Salcaema

Fig. 2. DAPI-banded chromosome X in humans (HSA), ring-tailed lemur (LCA), and black lemur (EMA), showing a completely different position of the centromere, indicated by arrows. EMA chromosome X is upside down to match marker order in humans and LCA.

Y was not accompanied by a rearrangement (14,15). In one of these cases, the chromosome was segregating in the family (15).

Remains oflnactivated Centromeres

Human chromosome 2 is the result of a telomere-telomere fusion between two ancestral acrocentric chromosomes that occurred after human-chimpanzee divergence (16), with one of the two centromeres becoming inactive. Indeed, stretches of alphoid sequences and clusters of segmental duplications, as remains of the inactivated centromere, are apparent at 2q21 (17,18). We have hypothesized that the evolutionary appearance of new centromeres, as part of a centromere-repositioning event, implies the inactivation of a functioning centromere. Studies on the evolutionary history of chromosome 6 provided evidence that remains of an inactivated centromere, which resulted from a centromere repositioning, are present in the human genome (19). Figure 3 summarizes the evolutionary history of chromosome 6, which is substantially conserved in evolution. Marker order analysis strongly suggested that the arrangement of the two New World monkeys (NWM), common marmoset (Callithrix jacchus [CJA]), and wooly monkey (Lagothrix lagothricha [LLA]) maintained the primate ancestral form. The marker arrangement of cat, chosen as the outgroup, supported this hypothesis. The centromeres of OWM and great apes, therefore, are evolutionary new centromeres. The major point of this analysis was the finding, in humans, of a duplicon cluster at 6p22.1 exactly corresponding to the region where the ancestral centromere was inactivated (Fig. 4). No remains of alphoid centromeric sequences, however, were detected. It can be hypothesized that, following the centromere inactivation, the pericentromeric duplicons dispersed in a relatively wide area (approx 10 Mb), whereas the large alphoid block, typical of active primate centromeres, completely disappeared. A second example of remains of an ancestral inactivated centromere is presented next.

Ancestral Centromeres and Neocentromeres

Centromeres are visible cytogenetically as the primary constriction and in primates are always associated with the presence of an array of a-satellite DNA (20). a-satellite, however, it is not an absolute requirement for centromere function, as evident in neocentromeres, which are devoid of satellite DNA.

The mechanisms underpinning neocentromere emergence is obscure. du Sart et al. (12) have hypothesized the existence of latent centromeres with a finite capacity for neocentromerization.

Fig. 3. Evolutionary history of human chromosome 6 established on the bases of marker order comparison among the orthologous chromosomes in great apes, Old World monkeys, New World monkeys, and prosimians. MFA, long-tailed macaque, Macaca fascicularis; PCR, silvered leaf-monkey, Presbytis cristata; CJA, common marmoset, Callithrix jacchus; LLA, wooly monkey, Lagothrix lagothricha; EMA, black lemur, Eulemur macaco; FCA, cat, Felix catus. N in a gray circle stands for a new centromere; A in the black circle stands for an ancestral centromere. PA, primate ancestor; CA, catarrhini ancestor; OA, OWM ancestor. (Adapted with permission from ref. 19.)

Fig. 3. Evolutionary history of human chromosome 6 established on the bases of marker order comparison among the orthologous chromosomes in great apes, Old World monkeys, New World monkeys, and prosimians. MFA, long-tailed macaque, Macaca fascicularis; PCR, silvered leaf-monkey, Presbytis cristata; CJA, common marmoset, Callithrix jacchus; LLA, wooly monkey, Lagothrix lagothricha; EMA, black lemur, Eulemur macaco; FCA, cat, Felix catus. N in a gray circle stands for a new centromere; A in the black circle stands for an ancestral centromere. PA, primate ancestor; CA, catarrhini ancestor; OA, OWM ancestor. (Adapted with permission from ref. 19.)

Our investigations of chromosome 15 evolution provided evidence of an obscure but intriguing relationship between neocentromeres and ancestral inactivated centromeres.

Chromosomes 14 and 15 in humans and in great apes are separate chromosomes. In OWM, they form a unique chromosome (chromosome 7 in macaque) (21). We used a panel of bacterial artificial chromosome (BAC) clones to track the evolutionary history of the 14/15 association in macaque (Macaca mulatta [MMU]), sacred baboon (Papio hamadryas [PHA]), long-tailed macaque (Macaca fascicularis [MFA]), and silvered-leaf monkey (Presbytis cristata [PCR]). All these species were found to share the 14/15 association, corresponding to chromosomes MMU7, PHA7, MFA7, and PCR5 (22). The results are summarized in Fig. 5. In all species these chromosomes appeared consistent with human chromosome14 and 15 being fused head-tail and retaining colinearity. The 14/15 association has been reported as ancestral to mammals (23). It can be deduced that in the ancestor to great apes, a noncentromeric fission event occurred between markers F and G, disrupting the 14/15 association and generating the present-day human

Fig. 4. Distribution of segmental duplications between chromosome 6 and other human chromosomes. Lines represent interchromosomal (gray) and intrachromosomal (black) alignments. Arrow indicates the cluster of intrachromosomal duplications at 6p22.1, which are the remains of an inactivated ancestral centromere. Note also that the pericentromeric region is relatively devoid of segmental duplications. This is in agreement with the finding that human the chromosome 6 centromere is an evolutionary neocentromere. (Courtesy of Dr. E. E. Eichler.)

Fig. 4. Distribution of segmental duplications between chromosome 6 and other human chromosomes. Lines represent interchromosomal (gray) and intrachromosomal (black) alignments. Arrow indicates the cluster of intrachromosomal duplications at 6p22.1, which are the remains of an inactivated ancestral centromere. Note also that the pericentromeric region is relatively devoid of segmental duplications. This is in agreement with the finding that human the chromosome 6 centromere is an evolutionary neocentromere. (Courtesy of Dr. E. E. Eichler.)

Fig. 5. The diagram shows the organization of human chromosomes 14/15 association in macaque (MMU,Macacamulatta). A fission event gave rise to chromosomes 14 and 15. The fission was followed by the emergence of two new centromeres and by the inactivation of the old centromere at 15q25. (Adapted with permission from ref. 22.)

Fig. 5. The diagram shows the organization of human chromosomes 14/15 association in macaque (MMU,Macacamulatta). A fission event gave rise to chromosomes 14 and 15. The fission was followed by the emergence of two new centromeres and by the inactivation of the old centromere at 15q25. (Adapted with permission from ref. 22.)

chromosomes 14 and 15. The ancestral centromere at 15q25 was inactivated and two new centromeres were generated in regions corresponding to their present-day locations in humans. Segmental duplication analysis showed a large cluster of intrachromosomal segmental duplications at 15q24-q26 (data not shown). This cluster very likely derives from the inactivated ances-

Fig. 6. Partial metaphases showing fluorescence in situ hybridization experiments using the human bacterial artificial chromosome clone RP11-182J1 (chr15:82,764,242-82,935,727; 15q25.2) on human chromosome 15 (left) and on macaque chromosome X.

tral centromere. Fluorescence in situ hybridization (FISH) experiments using BAC probes containing duplicons of this region flanked the centromere in macaque (Fig. 6), strongly indicating that they are dispersed remains of the ancestral inactivated centromere. The region encompassed by duplicons that flanked the ancestral centromere is approx 11 Mb in size. Similarly to the inactivated ancestral centromere at 6p22.1, no remains of alphoid sequences were found.

The structural organization of the centromeres of the present-day human chromosomes 14 and 15 are indistinguishable from any other human centromere. The conclusion is that the evolutionary new centromeres progressively acquired the complexity of a normal centromere, which very likely stabilize the centromere function.

Interestingly, one of the hotspot site of neocentromeres appearance in clinical cases is at 15q24-q26. We studied two of these neocentromeres that were found several megabases apart but inside the region encompassed by duplicons of the ancestral centromere.

Alonso et al. (24) have reported that neocentromeres mapping at the 13q32 hotspot span a 6.5-Mb region. Both sets of data disproved the hypothesis that the neocentromere hotspots refer to a single specific locus. Interestingly, however, both 15q25 neocentromeres mapped to duplicons that are the remains of the ancestral centromere, therefore establishing an intriguing but unclear correlation between duplicon dispersal and neocentromere emergence. If this correlation is correct, one would expect the appearance of neocentromeres at 6p22.1 and at 2q21, where abundant duplicon remains of an inactivated centromere are present. We think, however, that other important factors are involved in neocentromere appearance. It is worth noting that most of the acentric fragments are small in size and associated with phenotypes compatible with embryo development. It could be hypothesized that aneuploidies generated by other latent centromeres are incompatible with life and will never be detected among liveborn.

EVOLUTION OF CHROMOSOME 3: RECURRENT SITES FOR NEW CENTROMERE SEEDING

Because the evolutionary history of the neocentromere hotspot at 15q25 revealed a connection to an ancestral inactivated centromere, we undertook a comprehensive study of the evolutionary history of chromosome 3, harboring an additional neocentromere hotspot at 3q25-q26

Fig. 7. Summary of fluorescence in situ hybridization results delineating the evolutionary history of human chromosome 3 in primates. The hypothesized pericentric or paracentric inversions are indicated by square parentheses spanning the inverted segment. Some chromosomes are upside down to facilitate comparison. PA, primate ancestor; CA, catarrhini ancestor; OWM-A, Old World monkeys ancestor; NWM-A, New World monkeys ancestor; GA-A, great ape ancestor; N (in a gray circle), new centromere. The number that identifies the chromosome in each species is reported on top of the chromosome. Arrows with hatched lines point to species for which an intermediate ancestor has not been drawn. A white N in a gray circle indicates the chromosomes where new centromeres appeared during evolution. The black circles (HSA and OWM-A) and the black squares (GA-A and NWM-A) are positioned at loci where seeding of an evolutionary new centromere was detected. Cat (FCA) probes were identified using the following strategy. The sequence encompassed by each human bacterial artificial chromosome (BAC) was searched for conservation against mouse and rat genome. "Overgo" probes were designed on the most conserved region of each BAC and were then used to screen the cat BAC library RPCI-86. Cat probes, therefore, map to loci orthologous the loci encompassed by the corresponding human BACs. Species analyzed: great ape: common chimpanzee (Pan troglodytes, PTR), gorilla (Gorilla gorilla, GGO), Borneo (Pongopygmaeuspygmaeus, PPY-B) and Sumatra (Pongopygmaeus abelii, PPY-S) orangutan; Old World monkeys (OWM): rhesus monkey (Macaca mulatta, MMU, Cercopithecinae), sacred baboon (Papio hamadryas, PHA, Cercopithecinae), African green monkey (Cercopithecus aethiops, CAE, Cercopithecinae), silvered leaf-monkey (Presbytis cristata, PCR, Colobinae); New World monkeys (NWM): wooly monkey (Lagothrix lagothricha, LLA, Atelinae), common marmoset (Callithrix jacchus, CJA, Callitrichinae), dusky titi (Callicebus moloch, CMO, Callicebinae), squirrel monkey (Saimiri boliviensis, SBO); prosimians: ring-tailed lemur (Lemur catta, LCA). (Adapted with permission from ref. 25.)

(25). A panel of human BAC clones was used to reconstruct the order of chromosome 3 markers in representative extant primates (Fig. 7). The analysis clearly showed the emergence of new centromeres in OWM and great ape ancestors. Two human neocentromere cases at 3q25-26 were investigated. These two cases were found to have distinct but important implications toward a better understanding of new centromere emergence phenomenon both in evolution and in clinical cases. In the first case, the neocentromere emergence occurred at 3q26.1 as part of a complex rearrangement. The functional centromere was excised, and a neocentromere seeded in the acentric derivative chromosome. Human BAC clone RP11-498P15 was identified to be the closest to the neocentromere. This same clone (nc1 in Fig. 7) was found a few kilobases apart from the evolutionary new centromere that appeared in OWM ancestor. This coincidence suggested that the region at 3q26.1 maintained a latent centromere competence that was present before OWM/Hominidae divergence that occurred approx 25 million years ago (26). Furthermore, the new centromere that appeared in the great ape ancestor was seeded very close to marker G (Fig. 7). Interestingly, a new centromere was also seeded in this region in NWM ancestor (chromosome GF21 in Fig. 7) suggesting, again, that the same latent centromeric region was activated twice, in great ape and NWM ancestors. Pevzner and Tesler (27), comparing mouse and human evolutionary breakpoints, have suggested the "reuse" of breakpoint regions in evolution. Similarly, the examples of neocentromeres we have discovered, both in evolution and in clinical cases, seem to extend the "reuse" concept to centromeres.

Evolution ofChromosome 3: Centromeres and Telomeres

In four NWM species examined, human chromosome 3 is split into three chromosomes annotated as (I), (II), and (III) in the Fig. 7. Chromosomes (I) and (II) have an identical marker order in common marmoset (CJA), wooly monkey (LLA), and dusky titi (Callicebus moloch [CMO]), if centromeres are not considered. The position of the centromeres, on the contrary, appears peculiar: the centromere is located at the opposite telomere in both (I) and (II) chromosomes in CMO with respect to the ancestral form of CJA and LLA. We have found similar examples of telomere-centromere functional exchanges in the evolution of other chromosomes, for instance, chromosome 20 (28), and chromosomes 11 and 13 (Rocchi, unpublished results, 2005). Low-copy repeat-mediated exchanges may be responsible for centromere jumping between telomeres.

Evolutionary New Centromere Progression

A new centromere was seeded in OWM chromosome 6 during evolution (see Remains of Inactivated Centromeres; Fig. 3). The human BAC clone RP11-474A9 (L2 in Fig. 3), mapping at 6q24.3, yielded clear signals on both sides of the MFA6 centromere. Apparently, the new centromere recruited the huge amount of centromeric/pericentromeric sequences without affecting the displaced flanking sequences. The scenario accompanying the progression of the OWM new centromere that appeared on OWM chromosome 3 was different. We found that this centromere is encompassed by human K1 and K2 markers (BAC clones RP11-355I21 and RP11-418B12, respectively), located 436 kb apart in human. BAC clones spanning this gap did not yield any appreciable FISH signal in macaque, African green monkey (CAE), and PCR, whereas giving normal FISH signals in all great apes. Thus the 436-kb region appears lost in OWMs as a consequence of the neocentromerization event. This chromosomal trait was investigated for gene content. The human region (UCSC chr3:163,941,067-164,377,941) appears devoid of genes, harboring only two transcribed but not translated spliced RNAs (UCSC, July 2003 release). It can be hypothesized that the absence of functional constraint made possible the loss (or complete restructuring) of the region. On the contrary, the functional elements around the new centromere insertion on chromosome 6 in OWM (see above) behaved as a strong selective barrier to the degenerative activity of the evolving centromeric region.

Gene expression around the neocentromere that emerged at 10q25 (12,29) was investigated by Saffery et al. (30). The authors concluded that the neocentromere formation did not affect underlying gene expression. The likely scenario of the progression of an evolutionary new centromere appears as a competition between the degenerative activity of the centromere in the seeding area, and the selection acting against disturbance of functional elements present in the region. Nagaki et al. (31) have reported recently the organization of rice chromosome 8 centromere featuring an unusual organization, with genes embedded in the centromeric hetero-chromatin, as if the centromere were at an intermediate stage between new centromere seeding and a final stage in which all genes are excluded from the completely heterochromatized centromere, as occurs in the remaining rice centromeres.

Additional Euchromatic Duplicon Clusters ofPericentromeric Origin

Segmental duplications accumulate at greater density in pericentromeric and telomeric regions (6). However, large blocks of duplicons exist also in euchromatic locations. We have previously shown that duplicon clusters located, in humans, at euchromatic regions 6p22.1 and 15q25 are the remains of inactivated centromeres. Other chromosomal rearrangements can relocate pericentromeric duplicons in noncentromeric areas. One of the two breakpoints of both pericentric and paracentric evolutionary inversions falls, rather frequently, in pericentromeric regions (8,28,32). In these cases part of a pericentromeric duplicated region can be repositioned far apart from the centromere. Two nonpericentromeric olfactory receptor clusters lying on chromosome 3 are precisely located at domains identified by A2 and I1 markers (Fig. 7). The evolutionary history of chromosome 3 we have reconstructed indicated that these two clusters are examples of duplicons that were pericentromerically located during evolution (see marker organization in the primate ancestor [PA] in Fig. 7).

Most mutations of the CYP21 locus (steroid 21-hydroxylase gene) are owing to a close pseudogene (CYP21P) that triggers gene conversions during unequal crossovers (33). Both CYP21 and CYP21P are located inside the duplicon cluster at 6p22.1 that represents the remains of an evolutionary inactivated centromere (see above). Gene conversion, therefore, is an additional, indirect mechanism by which duplicon cluster contribute to human genome morbidity.

CENTROMERE REPOSITIONING "IN PROGRESS"

Most of human neocentromere cases arise as a consequence of a rearrangement that generated an acentric fragment. The presence of the acentric fragment as a supernumerary chromosome results in phenotypic abnormalities, which indirectly prevent the transmission of the neocentromere-bearing chromosome to successive generations. They, therefore, have no evolutionary relevance. Very recently, however, Amor et al. (34) and Ventura et al. (25) have documented two unprecedented neocentromere cases, at 4q21.3 and 3q24, respectively, that arose in an otherwise normal and mitotically stable chromosomes that were transmitted to subsequent generations. These individuals had normal phenotypes. They can be regarded as present-day episodes equivalent to new centromere appearance that occurred during the course of evolution and that were fixed in the population. Individuals with chromosomal changes not accompanied by clinical signs or reproductive problems usually escape detection. It can be reasonably supposed that the "normal" cases reported so far are only a minimal fraction of the existing ones.

In individuals heterozygous for the neocentromere, an odd number of meiotic exchanges internal to the chromosomal region encompassed by the two active centromeres leads to the formation of dicentric and acentric chromosomes. These problems could affect the fitness of heterozygous individuals and act as a negative selective pressure. Meiotic drive in females in favor of the repositioned chromosome is among the possible mechanisms favoring the fixation of the neocentromere. Meiotic drive favoring Robertsonian translocations has been reported by Pardo-Manuel de Villenna and Sapienza (35).

Was this article helpful?

0 0

Post a comment