M o l e C u l a r B i o l o G y

DNA fragility put into contextFragile sites are genomic regions prone to deletions or other alterations during DNA replication. The reason for the susceptibility of these sites to damage may be simple: they contain few replication initiation sites. See Letter p.120

k a y h u e B n e r

Agents that hinder DNA replication can often cause breaks at specific chromo-some regions commonly called fragile

sites. Two questions have been hotly debated concerning these chromosome regions: why are they so sensitive to DNA damage; and how does damage to genes encompassing such sites affect the biology of the damaged cells? A paper by Letessier et al.1 on page 120 of this issue presents a rather simple, yet ele-gant, answer to the first of these questions. The authors also report that, surprisingly, the sensitivity of specific fragile sites within a cell depends on the tissue or organ from which the cell originates — a conclusion that may provide clues to the answer to the second question.

During DNA replication, the enzyme heli-case breaks hydrogen bonds that hold the two DNA strands together, uncoiling the DNA spiral to form a structure called the replication fork. In humans, mice and other mammals, the propensity of common fragile sites to deletions and other rearrangements has been attributed to various features of the specific DNA sequences at fragile regions, which can span more than 1 million base pairs. For instance, DNA sequences at fragile sites might be prone to forming secondary structures that impair the movement of the replication fork, leading to its collapse and to DNA breaks2; fragile regions might replicate late in the DNA-replicative phase — even later in the presence of DNA-damaging agents — and so their replication remains incomplete3; or differences in the com-position of DNA-associated proteins between fragile and non-fragile regions4 may account for the susceptibility of the former to breakage.

Letessier and colleagues1 point to a different reason. The authors focus on FRA3B — the

most active fragile site in a type of human white blood cell called a lymphoblast and which lies within a tumour-suppressor gene known as FHIT (Fig. 1). They find that FRA3B is a fragile site not because the replication fork is slowed or stalled at this genomic region (locus), but because of the scarcity of replication initiation events there.

Indeed, the study shows that in lympho-blasts — but not in fibroblasts, which originate from connective tissue — initiation events are entirely absent from the central fragile region of FRA3B. Consequently, the replication of this large region within FHIT can be completed

Figure 1 | Damage-prone DNA. An image of normal chromosomes in a human lymphoblast. The FRA3B region (tagged by a green DNA probe flanking the gaps on both chromosomes 3) is a fragile site prone to breakage. Letessier et al.1 show that one reason for the fragility of this genomic region is a paucity of initiation sites for DNA replication. (Image adapted from ref. 10.)

most importantly, we will now start looking for more black holes in dwarf star-forming galaxies. Henize 2-10 reminds us to keep our eyes open and expect the unexpected. ■

Jenny E. Greene is in the Department of Astronomy, University of Texas, Austin, Texas 78712, USA.e-mail: jgreene@astro.as.utexas.edu

only by convergence of replication forks from the flanking regions (see Fig. 3a of the paper on page 122). The authors propose, therefore, that common chromosome fragile sites corre-spond to the initiation-poor regions that finish replication last in the given cell type.

To visualize DNA-replication dynamics in the fragile regions, Letessier et al. analysed stretched-out DNA strands, and searched for newly synthesized, fluorescently tagged FRA3B regions. They could thus examine both replication-fork speed and fork slow-ing, as well as mapping replication start and stop sites within FRA3B in lymphoblasts. The authors report that, within this fragile site, forks did not stall. Instead, the 700-kilobase core region of FRA3B, which contains the protein-coding portion of FHIT — the region most often mutated by deletion in cancers and precancerous lesions — showed no initiation events. By comparison, roughly ten initiation events occurred in similarly sized, non-fragile regions. These results suggest that a paucity of initiation events contributes to fragility.

A previous genome-wide analysis of replication timing5 in lymphoblasts and fibro-blasts revealed that, in these two cell types, the dynamics of replication (its timing and initia-tion patterns) are very different. For example, whereas initiation events occurred within the fragile sites in fibroblasts, they were consist-ently scarce in lymphoblasts. Do such differ-ences in initiation of replication between the two cell types — at least when grown in cul-ture — mean that the fragility of FRA3B also differs between them? According to Letessier and co-workers, the answer is indeed yes.

The authors find that fibroblasts do not exhibit the breaks at FRA3B that are charac-teristically seen in lymphoblasts. They also report similar findings for the second most active fragile region known in lymphocytes, FRA16D. Together, these data indicate that common chromosome fragile regions are “loci that correspond to the latest initiation-poor regions to complete replication in a given cell type”1. How such differential replication programs in cells of different tissue origin are established is unknown, but it cannot depend on DNA sequence because the DNA sequence of the FRA3B locus is the same in fibroblasts and lymphoblasts; likewise, the sequence of FRA16D is the same in the two cell types.

Common chromosomal fragile sites have previously been assessed in lymphoblasts,

7. Desroches, L.-B. & Ho, L. C. Astrophys. J. 690, 267–278 (2009).

8. Satyapal, S., Vega, D., Dudik, R. P., Abel, n. P. & Heckman, T. Astrophys. J. 677, 926–942 (2008).

9. Alvarez, M. A., Wise, J. H. & Abel, T. Astrophys. J. 701, L133–L137 (2009).

10. van der Marel, R. P. in Coevolution of Black Holes and Galaxies (ed. Ho, L. C.) 37–52 (Cambridge Univ. Press, 2004).

1. Reines, A. e., Sivakoff, G. R., Johnson, K. e. & Brogan, C. L. Nature 270, 66–68 (2011).

2. Johnson, K. e., Leitherer, C., Vacca, W. D. & Conti, P. S. Astron. J. 120, 1273–1288 (2000).

3. Allen, D. A., Wright, A. e. & Goss, W. M. Mon. Not. R. Astron. Soc. 177, 91–97 (1976).

4. Filippenko, A. V. & Ho, L. C. Astrophys. J. 588, L13–L16 (2003).

5. Barth, A. J. et al. Astrophys. J. 607, 90–102 (2004).6. Greene, J. e. & Ho, L. C. Astrophys. J. 670, 92–104

(2007).

4 6 | N A T U R E | V O L 4 7 0 | 3 F E b R U A R y 2 0 1 1

NEWS & VIEWSRESEaRch

© 2011 Macmillan Publishers Limited. All rights reserved

fibroblasts, some cancer cells of epithelial origin6,7 and even in cells of different lineages in a single study8. What has been lacking, however, is a systematic comparison of the frequency of breaks at known fragile sites in different cell types from a range of organs and tissues. Yet it is known that FHIT and WWOX — the two genes at the two most fragile loci in lymphoblasts — are also among the sequences most frequently altered by DNA deletion in precancerous and cancerous cells of epithelial origin (that is, cells from internal organs, not lymphoblasts); this is presumably due to expo-sure of cells within lung, colon and breast to agents that cause replicative stress6,7.

It has been argued that damage to genes at fragile sites, and the consequent loss of the genes’ expression, contributes to the selective growth of precancerous lesions and cancerous tumours. A counter-argument is that, because of the frequent deletions within the fragile sites, the loss of any associated gene expression is an unselected ‘passenger’ event in cancers9 and does not drive the expansive growth of the cancer cells.

A comparison of the frequency of breaks at known fragile sites in a range of cell types, including the epithelial cells of the lung, colon, breast and prostate — where most human can-cers originate — could shed light on several questions arising from the current study1. For instance, are FRA3B and FRA16D the most fragile regions in these cells, as they are in lym-phoblasts? If fibroblasts show infrequent DNA

damage at these sites, do they contain a differ-ent set of fragile sites, or do they show a low frequency of breaks across all chromosomes? Do the frequency and sites of breaks at fragile regions in cultured cells correspond to those in the same cell type within its organ of origin? Finally, if FRA3B is not the most fragile site in epithelial cells, does that strengthen the argu-ment that breaks in FHIT in precancerous cells and cancers have contributed to progressive growth of the lesions? Undoubtedly, Letessier and colleagues, and indeed other scientists, will be searching for the answers. ■

Kay Huebner is in the Department of Molecular Virology, Immunology and Medical Genetics, Ohio State University Comprehensive Cancer Center, Columbus, Ohio 43210, USA.e-mail: kay.huebner@osumc.edu

1. Letessier, A. et al. Nature 470, 120–123 (2011).2. Durkin, S. G. & Glover, T. W. Annu. Rev. Genet. 41,

169–192 (2007).3. Le Beau, M. M. et al. Hum. Mol. Genet. 7, 755–761

(1998).4. Jiang, y. et al. Hum. Mol. Genet. 18, 4501–4512

(2009).5. Hansen, R. S. et al. Proc. Natl Acad. Sci. USA 107,

139–144 (2010).6. Tsantoulis, P. K. et al. Oncogene 27, 3256–3264

(2008).7. Durkin, S. G. et al. Proc. Natl Acad. Sci. USA 105,

246–251 (2008).8. Turner, B. et al. Cancer Res. 62, 4054–4060 (2002).9. Bignell, G. R. et al. Nature 463, 893–898 (2010).10. Zimonjic, D. B. et al. Cancer Res. 57, 1166–1170

(1997).

T h e o r e T i C a l e C o l o G y

Waltz of the weevilThe aquatic plant Salvinia molesta is a widespread pest of waterways in the tropics and subtropics. A study of its control by a weevil in Australian billabongs sets a new standard in ecological time-series analysis. See Letter p.86

l e W i S T o n e

Waltzing Matilda, the bush ballad that became Australia’s unofficial national anthem, relates the exploits

of an out-of-control sheep shearer “camped by a billabong” (a small stagnant lake) in the Australian outback. On page 86 of this issue we read about the findings of an inspired group of theoretical ecologists1 and their models of out-of-control billabongs. Schooler et al.1 present a mathematical modelling study of an inva-sive plant species, the noxious weed Salvinia molesta, and its erratic large-scale outbreaks in four billabongs over a period of several decades. The authors’ skilfully executed modelling is an imaginative combination of nonlinear dynamics, statistical inference and stochastic time-series analysis.

The story of Salvinia has become a classic in the biological invasion literature2. This aquatic plant from South America is notorious for its rapid, almost uncontrollable growth, and since 1939 has become a pest in regions far from its home range. The weed is able to double its biomass every 3–4 days, generating thick mats of plant material that often cover entire water bodies (Fig. 1). Salvinia infestations can have devastating effects on lakes, billabongs and waterways, rendering them unusable for irrigation, as sources of drinking water or for sustaining local fish populations. In the past, entire villages in Papua New Guinea have been abandoned and the inhabitants relocated as a consequence of this out-of-control weed.

Salvinia first appeared in Australia’s Kakadu National Park in 1983 and, within months, biological control was implemented by the

3 F E b R U A R y 2 0 1 1 | V O L 4 7 0 | N A T U R E | 4 7

NEWS & VIEWS RESEaRch

© 2011 Macmillan Publishers Limited. All rights reserved