Intragene Higher Order Repeats in Neuroblastoma BreakPoint Family Genes Distinguish Humans from Chimpanzees

ic ,,1 Marija Rosandic ,2 Ivan Basar,1 and Ines Vlahovic 1 Vladimir Paar,*,1 Matko Glunc
1 2

Faculty of Science, University of Zagreb, Zagreb, Croatia Department of Internal Medicine, University Hospital Rebro, University of Zagreb, Zagreb, Croatia These authors contributed equally to this work. *Corresponding author: E-mail: paar@hazu.hr. Associate editor: James McInerney
Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014 Research article

Abstract
Much attention has been devoted to identifying genomic patterns underlying the evolution of the human brain and its emergent advanced cognitive capabilities, which lie at the heart of differences distinguishing humans from chimpanzees, our closest living relatives. Here, we identify two particular intragene repeat structures of noncoding human DNA, spanning as much as a hundred kilobases, that are present in human genome but are absent from the chimpanzee genome and other nonhuman primates. Using our novel computational method Global Repeat Map, we examine tandem repeat structure in human and chimpanzee chromosome 1. In human chromosome 1, we nd three higher order repeats (HORs), two of them novel, not reported previously, whereas in chimpanzee chromosome 1, we nd only one HOR, a 2mer alphoid HOR instead of human alphoid 11mer HOR. In human chromosome 1, we identify an HOR based on 39-bp primary repeat unit, with secondary, tertiary, and quartic repeat units, fully embedded in human hornerin gene, related to regenerating and psoriatric skin. Such an HOR is not found in chimpanzee chromosome 1. We nd a remarkable human 3mer HOR organization based on the ;1.6-kb primary repeat unit, fully embedded within the neuroblastoma breakpoint family genes, which is related to the function of the human brain. Such HORs are not present in chimpanzees. In general, we nd that humanchimpanzee differences are much larger for tandem repeats, in particularly for HORs, than for gene sequences. This may be of great signicance in light of recent studies that are beginning to reveal the large-scale regulatory architecture of the human genome, in particular the role of noncoding sequences. We hypothesize about the possible importance of human accelerated HOR patterns as components in the gene expression multilayered regulatory network. Key words: human brain evolution, chromosome 1, higher order repeats, NBPF genes, human hornerin gene, global repeat map.

Introduction
HumanChimpanzee Divergence and Regulatory Elements
Identication of genomic differences that set humans apart from chimpanzees is important from evolutionary, medical, and cultural perspectives (Dorus et al. 2004; Enard and Paabo 2004; Hayakawa et al. 2005; Haygood et al. 2007, 2010; Kleinjan and Lettice 2008; Portin 2008; Varki et al. 2008; Woolfe and Elgar 2008; Enard et al. 2009; Taft et al. 2010). The most striking trend in human evolution is the rapid increase in brain size over the past 34 My and the associated increase in cognitive capacity and complexity (Jerison 1975; Mekel-Bobrov et al. 2007). It could be expected that the genetic changes, potentially underlying human brain evolution, span a wide range from nucleotide substitutions to large-scale structural alteration of the genome (Vallender 2008; Vallender et al. 2008). Chimpanzees share nearly 99% of human DNA in genes, but given the substantial number of neutral mutations, only a small subset of the observed gene differences is likely to be responsible for the key phenotypic changes between humans and

chimpanzees (The chimpanzee sequencing and analysis consortium 2005). Developmental and evolutionary mechanisms underlying the characteristics that set the human brain apart from other human primates are poorly understood (Pollard et al. 2006a). Recently, computational search was performed for DNA sequences that have changed most since humans and chimpanzees diverged from their last common ancestor (Amadio and Walsh 2006; Pollard et al. 2006b; Prabhakar et al. 2006; Bird et al. 2007; Beniaminov et al. 2009). Among such human accelerated regions (HARs) the most pronounced is a 118-bp stretch on human chromosome 20, known as HAR1. It is highly conserved among species during most of vertebrate evolution, while changing signicantly after the chimpanzeehuman split: The human and chimpanzee HAR1 sequences differ by 18 bp. This rapid burst of base pair substitutions in HAR1 was interpreted as evidence that HAR1 acquired an important new function in humans and in particular may have altered human brain by inuencing the activity of a whole network of genes (Pollard et al. 2006a). On the other hand, many highly conserved noncoding sequences have been found (Dermitzakis

The Author 2011. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Mol. Biol. Evol. 28(6):18771892. 2011 doi:10.1093/molbev/msr009

Advance Access publication January 27, 2011

1877

Paar et al. doi:10.1093/molbev/msr009

MBE
sequences (King and Wilson 1975; Bush and Lahn 2008), and the chromosomal regions in man, that are gene-poor, harbor gene regulatory elements that have the ability to modulate gene expression over very long distances (Lettice et al. 2002). A possible functional role of noncoding sequences and in particular of repeats has been much discussed. It is of importance to investigate the role of repeat sequences in the regulatory mechanism, related to a general concept that the regulatory system of genomes is encoded in networks of repetitive sequence relationships (Britten and Kohne 1968; Britten and Davidson 1969, 1971; Davidson and Britten 1979). In this sense, the repetitive components may play a major architectonic role in higher order physical structuring. It was argued that in this way one could think about genomes as information storage systems with parallels to electronic information storage systems. From such a perspective, repetitive DNA sequences are important components of genomes, required for formatting coding information so that it can be accurately expressed and for formatting DNA molecules for transmission to new generations of cells. The cooperative nature of proteinDNA interactions provides another fundamental reason why repeated sequence elements are essential to format genomic DNA (Shapiro and von Sternberg 2005). This was accompanied by observation that tandem arrays are often the regions that vary most between related taxi (Shapiro and von Sternberg 2005). A probable importance of noncoding changes in the evolution of human traits was underscored, particularly for cognitive traits (Haygood et al. 2010). Recent studies are beginning to reveal the large-scale regulatory architecture of the human genome (Noonan and McCallion 2010).

et al. 2002; Bejerano et al. 2004, 2006; Ovcharenko 2008; Stephen et al. 2009). Half a century ago, Jacob and Monod (1961) discovered that specic noncoding sequences are required to activate genes in Escherichia coli and noted that mutations in these regulatory sequences might play a role in the evolution of organismal traits. Britten and Davidson (1969, 1971) proposed a model that discrete regulatory sequences are abundant in the genomes of higher eukaryotes and that increases in the number of regulatory elements and the complexity of gene regulatory networks drive increased biological complexity in evolution. King and Wilson (1975) proposed that phenotypic differences between humans and chimpanzees are mainly caused by regulatory changes in gene expression. Raff and Kaufman (1983) hypothesized that mutations affecting the function of regulatory genes are likely to underlie many evolutionary changes in morphology: changes in just a handful of regulatory interactions could cause large-scale morphological changes. Recent investigations are progressing along such general framework. Components of regulatory control in the human genome include proximal regulatory elements and long-range elements that can act across large genomic distances to inuence the spatial and temporal distribution of gene expression, with enhancers and repressors of gene activity often located in evolutionary conserved noncoding regions of the genome (Pennacchio and Rubin 2001; Caceres et al. 2003; Cooper and Sidow 2003; Gu JY and Gu X 2003; Nobrega et al. 2003; Wray et al. 2003; Khaitovich et al. 2004; Uddin et al. 2004; Heissig et al. 2005; Kleinjan and van Heyningen 2005; Evans et al. 2006; Maston et al. 2006; Pennacchio et al. 2006; King et al. 2007; Xie et al. 2007; Portin 2008; Visel et al. 2008; Wray and Babbitt 2008; Duret and Galtier 2009; Mercer et al. 2009; Gareld and Wray 2010; Haygood et al. 2010; Noonan and McCallion 2010; Xu et al. 2010). Recent analyses provide evidence of many thousands of distant-acting noncoding sequences, beyond the proximal regulatory sequences and thus are beginning to reveal the large-scale regulatory architecture of the genome. At the present time, however, identifying regulatory elements in the human genome still remains a major challenge (Woolfe and Elgar 2008; Noonan and McCallion 2010). Genomewide studies have shown that human genome in transcription produces many thousands of regulatory noncoding RNAs (Huttenhofer et al. 2005; Johnson et al. 2005; Costa 2007; Kapranov et al. 2007; Amaral et al. 2008; Mattick 2009; Mercer et al. 2009; Van Bakel and Hughes 2009; Taft et al. 2010; Xu et al. 2010). Transcriptome studies indicated that a large proportion of transcription takes place outside of the known gene boundaries (Mercer et al. 2009; Van Bakel and Hughes 2009). Recent results of Xu et al. (2010) showed that close to 40% of all transcripts expressed in the human brain map are within repetitive elements, whereas ,10% of the human brain transcriptome corresponds to nonrepetitive intergenic regions. Morphological and anatomical differences between humans and chimpanzees are mainly caused by differences in the regulation of function of genes arising from noncoding
1878

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

Neuroblastoma BreakPoint Family Genes

The recently described gene family neuroblastoma breakpoint family (NBPF) with an intricate genomic organization was expanded during recent primate evolution. There are indications that they are brain-related genes: At least one of them might suppress the development of neuroblastoma and possibly of other tumor types (Vandepoele et al. 2005, 2008, 2009; Gregory et al. 2006; Popesco et al. 2006; Vandepoele and van Roy 2007; Diskin et al. 2009). Previous studies of genomic sequences of human NBPF genes using standard bioinformatics tools (Altschul et al. 1990; Benson 1999; Gelfand et al. 2007) have reported tandem repeats in these genes, so-called NBPF repeats, based on diverging ;1.51.6-kb repeat unit (Vandepoele et al. 2005; Gregory et al. 2006; Gelfand et al. 2007; Warburton et al. 2008). Several reports have shown that the copy number of these genes is variable in humans (Tuzun et al. 2005; Redon et al. 2006). A remarkable reduction of NBPF copy number with respect to the human genome was found in other primates, and in mouse, the NBPF copies are absent (Vandepoele et al. 2005, 2009; Popesco et al. 2006).

Higher Order Repeats

Higher order repeats (HORs) have been extensively studied in human and nonhuman primate chromosomes (Warburton and Willard 1996). The best known prototypes

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

MBE

FIG. 1 Schematic illustrating of 11mer HOR. Seven HOR copies are denoted h1, h2,. . .,h7 and 11 constituent primary repeat monomers m1, m2,. . .,m11. Divergence div (hi, hj) between HOR copies is sizably smaller than divergence div(mk, ml) between primary repeat monomers within each HOR copy. Monomers within each HOR copy can signicantly diverge from each other, whereas divergence between HOR copies is sizably smaller. For a particular case of ;171-bp alpha satellite monomers in chromosome 1, monomers can be diverged ;35% from each other, whereas the HORs are highly similar (mutual divergence among HOR copies ,5%).

of HORs are alpha satellite (alphoid) arrays, located in the centromeric region of all human chromosomes (Manuelidis 1978; Wu and Manuelidis 1980; Willard 1985; TylerSmith and Brown 1987; Willard and Waye 1987a, 1987b; Warburton and Willard 1996; Jurka 2000; Alexandrov et al. 2003, 2006; Rudd and Willard et al. 2001; Rosandic 2004; Jurka et al. 2005; Paar et al. 2005, 2007; Warburton et al. 2008). Alpha satellite arrays consist of primary repeat units (alphoid monomers) of ;171 bp, tandemly arranged in a head-to-tail fashion. Individual alphoid monomers diverge by 2040% from each other. However, some stretches of alpha satellites are hierarchically organized into HORs, secondary repeat units with highly convergent HOR copies (divergence between HOR copies ,5%) (Warburton and Willard 1996). In that case, divergence between HOR copies is much smaller than divergence between monomers within each HOR copy. Figure 1 schematically describes the overall concept of HORs for an illustrative case of 11mer HOR. Most of the reported HORs are based on relatively short primary repeat units, like 171-bp alpha satellite or less. The longest primary repeat unit was reported by Warburton et al. (2008) as a ;3.3-kb repeat forming 18-kb HORs. Highly homogeneous arrays of HOR alpha satellite monomers are relatively recent additions to human genome (Warburton and Willard 1996; Alexandrov et al. 2001). An explanation for generating HORs involves unequal crossing-over between misaligned HOR units aligned on the register of homologous monomers. Unequal crossing-over, restricted to tandem sequences, explains the generation and local homogenization of HOR units and accounts for large size variation among HORs on homologous chromosomes (Southern 1975; Smith 1976; Willard and Waye 1987b; Warburton and Willard 1996; Alkan et al. 2004; Rudd et al. 2006). By the process of unequal crossingover, HORs enable rapid evolutionary development.

chromosome 1. In previous investigations, short stretches of human accelerated sequences (up to several hundreds of base pairs; HARs) have been identied, mostly corresponding to nonhuman multispecies conserved regions. Here, we search for HARs containing large tandem repeat and HOR units. We are asking whether a more complex genomic pattern, involving long-range correlations, is accelerated in brain-related human genes that could characterize the human brain evolution. As a case study, we investigate computationally the genomic sequence of chromosome 1 (National Center for Biotechnology Information [NCBI] Build 36.3 assembly) that contains the NBPF gene family. Here, we discover HORs within NBPF repeats in human chromosome 1, whereas we nd no HOR in arrays of NBPF repeats in chimpanzee chromosome 1. This is the rst time that HORs are found fully embedded within a gene. Furthermore, these HORs are based on an extraordinary large primary repeat unit. Besides the novel NBPF HOR pattern, highly accelerated in human chromosome 1 (in chimpanzee chromosome 1 this HOR pattern is missing), we nd two additional human HORs (one of them novel) and several tandem repeats with large repeat units, all showing signicant humanchimpanzee divergence, that is, the human accelerated pattern. We hypothesize on a possible role of this acceleration as possibly important components in the gene expression multilayered regulatory network.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

Materials and Methods

Key String Algorithm
In spite of powerful standard computational tools in bioinformatics, there are still difculties to identify and analyze long repeat units. For example, the Tandem Repeat Finder can identify tandem repeat units up to 2 kb (Benson 1999; Warburton et al. 2008). Here, we use a new robust approach, useful in particular for investigating very long and/or complex repeats. The Key String Algorithm et al. 2003, 2006; Paar et al. (KSA) framework (Rosandic 2005, 2007) is based on the use of a short sequence of nucleotides, referred to as key string, which cuts a given genomic sequence at each location where the key string appears within a given genomic sequence. The ensuing KSA fragments form the KSA length array that could be compared with an array of lengths of restriction fragments resulting from hypothetical complete digestion cutting genomic sequence at recognition sites corresponding to the KSA key string.

Global Repeat Map

The GRM algorithm is an extension of KSA framework and consists of the following steps: 1) GRM-Total module: computes the frequency versus fragment length distribution for a given genomic sequence by superposing results of consecutive KSA segmentations computed for an ensemble of all 8-bp key strings (48 5 65,536 key strings) (Paar et al. 2007). In the GRM diagram, each pronounced peak corresponds to one or more repeats at that length, tandem or dispersed. GRM computation is fast and can be easily
1879

Global Repeat Map Analysis of HumanChimpanzee Divergence

Here, we use our novel robust bioinformatics tool, the Global Repeat Map (GRM [see Methods], for identication of tandem repeats and HORs in human and chimpanzee

Paar et al. doi:10.1093/molbev/msr009

MBE

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

FIG. 2 Diagrammatic outline of ve steps in the process of applying the GRM algorithm.

executed for a human chromosome using PC. 2) GRMDom module: determines the dominant key string corresponding to the fragment length for each peak in the GRM diagram from step 1. An 8-bp key string (or a group of 8-bp key strings) that gives the largest frequency for a fragment length under consideration is referred to as a dominant key string. 3) GRM-Seg module: performs segmentation of a given genomic sequence into KSA fragments using dominant key string from the step 2. Any periodic segment within the KSA length array reveals the location of repeats and provides genomic sequences of the corresponding repeat copies. 4) GRM-Cons module: aligns all sequences of repeat copies from step 3 and constructs consensus sequence. 5) NW module: computes divergence between each repeat copy from step 3 and consensus sequence from step 4 using the Needleman Wunsch (Needleman and Wunsch 1970) algorithm. Diagram outlining the main steps in the GRM process is shown in gure 2. Code for GRM modules is available upon the request to the authors. Regarding the 8-bp choice of the key string size, using an ensemble of all r-bp key strings, the average length of KSA fragments is ;4r. With increasing length of key strings, the overall frequency of large fragment lengths increases. For an
1880

ensemble of all 8-bp key strings, from computed GRM diagrams, we can identify the primary and secondary repeat units as large as 100 kb. Characteristics of GRM are 1) robustness with respect to deviations from perfect repeats, that is, substitutions, insertions, and deletions; 2) straightforward and parameter-free identication of simple repeats (tandem and dispersed), applicable to very large repeat units (as large as several kilobases); 3) straightforward identication of HORs (secondary repeats) for very large primary and/or secondary repeat units; 4) straightforward determination of consensus lengths and consensus sequences for simple repeats and HORs; and 5) modest scope of computation using a PC. The GRM method is a straightforward method to provide a GRM in a single diagram identifying all pronounced repeats in a given sequence, without any prior knowledge of the sequences structure. Once the size of the repeat is determined, GRM provides in a straightforward way the location of the corresponding repeat arrays and their precise analysis. GRM is particularly useful for precise sequence analysis because the method does not involve any averaging procedure. It is also useful that the method is rather robust with respect to sizeable substitutions and indels. Once the consensus repeat unit is determined using

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

MBE

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

FIG. 3 GRM diagrams for human and chimpanzee chromosome 1 assemblies, Build 36.3 and Build 2.1, respectively. Horizontal axis: fragment length (length of repeat unit); vertical axis: frequency of appearance. Upper panel: human GRM diagram; lower panel: chimpanzee GRM diagram. (AD) GRM for the whole chromosome (except some smaller segments in inserts). (A) Insert of enlarged segments of human and chimpanzee GRM diagrams (fragment lengths 160180 bp) for contigs NT_077389.3 and NW_001230099.1, respectively (peak 171 bpupper

1881

Paar et al. doi:10.1093/molbev/msr009

MBE
of the remaining two copies, the divergence is sizably larger, 712%.) The GRM diagram of 1,410-bp HOR consensus sequence (g. 3F) reveals its internal repeat structure. First, there is a series of equidistant peaks at multiples of 39 bp (n 39 bp, n 5 1, 2, 3. . .). This shows that the 39-bp monomer is a primary repeat unit for the 1,410-bp HOR unit (pronounced peak in g. 3F). The consensus sequence of the 39-bp primary repeat unit is displayed in supplementary table S3A (Supplementary Material online). Among the peaks at multiples of 39 bp in gure 3F, the strongest peak is at ;0.7 kb, and the second strongest is at ;0.35 kb. This reveals three levels of HOR organization: nine 39-bp primary repeat units are organized into ;0.35-kb secondary repeat unit, two ;0.35-kb secondary repeat units are organized into ;0.7-kb tertiary HOR repeat unit, and nally, two ;0.7-kb tertiary repeat units are organized into the ;1.4-kb HOR quartic repeat units. Thus, the 1,410-bp repeat unit represents the quartic HOR repeat unit (g. 4). By using the consensus sequence, we nd that the average divergence between neighboring copies is gradually decreasing with increasing level of HOR organization, from ;32% between 39-bp copies (rst levelprimary repeat unit) to ;4% between 1,410-bp HOR copies (fourth levelquartic repeat unit) (see g. 4). Such hierarchy of divergences is a signature of HOR organization. Tandem repeats are of biological interest because they represent a rapidly evolving type of DNA sequences that can show profound differences even between closely related species, as are humans and chimpanzees, and may contribute to phenotypic differences. In the present case of 36mer HOR, the whole HOR array is embedded within the human hornerin (HRNR) gene that produces hornerin protein, expressed in regenerating and psoriatic skin (Takaishi et al. 2005). Comparing with positions within genomic sequence in the NCBI database, this HOR array is positioned within one lengthy exon in HRNR gene (g. 5). Using the structure of human hornerin protein, Takaishi et al. (2005) have deduced the corresponding amino acid sequence. The repetitive region was divided into segments. The smallest units of ;39-bp amino acids showed a moderate homology to each other. However, Takaishi et al. (2005) concluded that the analysis was compatible with the notion that unit of ;39 bp was at rst amplied 4-fold and then triplicated to form three segments in tandem, these being further amplied 6-fold, implying a {[(39 bp) 4] 3} 6 5 2.8-kb organization, which differs from the HOR organization {[(39 bp) 9] 2} 2 5 1.4 kb found here applying GRM to the Build 36.3 genomic assembly. In the corresponding GRM diagram (g. 3E), we clearly see

GRM, in the next step it could be well combined with basic local alignment search tool (Blast) search for dispersed units or their fragments. For very large repeat units, Tandem Repeat Finder has limitations, whereas GRM has no such size limitations.

Results and Discussion

GRM Diagrams and Repeat Patterns for Human and Chimpanzee Chromosomes 1
We compute the GRM diagrams for human chromosome 1 (Build 36.3) and chimpanzee chromosome 1 (Build 2.1) (g. 3). The GRM analysis of pronounced peaks leads to the identication of pronounced repeats with large repeat units (table 1, supplementary tables S1 and S2, Supplementary Material online). Of particular interest are three human HORs that distinguish humans from chimpanzees: 1) human 1,410-bp 36mer HOR based on 39-bp primary repeat unit, without a chimpanzee counterpart; 2) human 1,866-bp 11mer alphoid HOR based on ;171-bp alpha satellite primary repeat unit, with a chimpanzee counterpart 340-bp 2mer alphoid HOR; and 3) human ;4,770-bp 3mer HOR based on ;1.6-kb primary repeat unit, without the chimpanzee counterpart. Furthermore, using GRM, we nd eight tandem repeat units and some peaks that correspond to dispersed repeats (Alu, long interspersed nuclear elements [LINE]). In table 1, we compare human and chimpanzee GRM results and previous identication of repeats in human chromosome 1 (Warburton et al. 2008). We nd a number of GRM peaks corresponding to novel tandem repeat units. There is a signicant difference in the repeat pattern of human and chimpanzee chromosomes 1 (g. 3). The lengths of sequenced chromosome 1 assemblies of human Build 36.3 and chimpanzee Build 2.1 are similar, which is reected in a similar level of noise in the upper and lower panels of gure 3.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

Human-Specic 1,410-bp Quartic HOR Repeat Embedded in Human Hornerin Gene

In the GRM diagram for human chromosome 1 (g. 3), we identify ve copies of the 1,410-bp HOR repeat unit (start position in chromosome 1: 150452728, start position in contig NT_00487.18: 2676458) and determine the consensus HOR unit (supplementary table S3A, Supplementary Material online). The average divergence of HOR copies with respect to consensus is ;4%. (These ve HOR copies appear as two subsets: average divergence among the rst three copies is ;1%, whereas for pairs involving one or two

panel and peaks 169 and 171 bplower panel). (B) Upper panel: peak at 1,866 bp: 11mer alphoid HOR in the centromeric region. Peak at 2,241 bp: tandem of 17 2,241-bp copies (divergence with respect to consensus ,1%). Peak at 1,410 bp: tandem of ve copies. Insert of enlarged segment of chimpanzee GRM diagram (fragment lengths 9001,100 bp) for contig NW_001229599.1. (C): Peaks at 4,770 bp and 3,183, 3,193 bp correspond to 3mer HOR and its 2mer subset, respectively (see the text). Peak at 3,731 bp: tandem of two copies (consensus length 3,731 bp, divergence 0.01%). Peak at 7,380: tandem of four copies (divergence ,0.5%). (E) GRM diagram for genomic segment 02.7 Mb in human contig NT_113799.1. (F) GRM diagram for 1,410-bp human consensus HOR.

1882

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

Table 1 Large Repeat Units in Human and Chimpanzee Chromosome 1: GRM Computation and Previous Identication for Humans and GRM Computation for Chimpanzees.
GRM Human 135 bp, Signature of Alu 166 bp, Signature of Alu ;171 bp, PRU;171 bp (alpha satellite, div. ;20%) ;310 bp, Signature of Alus in tandem 342 bp, 2 3 171 PRU 513 bp, 3 3 171 PRU 684 bp, 4 3 171 PRU 972 bp, PRU 5 972 (TA, 11 copies, div. ;5%) 1,410 bp, HOR 36mer (TA, 5.5 copies, div. ;3%) (QRU;1,410 bp, TRU;0.7 kb, SRU;0.35 kb, PRU39 bp) ;1,600 bp, PRU;1.6 kb (5 families, 165 copies in 13 arrays) 1,866 bp, HOR 11mer alphoid (SRU 5 1,866 bp, PRU 5 171 bp, div. ;4%) 2,241 bp, PRU 5 2,241 bp (TA, 17 copies, div ;0.4%) 3,1833,193 bp, 2 31.6-kb PRU ;3,416 bp, PRU (TA 3.25 copies, div. ;1%) ;3,732 bp, 2 3 1,866 bp 4,7204,770 bp, HOR 3mer (SRU 5 4,770 bp, PRU;1.6 kb) Previous Identicationa Human GRM Chimpanzee 135 bp, Signature of Alu 166 bp, Signature of Alu 169 bp, PRUalpha satellite 171 bp, PRUalpha satellite ;310 bp, Signature of Alus in tandem 340 bp, HOR 2mer (SRU 5 340 bp, PRU 5 169 1 171) 680 bp, 2 3 340 PRU 972 bp, PRU (TA, 9.8 copies in two arrays, div. ;5%)

;171 bp

342 bp, 2 3 171 PRU 513 bp, 3 3 171 PRU 684 bp, 4 3 171PRU 972, PRU 5 972 bp (TA, 11 copies) ;1.4 kb (TA, 5.6 copies) ;1.5 kb (84 copies in 3 arrays, div. <5%) 1,866 bp, HOR 11mer alphoid (SRU 5 1,866 bp, PRU 5 171 bp) 2.5 kb (TA, 16 copies, div. <1%)

3.4 kb (TA, four copies, div. <2%)

3,229 bp, PRU (TA, two copies, div. ;1%) 3,451 bp, PRU (TA, two copies, div. ;1%)

4,972 bp, PRU (TA, two copies, div. ;1%) 5,408 bp, PRU (TA, two copies, div. ;2%) 6,2776,305 bp (TA, ;13 copies in 6 arrays, div. ;5%) 6,461 bp, PRU (TA, two copies, div. ;2%) 7,380 bp, PRU 5 7,380 (TA, 4.4 copies, div.;0.5%) 7.4 kb (TA, 4.5 copies, div. <1%) 9,734 bp, two 7,295-bp dispersed repeats (div. ;1%) 10,828 bp, PRU (TA, two copies, div. ;5%) 10.1 (TA, three copies, div.<9.7%) 11,602 bp, two ;7-kb dispersed repeats (div. ;1%) 13,255 bp, two 3,259-bp dispersed repeats (div. ;1%) 13,745 14,264 14,618 17,095 18,555 bp, bp, bp, bp, bp, two ;6.1-kb dispersed LINE1 copies (div. ;5%) two ;6-kb dispersed parts of L1HS (div. ;5%) two ;4.7-kb dispersed parts of L1HS (div. ;5%) PRU 18,555 bp (TA, three copies, div. ;2%)b PRU 18,555 bp (TA, three copies, div. ;2%)b

18.6 kb (TA, three copies, div. <3%) 38.8 kb (TA, ve copies)

NOTE.PRU, primary repeat unit; SRU, secondary repeat unit; TRU, tertiary repeat unit; QRU, quartic repeat unit; TA, tandem array. a Previously reported results for large repeats in chromosome 1 are from tables 1 and 2 in Warburton et al. (2008). b Both the 18,555- and 17,095-bp peaks are generated by a tandem of three ;18.6-bp copies. In the third copy, the rst ;1.5-bp bases are deleted, and therefore, the third copy is shortened to 7,095 bp. Thus, the rst and second copies generate the peak at 18,555 bp, whereas the second and third are at 17,095 bp.

MBE

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

1883

Paar et al. doi:10.1093/molbev/msr009

MBE

FIG. 4 Schematic illustrating hierarchical structure of 1,410-bp quartic HOR. Nine 39-bp primary repeat units denoted m01,. . .,m09 (divergence ;30%) (bottom) form a ;0.35-kb secondary repeat unit; two ;0.35-kb secondary repeat units denoted n01 and n02 (divergence ;21%) form a ;0.7-kb tertiary repeat unit; two ;0.7-kb tertiary repeat units denoted v01 and v02 (divergence ;14%) form ;1.4-kb quartic HOR repeat unit (HOR copies average divergence ;4%) (top). In chimpanzee chromosome 1, such HOR organization is absent.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

GRM peaks at 39, ;0.35, ; 0.7, and ; 1.4-kb (table 2), in accordance with our annotation of 1.4-kb quartic HOR, whereas there is no peak at 2.8 kb predicted by Takaishi et al. (2005). By using Tandem Repeat Finder, a tandem repeat unit of 1.4 kb was previously detected (Warburton et al. 2008), but the HOR structure was not found. Because the 1.4-kb repeat is within a coding region (g. 5), it is of evolutionary interest to compare the resultant nucleotide structure of this highly repetitive gene, that is, how variation between repeat copies affects the protein structure. Using the EMBOSS Transeq code (Williams 2002) for translating nucleotide into protein sequence, we nd that the variation within the genomic sequence mostly causes nonsynonymous changes, predicting a variable protein domain structure (see supplementary table S3B, Supplementary Material online). In chimpanzee GRM diagram, we nd no counterparts of 1.4-kb human HOR.

Human 11mer and Chimpanzee 2mer Alphoid HORs

For 11mer alphoid HOR in chromosome 1 (Waye et al. 1987), the consensus length obtained using the GRM algorithm is 1,866 bp (g. 3), in accordance with the precise value from previous bioinformatics studies (Paar et al. et al. 2006; Warburton et al. 2008). The array 2005; Rosandic of alpha satellite monomers constituting HOR gives rise to

GRM peaks at multiples of alpha satellite monomer length 171:171 bp, 2 171 5 342 bp, 3 171 5 513 bp,. . . with decreasing frequency (g. 3). The next higher alphoid GRM fragment length corresponds to a tandem of two 11mer HOR copies, that is, 2 1,866 bp 5 3,732 bp (table 1). For the chimpanzee chromosome 1, we nd a different alphoid HOR pattern: Instead of the 11mer HOR, we obtain the 2mer HOR. This 2mer HOR is based on two alpha satellite monomers of the lengths 169 and 171 bp (mutual divergence ;25%). Thus, the length of the 2mer HOR unit is 340 bp; this is the length of a peak in the chimpanzee GRM diagram. Accordingly, the next pronounced alphoid GRM fragment length corresponds to a tandem of two 2mer HOR copies (mutual divergence ,1%), that is, 2 340 5 680 bp, whereas there is no pronounced peak at ;3 171 5 513 bp. We compare the alphoid HOR units of human chromosome 1 (11mer) and chimpanzee chromosome 1 (2mer) (supplementary table S4, Supplementary Material online); the average divergence between the 169and 171-bp chimpanzee monomers and m01m11 human monomers is 27%. The present result for 2mer HOR in chimpanzee is in accordance with previous observations that some early primate alpha satellites have a dimeric organization (Alexandrov et al. 2001; Schueler and Sullivan 2006). It is not surprising that the alphoid HORs from human and chimpanzee chromosomes 1 are different, because different alphoid HORs can

FIG. 5 Human 1,410-bp HOR (ve copies) embedded within an exon in the human hornerin gene. Upper strip: tandem array of ve HOR copies w01, w02, w03, w04, and w05 (start and end positions within chromosome 1 are given). Lower strip: exon (shadowed) and intron (blank) positions in Human Hornerine Gene. HOR array is fully embedded within one exon.

1884

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

Table 2 Scheme of Quartic HOR Organization Based on 39-bp Primary Repeat Unit Embedded within Human Hornerin Gene in Human Chromosome 1.
Repeat Unit Primary Secondary Tertiary Quartic Structure Monomer 9 3 39 bp 2 3 0.35 kb 2 3 0.7 kb Length 39 bp ;0.35 kb ;0.7 kb ;1.4 kb

MBE

be found on homologous chromosomes on these species, representing a rapid expansion since speciation.

Human-Specic 4,770-bp 3mer HOR in NBPF Genes in NT_113799.1

Performing GRM computation for Build 36.3 genomic assembly of human chromosome 1, we obtain an interesting new result: The two GRM peaks, at 4,770 and 3,193 bp, correspond to HORs based on the ;1.6-kb primary repeat unit in the NBPF genes. The largest contribution to the 4,770-bp peak arises from the contig NT_113799.1 in chromosome 1. The GRM diagram for the 4,770-bp consensus sequence shows two peaks, at ;1.6 and ;3.2 kb (g. 6A). In a further step, the GRM diagram computed for the 3.2-kb peak shows only one pronounced peak, at ;1.6-kb (g. 6B). This shows that the ;1.6-kb peak corresponds to the primary repeat unit for the 4,770- and 3,193-bp peaks. Using GRM, we determine a dominant key string CACTGACC for segmentation of the contig NT_113799.1 into a maximum number of the ;1.6-kb fragments. Using this key string, we obtain an array of KSA fragment lengths (supplementary table S5, Supplementary Material online). Aligning the resulting length arrays, we see that the three ;1.6-kb fragments, representing monomers, form a tandemly repeating 3mer: It consists of three ;1.6-kb NBPF monomers, belonging to three monomer families, with consensus lengths 1,623, 1,593, and 1,554 bp, respectively. Consensus length for each family is determined as the most frequent length for primary repeat monomers belonging to that family. Explicit relations between consensus sequence and copies are illustrated in Supplementary gure S13, Supplementary Material online. The corresponding basic consensus monomers are denoted as m01, m02, and m03, respectively (see consensus sequences in supplementary table S6, Supplementary Material online). Finally, the computed GRM diagram for each of three ;1.6kb basic monomers is without any pronounced peak, reecting its monomer character, that is, the absence of any internal repeat structure. The broadened peak consisting of a cluster of scattered ;1.6-kb lengths in the human GRM diagram (g. 3B) is due to cases with the presence of the same key string in both m01 and m02 and/or in both m02 and m03 within the 3mer m01m02m03. A peak at ;3.2 kb appears in the case when the same key string is present in m01 and m03 but not in m02. Table 3 shows the detailed NBPF monomer structure of the 3mer HOR. It is seen that the m03 monomer is deleted from 3mer HOR copies no. 68. This mono-

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

FIG. 6 (A) GRM diagram of the 4,770-bp 3mer HOR consensus sequence (m01m02m03). (B) GRM diagram of the 3.2-kb 2mer HOR consensus sequence (one monomer from 3mer HOR is missing).

mer deletion, resulting in the appearance of NBPF dimers within the NBPF array, contributes to the ;3.2-kb peak in the GRM diagram. Divergence between the three consensus monomers is 19% (m01 vs. m02), 15% (m01 vs. m03) and 20% (m02 vs. m03). On the other hand, the average divergence between the 3mer HOR copies is mostly ,0.5%, that is, the 3mer copies are highly identical. This substantially smaller divergence between HOR copies than between constituting monomers is a signature of well-developed HOR pattern (see Introduction). Thus, the 4,770-bp repeat unit corresponds to strongly convergent HORs based on three diverging monomers m01, m02, and m03. This HOR array consists of a tandem of 17 highly identical HOR copies (in the interval in chromosome 1: 146618386146694662). (We nd no NBPF monomers outside the HOR array.) This HOR is fully embedded within the NBPF20 gene (NM_001037675.2). (Note that the Build 36.3 sequence is reverse complement to the sequence of NBPF20 gene from NCBI database.) All 17 NBPF HOR copies (horizontal stretches) are displayed by aligning the constituent NBPF monomers and the corresponding regions of exons are shown below HOR copies (vertical bars; g. 7). An example of divergence between segments at position of exons in the 15th and 16th HOR copies and between the corresponding protein sequences is shown in supplementary table S9 (Supplementary Material online). The 4,770-bp 3mer HOR was not reported previously. However, the self-similarity dot plot of the NBPF region can also detect the presence of this HOR.

Human-Specic 4,770-bp 3mer HOR in NBPF Genes in other Contigs in Chromosome 1

We perform a similar search and analysis of the 4,770-bp 3mer HORs in other contigs in human chromosome 1
1885

Paar et al. doi:10.1093/molbev/msr009

MBE
(supplementary table S7, Supplementary Material online). Most of the NBPF repeats in other contigs are organized in HORs; at the edges of NBPF regions, increased divergence and/or truncation appears in HOR copies. HOR arrays or their segments are present in 14 distinct arrays in 10 contigs (supplementary table S8, Supplementary Material online). Only three NBPF HOR tandem arrays are of signicant size, located in contigs NT_113799.1, NT_079497.3, and NT_004434.18: They contain 17, 16, and 12 NBPF HOR copies, respectively (g. 8). All three HOR arrays are highly similar: The average divergence between 17 HOR copies in NT_113799.1 and 16 HOR copies in NT_079497.3 is 0.6%, and between 17 HOR copies in NT_113799.1 and 12 HOR copies in NT_004434.18, the average divergence is 2%.
Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

Absence of NBPF HORs in Nonhuman Primates

To search for NBPF HORs in nonhuman primate genomes, we used human NBPF consensus monomers m01, m02, m03 (supplementary table S6, Supplementary Material online) for Blast (Altschul et al. 1990) search against the chimpanzee, orangutan, and rhesus macaque genomic sequences (supplementary tables S10S12, Supplementary Material online). The arrays of NBPF repeats in chimpanzees and other nonhuman primates are not organized into HORs. This is shown by the absence of pronounced GRM peaks at ;4,770 bp (g. 3).

Human and Chimpanzee GRM Diagrams and Tandem Repeats without Higher Order Organization
For most of monomeric tandem repeats, there is a sizeable difference between human and chimpanzee chromosomes 1 (see table 1). For example, pronounced repeat units at 2,241, 7,380, and 18,555 bp in human chromosome 1 are absent in chimpanzee chromosome 1. On the other hand, a sizeable 9,734-bp GRM peak in chimpanzee chromosome 1 is absent in human chromosome 1. The exception is, for example, a tandem based on the 972-bp repeat unit that is present in both humans and chimpanzees. The corresponding chimpanzee peak in an overall GRM diagram of the whole chromosome 1 is weak, almost immersed into noise (weak peak in the lower panel of g. 3B). To reduce the noise, we additionally compute the GRM diagram for the contig NW_001229599.1 that contains tandem array based on the 972-bp repeat unit.
genomic sequence that are .98% identical to the sequences corresponding to exons, but belong to NCBI introns, are denoted a, b,. . .,h. Note that the Build 36.3 sequence is reverse complement with respect to the NBPF20 gene from NCBI. For each HOR copy, the start position is shown (0 is assigned to the start of the rst HOR copy). (Position zero corresponds to the position 146618386 in chromosome 1.) The NBPF gene starts at the position of the rst HOR copy (top) and extends beyond the last HOR copy in the gure (bottom). Between the end of the last HOR copy and the end of NBPF20 gene, there are a number of additional exons (not shown in gure).

FIG. 7 Aligned human 4,770-bp 3mer HOR copies in human NBPF20 gene from contig NT_113799.1. Each horizontal stretch represents a 3mer HOR copy with its three constituting monomers m01, m02, and m03. Below a stretch the corresponding regions of exons in the NBPF gene are shown (vertical bars). Exons reported in the NCBI database are denoted a, b,. . ., h. The segments from Build 36.3

1886

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

Table 3 NBPF Monomer Repeat Structure of 3mer HORs in Contig NT_113799.1 in Chromosome 1.
Consensus Monomera m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02g m13h m02g m13h m02g m13h m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03 m01 m02 m03
a

MBE
HOR Lengthd (bp) HOR Copy No.e HOR Divergencef (%)

Ch1 Start Position 146618386 146619994 146621581 146623135 146624758 146626345 146627899 146629522 146631105 146632659 146634282 146635877 146637431 146639056 146640649 146642195 146643820 146645411 146647011 146648604 146650204 146651797 146653397 146654990 146656544 146658167 146659760 146661314 146662937 146664530 146666084 146667707 146669290 146670844 146672467 146674060 146675618 146677241 146678826 146680380 146682003 146683594 146685152 146686775 146688346 146689900 146691523 146693116

Contig Start Position 75723 77331 78918 80472 82095 83682 85236 86859 88442 89996 91619 93214 94768 96393 97986 99532 101157 102748 104348 105941 107541 109134 110734 112327 113881 115504 117097 118651 120274 121867 123421 125044 126627 128181 129804 131397 132955 134578 136163 137717 139340 140931 142489 144112 145683 147237 148860 150453

Monomer Lengthb (bp) 1,608 1,587 1,554 1,623 1,587 1,554 1,623 1,583 1,554 1,623 1,595 1,554 1,625 1,593 1,546 1,625 1591 1,600 1,593 1,600 1,593 1,600 1,593 1,554 1,623 1,593 1,554 1,623 1,593 1,554 1,623 1,583 1,554 1,623 1,593 1,558 1,623 1,585 1,554 1,623 1,591 1,558 1,623 1,571 1,554 1,623 1,593 1,546

Monomer Divergencec (%) 8.9 0.7 0.1 0.1 0.6 0.1 0.1 0.8 0.1 0 0.2 0.1 0.3 0 0.8 0.1 0.2 0.1 0.1 0 0.1 0 0.2 0.1 0.1 0.2 0.1 0.1 0.2 0.1 0.1 0.8 0.1 0.1 0.3 0.3 0.1 0.7 0 0.1 0.2 0.3 0.1 1.5 0.1 0.1 0.1 1

4,749

3,3

4,764

0,3

4,760

0,3

4,772

0,1
Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

4,764 3,216 3,193 3,193

5 6 7 8

0,4 0,1 0,1 0

4,747

0,1

4,770

10

0,1

4,770

11

0,1

4,760

12

0,3

4,774

13

0,2

4,762

14

0,3

4,772

15

0,2

4,748

16

0,6

4,762

17

0,4

Consensus monomers m01 (1,623 bp), m02 (1,593 bp), and m03 (1,554 bp) (see supplementary table S2, Supplementary Material online). b Length of monomer copy. c Divergence of monomer copy with respect to consensus monomer. d Length of HOR copy (consensus length 4,770 bp). e Ordinal number of HOR copy in tandem array. f Divergence of HOR copy with respect to consensus HOR. g Monomer m03 in HOR copies No. 68 is missing. h Monomer m01 in HOR copies no. 79 is substituted by hybrid m13 (1,600 bp). Divergence of m13 with respect to consensus monomers m01, m02, and m03 is 7%, 22%, and 9%, respectively. The m13 sequence exhibits a specic hybrid structure: the rst 845 bases in m13 are identical to the rst 841 bases from m03 (additionally, a 4-bp insertion AGAG is inserted into m03 after position 581), whereas the last 821-bp segment in m13 is identical to the last 821-bp segment from m01. Divergence between m13 and m01 (7%) is smaller than divergence between m01 and m03 but larger than the mutual divergence between different copies of m01 or of m03 (,0.5%). Considering the three 1,600-bp copies as belonging to the m01 family, divergence of the corresponding three HOR copies with respect to consensus HOR is still ,5%, whereas for the m13 classication of these three monomer copies, divergence between HOR copies is reduced to 0.5%.

1887

Paar et al. doi:10.1093/molbev/msr009

MBE

FIG. 8 Schematic illustrating three NBPF 3mer HOR copies based on the ;1.6-bp monomers in human chromosome 1. In chimpanzee chromosome 1, this HOR organization is absent.

In this case with a shorter sequence, the noise level at the position of the 972-bp peak is reduced (as seen from insert in the lower panel in g. 3B). For short repeat units ,20 bp, the corresponding human and chimpanzee GRM peaks are similar. In the length interval from 20 to 100 bp, the most pronounced peak in the human GRM diagram is at 39 and 40 bp, and in chimpanzee, it is at 32 bp. We nd weaker peaks at multiples of 39 bp (78, 117 bp,. . .) and 40 bp (80, 120 bp,. . .) in human and at multiples of 32 bp (64, 96 bp,. . .) in chimpanzee GRM diagrams. The human 39-bp repeat unit is the primary repeat unit for the ;1,410-bp HOR, and 40 bp is a tandem repeat unit of another array.

two Alu sequences; because of the variable length of the poly-A tail, this GRM peak is broadened in the interval of about a dozen base pairs, centered at ;310 bp.
Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

Statistical Signicance of Peaks in the GRM Diagram

An important question can be raised about the statistical signicance of the peaks appearing in the GRM diagram, because they arise by applying an ensemble of a large number (48 5 65,536) of key strings. We use three strong tests of this signicance. First, we compute the corresponding GRM diagrams for articial sequences, constructed using a random number generator, of the same nucleotide composition as in the genomic sequences under study. We compute the GRM diagram for a pseudochromosome 1, constructed in such a way (g. 9). This randomly constructed sequence from a given number of nucleotides does not produce any signicant peak above the level of noise in the GRM diagram. The second test of statistical signicance is provided by the steps of the GRM algorithm (see Methods): For each length at position of a peak, we determine the dominant key string and determine the corresponding repeat array giving rise to this peak. In this way, we nd explicitly that to each GRM peak corresponds a concrete repeat pattern. The third test of statistical signicance is the identication of previously known HORs in human chromosomes (e.g., alphoid 11mer HOR in chromosome 1, alphoid 18mer HOR in chromosome 7, and so on). In this way, we identify GRM peaks for all previously known human alphoid HORs. Additionally, for the Build 36.3 assembly of human chromosome 1, we nd peaks corresponding to all repeats previously identied in human chromosome 1 using standard bioinformatics tools (Warburton et al. 2008) and some novel peaks reported for the rst time in this work.

GRM Signature of Alu Sequences

The pronounced GRM peaks at 135, 166, and ;310 bp in the GRM diagram (g. 3) are signature of Alu sequences. Human Alu elements are dimeric structures of about 300 bp that comprise two similar but nonidentical monomers (Weiner et al. 1986; Britten et al. 1988; Jurka 1995, 2004; Schmid 1996). The longer monomer (on the right) contains an insert of 31 nt that is absent in the shorter monomer (on the left). The two monomers are connected by a short internal poly-A tract. The dimer sequence is followed by a long terminal poly-A tail of variable length. The 135-bp GRM fragment length corresponds to a distance between the starts of the left and right monomers. The 166-bp GRM fragment corresponds to the distance from the start of the segment of right monomer after the 31bp insert and the start of a similar segment in the left monomer (thus, this distance is 135 31 bp 5 166 bp). The ;310-bp length corresponds to a tandem of

Human Accelerated NBPF and HRNR HORsComponents of Genomic Regulatory Network?

Previously, it was found that the number of NBPF monomer repeats is related to the evolutionary level of higher primates, showing a gradual increase with evolutionary development (Vandepoele et al. 2005, 2009; Popesco et al. 2006), but the NBPF HOR pattern was not identied. Our novel GRM results for human chromosome 1 show that the ;1.6-kb NBPF monomers, mutually diverging 2025%,

FIG. 9 GRM diagram for pseudochromosome 1 generated by a random number generator.

1888

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

Table 4 Evolutionary Acceleration of NBPF Monomer and HOR Pattern.
Human Chimpanzeee Orangutanf Rhesus macaqueg Mouseh
a b

MBE
No. of Tandem HOR Copiesc 47 0 0 0 0

Monomer Copiesa 165 48 17 7 0

Total No. Of HOR Copiesb 57 14 7 2 0

Total number of NBPF monomers. Number of all NBPF HOR copies (both tandemly organized and dispersed). c Total number of NBPF HOR copies organized in tandem repeats (two or more HOR copies in tandem array). d Genomic assembly NCBI Build 36.3. e Genomic assembly NCBI Build 2.1. Most of the contigs containing NBPF repeats are in chromosome 1, whereas eight monomers and two distinct HOR copies are in contigs not yet assigned to the chromosome. f Genomic assembly WUSTL Pongo_albelii-2.0.2. All contigs containing NBPF repeats are in chromosome 1. g Genomic assembly NCBI Build 1.1. Most of the contigs containing NBPF repeats are in chromosome 1, whereas the one monomer is in a contig not assigned to the chromosome. h Genomic assembly NCBI Build 2.1.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

are in fact organized into highly identical 3mer HORs (4,770-bp consensus). With respect to previously known HORs in the human genome, this HOR pattern is interesting in several aspects: 1) The repeat unit is much longer than most of the primary repeat units identied so far. 2) Most of the NBPF repeat monomers are organized into 3mer HORs. 3) There are three distinct major NBPF HOR arrays, having practically the same consensus sequence. 4) NBPF HOR is fully embedded within a gene. Because the NBPF genes might suppress the development of neuroblastoma and possibly other tumor types (Vandepoele et al. 2005, 2008, 2009; Gregory et al. 2006; Popesco et al. 2006), there is an interesting question on a possible relation of tumor dynamics to this peculiar 4,770-bp HOR pattern. Comparison of the appearance of NBPF monomers and HOR copies and of tandem repeats of NBPF HOR copies in chromosome 1 of humans and some nonhuman vertebrates is shown in table 4. The tandem repeat of the NBPF HOR copy pattern shows a discontinuous jump in the evolutionary step from chimpanzee to the human genome: from total absence of tandem repeats of NBPF HOR copies in chimpanzees to 47 tandem repeat HOR copies in humans. This human accelerated HOR pattern (referred to as HAHOR) is one of the factors that distinguish humans from nonhuman primates. Another case of HOR being fully embedded within a gene is a 1.4-kb quartic HOR based on 39-bp primary repeat unit. This HOR is fully embedded within a single large exon in the human hornerin gene (HRNR), related to human skin. It is of substantial interest evolutionary to compare the resultant amino acid structure of this highly repetitive gene. We show that the presence of NBPF HOR and HRNR HOR tandem repeats is an exclusively human specic. We hypothesize that this may be related to the possible regulatory role of the HOR pattern. In both of these within-the gene HORs, we nd no HOR counterparts in the chimpanzee genome: These HORs represent a human accelerated pattern. It seems of interest to look systematically for higher order repeats (HORs), particularly in or near brain-related and other genes and for their possible regulatory function. More generally, it is an intriguing question related to a possible genomic regulatory network (GRN) with vari-

ous components including known gene regulators and as yet hidden regulatory elements at different levels of coordination and hierarchical organization. In this sense, it is challenging whether some types of accelerated sequence may be functional, acting dynamically within a network of genes, gene regulators, and higher order GRN components. Small differences among repeat copies, even when these copies exhibit a high degree of mutual similarity, might have a signicant impact on the function of the multilayer regulatory network (circuitry). Their inuence on gene expression may be direct, as in the case of some known HAR regulators, or indirect along a hierarchy of possible higher level organization. In this connection, it should be noted that growing evidence from the eld of developmental genetics is supporting the hypothesis that small differences in the time of activation or in the level of activity of even a single gene could have important evolutionary consequences owing to the extensive consequences that changes in regulatory interactions could have on developmental processes (Gareld and Wray 2010). In general, one could argue that because any vanishingly small input applied at critical points in nonlinear dynamical systems might cause a nite response (Otto 1993; Hilborn 1994; Alligood et al. 1997; Zak et al. 1997), a system could be controlled by a microdynamical device that operates by sign strings, as a genetic code. Tandem repeats and HORs are the expected result of rapid expansion, given models of unequal crossing-over. They are of interest in the framework of a large-scale regulatory architecture of genome as a rapidly evolving type of DNA sequence that can show a profound difference between closely related species and may contribute to phenotypic differences. This may shed more light on humanchimpanzee evolutionary divergence from the point of view of long-range higher order periodicity.

Supplementary Material
Supplementary tables S1S12 and Supplementary gure S13 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
1889

Paar et al. doi:10.1093/molbev/msr009

MBE
Dermitzakis ET, Reymond A, Lyle R, et al. (11 co-authors). 2002. Numerous potentially functional but non-genic conserved sequences on human chromosome 21. Nature 420:578582. Diskin SJ, Hou C, Glessner JT, et al. (27 co-authors). 2009. Copy number variation at 1q21.1 associated with neuroblastoma. Nature 459:987991. Dorus S, Vallender EJ, Evans PD, Anderson JR, Gilbet SL, Mahowald M, Wyckoff GJ, Malcom CM, Lahn BT. 2004. Accelerated evolution of nervous system genes in the origin of Homo sapiens. Cell 119:10271040. Duret L, Galtier N. 2009. Biased gene conversion and the evolution of mammalian genomic landscapes. Annu Rev Genom Hum Genet. 10:285311. Enard W, Gehre S, Hammerschmidt K, et al. (56 co-authors). 2009. A humanized version of Foxp2 affects cortico-basal ganglia circuits in mice. Cell 137:961971. Enard W, Paabo S. 2004. Comparative primate genomics. Annu Rev Genom Hum Genet. 5:351378. Evans PD, Vallender EJ, Lahn BT. 2006. Molecular evolution of the brain size regulatory genes CDK5RAP2 and CENPJ. Gene 375:7579. Gareld DA, Wray GA. 2010. The evolution of gene regulatory interactions. BioScience 60:1523. Gelfand Y, Rodriguez A, Benson G. 2007. TRDBthe tandem repeats database. Nucleic Acids Res. 35:D80D87. Gregory SG, Barlow KF, McLay KE, et al. (178 co-authors). 2006. The DNA sequence and biological annotation of human chromosome 1. Nature. 441:315321. Gu JY, Gu X. 2003. Induced gene expression in human brain after the split from chimpanzee. Trends Genet. 19:6365. Hayakawa T, Altheide TK, Varki A. 2005. Genetic basis of human brain evolution: accelerating along the primate speedway. Dev Cell. 8:24. Haygood R, Babbitt CC, Fedrigo O, Wray GA. 2010. Contrasts between adaptive coding and noncoding changes during human evolution. Proc Natl Acad Sci U S A. 107:78537857. Haygood R, Fedrigo O, Hanson B, Yokoyama KD, Wray GA. 2007. Promoter regions of many neural- and nutrition-related genes have experienced positive selection during human evolution. Nat Genet. 39:11401144. Heissig F, Krause J, Bryk J, Khaitovich P, Enard W, Paabo S. 2005. Functional analysis of human and chimpanzee promoters. Genome Biol. 6:R57. Hilborn RC. 1994. Chaos and nonlinear dynamics. Oxford: Oxford University Press. Huttenhofer A, Schattner P, Polacek N. 2005. Non-coding RNAs: hope or hype? Trends Genet. 21:289297. Jacob F, Monod J. 1961. Genetic regulatory mechanisms in synthesis of proteins. J Mol Biol. 3:318356. Jerison H. 1975. Evolution of brain and intelligence. New York: Academic Press. Johnson JM, Edwards S, Shoemaker D, Schadt EE. 2005. Dark matter in the genome: evidence of widespread transcription detected by microarray tiling experiments. Trends Genet. 21: 93102. Jurka J. 1995. Origin and evolution of Alu repetitive elements. In: Maraia RJ, editor. The impact of short interspersed elements (SINEs) on the host genome. RG Landes, Austin, Texas: Springer. p. 2542. Jurka J. 2000. Repbase update: a database and an electronic journal of repetitive elements. Trends Genet. 9:418420. Jurka J. 2004. Evolutionary impact of human Alu repetitive elements. Curr Opin Genet Dev. 14:603608. Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J. 2005. Repbase update, a database of eukaryotic repetitive elements. Cytogenet Genome Res. 110:462467.

Acknowledgments
We thank two anonymous reviewers for valuable comments and suggestions, contributing signicantly to the , J. quality of our manuscript. We also thank D. Ugarkovic Brajkovic, and K. Vlahovicek for useful discussion.

References
Alexandrov IA, Kazakov A, Tumeneva I, Shepelev V, Yurov Y. 2001. Alpha-satellite DNA of primates: old and new families. Chromosoma 110:253266. Alkan C, Eichler EE, Bailey JA, Sahinalp SC, Tuzun E. 2004. The role of unequal crossover in alpha-satellite DNA evolution: a computational analysis. J Comp Biol. 11:933944. Alligood KT, Sauer TD, Yorke JA. 1997. Chaos. An introduction to dynamical systems. New York: Springer. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. J Mol Biol. 215:403410. Amadio JP, Walsh CA. 2006. Brain evolution and uniqueness in the human genome. Cell 126:10331035. Amaral PP, Dinger ME, Mercer TR, Mattick JS. 2008. The eukaryotic genome as an RNA machine. Science 319:17871789. Bejerano G, Lowe CB, Ahituv N, King B, Siepel A, Salama SR, Rubin EM, Kent WJ, Haussler D. 2006. A distal enhancer and an ultraconserved exon are derived from a novel retroposon. Nature 441:8790. Bejerano G, Pheasant M, Makunin I, Stephen S, Kent WJ, Mattick JS, Haussler D. 2004. Ultraconserved elements in the human genome. Science 304:13211325. Beniaminov A, Westhof E, Krol A. 2009. Distinctive structures between chimpanzee and human in a brain noncoding RNA. RNA Biol. 6:113121. Benson G. 1999. Tandem repeats nder: a program to analyze DNA sequences. Nucleic Acids Res. 27:573580. Bird CP, Stranger BE, Liu M, Thomas DJ, Ingle CE, Beazley C, Miller W, Hurles ME, Dermitzakis ET. 2007. Fast-evolving noncoding sequences in the human genome. Genome Biol. 8:R118. Britten RJ, Baron WF, Stout DB, Davidson EH. 1988. Sources and evolution of human Alu repeated sequences. Proc Natl Acad Sci U S A. 85:47704774. Britten RJ, Davidson EH. 1969. Gene regulation for higher cellsa theory. Science 165:349357. Britten RJ, Kohne DE. 1968. Repeated sequences in DNA. Science 161:529540. Britten RJ, Davidson EH. 1971. Repetitive and nonrepetitive DNA sequences and a speculation on the origins of evolutionary novelty. Quart Rev Biol. 46:111138. Bush EC, Lahn BT. 2008. A genome-wide screen for noncoding elements important in primate evolution. BMC Evol Biol. 8:17. Caceres M, Lachuer J, Zapala MA, Redmond JC, Kudo L, Geschwind DH, Lockhart DJ, Preuss TM, Barlow C. 2003. Elevated gene expression levels distinguish human from nonhuman primate brains. Proc Natl Acad Sci U S A. 100: 1303013035. The chimpanzee sequencing and analysis consortium. 2005. Initial sequence of the chimpanzee genome and comparison with the human genome. Nature 437:6987. Cooper GM, Sidow A. 2003. Genomic regulatory regions: insights from comparative sequence analysis. Curr Opin Genet Dev. 13:604610. Costa FF. 2007. Non-coding RNAs: lost in translation? Gene 386:110. Davidson EH, Britten RJ. 1979. Regulation of gene expression: possible role of repetitive sequences. Science 204:10521059.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

1890

Intragene HORs Distinguish Humans from Chimpanzees doi:10.1093/molbev/msr009

Kapranov P, Willingham AT, Gingeras TR. 2007. Genome-wide transcription and the implications for genomic organization. Nat Rev Genet. 8:413423. Khaitovich P, Muetzel B, She XW, et al. (15 co-authors). 2004. Regional patterns of gene expression in human and chimpanzee brains. Genome Res. 14:14621473. King DC, Taylor J, Zhang Y, Cheng Y, Lawson HA, Martin J, Chiaromonte F, Miller W, Hardison RC. 2007. Finding cisregulatory elements using comparative genomics: some lessons from ENCODE data. Genome Res. 17:775786. King MC, Wilson AC. 1975. Evolution at two levels in humans and chimpanzees. Science 188:107116. Kleinjan DA, Lettice LA. 2008. Long-range gene control and genetic disease. Adv Genet. 61:339388. Kleinjan DA, van Heyningen V. 2005. Long-range control of gene expression: emerging mechanisms and disruption in disease. Am J Hum Genet. 76:832. Lettice LA, Horikoshi T, Heaney SJH, et al. (21 co-authors). 2002. Disruption of long-range cis-acting regulator for Shh causes preaxial polydactyly. Proc Natl Acad Sci U S A. 99:75487553. Manuelidis L. 1978. Complex and simple sequences in human repeated DNAs. Chromosoma 66:121. Maston GA, Evans SK, Green MR. 2006. Transcriptional regulatory elements in the human genome. Annu Rev Genom Hum Genet. 7:2959. Mattick JS. 2009. The genetic signatures of noncoding RNAs. PLoS Genet. 5:e1000459. Mekel-Bobrov N, Posthuma D, Gilbert SL, et al. (22 co-authors). 2007. The ongoing adaptive evolution of ASPM and microcephalin is not explained by increased intelligence. Hum Mol Genet. 16:600608. Mercer TR, Dinger ME, Mattick JS. 2009. Long non-coding RNAs: insights into functions. Nat Rev Genet. 10:155159. Needleman SB, Wunsch CD. 1970. A general method applicable to the search for similarities in the amino acid sequence of two proteins. J Mol Biol. 48:443453. Nobrega MA, Ovcharenko I, Afzal V, Rubin EM. 2003. Scanning human gene deserts for long-range enhancers. Science. 302:413. Noonan JP, McCallion AS. 2010. Genomics of long-range regulatory elements. Annu Rev Genom Hum Genet. 11:123. Otto E. 1993. Chaos in dynamical systems. Cambridge: Cambridge University Press. Ovcharenko I. 2008. Widespread ultraconservation divergence in primates. Mol Biol Evol. 25:16681676. M, Glunc ic M. 2007. Consensus higher Paar V, Basar I, Rosandic order repeats and frequency of string distributions in human genome. Curr Genom. 8:93111. M, Glunc ic M, Basar I, Pezer R, DurajlijaPaar V, Pavin N, Rosandic inic S. 2005. ColorHORnovel graphical algorithm for fast scan Z of alpha satellite higher-order repeats and HOR annotation for GenBank sequence of human genome. Bioinformatics 21:846852. Pennacchio LA, Ahituv N, Moses AM, et al. (19 co-authors). 2006. In vivo enhancer analysis of human conserved non-coding sequences. Nature 444:499502. Pennacchio LA, Rubin EM. 2001. Genomic strategies to identify mammalian regulatory sequences. Nat Rev Genet. 2:100109. Pollard KS, Salama SR, King B, et al. (13 co-authors). 2006a. Forces shaping the fastest evolving regions in the human genome. PLoS Genet. 2:15991611. Pollard KS, Salama SR, Lambert N, et al. (16 co-authors). 2006b. An RNA gene expressed during cortical development evolved rapidly in humans. Nature 443:167172. Popesco MC, Maclaren EJ, Hopkins J, Dumas L, Cox M, Meltesen L, McGavran L, Wyckoff GJ, Sikela JM. 2006. Human lineage-

MBE

specic amplication, selection, and neuronal expression of DUF1220 domains. Science 313:13041307. Portin P. 2008. Evolution of man in the light of molecular genetics: a review. Part II. Regulation of gene function, evolution of speech and of brains. Hereditas 145:113125. Prabhakar S, Noonan JP, Paabo S, Rubin EM. 2006. Accelerated evolution of conserved noncoding sequences in humans. Science 314:786. Raff RA, Kaufman TC. 1983. Embryos, genes, and evolution: the developmental-genetic basis of evolutionary change. New York: Macmillan. Redon R, Ishikawa S, Fitch KR, et al. (43 co-authors). 2006. Global variation in copy number in the human genome. Nature 444:444454. M, Paar V, Basar I. 2003. Key-string segmentation Rosandic algorithm and higher-order repeat 16mer (54 copies) in human alpha satellite DNA in chromosome 7. J Theor Biol. 221:2937. M, Paar V, Basar I, Glunc ic M, Pavin N, Pilas I. 2006. CENPRosandic B box and pJa sequence distribution in human alpha satellite higher-order repeats (HOR). Chromosome Res. 14:735753. Rudd MK, Willard HF. 2004. Analysis of the centromeric regions of the human genome assembly. Trends Genet. 20:529533. Rudd MK, Wray GA, Willard HF. 2006. The evolutionary dynamics of alpha-sarellite. Genome Res. 16:8896. Schmid CW. 1996. Alu: structure, origin, evolution, signicance, and function of one-tenth of human DNA. Prog Nucleic Acid Res Mol Biol. 53:283319. Schueler MG, Sullivan BA. 2006. Structural and functional dynamics of human centromeric chromatin. Annu Rev Genomics Hum Genet. 7:301313. Shapiro JA, von Sternberg R. 2005. Why repetitive DNA is essential to genome function. Biol Rev. 80:227250. Smith GP. 1976. Evolution of repeated DNA sequences by unequal crossover. Science 191:528535. Southern EM. 1975. Long range periodicities in mouse satellite DNA. J Mol Biol. 94:5169. Stephen S, Pheasant M, Makunin IV, Mattick JS. 2009. Large-scale appearance of ultraconserved elements in tetrapod genomes and slowdown of molecular clock. Mol Biol Evol. 25:402408. Taft RJ, Pang KC, Mercer TR, Dinger M, Mattick JS. 2010. Non-coding RNAs: regulators of disease. J Pathol. 220:126139. Takaishi M, Makino T, Marohashi M, Huh N. 2005. Identication of human hornerin and its expression in regenerating and psoriatic skin. J Biol Chem. 280:46964703. Tuzun E, Sharp AJ, Bailey JA, et al. (12 co-authors). 2005. Fine-scale structural variation of the human genome. Nat Genet. 37: 727732. Tyler-Smith C, Brown WRA. 1987. Structure of the major block of alphoid satellite DNA on the human Y chromosome. J Mol Biol. 195:457470. Uddin M, Wildman DE, Liu G, Xu W, Johnson RM, Hof PR, Kapatos G, Grossman LI, Goodman M. 2004. Sister grouping of chimpanzees and humans as revealed by genome-wide phylogenetic analysis of brain gene expression proles. Proc Natl Acad Sci U S A. 101:29572962. Vallender EJ. 2008. Exploring the origins of the human brain through molecular evolution. Brain Behav Evol. 72:168177. Vallender EJ, Mekel-Bobrov N, Lahn BT. 2008. Genetic basis of human brain evolution. Trends Neurosci. 31:637644. Van Bakel H, Hughes TR. 2009. Establishing legitimacy and function in the new transcriptome. Brief Funct Genomic Proteomic. 8:424436. Vandepoele K, Andries V, Van Roy F. 2009. The NBPF1 promoter has been recruited from the unrelated EVI5 gene before simian radiation. Mol Biol Evol. 26:13211332.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

1891

Paar et al. doi:10.1093/molbev/msr009

Vandepoele K, Andries V, Van Roy N, Staes K, Vandesompele J, Laureys G, De Smet E, Berx G, Speleman F, Van Roy F. 2008. A constitutional translocation t(1;17) (p36.2;q11.2) in a neuroblastoma patient disrupts the human NBPF1 and ACCN1 genes. PLoS ONE. 3:e2207. Vandepoele K, van Roy F. 2007. Insertion of an HERV(K) LTR in the intron of NBPF3 is not required for its transcriptional activity. Virology. 362:15. Vandepoele K, Van Roy N, Staes K, Speleman F, Van Roy F. 2005. A novel gene family NBPF: intricate structure generated by gene duplications during primate evolution. Mol Biol Evol. 22: 22652274. Varki A, Geschwind DH, Eichler EE. 2008. Explaining human uniqueness: genome interactions with environment, behaviour and culture. Nat Rev Genet. 9:749763. Visel A, Akiyama JA, Shoukry M, Afzal V, Rubin EM, Pennacchio LA. 2008. Functional autonomy of distant-acting human enhancers. Genomics 93:509513. Warburton PE, Hasson D, Guillem F, Lescale C, Jin X, Abrusan G. 2008. Analysis of the largest tandemly repeated DNA families in the human genome. BMC Genomics. 9:533. Warburton PE, Willard HF. 1996. Evolution of centromeric alpha satellite DNA: molecular organization within and between human and primate chromosomes. In: Jackson M, Strachan T, Dover G, editors. Human genome evolution. Oxford: BIOS Scientic Publishers. p. 121145. Waye JS, Durfy SJ, Pinkel D, Kenwrick S, Patterson M, Davies KE, Willard HF. 1987. Chromosome-specic alpha satellite DNA from human chromosome 1: hierarchical structure and genomic organization of a polymorphic domain spanning several hundred kilobase pairs of centromeric DNA. Genomics 1: 4351. Weiner AM, Deininger PL, Efstratiaadis A. 1986. The reverse ow of genetic information: pseudogenes and transposable elements

MBE
derived from nonviral cellular RNA. Annu Rev Biochem. 55: 631661. Willard HF. 1985. Chromosome-specic organization of human alpha satellite DNA. Am J Hum Genet. 37:524532. Willard HF, Waye JS. 1987a. Chromosome-specic subsets of human alpha satellite DNA: analysis of sequence divergence within and between chromosomal arrays and evidence for an ancestral repeat. J Mol Evol. 25:207214. Willard HF, Waye JS. 1987b. Hierarchical order in chromosomespecic human alpha satellite DNA. Trends Genet. 3:192198. Williams G. 2002. EMBOSS Transeq. Available from: www.ebi.ac.uk/ Tools/emboss/transeq/index.html. Woolfe A, Elgar G. 2008. Organization of conserved elements near key developmental regulators in vertebrate genomes. Adv Genet. 61:307338. Wray GA, Babbitt CC. 2008. Enhancing gene regulation. Science 321:13001301. Wray GA, Hahn MW, Abouheif E, Balhoff JP, Pizer M, Rockman MV, Romano LA. 2003. The evolution of transcriptional regulation in eukaryotes. Mol Biol Evol. 20:13771419. Wu JC, Manuelidis L. 1980. Sequence denition and organization of a human repeated DNA. J Mol Biol. 142:363386. Xie X, Mikkelsen TS, Gnirke A, Lindblad-Toh K, Kellis M, Lander ES. 2007. Systematic discovery of regulatory motifs in conserved regions of the human genome, including thousands of CTCF insulator sites. Proc Natl Acad Sci U S A. 104:71457150. Xu AG, He L, Li Z, et al. (15 co-authors). 2010. Intergenic and repeat transcription in human, chimpanzee and macaque brains measured by RNA-Seq. PLoS Comput Biol. 6:e1000843. Zak M, Zbilut JP, Meyers RE. 1997. From instability to intelligence. Complexity and predictability in nonlinear dynamics. Berlin (Germany): Springer.

Downloaded from http://mbe.oxfordjournals.org/ at Banaras Hindu University on January 21, 2014

1892