(A) Average abundance and length of L1 (or Alu) sequences along a 6-Mb genomic segment spanning the 10q25 neocentromere, encompassing the 330-kb CENP-A-binding domain (shaded in yellow; see Figure 2B for distribution of the seven CENP-A-binding sub-clusters within this domain) and the 3.5-Mb increased chromosomal scaffold (S/MAR)-attachment domain (shaded in grey) [17] analyzed using a 50-kb window. (B) Distribution of full-length L1s (green bars) as identified using the online L1Base program (http://l1base.molgen.mpg.de/).

(A) A schematic diagram showing the FL-L1b target sites for the three independent primer sets (FL-L1b.1, FL-L1b.2, and FL-L1b.3; together spanning a genomic region of 415 bp) used in RT-PCR assays. (B) Positive transcription of FL-L1b was seen with all 3 primer sets in CHOK1-N10 and CHOK1-M10 cell lines, but not in CHOK1 #8 cell line after 40 cycles of RT-PCR amplification. CHOK1 #8, CHOK1-N10 and CHOK1-M10 were hamster-human monochromosomal hybrid lines containing unrelated normal human chromosome 10, progenitor normal human chromosome 10, and mardel(10) chromosome, respectively. The transcription of positive control β-actin gene was detected in all the cell lines. ‘¢’ = genomic DNA control. (C) Quantitative RT-PCR results (mean for n = 3, with standard error of the mean, SEM), indicating no significant difference in FL-L1b transcription in CHOK10-N10 and CHOK1-M10. Relative FL-L1b transcription levels in CHOK1-N10 and CHOK1-M10 were compared to that of CHOK1 #8, which was used as a normalization control. ΔC_T = C_T [test segment]−C_T [β-actin control]. (D) RNA-ChIP-qPCR analysis. ChIP was performed using an anti-CENP-A antibody followed by quantitative RT-PCR analysis. FL-L1b RNA was significantly enriched (P<0.05) in the CENP-A-bound fraction in CHOK1-M10 but not in CHOK1-N10. None of the negative controls (18S rDNA, 5S rDNA, and β-actin ACTB RNAs) was enriched in the precipitated fractions. Relative binding values (mean for n = 5, with SEM) on the Y-axis represent the fold-enrichment of FL-L1b RNA CHOK1-M10 compared to that of CHOK1-N10.

(A) A total of six FL-L1s (green bars) were found, with a cluster of four of these elements (a–d) localizing within and near the CENP-A-binding region (shaded in yellow) [9]. Open bars denote long truncated L1s; hatched bars denote truncated L1s containing an internal promoter sequence; black bars denote orphan L1 promoter sequences. The positions of 13 transcribed genes (black arrows/arrowheads) are shown [17], of which 11 are found within the 3.5-Mb S/MAR domain (shaded in grey). (B) Close-up view of the CENP-A-binding and immediately surrounding region, showing the distribution of the seven CENP-A-binding clusters in relation to the ATRNL1 gene and FL-L1a–d. Note that FL-L1a–c reside within the CENP-A-binding clusters (red boxes) [18], with the actively transcribed FL-L1b being localized within the largest, central cluster.

(A) Efficient RNAi knockdown of FL-L1b in CHOK1-M10. CHOK1-M10 cells were transfected with FL-L1b-specific siRNA oligonucleotide duplexes, FL-L1b siRNA#1 and siRNA#2, at a final concentration of 25 nM. Following 48 hours post transfection, FL-L1b RNA levels were assayed by quantitative RT-PCR analysis using primer sets FL-L1b.1 (black bars) and FL-L1b.3 (white bars). Both siRNA duplexes effectively knocked-down the transcription level of FL-L1b (P<0.05, indicated by the asterisks) by >80% (mean for n = 3, with SEM), compared to the transfection-reagent-only and Stealth siRNA low GC negative controls. (B) The structural integrity of 10q25 neocentromere in CHOK1-M10 cells after 24 hours post siRNA transfection was investigated by dual immunofluorescence/FISH analysis using an anti-CENP-A antiserum (CREST6) and a BAC probe (RP11-359H22) specific for the 10q25 neocentromeric region of mardel(10). After FL-L1b knockdown, the mean fluorescence intensity of CREST6 at 10q25 neocentromere was significantly reduced (by 20–30%; P<0.001, indicated by the asterisks). Each spot in the combined scatter/box plot represented the relative amount of CENP-A protein at 10q25 neocentromere in one metaphase spread, which was calculated by an average of 5 measurements normalized against the average signals on 15–20 CHO centromeres within each spread. (C) Zeocin kill-curve analysis. (i) Addition of 200 µg/ml Zeocin effectively killed CHOK1-N10 cells but did not affect the growth of CHOK1-M10 cells (n = 4, with SEM); CHOK1-M10 cells were resistant to Zeocin because the mardel(10) chromosome had been tagged with a Zeocin resistance gene. The majority of the CHOK1-N10 cells (>80%; P<0.05) were killed by 48 hours post Zeocin selection (n = 4, with SEM). % cell survival = total cell number under Zeocin selection/total cell number without Zeocin selection. (ii) A 40–50% reduction in cell viability (mean for n = 4, with SEM) was observed following knockdown of FL-L1b RNA and 48 hours of Zeocin selection. The differences in the % of cell survival in the L1b knockdown samples were statistically significant (P<0.05; indicated by the asterisks). (D) The stability of the neocentromeric mardel(10) chromosome or normal human chromosome 10 in the various hybrid cell lines was calculated as a percentage of the total number of cells containing a positive FISH signal of BAC RP11-359H22 after 48 hours of RNAi knockdown. FL-L1b transcription knockdown resulted in significantly reduced stability of the mardel(10) chromosome (mean for n = 4, with SEM) but not of the normal human chromosome 10 in both the hamster and mouse hybrid cell lines (P<0.01; indicated by the asterisks). (E) Quantitative RT-PCR analysis of 13 genes found within or surrounding the 10q25 neocentromeric chromatin (see Figure 2A) in CHOK1-M10 was carried out 24 hours post FL-L1b RNA knockdown. While most of the genes were unaffected, the transcriptional activities of 2 neighboring genes, TRUB1 and ATRNL1 (mean for n = 4, with SEM), were significantly reduced (by 60–70%; P<0.05, indicated by the asterisks).