DNA Structures, Repeat Expansions and Human Hereditary Disorders Sergei M Mirkin
Total Page:16
File Type:pdf, Size:1020Kb
DNA structures, repeat expansions and human hereditary disorders Sergei M Mirkin Expansions of simple DNA repeats are responsible for more The genetic nature of these disorders was finally under- than two dozen hereditary disorders in humans, including stood upon cloning and characterizing the fragile X muta- fragile X syndrome, myotonic dystrophy, Huntington’s disease, tion in 1991 [3–5]. This mutation appeared to be caused various spinocerebellar ataxias, Friedreich’s ataxia and others. by the progressive intergenerational expansion of a sim- 0 During the past decade, it became clear that unusual structural ple DNA repeat, (CGG)n, located within the 5 -untrans- features of expandable repeats greatly contribute to their lated region (UTR) of the FMR1 gene. This striking instability and could lead to their expansion. Furthermore, DNA discovery was soon followed by the demonstration of replication, repair and recombination are implicated in the (CAG)n repeat expansions in spinobulbar muscular atro- formation of repeat expansions, as shown in various phy [6], (CTG)n repeat expansions in myotonic dystrophy experimental systems. The replication model of repeat [7–9] and (GAA)n repeat expansions in Friedreich’s ataxia expansion stipulates that unusual structures of expandable [10], among others. As of today, more than two dozen repeats stall replication fork progression, whereas extra human hereditary disorders are linked to simple repeat repeats are added during replication fork restart. It also expansions. explains the bias toward repeat expansion or contraction that was observed in different organisms. In all cases, repeats are stably inherited until their lengths exceed a threshold of approximately 100–200 bp. Beyond Addresses this threshold, they start to expand during intergenera- Department of Biochemistry and Molecular Genetics, University of tional transmission, making up to several thousand copies Illinois at Chicago, Chicago, IL 60607, USA in the course of a few generations for some diseases. The Corresponding author: Mirkin, Sergei M ([email protected]) progressive character of repeat expansions provided the first clue to understanding the hereditary pattern described above [11]. The earlier onset and severity of Current Opinion in Structural Biology 2006, 16:1–8 the disease, and the probability of the repeat’s expansion This review comes from a themed issue on increase with its length, accounting for genetic anticipa- Nucleic acids tion and the Sherman paradox, respectively. During the Edited by Anna Marie Pyle and Jonathan Widom past decade, it became clear that these genetic phenom- ena are probably grounded in the unusual structural characteristics of expandable repeats. This review 0959-440X/$ – see front matter attempts to describe how the formation of unusual # 2006 Elsevier Ltd. All rights reserved. DNA structures by expandable repeats during DNA replication and/or repair could lead to their expansion, DOI 10.1016/j.sbi.2006.05.004 resulting in disease. Main characteristics of repeat expansions Introduction Originally, expansions were limited to trinucleotide The phenomenon of genetic anticipation was first (CGG)n(CCG)n [4,5], (CAG)n(CTG)n and (GAA)n described in 1918 for the human hereditary disorder (TTC)n repeats. It is now clear that tetrameric (CCTG)n myotonic dystrophy. This disease appeared to have ear- (CAGG)n [12], pentameric (AATCT)n(AGATT)n [13] lier onset and increased severity as the mutant gene was and even dodecameric (C4GC4GCG)n(CGCG4CG4)n transmitted from one generation to the next [1]. Genetic [14] repeats can also expand, leading to human disease. anticipation was subsequently detected for other heredi- For each disease, expansions are limited to just one repeat tary neurological disorders, including Huntington’s dis- of a given gene. Thus, repeat expansions are not caused ease, Friedreich’s ataxia and different spinocerebellar by mutations in the trans-acting factors involved in DNA ataxias. The penetrance of mutations responsible for replication, repair or recombination, which would induce these diseases also appeared to increase in successive general microsatellite instability [15]. The primary events generations. The latter trend, first described for fragile leading to repeat expansion occur in cis. X syndrome, became known as the Sherman paradox [2]. Neither genetic anticipation nor the Sherman The major characteristics of different repeat expansions paradox could be explained in terms of classic Mendelian are fundamentally similar. A repeat starts to expand when genetics and were often ruled out as a result of ascertain- its length exceeds a threshold of roughly 100–200 bp. ment bias. Normal alleles usually have much shorter repetitive runs. www.sciencedirect.com Current Opinion in Structural Biology 2006, 16:1–8 COSTBI 356 2 Nucleic acids The so-called ‘long normal alleles’ contain long repeats Figure 1 with several stabilizing interruptions. In expanding alleles, the interruptions at one end of a repeat are usually absent, creating lengthy homogenous repetitive runs [16]. Thus, there exists a link between the repeat’s integrity and its propensity to expand. The longer the repeat becomes, the more likely it is that it will expand further. Although expansions occur with mutation frequency for A bumpy ride — the formation of ‘slip outs’ by the nascent DNA strand (red) during DNA replication can lead to small-scale repeat repeats near the threshold length, long repeats expand expansions [18]. with near 100% probability. Finally, expansions are large- scale events, such that dozens or even hundreds of repeats could be added during a single transmission step. This both Watson–Crick (WC) and non-WC base pairs indicates that significant DNA synthesis is needed to (Figure 2a). These hairpins contain different mismatches, generate these extra repeats. contributing to their stability in the following order: CGG > CCG CTG > CAG [20]. Formation of hair- Structures of expandable DNA repeats pin-like structures by d(GAA)n and d(TTC)n repeats Which properties of the repeats could account for their was also later proposed [25]. propensity to expand? The simplest explanation was that the repetitive nature of expandable elements leads to Individual strands of expandable repeats can fold into occasional strand slippage during DNA replication other unusual DNA conformations as well. For example, [17,18]. An unrepaired ‘slip out’ in the nascent DNA strand single-stranded d(CGG)n repeats fold into a peculiar would convert into an expanded repeat after a second tetrahelical structure stabilized by intertwined G-quartets, round of replication (Figure 1). This hypothesis could as shown in Figure 2b[26]. not adequately explain several characteristics of repeat expansions, however. First, not every repeat expands. The structure-forming potential of expandable repeats Second, strand slippage usually causes limited repeat drastically changes the outcome of DNA strand slippage. length polymorphism, rather than large-scale expansion. Denaturing and renaturing repeat-containing duplexes Third, the bias toward repeat expansion in humans remains leads to the formation of unusually stable slip-stranded unaccounted for. DNAs, in which the ‘loop outs’ form hairpin-like struc- tures (Figure 2c). These stable hairpins kinetically By the mid 1990s, it became evident that expandable trap repetitive DNA in the otherwise unfavorable slip- repeats have unusual structural properties [19]. In a stranded conformation [27]. pioneering study [20], soon supported by others [21–24], d(CGG)n, d(CGG)n, d(CTG)n and d(CAG)n stretches were The (GAA)n(TTC)n repeat belongs to the group of shown to fold into hairpin-like structures, comprising so-called homopurine–homopyrimidine mirror repeats, Figure 2 Unusual DNA structures formed by expandable repeats. (a) Imperfect hairpins composed of (CNG)n repeats. (b) G-quartets composed of (CGG)n repeats. (c) Slip-stranded DNA. (d) Various triplexes formed by (GAA)n repeats (only one possible conformation of sticky DNA is shown). In the repeats, purines are red and pyrimidines are green; flanking DNA is shown in black. Current Opinion in Structural Biology 2006, 16:1–8 www.sciencedirect.com DNA repeat expansions and disease Mirkin 3 which form intramolecular triplexes (H-DNA) under the is DNA unwinding or even complete strand separation. influence of negative supercoiling [28]. Apparently, this The main cellular process involving DNA strand separa- particular repeat can form a variety of three-stranded tion is DNA replication. During replication fork progres- DNA structures (Figure 2d). Individual repetitive runs sion, a portion of the lagging strand template, called the were shown to adopt either H-y (pyrimidine/purine/ Okazaki initiation zone (OIZ), remains transiently single pyrimidine triplex) [29] or H-r (pyrimidine/purine/purine stranded to ensure coordinated syntheses of the leading triplex) [30] conformations under physiological condi- and lagging DNA strands. It was plausible to suggest, tions. Two distant directly repeated (GAA)n(TTC)n therefore, that expandable repeats could fold into unusual tracts within the same plasmid also form a composite secondary structures while within the OIZ of the lagging triplex structure, called ‘sticky DNA’ [30,31]. The fine strand template. structure of sticky DNA remains unknown, but it might be reminiscent of the composite triplex structure of The first support for the lagging strand hypothesis came the two distant homopurine–homopyrimidine runs, as from a landmark