DNA quadruplexes and dynamical genetics

Medical Hypotheses (2001) 57(1), 103–111 © 2001 Harcourt Publishers Ltd doi: 10.1054/mehy.2001.1291, available online at http://www.idealibrary.com on

DNA quadruplexes and dynamical genetics V. De Fonzo,1 E. Bersani,1 F. Aluffi-Pentini,2 V. Parisi3 1

EuroBioPark, c/o Parco Scientifico, Università di Roma ‘Tor Vergata’, Rome, Italy Dipartimento di Metodi e Modelli Matematici, Università di Roma ‘La Sapienza’, Rome, Italy 3 Sezione INFM, Dipartimento di Fisica, Università di Roma ‘Tor Vergata’, Rome, Italy 2

Summary In a recent paper, we have put forward the hypothesis that there exist smart purposive mechanisms – tandem repeat length managers – which regulate the length of some tandem repeat, or cause rearrangements, and are almost always driven by some variable number tandem repeat. We have called the framework in which such mechanisms act ‘dynamical genetics’. The purpose of this paper is to contribute to lay the foundations of a molecular study of the above mechanisms, by proposing a hypothesis, based on various kinds of supporting evidence and plausibility arguments, about the special importance of DNA quadruplexes for dynamical genetics, and by considering the involved enzymes. This hypothesis states that a tandem repeat length manager acts almost always by monitoring a DNA tract that has the characteristics of being a variable number tandem repeat and/or forming a DNA quadruplex, and that it is almost always driven by at least one of them. © 2001 Harcourt Publishers Ltd

INTRODUCTION In a recent paper (1), we have put forward the hypothesis that, in many more cases than is generally accepted, there exist smart purposive mechanisms that monitor and regulate the length of some ‘tandem repeats’ (TRs), or that also cause rearrangements, point mutations and apoptosis, and are almost always driven by some ‘variable number tandem repeat’ (VNTR). We denoted collectively such mechanisms, which we consider of paramount importance, as ‘tandem repeat length manager’ (TRLM). Both for the sake of simplicity and due to the overwhelming importance of TRs, we shall continue to use the term TRLM also in the possible rare cases of smart and purposive DNA editing where TRs are not known to be directly involved. The purpose of the present paper is to contribute to lay the foundations of a molecular study of TRLMs. We

put forward the basic hypothesis about the special importance of TRLMs of the anomalous three-dimensional nucleic acid structure known as DNA quadruplex, and also we consider the involved enzymes. The paper is organized as follows. In the first three sections, we discuss the nature of DNA quadruplexes, we formulate the basic hypothesis, and we discuss the relations between quadruplexes and TRLMs. In the following sections, we separately consider the role of DNA quadruplexes in rearrangement, VNTR-associated diseases, ageing, Creutzfeldt–Jakob disease, diseases due to viruses and parasites. We finally discuss the biochemical basis of TRLMs, the role of quadruplexes and triple helices, some possible further roles of DNA quadruplexes, and we draw some synthetic conclusions. ABOUT DNA QUADRUPLEXES

Received 4 August 2000 Accepted 5 January 2001

While, in the Watson–Crick double-helix model of DNA, the guanine base interacts with cytosine forming the standard G-C pairs, there exist anomalous structures in

Correspondence to: Valerio Parisi, Sezione INFM, Dipartimento di Fisica, Università di Roma ‘Tor Vergata’, Via della Ricerca Scientifica 1, I-00133 Rome, Italy. Phone: +39 06 7259 4896; Fax: +39 06 2023507; E-mail: [email protected]

This work was partially supported by INFM, Istituto Nazionale di Fisica della Materia, Rome, Italy.

103

104

De Fonzo et al.

which four Hoogsteen-bonded G bases form square, planar arrangements (2) that link four DNA tracts belonging to one or more molecules (3). Such planar arrangements are called G-quartets (or guanine tetrads) and, with suitable concentration of monovalent cations, are very stable. A structure containing some (often two or three) stabilizing G-quartet planes, and possibly also less stable (or even destabilizing) similar tetrad arrangements of different bases, is often called DNA quadruplex, and simply quadruplex. It is important to recall once and for all that, while the main features of a DNA quadruplex have been till now thoroughly studied in vitro, any feature, and even the formation, of quadruplexes in vivo is indirectly argued and often only presumptive. The most important sites where putative quadruplexes form, in mammalian cells, lie in immunoglobulin heavy chain switch regions, telomeres and rDNA. In recent years, the idea has grown stronger that G-rich DNA regions play fundamental roles in specific cellular functions (see Simonsson and Sjöback (4) and references therein). It is important to note that quadruplexes may have different 3D structure (for instance the diagonal and adjacent loops (5)), besides the obvious possibility of having different composition. Moreover, since different quadruplex structures have different features, even if the structure and characteristics of G-quartets are well conserved, we think it obvious that the behaviour specificity of a quadruplex must be due to the neighbouring bases, which often protrude outwards. An example of behaviour specificity is the fact that there exist specific enzymes that allow the folding and unfolding of different types of quadruplexes. A wellknown case is provided by two helicases involved in different diseases, the Bloom and the Werner syndromes: the cause of each disease appears to be the lack of the corresponding helicase (respectively named BLM and WRN) apt to unfold specific quadruplex structures (6,7). Another example is given by the G-rich VNTRs inside the immunoglobulin heavy chain switch regions, which, although similar, are different and well characterized for each class and appear to form different types of quadruplexes. A third example is provided by aptamers. Aptamers are oligonucleotides (DNA or RNA) that specifically bind to some enzymatic sites or to other proteins. Usually aptamers form DNA or RNA quadruplexes: it is conceivable that G-quartets give an aptamer the rigidity required for molecular docking specificity. We note that the variety of quadruplex behaviour supports our thesis that quadruplexes drive the many different mechanisms, having different functions, that act in the cell; in fact it appears to us difficult to envisage how this could happen if all quadruplexes were equivalent. Medical Hypotheses (2001) 57(1), 103 –111

THE BASIC HYPOTHESIS We now explicitly state our basic hypothesis, only briefly outlined elsewhere (8). Our hypothesis states that the smart and purposive mechanisms that manage the VNTRs, or cause useful point mutation or recombination events (i.e. mechanisms of the TRLM type), act almost always by monitoring a DNA tract. These tracts have one or both the characteristics of being a VNTR and of forming a quadruplex. We think that almost always a TRLM is driven by a VNTR and/or a quadruplex. We collect in the following paragraphs a number of arguments providing various kinds of support to our hypothesis. QUADRUPLEXES AND TRLMS The first supporting evidence in favour of our basic hypothesis is the fact that many TRLMs identified up to now are commonly believed to be regulated by VNTRs or quadruplexes. We first give a list of mechanisms, in human beings, which are certainly of the TRLM type: 1. The VNTR management occurring in the regulation of telomeres’ length (9): telomeres’ length decreases during every DNA duplication and is increased again by telomerase action or by other mechanisms. It is well known that telomeres contain quadruplexforming VNTRs, and that telomere length regulation is due to the inhibition of telomerase by suitable quadruplexes (10). 2. The homologous recombination occurring in the immunoglobulin heavy chain class switching (11): each genome region where class switching occurs (switch region) contains a VNTR that is different for each class type and forms quadruplex structures (12). 3. The regulation of the number of rDNA transcriptional units (13): another important case of smart and purposive modifications of DNA is the management of the number of rDNA units (we note that, at least in yeast, it has been demonstrated that the dynamics of these units is certainly involved in ageing (14), see below). Since such units are adjacent and their number varies according to their function (15), we can speak of a VNTR in a right albeit uncommon way, indeed the repeated sequence does not consist of a few bases but is composed by a whole gene, with an adjacent short tract named intergenic spacer (IGS). It is important to note that IGS contains microsatellites and sequences that can form quadruplex structures (13,16); some people are starting to think that the VNTRs are used to manage the number of genes (17). © 2001 Harcourt Publishers Ltd

DNA quadruplexes and dynamical genetics

4. The homogenization of rDNA genes (and possibly also of other gene clusters) (18): the various copies of rDNA remain exactly identical because of some mechanisms, which might be considered as purposive adaptive mutations, although they would tend to diverge functionally due to random point mutations. It is generally accepted that such a homogenization phenomenon is performed by a gene conversion, that is a special case of rearrangement, whereby a gene of the cluster is copied over another gene, obtaining a uniformization (concerted evolution) (19). 5. The chromosomal crossing over during pachytene: the crossing over between two homologous chromosomes during the prophase of meiosis I has a fundamental importance and is the best known case of DNA rearrangement. It occurs due to chromosome pairing in a number of sites; in particular during zygotene the sites start pairing and during pachytene the crossing over occurs because of the formation of quadruplexes (12) in sequences that are probably also VNTRs. 6. The V(D)J recombination (20): the mechanisms of somatic site-specific recombination in immunoglobulin gene arrangement produce the genetic diversity in immunoglobulin and T cell receptor genes: they are the better known case of purposive DNA editing and so of TRLMs. 7. The homogenization and the hypermutation in immunoglobulin genes (21): in immunoglobulin genes there are some constant regions, probably protected by homogenization phenomena (17), and some hypervariable regions (22): both phenomena are often believed to be due to some specific gene conversions and are obviously of essential relevance.

as distinct from the purposive ones described in the previous section. It is well known that, in yeast, a quadruplex can induce genetic rearrangements (23) and often such recombination events do not occur in the putative quadruplex sequence but only nearby. Moreover, it is known that G-rich minisatellites are often recombination hotspots (24,25) and can strongly stimulate in vitro recombination (26). Since many minisatellites are G-rich (27) with a pattern that is likely to form quadruplexes and since recombination hotspots have been found near G-rich sequences, such as the human insulin gene (28) that contains a well-known mini-satellite (29), it is likely that quadruplexes induce recombinations.

QUADRUPLEXES AND VNTR-ASSOCIATED DISEASES Our idea, first proposed in De Fonzo et al. (1), is that TRLMs usually operate correctly in monitoring and controlling some VNTRs and that, in the rare cases of malfunctioning, they cause a disease (the already known cases of such diseases are typically neurodegenerative). We now add that, since in many of these cases the VNTR seems to form quadruplexes, this appears to be a supporting evidence to our hypothesis that also the correctly operating TRLMs monitor and control quadruplexes. We list below the cases were, as far as we know, disease-associated VNTRs are commonly believed to form quadruplex structures (as suggested by in vitro experiments) in the gene regulation region. ●

We now summarize the above points from the point of view of our basic hypothesis. We note that for the above mechanisms it is well known that: Quadruplexes are believed to be certainly present in cases 1, 2, 3, 4, 5, and the fundamental role we assign to them is as yet generally accepted only in cases 1, 2, 5. VNTRs are certainly involved in cases 1, 2, 3, 4 and almost certainly in case 5; moreover, in cases 1 and 3, VNTRs are certainly the driving factor. Case 3 is more complicated, as discussed above. Although in cases 6 and 7 no involvement of quadruplexes or VNTRs has as yet been found, we feel that the above facts, as a whole, do already provide a good support to our hypothesis.

QUADRUPLEXES AND REARRANGEMENT In this section we consider genetic rearrangement in what are commonly believed to be non-purposive cases, © 2001 Harcourt Publishers Ltd

105

● ●

Fragile X syndrome (30). In this case it has even been recently suggested (31) that the VNTR expansion causing the disease is instigated just by the excessive structural stability of a quadruplex. This hypothesis is based on the experimental fact that quadruplex stability diminishes when CGG repeat is interrupted by AGG triplets, and neither expansion nor disease occurs. We note that this is in close agreement with our views on the role of quadruplexes in TRLMs. Insulin-dependent diabetes mellitus (32,33). Progressive myoclonus epilepsy 1 (34).

Moreover, we think that also in other diseases due to a CGG repeat expansion, such as oculopharyngeal muscular dystrophy (35) and cleidocranial dysplasia (36), the VNTR region may form quadruplexes (in a number of cases the VNTR involved has been associated with a triple helix structure). Cases where quadruplexes are involved (independently from belonging to a VNTR) are the Bloom syndrome and the Werner syndrome, already discussed above (6,7) Medical Hypotheses (2001) 57(1), 103 –111

106

De Fonzo et al.

and retinoblastoma. Murchie and Lilley (37) showed that the 5′ region (including the promoter) of the retinoblastoma susceptibility gene contains G-rich sequences that form in vitro quadruplexes. The authors point out that a very similar consensus site is present in regulatory sequences of various genes. Moreover, they suggest that the mechanism involved here is similar to that of the immunoglobulin switch regions. It has been further proposed that formation of quadruplexes occurs also in the coding part of the c-Ha-ras gene involved in various cancer types (38). We think that the presence (or absence) of quadruplex structures is a decisive factor in some neoplastic transformations. A further support in favour of the importance of quadruplexes in many diseases is provided by Usdin (39), who shows that NGG-triplet repeats form similar intrastrand structures, among which are quadruplexes, and considers the quadruplexes formed by AGG and TGG repeats as possible causes of triplet expansion diseases, as yet to be identified. QUADRUPLEXES AND VNTR IN AGEING Ageing in man is a complex phenomenon, still not satisfactory explained, which appears to depend on several causes: although the picture is still confused and far from complete, we are convinced that quadruplexes and VNTRs play a fundamental role in ageing. We first recall that (as already discussed above) telomeres contain VNTRs that form quadruplexes (3); and the most popular cause of ageing is the shortening of telomeres. Another VNTR-related possible cause of ageing (which we already proposed (1) on the base of the similarity with many triplet-associated neurodegenerative diseases) is the variation of the length of VNTRs that code for the polyglutamine tracts in transcription factors (with regard to transcription factors, and independently from ageing, a possible important role of quadruplexes will be discussed later on). We recall, due to its special importance, the TATA-binding protein, an important general transcription initiation factor with ubiquitous expression, including in the central nervous system: it normally has a polymorphic poly-CAG tract and has been recently involved in triplet-associated diseases (40). A further clue to the role of VNTRs and quadruplexes in human ageing is given by the corresponding situation in the case of Saccharomyces cerevisiae like all eukaryotes it shares many mechanisms with man, and is a much simpler and better understood case. In this yeast, one of the most important causes of ageing appears to be the migration of some enzymes which process DNA (in our parlance, TRLMs), specially the Sir proteins, and which regulate the gene silencing, towards other chromosomal Medical Hypotheses (2001) 57(1), 103 –111

regions. More specifically, such enzymes, which in a normal cell interact with telomeres, after a number of buddings migrate instead in nucleolus interacting with rDNA. The migration allows a shortening of telomeres and somehow causes nucleolus destructuring and ultimately leads to ageing (41). We are convinced that it cannot be by chance that both telomeres and nucleolar structure are founded on quadruplex structures; and we feel that it is not unlikely that an analogous albeit more complex process occurs also in humans. QUADRUPLEXES AND CREUTZFELDT–JAKOB DISEASE In a previous paper (1), we advanced the hypothesis that the contagion of the Creutzfeldt–Jakob Disease (CJD) is not due only to the prion protein (PrP) (42) but is mainly due to the VNTRs of the nucleic acids contained in the prion and trapped in the protein matrix. In the framework of the above hypothesis, the contagion occurs through a complicated mechanism where anomalous threedimensional structures of nucleic acids play a fundamental role. However, for brevity’s sake, all the necessary details and supporting arguments (including the explanation of many other as yet unclear phenomena about CJD) have been deferred for a forthcoming paper. We only explicitly describe here briefly, within the scope of the present paper, only the aspects of our theory dealing with quadruplexes, namely our conviction that the above anomalous three-dimensional structures are RNA (or even DNA) quadruplexes. In more detail: we think that the repeated sequence of RNA forms quadruplexes and subsequently binds to the prion protein favouring its transition to the scrapie form and exploiting it as a carrier to perform contagion. Arguments supporting the role of quadruplexes in CJD, already touched upon in (1), are the following: ●

●

From an RNA library, some sequences (aptamers) have been selected that specifically bind to prion protein: and it appears that they have a quadruplex structure (43); The degree of conservation of bases in the VNTR tract, rich in C and G, of the PrP-mRNA (unlikely a chance effect) is remarkable; generally this fact induces to think that the exactness of the sequence is more important in favouring a peculiar threedimensional structure of mRNA rather than in coding for the protein. Moreover thermodynamic calculations (44) suggest that the VNTR tract of the PrP-mRNA forms a structure called hairpin-C. We think that such a well-conserved sequence forms RNA quadruplexes.

A further argument is as follows. It is well known that the contagion agent withstands both exceptionally © 2001 Harcourt Publishers Ltd

DNA quadruplexes and dynamical genetics

adverse chemical–physical conditions and the action of enzymes, such as nucleases, apt to degrade normal nucleic acids (45). For this reason, many authors do not consider nucleic acids as likely contagion agents because they state that the possible protection provided by the almost indestructible scrapie protein (1) is not sufficient. We note instead that the CJD theory we proposed assumes the presence of quadruplexes that, in addition to the above protein protection, enjoy a great thermodynamic stability, and are not degraded by agents such as nuclease (46), that could instead easily degrade a standard nucleic acid structure. QUADRUPLEXES IN DISEASES DUE TO VIRUSES AND PARASITES We shall now examine the role of quadruplexes in phenomena such as virus integration in host genome and in parasites escaping from the host immune surveillance. We list the most important cases in which quadruplexes are thought important for viruses and parasites. ●

●

●

●

In the case of human immunodeficiency virus type 1 (HIV-1), a quadruplex is needed for dimerization (47). Moreover, it is known that an oligonucleotide forming quadruplex structures is a powerful inhibitor of the viral integrase (48). We think that integrase interacts with the quadruplexes of the virus and that the above phenomenon is an example of competitive enzymatic inhibition. The simian virus 40 (SV40) not only appears to form quadruplexes in its genome (49) but also encodes the T-antigen helicase that unwinds the quadruplexes in mouse immunoglobulin switch region (50) and some DNA triple helices (51). Within the Epstein–Barr virus (EBV) genome, there are G-rich recombinogenic sequences that are specifically targeted by the human cellular protein Sp1. It has been suggested that Sp1 binds to or induces the formation of quadruplexes (52,53) and that these structures are recombination intermediates (12). The Sp1 protein binds to the G-rich sequences in the virus terminal TRs involved in the interconversion between linear and circular forms; such a conversion is needed for virus integration and this suggests that the conservation of recombinogenic motifs in viruses has a functional role (53). We note that many virus integration sequences contain human telomere-like simple repeats (54,55) and that the EBV infection induces the immunoglobulin switching from class M to class E (56). We think that in both cases the putative quadruplexes play a pivotal role. The adeno-associated virus type 2 (AAV2) specifically integrates in a region (AAVS1) of chromosome 19 (57) and is the only virus of eukaryotes capable of targeted

© 2001 Harcourt Publishers Ltd

●

107

integration. Both the viral DNA terminal repeat region and the AAVS1 form stem-hairpin loop structures whose pairing is a potential candidate for quadruplex formation on the pathway to integration (58). In the case of Trypanosoma brucei, two proteins – ST-1 and ST-2 – have been identified with complementary binding features: ST-1 binds with high affinity to a Crich single strand (59) while ST-2 binds to a G-rich single strand (60). It has been suggested that ST-1 and ST-2 cooperate in regulating structures of the quadruplex type (60), and thus to favour the gene conversion causing antigenic variation (61) to escape the host immune system.

We further consider here the L1 retrotransposon, a mobile element, widespread integrated in the mammalian genome but rarely involved in a disease. L1 is similar to a retrovirus in that it codes for the reverse transcriptase, although it lacks the essential sequences (long terminal repeats) needed by retroviruses to integrate and does not code for the integrase. The L1 elements of different mammalian species exhibit a great sequence variability but all of them contain a highly conserved G-rich sequence. Howell and Usdin observed in vitro (62) that such G-rich tract can form intrastrand quadruplexes. They also suggested that retrotranscription and integration are induced just by this structure. We have already noted also that the HIV retrovirus integration is sensitive to quadruplexes. We now list two cases in which we think that quadruplexes are involved and important for viruses, or parasites, although to our knowledge this does not yet appear to be recognized. ●

●

The herpes simplex virus type 1 (HSV1) (63) contains a C-rich VNTR based on a motif of 15 nucleotides. Although it has not even been observed in vitro, we think that it is likely that such a sequence forms a quadruplex, allowing the virus to integrate by recombination in host DNA; Mycobacterium tuberculosis contains highly polymorphic GC-rich-repetitive sequences (PGRS) (64) spread over all the chromosome. It appears plausible that such VNTRs, so rich in G, could form quadruplexes and could be associated to genome plasticity and therefore to escaping the host immune system.

THE BIOCHEMICAL BASIS OF TANDEM REPEAT LENGTH MANAGERS Until now, we have not delved into the biochemical basis of TRLMs (except for the quadruplex structure). Since, nearly always, useful purposive cell mechanisms and metabolic reactions are performed by proteic Medical Hypotheses (2001) 57(1), 103 –111

108

De Fonzo et al.

enzymes, it is natural to think that also TRLM actions are nearly always performed by proteic enzymes, possibly associated with oligonucleotides (as in telomerase). Some TRLM-performing enzymes have already been observed, particularly in man and in S. cerevisiae. A good example is provided by telomerase and by various enzymes that cooperate with it in regulating the telomere length. Many other examples, although the picture is still incomplete and rapidly evolving, are provided by the enzymes involved in: ●

●

●

●

monitoring or controlling VNTRs, such as those binding to minisatellites (65,66) or to centromeric satellites (67) (e.g. Msbp-1 and CENP-B); recombination phenomena (e.g. Ku, RAG1 or RAG2), for example in central nervous system (68,69); point mutations, such as those involved in mismatch repair (e.g. Msh2) (70); apoptotic events (e.g. DNA-PK, and its Ku subunits) (71), caused by mechanisms we called ‘TRLM killer’ (1).

While, in simple metabolic reactions, usually a single enzyme catalyses a single reaction, a TRLM is more complicated and needs a complex of many enzymes. For example, in the case of yeast telomere length management, complexes such as Mre11/Rad50/Xrs2 and Sir proteins (72) are involved; the Mre11/Rad50/Xrs2 complex is also involved in – neither mitotic nor meiotic – CAG trinucleotide repeat expansions in a yeast model (73), a phenomenon similar to what we already hypothesized for CAG instability in human neurodegenerative diseases (1). Moreover, generally such complexes are built in a modular way, in the sense that a single enzyme has a share in different complexes and is involved in different TRLMs. An example is found in the Ku proteins that participate in different recombination events and in telomere silencing (74,75). It has been observed that some of these enzymes exhibit such a similarity of sequences or 3D structures that it is natural to hypothesize that they originate from the same ancestral protein; examples are provided by telomerase with reverse transcriptase and RAG1 with transposase. If we hypothesize that, among the inherited features, are the implemented mechanism and the characteristic (VNTR and quadruplexes) of DNA involved, it is natural to deduce that if two enzymes have parallel roles in different TRLMs, also the involved VNTRs and quadruplexes should be similar. We note that some of the enzymes involved in TRLM: ●

●

are selectively activated or inhibited by suitable quadruplexes, such as telomerase and integrase; form or unfold selectively some quadruplex, such as helicases of the RecQ family and resolvase.

Medical Hypotheses (2001) 57(1), 103 –111

It is therefore natural for us to suppose that similar parallelisms occur more frequently than observed as yet. ON TRIPLE HELICES In this section, we collect a number of reasons for reconsidering the widespread implicit belief that triple helices play almost always a major role in causing recombinations and point mutations. ●

●

●

●

●

In many cases of useful purposive TRLM action (see above), a specific quadruplex has been found involved, while triple helices have been often found involved only in useless (or even harmful) recombinations and point mutations. It is well known that DNA can reversibly assume in vitro different 3D structures depending on various factors: in particular the same tract can form a quadruplex or a triple helix (76) (selectively stabilized by different environments, e.g. by monovalent or bivalent ions respectively). It is likely that such structural transitions occur also in vivo, and we think that the transition outcome plays a fundamental role. Moreover we think it likely that in vivo the triple helix (sometimes observed in vitro) can sometimes be a temporary (and non interacting with TRLMs) transition from a quadruplex structure. A triple helix structure is often assumed when an anomalous DNA folding is observed without a sufficient number of guanines to form the G-quartet planes. It is, however, important to observe that quadruplex structures have been exhibited in which some stabilizing planes have some guanines (58,62) or even all the guanines (77) supplanted by other bases. When one of the two DNA strands folds into an anomalous structure, it often occurs that the other strand folds into a different anomalous structure: for example a C-rich strand can form triple helices (or other anomalous structures, such as i-motifs (78)) while the other, G-rich, can assume quadruplex structures (as well illustrated in Figure 5 from Simonsson et al. (79)). When a mutation hotspot is observed near C-rich tracts that are suspected to form triple helices, it is often concluded that triple helices induce the mutations. We think instead, in agreement with our views, that such mutations are rather induced by quadruplexes in the other strand. The best known case of VNTR-associated diseases is the case of repeated CAG usually associated to neurodegenerative diseases; in this case a triple helix structure has often been suggested. We only note, however, that a quadruplex structure might still be possible in spite of the scarcity of guanines, but forming several non-standard G-quartet planes. We recall that tetrad arrangements of different bases, © 2001 Harcourt Publishers Ltd

DNA quadruplexes and dynamical genetics

●

even less standard than those proposed for L1 retrotransposon (62), such as C-tetrads (49) or Atetrads (80), have been observed. We think that also CAG repeats form quadruplexes, although to our knowledge this is the first time that such structures have been proposed for this case. We note that if the CAG repetition is interrupted by some other triplets neither expansion nor disease occurs (81), as in the case of CGG repeats in X fragile (see above). Therefore, we deem it reasonable to think that also for CAG repeats the interruption decreases the stability of the quadruplex we are hypothesizing, and that the cause of the disease is the excessive stability of such a quadruplex. We finally note that, as already stated (8), one of the main causes of polycystic kidney disease type 1 could be the action of TRLMs that expand quadruplex-forming VNTRs, and successively cause rearrangements. In particular, we proposed as guilty the VNTR, in intron 1, whose polymorphism we indicated in (82). We think that the triple helix often hypothesized in intron 21, and supposedly the cause of the disease, could be at most a secondary indirect cause, also since no VNTR has ever been claimed therein.

FURTHER ROLES OF QUADRUPLEXES We recall the central role that we assigned to quadruplexes in dynamical genetics, that is in controlling the smart and purposive modifications of DNA. We think, moreover, based on recent observations (79), that folding and unfolding of quadruplexes in transcription factor promoters has an important regulating role, until recently completely ignored, for cell metabolism and physiology (and so beyond the scope of the present paper), through the modulation of the transcription factor synthesis. (About VNTRs in transcription factors see above and De Fonzo et al. (1).) Simonsson et al. (79) envisioned a scenario where, in the case of c-myc, the formation of quadruplexes in the promoter drives the gene transcription, and they observed sequences required to form quadruplexes also in the promoters of other important transcription factors (oncogenes). Moreover, we note that also in the case of Ku (83) (involved in TRLMs), G-rich regions have been observed in the promoter. CONCLUSIONS To summarize: In the crucial loci of DNA where TRLMs are acting, often quadruplex-forming VNTRs have been seen or at least hypothesized, which interact with suitable enzymes. In our opinion, this provides a molecular basis for dynamical genetics. © 2001 Harcourt Publishers Ltd

109

The distilled wisdom of the paper could be expressed – not without some severe oversimplification – in the following cliché: when DNA follows the rules of classical genetics, it usually folds into the classical Watson–Crick double helix structure, while when it interacts with TRLMs (dynamical genetics), it usually forms anomalous structures, among which quadruplexes have a great importance. By the way, such a formulation agrees with the general rule that the biochemical behaviour of macromolecules is determined mainly by their 3D structure. Our work originates from an attempt to give some logical order to a great wealth of experimental works that is, however, still insufficient and sparse; therefore, we hope that many more experiments will be performed, with the aim of improving our knowledge on the interaction between quadruplexes and DNA dynamics, and so confirming or disproving our hypotheses. REFERENCES 1. De Fonzo V., Bersani E., Aluffi-Pentini F., Parisi V. A new look at the challenging world of tandem repeats. Med Hypotheses 2000; 54: 750–760. 2. Gellert M., Lipsett M. N., Davies D. R. Helix formation by guanylic acid. Proc Natl Acad Sci USA 1962; 48: 2013–2018. 3. Williamson J. R. G-quartet structures in telomeric DNA. Annu Rev Biophys Biomol Struct 1994; 23: 703–730. 4. Simonsson T., Sjöback R. DNA tetraplex formation studied with fluorescence resonance energy transfer. J Biol Chem 1999; 274: 17379–17383. 5. Weitzmann M. N., Woodford K. J., Usdin K. The development and use of a DNA polymerase arrest assay for the evaluation of parameters affecting intrastrand tetraplex formation. J Biol Chem 1996; 271: 20958–20964. 6. Sun H., Karow J. K., Hickson I. D., Maizels N. The Bloom’s syndrome helicase unwinds G4 DNA. J Biol Chem 1998; 273: 27587–27592. 7. Fry M., Loeb L. A. Human Wemer syndrome DNA helicase unwinds tetrahelical structures of the fragile X syndrome repeat sequence d(CGG)n. J Biol Chem 1999; 274: 12797–12802. 8. Bersani E., De Fonzo V., Aluffi-Pentini F., Parisi V. On new hypotheses about autosomal dominant polycystic kidney disease type 1. Hum Genet (submitted). 9. Marcand S., Gilson E., Shore D. A protein-counting mechanism for telomere length regulation in yeast. Science 1997; 275: 986–990. 10. Zahler A. M., Williamson J. R., Cech T. R., Prescott D. M. Inhibition of telomerase by G-quartet DNA structures. Nature 1991; 350: 718–720. 11. Maizels N. Immunoglobulin class switch recombination: will genetics provide new clues to mechanism? Am J Hum Genet 1999; 64: 1270–1275. 12. Sen D., Gilbert W. Formation of parallel four-stranded complexes by guanine-rich motifs in DNA and its implications for meiosis. Nature 1988; 334: 364–366. 13. Gonzalez I. L., Sylvester J. E. Complete sequence of the 43-kb human ribosomal DNA repeat: analysis of the intergenic spacer. Genomics 1995; 27: 320–328. 14. Johnson F. B., Sinclair D. A., Guarente L. Molecular biology of aging. Cell 1999; 96: 291–302.

Medical Hypotheses (2001) 57(1), 103 –111

110

De Fonzo et al.

15. Kuo B. A., Gonzalez I. L., Gillespie D. A., Sylvester J. E. Human ribosomal RNA variants from a single individual and their expression in different tissues. Nucleic Acids Res 1996; 24: 4817–4824. 16. Hanakahi L. A., Sun H., Maizels N. High affinity interactions of nucleolin with G-G-paired rDNA. J Biol Chem 1999; 274: 15908–15912. 17. Bailey A. D., Pavelitz T., Weiner A. M. The microsatellite sequence (CT)n × (GA)n promotes stable chromosomal integration of large tandem arrays of functional human U2 small nuclear RNA genes. Mol Cell Biol 1998; 18: 2262–2271. 18. Kobayashi T., Heck D. J., Nomura M., Horiuchi T. Expansion and contraction of ribosomal DNA repeats in Saccharomyces cerevisiae: requirement of replication fork blocking (Fob1) protein and the role of RNA polymerase I. Genes Dev 1998; 12: 3821–3830. 19. Liao D. Concerted evolution: molecular mechanism and biological implications. Am J Hum Genet 1999; 64: 24–30. 20. Lewis S. M. The mechanism of V(D)J joining: lessons from molecular, immunological, and comparative analyses. Adv Immunol 1994; 56: 27–150. 21. Kelsoe G. V(D)J hypermutation and receptor revision: coloring outside the lines. Curr Opin Immunol 1999; 11: 70–75. 22. Harris R. S., Kong Q., Maizels N. Somatic hypermutation and the three R’s: repair, replication and recombination. Mutat Res 1999; 436: 157–178. 23. Liu Z., Gilbert W. The yeast KEM1 gene encodes a nuclease specific for G4 tetraplex DNA: implication of in vivo functions for this novel DNA structure. Cell 1994; 77: 1083–1092. 24. Wahls W. P., Wallace L. J., Moore P. D. Hypervariable minisatellite DNA is a hotspot for homologous recombination in human cells. Cell 1990; 60: 95–103. 25. Wahls W. P., Moore P. D. Recombination hotspot activity of hypervariable minisatellite DNA requires minisatellite DNA binding proteins. Somat Cell Mol Genet 1998; 24: 41–51. 26. Boán F., Rodríguez J. M., Gómez-Márquez J. A non-hypervariable human minisatellite strongly stimulates in vitro intramolecular homologous recombination. J Mol Biol 1998; 278: 499–505. 27. Jeffreys A. J., Wilson V., Thein S. L. Hypervariable ‘minisatellite’ regions in human DNA. Nature 1985; 314: 67–73. 28. Chakravarti A., Elbein S. C., Permutt M. A. Evidence for increased recombination near the human insulin gene: implication for disease association studies. Proc Natl Acad Sci USA 1986; 83: 1045–1049. 29. Bell G. I., Karam J. H., Rutter W. J. Polymorphic DNA region adjacent to the 5′ end of the human insulin gene. Proc Natl Acad Sci USA 1981; 78: 5759–5763. 30. Fry M., Loeb L. A. The fragile X syndrome d(CGG)n nucleotide repeats form a stable tetrahelical structure. Proc Natl Acad Sci USA 1994; 91: 4950–4954. 31. Weisman-Shomer P., Cohen E., Fry M. Interruption of the fragile X syndrome expanded sequence d(CGG)(n) by interspersed d(AGG) trinucleotides diminishes the formation and stability of d(CGG)(n) tetrahelical structures. Nucleic Acids Res 2000; 28: 1535–1541. 32. Kennedy G. C., German M. S., Rutter W. J. The minisatellite in the diabetes susceptibility locus IDDM2 regulates insulin transcription. Nat Genet 1995; 9: 293–298. 33. Hammond-Kosack M. C., Docherty K. A consensus repeat sequence from the human insulin gene linked polymorphic region adots multiple quadriplex DNA structures in vitro. FEBS Lett 1992; 301: 79–82. 34. Pataskar S. S., Brahmachari K. S. Conformational Polymorphism Observed in the Dodecamer Repeat Sequence Responsible For Enhanced Myoclonus Epilepsy [Epm1]. The Eleventh

Medical Hypotheses (2001) 57(1), 103 –111

35.

36.

37.

38.

39.

40.

41. 42. 43.

44.

45.

46.

47. 48.

49.

50.

51.

52.

53.

54.

Conversation in the Discipline Biomolecular Stereodynamics, June 15–19, 1999. The University at Albany, Albany, New York. Brais B., Bouchard J.-P., Xie Y.-G., et al. Short GCG expansions in the PABP2 gene cause oculopharyngeal muscular dystrophy. Nat Genet 1998; 18: 164–167. Mundlos S., Otto F., Mundlos C. et al. Mutations involving the transcription factor CBFA1 cause cleidocranial dysplasia. Cell 1997; 89: 773–779. Murchie A. I., Lilley D. M. Retinoblastoma susceptibility genes contain 5′ sequences with a high propensity to form guaninetetrad structures. Nucleic Acids Res 1992; 20: 49–53. Smith S. S., Baker D. J., Jardines L. A. A G4-DNA/B-DNA junction at codon 12 of c-Ha-ras is actively and asymmetrically methylated by DNA (cytosine-5) methyltransferase. Biochem Biophys Res Commun 1989; 160: 1397–1402. Usdin K. NGG-triplet repeats form similar intrastrand structures: implications for the triplet expansion diseases. Nucleic Acids Res 1998; 26: 4078–4085. Koide R., Kobayashi S., Shimohata T. et al. A neurological disease caused by an expanded CAG trinucleotide repeat in the TATA-binding protein gene: a new polyglutamine disease? Hum Mol Genet 1999; 8: 2047–2053. Sinclair D. A., Mills K., Guarente L. Extrachromosomal rDNA circles – a cause of aging in yeast. Cell 1997; 91: 1033–1042. Prusiner S. B. Prions. Proc Natl Acad Sci USA 1998; 95: 13363–13383. Weiss S., Proske D., Neumann M. et al. RNA Aptamers Specifically Interact with the prion protein PrP. J Virol 1997; 71: 8790–8797. Luck R., Steger G., Riesner D. Thermodynamic prediction of conserved secondary structure: application to the RRE element of HIV, the tRNA-like element of CMV and the mRNA of prion protein. J Mol Biol 1996; 258: 813–826. Neary K., Caughey B., Ernst D., Race R. E., Chesebro B. Protease sensitivity and nuclease resistance of the scrapie agent propagated in vitro in neuroblastoma cells. J Virol 1991; 65: 1031–1034. Bishop J. S., Guy-Caffey J. K., Ojwang J. O. et al. Intramolecular G-quartet motifs confer nuclease resistance to a potent anti-HIV oligonucleotide. J Biol Chem 1996; 271: 5698–5703. Awang G., Sen D. Mode of dimerization of HIV-1 genomic RNA. Biochemistry 1993; 32: 11453–11457. Mazumder A., Neamati N., Ojwang J. O., Sunder S., Rando R. F., Pommier Y. Inhibition of the human immunodeficiency virus type 1 integrase by guanosine quartet structures. Biochemistry 1996; 35: 13762–13771. Patel P. K., Bhavesh N. S., Hosur R. V. NMR Observation of a Novel C-Tetrad in the Structure of the SV40 Repeat Sequence GGGCGG. Biochem Biophys Res Commun 2000; 270: 967–971. Baran N., Pucshansky L., Marco Y., Benjamin S., Manor H. The SV40 large T-antigen helicase can unwind four stranded DNA structures linked by G-quartets. Nucleic Acids Res 1997; 25: 297–303. Kopel V., Pozner A., Baran N., Manor H. Unwinding of the third strand of a DNA triple helix, a novel activity of the SV40 large Tantigen helicase. Nucleic Acids Res 1996; 24: 330–335. Sun R., Spain T. A., Lin S. F., Miller G. Autoantigenic proteins that bind recombinogenic sequences in Epstein-Barr virus and cellular DNA. Proc Natl Acad Sci USA 1994; 91: 8646–8650. Sun R., Spain T. A., Lin S. F., Miller G. Sp1 binds to the precise locus of end processing within the terminal repeats of EpsteinBarr virus DNA. J Virol 1997; 71: 6136–6143. Gompels U. A., Nicholas J., Lawrence G., et al. The DNA sequence of human herpesvirus-6: structure, coding content, and genome evolution. Virology 1995; 209: 29–51.

© 2001 Harcourt Publishers Ltd

DNA quadruplexes and dynamical genetics

55. Kishi M., Bradley G., Jessip J., Tanaka A., Nonoyama M. Inverted repeat regions of Marek’s disease virus DNA possess a structure similar to that of the a sequence of herpes simplex virus DNA and contain host cell telomere sequences. J Virol 1991; 65: 2791–2797. 56. Li M. J., Maizels N. Activation and targeting of immunoglobulin switch recombination by activities induced by EBV infection. J Immunol 1999; 163: 6659–6664. 57. Samulski R. J., Zhu X., Xiao X et al. Targeted integration of adeno-associated virus (AAV) into human chromosome 19. EMBO J 1991; 10: 3941–3950. 58. Kettani A., Bouaziz S., Gorin A., Zhao H., Jones R. A., Patel D. J. Solution structure of a Na+ cation stabilized DNA quadruplex containing G.G.G.G and G.C.G.C tetrads formed by G-G-G-C repeats observed in adeno-associated viral DNA. J Mol Biol 1998; 282: 619–636. 59. Eid J. E., Sollner-Webb B. ST-1, a 39-kilodalton protein in Trypanosoma brucei, exhibits a dual affinity for the duplex form of the 29-base-pair subtelomeric repeat and its C-rich strand. Mol Cell Biol 1995; 15: 389–397. 60. Eid J. E., Sollner-Webb B. ST-2, a telomere and subtelomere duplex and G-strand binding protein activity in Trypanosoma brucei. J Biol Chem 1997; 272: 14927–14936. 61. Donelson J. E. Mechanisms of antigenic variation in Borrelia hermsii and African trypanosomes. J Biol Chem 1995; 270: 7783–7786. 62. Howell R., Usdin K. The ability to form intrastrand tetraplexes is an evolutionarily conserved feature of the 3′ end of L1 retrotransposons. Mol Biol Evol 1997; 14: 144–155. 63. Umene K., Watson R. J., Enquist L. W. Tandem repeated DNA in an intergenic region of herpes simplex virus type 1 (Patton). Gene 1984; 30: 33–39. 64. Poulet S., Cole S. T. Characterization of the highly abundant polymorphic GC-rich-repetitive sequence (PGRS) present in Mycobacterium tuberculosis. Arch Microbiol 1995; 163: 87–95. 65. Collick A., Dunn M. G., Jeffreys A. J. Minisatellite binding protein Msbp-1 is a sequence-specific single-stranded DNA-binding protein. Nucleic Acids Res 1991; 19: 6399–6404. 66. Wahls W. P., Swenson G., Moore P. D. Two hypervariable minisatellite DNA binding proteins. Nucleic Acids Res 1991; 19: 3269–3274. 67. Choo K. H. A. The Centromere. New York: Oxford University Press, 1997. 68. Bakalkin G., Yakovleva T., Hurd Y. L., Nussenzweig A., Li G. C., Terenius L. Autoantigen Ku in the brain. Developmentally regulated expression and subcellular localization. Neuroreport 1998; 9: 2147–2151.

© 2001 Harcourt Publishers Ltd

111

69. Alt F. W., Rathbun G., Yancopoulos G. D. DNA recombination in the brain? Curr Biol 1991; 1: 3–5. 70. de Wind N., Dekker M., Berns A., Radman M., te Riele H. Inactivation of the mouse Msh2 gene results in mismatch repair deficiency, methylation tolerance, hyperrecombination, and predisposition to cancer. Cell 1995; 82: 321–330. 71. McConnell K., Dynan W. S. The DNA-dependent protein kinase: a matter of life and (cell) death. Curr Opin Cell Biol 1996; 8: 325–330. 72. Haber J. E. The Many Interfaces of Mre11. Cell 1998; 95: 583–586. 73. Richard G. F., Goellner G. M., McMurray C. T., Haber J. E. Recombination-induced CAG trinucleotide repeat expansions in yeast involve the MRE11-RAD50-XRS2 complex. EMBO J 2000; 19: 2381–2390. 74. Tuteja R., Tuteja N. Ku autoantigen: a multifunctional DNAbinding protein. Crit Rev Biochem Mol Biol 2000; 35: 1–33. 75. Cooper M. P., Machwe A., Orren D. K., Brosh R. M., Ramsden D., Bohr V. A. Ku complex interacts with and stimulates the Werner protein. Genes Dev 2000; 14: 907–912. 76. Olivas W. M., Maher L. J. Competitive triplex/quadruplex equilibria involving guanine-rich oligonucleotides. Biochemistry 1995; 34: 278–284. 77. Patel P. K., Hosur R. V. NMR observation of T-tetrads in a parallel stranded DNA quadruplex formed by Saccharomyces cerevisiae telomere repeats. Nucleic Acids Res 1999; 27: 2457–2464. 78. Gilbert D. E., Feigon J. Multistranded DNA structures. Curr Op in Struct Biol 1999; 9: 305–314. 79. Simonsson T., Pecinka P., Kubista M. DNA tetraplex formation in the control region of c-myc. Nucleic Acids Res 1998; 26: 1167–1172. 80. Patel P. K., Koti A. S., Hosur R. V. NMR studies on truncated sequences of human telomeric DNA: observation of a novel Atetrad. Nucleic Acids Res 1999; 27: 3836–3843. 81. Djian P., Hancock J. M., Chana H. S. Codon repeats in genes associated with human diseases: fewer repeats in the genes of nonhuman primates and nucleotide substitutions concentrated at the sites of reiteration. Proc Natl Acad Sci USA 1996; 93: 417–421. 82. De Fonzo V., Bersani E., Aluffi-Pentini F., Castrignanò T., Parisi V. Are only repeated triplets guilty? J Theor Biol 1998; 194: 125–142. 83. Ludwig D. L., Chen F., Peterson S. R., Nussenzweig A., Li G. C., Chen D. J. Ku80 gene expression is Sp1-dependent and sensitive to CpG methylation within a novel cis element. Gene 1997; 199: 181–194.

Medical Hypotheses (2001) 57(1), 103 –111