Identification of the shortest telomeres. BJ metaphases at PD84 were serially hybridized to both C- and G-rich telomeric probes to maximize the detection of small telomeric regions. (A) Telomeres so short that they failed to give an observable signal (signal-free ends) were identified by CGH analysis. (B) The frequency of signal-free ends per 100 metaphases is shown for all telomeric ends from an analysis of 403 metaphases. BACs within 100 kb of the ends of the telomeres indicated by vertical arrows were used in the analysis shown in .

Colocalization of short telomeres with γH2AX/53BP1 foci at senescence. (A) BJ cells at PD 83 were first stained with γH2AX/53BP1 and then serially hybridized in pairs to four different BAC probes, The probes to 6pter (41% signal-free ends), 9qter (7.7% signal-free ends), and 7qter (0.2% signal-free ends) were within 100 kb of the telomere, whereas that to 9q21 was 62 Mb from the telomere. Colored arrows highlight the position of the small but intense BAC signals. The same cell appears in all three panels and shows colocalization of both 6pter (middle) and 9qter (right) to the same γH2AX/53BP1 focus (left, digitally overlaid onto the other two images). (B) The frequency of colocalization with markers of DNA damage foci. The ends having shorter telomeres (6pter, 17qter, and 9qter) but not the ends having longer telomeres (7qter and 17pter) or the internal locus (9q21) showed frequent colocalization with markers of DNA damage foci. Data are for at least 900 DNA damage foci per BAC on three slides, ±SD between slides. The % colocalization was calculated as the percent of all γH2AX/53BP1 foci that colocalized with the specific BAC probe. The γH2AX/53BP1 positive foci in near-senescent BJ cells were distributed as follows: 43% of the cells had a single focus, 35% had two, 8% had three, 6% had four, and 8% had five or more foci. Correction for the number of colocalized foci per cell would result in a doubling of the % colocalization if expressed per cell instead of the data per focus as shown. (C) Presence of multiple short ends colocalizing with damage foci in single cells. The data in B have been reorganized to show the frequency with which a cell showing colocalization of one particular short telomere (e.g., 6pter) with a DNA damage focus also showed colocalization of other probes to DNA damage foci within the same cell. The short telomeres are shown in red.

The same telomeres are short in normal BJ and BJ cells expressing E6+E7. BJ cells infected with a retrovirus expressing human papilloma virus 16 proteins E6+E7 at about PD 65 were analyzed for the presence of telomeres with signal-free ends at PD 88.5 using the same techniques as for BJ cells in . The results of both analyses demonstrate that E6+E7 has neither affected the telomere shortening rate nor the telomeres showing signal-free ends in cells with abrogated cell-cycle checkpoints, because both cell types show the same frequencies and distribution.

Chromosomal end-associations in BJ E6+E7 cells. The frequency of end-associations for specific telomeres is shown for BJ foreskin fibroblasts expressing E6+E7 at PD 84–91, the approximate time when M1 would have occurred in parallel cultures without E6+E7. The axes are arranged according to the frequency of signal-free ends per 100 metaphases shown in . Each event is plotted twice, because an association between telomeres A and B is equivalent to an association between B and A. The dotted diagonal line reflects this line of symmetry. Almost all of the associations either involve short-short associations (darker gray box at the top left) or associations between the 10 shortest and other chromosomes (lighter gray boxes). There is no significant bias of one or a few specific ends. The distribution as plotted does not distinguish between homologues of the 22 autosomal chromosomes, so there are only 48 ends along each axis. If separated into homologues, the 10 shortest would represent 10 of 92 ends (see ).

Telomeric sequence variants distinguish parental chromosomes. (A) Telomeres were hybridized with probes for the telomeric repeat [(TTAGGG)₃-FITC] or a Cy3-labeled sequence variant (TGAGGG)₃ often found at the very base of the telomere. The top chromosome 8 shows a strong telomeric signal on the q arm but a weak variant repeat signal, whereas the lower chromosome 8q from the same cell exhibits no TTAGGG fluorescence (thus a signal-free end) but a strong sequence variant signal, thus allowing the parental homologues to be determined. (B) In cases where the parental homologues could be determined by polymorphisms in the amount of sequence variants, the signal-free ends were always present on one but not both parental telomeres. The probability of finding this pattern in 3 metaphases is 0.25, in 2 metaphases is 0.5, and for all 18 metaphases on 7 ends shown is 0.0005.