfamily 1A and 2A, these are thought to be involved in DNA repair. On the other hand, the AlkB domain of hALKBH8, belonging to subfamily 2A, is involved in removal of tRNA modifications. However, the amino-acid composition of the nucleotide recognition lid, and thus indirectly the classification into subfamilies, is insufficient to predict substrate specificity.

Many aspects of the AlkB family have been reviewed over the past two decades, such as RNA repair [[33], [34], [35]], DNA damage repair [[36], [37], [38]], base-flipping mechanisms [39], reaction mechanism [40], cellular functions [41] and substrate promiscuity [42]. Lastly, Lu et al. [43] and Shen et al. [44] compare the TET family of dioxygenases with the AlkB family. In this review, we aim to give a concise, state-of-the-art view on the structural features contributing to the substrate specificity towards different nucleic acid modifications.