HMGB1 gene knockout in mouse embryonic ...

19 downloads 3788 Views 615KB Size Report
Apr 11, 2012 - length) and actin (~90 bp) in ULTRAhyb solution (Ambion) at 55 and ...... Woo SH, An S, Lee HC, Jin HO, Seo SK, Yoo DH, Lee KH, Rhee CH,.
Chromosoma DOI 10.1007/s00412-012-0373-x

RESEARCH ARTICLE

HMGB1 gene knockout in mouse embryonic fibroblasts results in reduced telomerase activity and telomere dysfunction Eva Polanská & Zuzana Dobšáková & Martina Dvořáčková & Jiří Fajkus & Michal Štros

Received: 24 January 2012 / Revised: 11 April 2012 / Accepted: 11 April 2012 # Springer-Verlag 2012

Abstract Telomere repeats are added onto chromosome ends by telomerase, consisting of two main core components: a catalytic protein subunit (telomerase reverse trancriptase, TERT), and an RNA subunit (telomerase RNA, TR). Here, we report for the first time evidence that HMGB1 (a chromatin-associated protein in mammals, acting as a DNA chaperone in transcription, replication, recombination, and repair) can modulate cellular activity of mammalian telomerase. Knockout of the HMGB1 gene (HMGB1 KO) in mouse embryonic fibroblasts (MEFs) results in chromosomal abnormalities, enhanced colocalization of γ-H2AX foci at telomeres, and a moderate shortening of telomere lengths. HMGB1 KO MEFs also exhibit significantly (>5-fold) lower telomerase activity than the wild-type MEFs. Correspondingly, enhanced telomerase activity is observed upon overexpression of HMGB1 in MEFs. HMGB1 physically interacts with both TERT and TR, as well as with active telomerase complex in vitro. However, direct interaction of HMGB1 with telomerase is most likely

Communicated by Jan Karlseder Electronic supplementary material The online version of this article (doi:10.1007/s00412-012-0373-x) contains supplementary material, which is available to authorized users. Z. Dobšáková : M. Dvořáčková : J. Fajkus (*) CEITEC - Central European Institute of Technology, Masaryk University, Brno, Czech Republic e-mail: [email protected] E. Polanská : Z. Dobšáková : J. Fajkus : M. Štros (*) Academy of Sciences of the Czech Republic, Institute of Biophysics, Brno, Czech Republic e-mail: [email protected]

not accountable for the observed higher telomerase activity in HMGB1-containing cells, as revealed from the inability of purified HMGB1 protein to stimulate telomerase activity in vitro. While no transcriptional silencing of TERT is observed in HMGB1 KO MEFs, levels of TR are diminished (~3-fold), providing possible explanation for the observed lower telomerase activity in HMGB1 KO cells. Interestingly, knockout of the HMGB2 gene elevates telomerase activity (~3-fold) in MEFs, suggesting that the two closely related proteins of the HMGB family, HMGB1 and HMGB2, have opposite effects on telomerase activity in the cell. The ability of HMGB1 to modulate cellular activity of telomerase and to maintain telomere integrity can help to understand some aspects of the protein involvement in chromosome stability and cancer.

Introduction Mechanisms involved in the regulation of maintenance of chromosome ends, the telomeres, are the subject of intensive investigation, namely due to the known links of telomere biology to the problems of genome stability, cancer, and aging (Fajkus et al. 2002). Telomere repeats are added onto chromosome ends by telomerase, a ribonucleoprotein consisting of two essential components: a catalytic protein subunit telomerase reverse trancriptase (TERT), and an RNA subunit, telomerase RNA (TR). Numerous factors participating in regulation of telomere synthesis by telomerase have been identified, including telomere- and telomeraseassociated proteins (Collins 2008). Telomere function is apparently tightly linked to chromatin architecture, and analysis of the contribution of other factors involved in chromatin dynamics is of special interest to current telomere biology. Many of the chromatin structural

Chromosoma

changes are mediated by the large and diverse superfamily of high mobility group (HMG) proteins, including the HMGBtype proteins (Štros et al. 2007; Thomas and Travers 2001). There are four canonical HMGB proteins in human and mice: HMGB1, HMGB2, HMGB3, and HMGB4 (Štros 2010). HMGB1 is an abundant and ubiquitous chromatinassociated protein in mammals, acting as a DNA chaperone in transcription, replication, recombination, and repair (Agresti and Bianchi 2003). In addition to the nuclear role of HMGB1 (Štros 2010), protein can be secreted into the extracellular milieu acting as a signaling molecule by binding to several cell receptors involved in inflammation, cell migration to promote tumor growth and metastasis (Taguchi et al. 2000). The importance of HMGB1 for life is illustrated by the phenotype of HMGB1 knockout (KO) mice, which die 24 h after birth (Calogero et al. 1999). Unlike the HMGB1 KO mice, the mouse embryonic fibroblasts (MEFs) with HMGB1 knockout are viable (Calogero et al. 1999) but exhibit striking chromosomal aberrations (Giavara et al. 2005), suggesting the importance of HMGB1 in promoting genomic stability. Despite very similar DNA binding properties of HMGB1 and HMGB2 in vitro [reviewed by Štros (2010)], different phenotypes of HMGB1 and HMGB2 KO mice might suggest that the HMGB proteins may not be equivalent in their functions in vivo (Calogero et al. 1999; Ronfani et al. 2001). In this paper, we report that knockout of the HMGB1 gene in mouse embryonic fibroblasts results in reduced telomerase activity and telomere dysfunction. We also provide evidence for distinct roles of two closely related proteins of the HMGB family, HMGB1 and HMGB2, in modulation of telomerase activity.

Telomerase activity assays Cells were collected at ~80 % confluence, and extracted with CHAPS lysis buffer from the TRAPeze®XL Telomerase Detection Kit (Millipore) using the manufacturer’s recommendations. Telomerase activity was also measured in cellular lysates prepared in CHAPS-free buffer as detailed in Fernandez-Capetillo et al. (2004). Subsequent telomerase activity assays were performed in a dual-color real-time PCR mode using a Rotorgene 3000 (Corbett Research) as we described previously (Fajkus et al. 2003). Absolute quantification was performed using dilutions of the control template TSR8 (0.1–1.0 amol). Relative TRAP quantification was performed using several internal control templates. The internal control template TSK2 (used in the TRAPeze®XLTelomerase Detection Kit) was amplified by the K2 and TS primers from the kit to generate a 56-bp product. Some of the telomerase activity assays were also carried out using primers TS and NT to amplify TSNT oligonucleotide as an internal control template (Herbert et al. 2006). Alternatively, a conventional TRAP assay using a TRAPeze® Telomerase detection kit was used and the products were resolved on 10 % polyacrylamide gels, followed by SYBR Green staining (Roche) as described (Fajkus et al. 2003; Malaska et al. 2000). The impact of purified HMGB1 [5–200 ng, recombinant HMGB1 isolated from Escherichia coli or from mammalian cells (Štros et al. 2007)] on telomerase activity in vitro was studied using cellular lyses (150–300 ng of total proteins) containing protease inhibitors and Protector RNase inhibitor (Roche, 1 U/μl) and increasing amounts of HMGB1 protein (typically, 25– 150 ng). Protein concentrations of cellular lysates were determined using Bio-Rad protein microassay with a 96-wells microplates reader (Sunrise) and BSA as a standard.

Materials and methods Analysis of TERT and TR expression in MEFs Cells MEFs were either wild-type (HMGB1+/+) or cells derived from KO mice (HMGB1−/−) and immortalized with SV 40 large T antigen (kindly provided by M.E. Bianchi, Italy) (Calogero et al. 1999). Other HMGB1−/− or HMGB2−/− MEFs (immortalized by the 3T3 protocol) were purchased from HMGBiotech, Italy. Overexpression of HMGB1 was achieved using plasmid p-mHMG1 containing the genomic mouse HMGB1 with its own promoter, enhancer, exons, and introns (Falciola et al. 1997; purchased from HMGBiotech, Italy). MEF cells were transfected by Nucleofector II Device Amaxa set at program T020 (Nucleofector Kit 1, Lonza). The cells were analyzed ~44–48 h post nucleofection. The efficiency of nucleofection was monitored by flow cytometry of MEF cells transfected with plasmid pmaxGFP Vector (Lonza) overexpressing the enhanced green fluorescent protein eGFP.

Expression of the mouse telomerase subunit mTERT (RNA) was analyzed in triplicates by real-time RT-PCR using mGAPDH as normalizer. SuperScript III One Step RTPCR with Platinum Taq DNA polymerase (Invitrogen) was used with the following primers: mTERT-Fw: 5′-TGGTGGAGGTTGTTGCCAA-3′ mTERT-Rev: 5′-CCACTGCATACTGGCGGATAC-3′ mGAPDH-Fw: 5′-GAACGGGAAGCTCACTGGC-3′ mGAPDH-Rev: 5′-ACCACCTTCTTGATGTCATCA TACTT-3′ The following conditions for real-time qRT-PCR of mTERT and mGAPDH were used on a Rotorgene 3000 cycler: 50 °C/30 min (reverse transcription), 94 °C/2 min (initial denaturation), 40 cycles of 94 °C/15 s, 55 °C/30 s and 68 °C/1 min, which was followed by a final extension step of 68 °C/5 min.

Chromosoma

qRT-PCR or semiq RT-PCR of mTR was carried out using cDNAs synthesized from total RNA (isolated by RNeasy Mini Kit, Qiagen) using M-MuLV (NEB) reverse transcriptase using the manufacturer instructions. The following mTR primers were used: mTR-Fw: 5′-AGTCCCGTACCCGCCTACAGG CCGC-3′ mTR-Rev: 5′-GCCCCGCGGCTGACAGAGG-3′ PCR conditions for amplification of cDNA were: 95 °C/ 15 min (initial denaturation), up to 39 cycles of 95 °C/20 s, 58 °C 30 s, and 72 °C/35 s, followed by a final extension step at 72 °C for 2 min (ThermoStart Taq DNA polymerase, ABgene). PCR reactions were performed at different amounts of cDNAs (typically 25–50 ng of the total RNA) using GAPDH or β-2-microglobulin as normalizers for HMGB1−/− or HMGB2−/− MEFs, respectively. The following primers for β-2-microglobulin were used: β-2-MG_Fw: 5′-GGCCTGTATGCTATCCAGAA-3′ β-2-MG_Rev: 5′-GAAAGACCAGTCCTTGCTGA-3′ Northern blotting Total RNA was prepared from MEFs at ~80–90 % confluence using Tri reagent as detailed by the manufacturer (Sigma). Typically, 25 μg of total RNA was used for Northern blots. RNA was separated by electrophoresis on 1.2 % agarose gels containing formaldehyde/MOPS at 120 V for 4–5 h. After electrophoresis, the gels were photographed under UV illumination to assess the quality of RNA (28S and 18S RNA). RNA was transferred to Hybond-N-Plus membranes (GE Healthcare), and the RNA blots were hybridized with 32P-labeled DNA probes against mTR (full length) and actin (~90 bp) in ULTRAhyb solution (Ambion) at 55 and 42 °C, respectively, according to the manufacturer instructions. RNA levels were quantified by Storm PhosphorImager using ImageQuant software (GE Healthcare). Loading RNA controls were 18S RNA (ethidium bromide) and actin.

were cut into aliquots corresponding to ca.106 cells and placed in a 15 ml Falcon tube with 2 ml of lysis buffer [0.5 M EDTA (pH 8.0); 10 mM Tris–HCl (pH 8.0); 1.0 % lauroylsarcosine]. After 20 min at room temperature, the lysis buffer was removed and replaced by 1 ml of the same buffer. Proteinase K was added to a final concentration of 500 μg/ml and incubation was performed at 50 °C for 24 h. Then the blocks were washed with 10 ml of TE buffer (three times for 20 min) and then with 2 ml of TE with 0.2 mM PMSF (2×30 min) to remove remainders of proteinase activity. Finally, the blocks were washed with 10 ml of 0.1× TE to remove PMSF and its hydrolytic products. Prior to restriction enzyme digestion, DNA plugs were incubated in 1× TaqI restriction enzyme buffer (Takara) for 30 min, and then digested overnight at 55 °C with 50 U of TaqI in 300 μl of fresh buffer. Pulsed-field gel electrophoresis was run on a 1 % agarose gel in 0.5× TBE, using a Gene Navigator System (GE Healthcare) for 23 h at 6 V/cm at a constant pulse time of 5 s (Hemann and Greider 2000). Running temperature was kept at 12 °C. Following electrophoresis and ethidium bromide staining, in-gel hybridizations were performed, first with the native gel (to visualize telomeric G-overhang signal) and subsequently with the same gel after denaturation as described in Hemann and Greider (2000). Briefly, the native gel was dried down on filter paper for 1 h at 50 °C, prehybridized for 1 h at 55 °C in 20 mM NaH2PO4, 0.1 % SDS, 5× Denhardt’s reagent and 5× SSC, and hybridized with the 32P-end-labeled oligonucleotide probe (CCCTAA)4 for 3 h at 55 °C in 5 ml prehybridization solution. After hybridization, the gel was washed three times for 20 min in 3× SSC at room temperature and three times for 20 min in 3× SSC/0.1 % SDS at 58 °C. Image analysis was performed using a Storm Phosphorimager and ImageQuant software (Amersham Biosciences). The signal intensity was normalized to ethidium bromide fluorescence as a loading control. Following visualization, the dried gel was alkalidenatured in 0.6 M NaCl, 0.2 M NaOH for 1 h, neutralized in 1.5 NaCl, 0.5 M Tris.HCl pH 7.4 for 1 h, rinsed in distilled water for 30 min and reprobed to yield the total signal of telomeric sequences. Fluorescence in situ hybridization

Analysis of G-overhangs and terminal restriction fragments High-molecular-weight DNA was prepared in agarose plugs (Hemann and Greider 2000). Briefly, MEFs (SV40-T) cells (ca. 5×106) were washed twice in PBS at room temperature and the pellet was resuspended in a minimal volume of PBS (100–200 μl) using a wide-bore tip. Then, the suspension was briefly incubated at 40 °C, followed by addition of 1 volume of 1.6 % (w/v) low-melting-point agarose in PBS (premolten in a boiling bath and then incubated at 40 °C). The suspension was mixed by careful pipetting using a widebore tip and transferred to the sample mold (Bio-Rad). Plugs

Wild-type and HMGB1−/− MEFs (SV40-T) were compared by in situ analysis. The cells were treated with 0.01 % colcemid for 1 h before harvesting, twice washed in 1× PBS, lysed using 0.075 M KCl at 37 °C for 25 min, and finally spun down at 1,200 g. The nuclei obtained were briefly washed in Carnoy’s fixative (MetOH/glacial acetic acid, 3:1) five times, followed by fixation in the appropriate volume of the same solution. Fluorescence in situ hybridization (Q-FISH) was performed with a Cy3-conjugated PNA probe according to the manufacturer’s protocol (DAKO). Briefly, samples were fixed in 3.7 % formaldehyde, washed

Chromosoma

in tris-buffered saline and treated with proteinase K solution at 37 °C for 10 min. The samples together with probes were denatured at 80 °C for 3 min, hybridized at room temperature for 2 h, followed by washing at 65 °C for 5 min. The samples were finally stained with 2 μg/ml DAPI in Vectashield (Vector Laboratories). On average, 50–60 metaphases have been analyzed. Image acquisition, telomere length measurement, and cytogenetic analysis Digital images were captured by Olympus BX61 connected to a CCD camera (Zeiss). For image analysis, the ISIS software (Meta) was used using 120 ms as an exposure time. About 600 chromosomes (corresponding to 15 cells) from wild type or HMGB1−/− MEF cells were analyzed. Fluorescence intensity of particular telomeric signals was analyzed by using TFL telo software (Poon and Lansdorp 2001). Acquired data were processed using Excell software to make appropriate histograms and perform statistical analysis. Telomere lengths (x-axis) were expressed in arbitrary fluorescence units. About 57 cells from each sample were scored for chromosomal aberration, and analysis was performed in TFL telosoftware in DAPI channel. Presence of telomeric spots was checked in each of the chromosomal aberrations.

identical for all nuclei (≥100 nuclei were analyzed for each cell line). Percentage of cells with ≥5 colocalization events (γH2AX foci at telomeres) was analyzed by the Aquiarium colocalization software. Pull-down assays GST or GST-fused HMGB1 and domains were used in pulldown assay as described (Štros et al. 2002). Purified proteins (~2–4 μg) were bound to glutathione-agarose beads in PD buffer [20 mM Tris.HCl pH 8.0, 0.2 M NaCl, 0.1 mM EDTA, 5 mM MgCl2, 0.1 % Triton X-100, 0.1 % NP-40, 10 % (v/v) glycerol and 1 mM DTT] and rocked for 3 h with active telomerase reconstituted in vitro with [35S]-TERT [synthesized in vitro in the TNT Quick Coupled Transcription/Translation System (Promega) using plasmid HA-TERT-pcDNA3.1-Zeo] and in vitro transcribed TR (Garforth et al. 2006; Jaouen et al. 2005). Some of the samples were pre-incubated with DNase I or RNase (~5 μg/sample) at RT for 60 min. Proteins associated with washed beads were resolved on SDS/10 % PAAG and analyzed on a PhosphorImager (GE Healthcare). Telomerase activity associated with the beads was determined by the TRAP assay as indicated above.

Results Immunofluorescence microscopy MEFs cells were plated into the SuperFrost microscope slides (Menzel–Gläser) and cultured overnight to ~30–50 % confluency. The slides were rinsed with PBS, fixed in 4 % (v/v) paraformaldehyde for 10 min at room temperature and permeabilized with 0.2 % Triton X-100 for 14 min. Before immunostaining, the slides were blocked with 7 % FBS and 2 % BSA in PBS, followed by overnight incubation with rabbit α-telomerase antibody (sc-7212, Santa Cruz). For colocalization experiments, rabbit α-POT1 (H-200) antibody (sc-33789, Santa Cruz) and mouse γ-H2AX phospho 139 S antibody (05–636, Cell Signaling) were used. After extensive washing with PBS, the cells were blocked with 5 % donkey serum in PBS, followed by incubation with either secondary mouse IgG antibodies conjugated with FITC (715-095-150, Jackson ImmunoResearch) and/or donkey IgG antibodies conjugated with Cy3 (711-165-152, ImmunoResearch). DNA staining was with TO-PRO 3 iodide (Molecular Probes). Samples were mounted in Vectashield (Vector Laboratories) and analyzed using the confocal fluorescence microscope Olympus IX81 equipped with Fluoview 500 confocal unit. 3D-image visualization (one focal plane per nucleus) was done using Aquiarium software (http://cbia.fi.muni.cz/acquiarium.html). For colocalization experiments with POT1 (Cy3), threshold was set for each nucleus to obtain ~80 most intensive signals (telomeres), while for γ-H2AX, threshold was

Lack of HMGB1 results in increased chromosome instability in murine cells Recent report indicated that primary mouse embryonic fibroblasts (MEFs) bearing a homozygous deletion of the HMGB1 gene (designated as HMGB1−/− MEFs or HMGB1 KO MEFs) displayed striking chromosomal instability (Giavara et al. 2005). We have also observed increased numbers of chromosomal aberrations in the HMGB1 KO MEFs immortalized by the SV40 large T antigen [designated as HMGB1 KO MEFs (SV40), these cells were used throughout this study] (Fig. 1). The most common chromosomal aberrations were Robertsonian-like fusions (~20 %), minichromosomes (~5 %), and end-to-end fusions (~7 %). Interestingly, no telomeric signals were detected at the fused sites. Our results complement previous observations of increased chromosome instability in HMGB1 KO primary MEFs (Giavara et al. 2005), suggesting that similar chromosomal instability could be also observed in HMGB1 KO MEFs (SV40). Slight differences in the percentage of Robertsonian like fusions, and end-to-end fusions between the HMGB1 KO primary (Giavara et al. 2005) and the HMGB1 KO MEFs (SV40; Fig. 1) may be related to the immortalization protocol and/or to normal fluctuations of these types of chromosomal instabilities (as well as to different number of metaphases analyzed).

Chromosoma

Fig. 1 HMGB1 deficiency results in chromosomal aberrations in MEFs. HMGB1 deficiency results in chromosome instability in MEFs immortalized by SV40 large T antigen. FISH with a Cy3-labeled telomeric PNA probe (red) in HMGB1−/− MEFs (a–e); chromosomes are stained with DAPI (blue). Arrows chromosomal aberrations such as Robertsonian-like fusions and end-to-end fusions (notice the absence of telomeric signals at sites of fused chromosomes). A total of 57 metaphases were analyzed

tracts present in Mus musculus, often resulting in a long heterogenous smear when assessed by Southern hybridization. To obtain information on possible changes at the level of individual telomeres, a complementary approach using Q-FISH analysis was applied. In this analysis, the HMGB1−/− MEFs show considerably higher heterogeneity of telomere lengths compared to the parental MEF cells (Fig. 2b). The relatively wider distribution of telomere lengths in the HMGB1 KO MEFs is accompanied by a rare appearance of highly extended telomeres. The inclusion of cases of rarely extended telomeres into quantitative evaluation results in a slight increase of the average telomere lengths (about 4 % of the maximum acquired telomere fluorescence signal, p< 0.01). Increased incidence of chromosome ends lacking telomere signal was also observed in HMGB1−/− MEFs (7 % compared to 4 % in wt cells). In summary, knockout of the HMGB1 gene in MEFs results in higher heterogeneity of telomere lengths and general telomere shortening (detected by both TRF and Q-FISH analyses). HMGB1 knockout results in enhanced DNA damage response at telomeres

Telomere length is deregulated in HMGB1-knockout MEFs A possibility that increased chromosomal instability in HMGB1 KO MEFs could be related to telomere alternations (Murnane 2006) was investigated by comparing changes in telomere lengths in parental and HMGB1 KO MEFs by two independent techniques: (1) terminal restriction fragments (TRF) or (2) Q-FISH analysis. Data obtained from TRF analysis revealed that HMGB1 deficiency exhibited no effect on telomeric G-overhang integrity (Fig. 2a, native gel) but the average telomere lengths evaluated from this pattern were 22.5 and 16.7 kb for parental and HMGB1 KO MEFs, respectively, indicating a shortening of telomeres by ~6 kb (26 % decrease) in the HMGB1-defficient cells. Reprobing of the gel after denaturation revealed a total signal of telomeric sequences (Fig. 2a, denatured gel) showing new higher-molecular-weight fragments (marked by asterisks, the signals likely represent fragments of telomeric sequence without G-overhangs), in addition to the fragments detected in the previous hybridization in native gel. The mean telomere lengths evaluated from the denatured gel is therefore shifted towards higher values (59 and 48 kb in the wt and HMGB1 KO MEFs, respectively), indicating a total decrease of telomere lengths in HMGB1-deficient MEFs by ~11 kb (19 %). No changes in the structure of telomeric chromatin in the HMGB1 KO MEFs have been detected by micrococcal nuclease digestion and hybridization with telomeric probe (Electronic supplementary material (ESM) Fig. S1). Monitoring telomere length of mouse telomeres is a technical challenge due to the extremely long telomeric

The observed chromosome instability (Fig. 1) and decreased telomere length (Fig. 2) in HMGB1 KO MEFs prompted us to assess telomere dysfunction-induced foci [TIF, (Takai et al. 2003)]. It is assumed that dysfunctional (uncapped) telomeres resemble damaged DNA as revealed by association with a number of DNA damage response factors including γ-H2AX (Paull et al. 2000). To test this, we have studied colocalization of telomeres (detected by POT1 antibody) with DNA damage associated protein γ-H2AX. Dual staining revealed that a number of γ-H2AX foci colocalized with POT1 in MEFs and therefore represented telomeric loci. The percentage of cells with at least five γ-H2AX foci at telomeres was 8.9 and 27 % for wt and HMGB1−/− MEF cells (SV40), respectively. Although immortalization protocol affected the exact percentage of colocalized MEFs γ-H2AX foci at telomeres in SV40 or 3T3 MEFs, the HMGB1-deficient MEFs exhibited (regardless of the immortalization protocol) ~3–5-fold increase in the proportion of cells with more than five colocalization events with telomeres compared to the wt MEF cells (Fig. 3d). Detection of γ-H2AX positive regions at telomeres in wild-type MEFs (Fig. 3d) could be a consequence of a number of cellular events including dysfunctional telomeres in tumor cells [reviewed in Fernandez-Capetillo et al. (2004)]. While γH2AX foci were present in both wt and HMGB1−/− MEF cells, enhanced colocalization of the γ-H2AX foci with telomeres was observed only in HMGB1−/− MEF cells (Fig. 3d), indicating that the lack of HMGB1 protein in mouse embryonic fibroblasts resulted in increased telomere uncapping.

Chromosoma Fig. 2 HMGB1 deficiency results in deregulation of telomere length in MEFs. a TRF analysis. HMGB1−/− and HMGB1+/+ MEFs (SV40-T) were digested by TaqI and TRFs run on pulsed-field gel electrophoresis, stained with ethidium bromide (EtBr), and hybridized with the 32P-labeled telomeric probe (CCCTAA)4. Asterisks high-molecular-mass TRFs which is visible only after denaturation. b Q-FISH analysis. Histogram of telomere lengths were evaluated from relative fluorescence of individual telomeres (RFU) determined by Q-FISH in HMGB1−/− and parental HMGB1+/+ MEFs. Each histogram represents 600 chromosomes. Values on x axis are displayed in arbitrary fluorescence units (RFU/1000) and are segmented into categories representing intervals of 30 units. Y-axis indicates frequencies (the number of events in each signal category). Evaluation of results shows 40 units difference between medians in HMGB1−/− and parental HMGB1+/+ MEFs. Statistical analysis shows the difference as significant at the level of p10-fold) in nuclear extracts from the HMGB1 KO MEFs as compared to the wt MEFs (Fig. 4). As telomerase is often activated in cells immortalized by SV40 large T antigen (Foddis et al. 2002; Wang and Zhu 2004), we have also measured telomerase activity in the wt MEFs immortalized by a different protocol, the cell passage 3T3. As shown in Fig. 4, telomerase activity in parental MEFs was markedly lower (~10-fold) in cells immortalized by the 3T3 protocol as compared to cells immortalized by the SV40 large T antigen. However, regardless the immortalization method used, the HMGB1-knockout in these MEF cells resulted in ~5- and ~10-fold decrease in telomerase activity.

The latter data strongly suggested that the observed lower telomerase activity in the HMGB1 KO MEFs relative to the wild-type MEFs (Fig. 4) was a consequence of the lack of HMGB1 protein in these cells. Thus, HMGB1 enhances telomerase activity in mouse embryonic fibroblasts. HMGB1 interacts with active telomerase complex To determine whether HMGB1 can interact with active telomerase, pull-down experiments were carried out with telomerase reconstituted from in vitro translated [ 35S]mTERT (reticulocyte lysate, RRL) and in vitro transcribed mTR. As shown in Fig. 5a, [35S]-mTERT was found to be associated with HMGB1. HMGB1 consists of two HMGB1 domains, A and B (termed HMG-boxes), and an acidic Ctail (ESM Fig. S2). Similar pull-down experiments with HMGB1 lacking the C-tail (HMGB1-ΔC) revealed binding to mTERT, indicating that the HMG-boxes of HMGB1, and not the acidic C-tail, were responsible for the interaction of HMGB1 with mTERT (Fig. 5a). Association of mTERT

Chromosoma Fig. 3 HMGB1-deficiency in MEFs results in increased DNA damage response at telomeres. Representative images of POT1 (red), γ-H2AX (green), DNA (blue) and merge (POT1/γH2AX/DNA, lower right panels in a–c) immunofluorescence of the whole nuclei of the wild-type MEFs (a) and the HMGB1−/− MEFs (b). c 3D-projections (one focal plane per nucleus) of the HMGB1−/− MEFs. Colocalization of γ-H2AX foci with telomeres (POT1) is indicated in 3D images by arrowheads. a–c show images of MEF cells immortalized by SV40 large T antigen. d Percentage of cells with ≥5 colocalization events (γ-H2AX foci at telomeres) in MEF cells immortalized by the SV40 large T antigen (SV40) or the 3T3 passage protocol (3T3). In each of the MEF cell lines, ≥ 100 nuclei were analyzed by the Acquarium software. Inset in d depicts enlarged colocalization events shown in c (indicated by arrowheads). Scale bar 1.5 μm

with the RNA component of mouse telomerase, mTR, into active telomerase complex was not a prerequisite for

Fig. 4 HMGB1 enhances telomerase activity in MEFs. Telomerase activity assays (TRAP) of cellular extracts from HMGB1+/+ and HMGB1−/− MEFs immortalized by SV40 large T antigen (SV40-T) or by the 3T3 protocol (3T3 cells) were performed and quantified in a dual-color real-time PCR mode (left) or by the conventional TRAP assay and gel electrophoresis (right). Arrowhead indicates amplification of TRAP internal control template. 1 and 3 HMGB1+/+ MEFs; 2 and 4 HMGB1−/− MEFs. Telomerase activity is expressed relative to the activity in HMGB1+/+ MEFs which is regarded 100 %. Results were calculated from at least three independent experiments, each in quadruplicates

HMGB1 binding to mTERT as HMGB1 could interact with isolated mTERT in the absence of mTR (ESM Fig. S2). HMGB1 interaction with mTERT was observed irrespective the presence or absence of DNase I, demonstrating that the physical interaction of HMGB1 with mTERT was not mediated via DNA that could be present in the RRL and/or associated with HMGB1. In addition, HMGB1 could also bind to isolated mTR (telomerase RNA component) as revealed by electrophoretic gel-shift assay (ESM Fig. S2). Having detected binding of HMGB1 to mTERT in vitro (Fig. 5a, top), TRAP assay was used to demonstrate association of telomerase activity with HMGB1 or HMGB1 lacking the acidic C-tail (Fig. 5a, low). Higher telomerase activity (and correspondingly higher amount of [35S]mTERT) associated with HMGB1-ΔC, as compared to the full-length HMGB1, could indicate that the acidic C-tail of HMGB1 downregulated the ability of the two HMGB1 domains to interact with active telomerase. Interestingly, only background activity of telomerase was associated with individual HMG boxes of HMGB1 regardless their distinct preferences for mTERT binding (Fig. 5a). These data demonstrated that the interaction of active telomerase with HMGB1 in vitro required two (covalently linked) HMGboxes of the protein. Importantly, assembly of mTERT with mTR into active telomerase ribonucleoprotein complex could enhance (~3-fold) the affinity of mTERT for HMGB1,

Chromosoma Fig. 5 Association of HMGB1 with active telomerase. a HMGB1 associates with active telomerase reconstituted in vitro from [35S]-TERT and TR in reticulocyte lysate. [35S]-TERT (autoradiography) and telomerase activity (TRAP) associated with GST-fused HMGB1 or domains by the in vitro pull-down assay (PD). Equal loading of GST proteins after pull-down is shown by gel staining (Coomassie). Arrowhead indicates amplification of TRAP internal control template. b Enhanced affinity of mTERT for HMGB1 in active telomerase as revealed by the PD assay in the absence or presence of RNase

as revealed by pre-incubation of the reconstituted telomerase with RNase before the pull-down assay (Fig. 5b). To find out whether HMGB1 could interact with active telomerase in MEFs, immunoprecipitation (IP) of cellular lysates was carried out. Using western blotting/antibody detection, we were unable to detect unambiguously telomerase (TERT protein) that would coimmunoprecipitate with HMGB1 (ESM Fig. S4). However, the TRAP assay consistently detected ~4-fold higher telomerase activity coimmunoprecipiting with HMGB1 relative to the control IP experiment (ESM Fig. S4; see “Discussion” section). Effect of HMGB1 on telomerase activity in vivo and in vitro The involvement of HMGB1 in modulation of cellular activity of telomerase was also studied by transient (plasmid-based) overexpression of HMGB1 in MEFs. As shown in Fig. 6c, levels of the overexpressed mouse (HA-tagged) HMGB1 protein were similar to those of the endogenous mouse HMGB1 protein. TRAP assay revealed that overexpression of HMGB1 in MEFs resulted in >3-fold enhancement of telomerase activity relative to the mock-transfected cells (Fig. 6a, b).

In order to find out whether HMGB1 could directly stimulate telomerase activity, purified HMGB1 protein was added either to cellular lysates from MEFs or telomerase reconstituted in vitro. Repeated TRAP assays could not unambiguously detect any significant enhancement of telomerase activity upon addition of HMGB1 protein (either recombinant from E. coli or isolated from mammalian cells, data not shown). Thus, the observed higher activity of telomerase in the wt MEFs relative to the HMGB1 KO MEFs is most likely not a consequence of a direct stimulatory effect of HMGB1 on telomerase due to direct interaction of HMGB1 protein with telomerase (see “Discussion” section). Downregulation of telomerase RNA component mTR in HMGB1-knockout MEFs Regulation of telomerase activity occurs primarily at the level of transcription of its two core subunits, TERT and TR (Collins 2008). To find out whether the decreased activity of telomerase in HMGB1 KO MEFs (Fig. 4a) was due to decreased cellular levels of mTERT, real-time qRT-PCR was used to analyze levels of mTERT. The data obtained

Chromosoma

Fig. 6 Overexpression of HMGB1 in MEFs stimulates telomerase activity. a Telomerase activity in wild-type MEFs (SV40-T) upon (plasmid directed) overexpression of HA-tagged mouse HMGB1 (HA-HMGB1). Telomerase activity was quantified by the TRAP assay in a dual-color real-time PCR mode (telomerase activity is expressed relative to that in mock-transfected parental MEFs which is regarded 100 %) or by conventional TRAP assay analyzed by gel

electrophoresis (arrowhead, internal TRAP assay standard of 56 bp). b Western blotting and antibody detection of cellular lysates [mocktransfected (control) or HA-HMGB1-transfected MEFs] with HMGB1 and β-actin antibodies (loading control). 1 Mock–transfection, 2 transfection with plasmid encoding mouse HMGB1 protein (HA-HMGB1). Endogenous mouse HMGB1 protein, endo-HMGB1

revealed indistinguishable expression levels of mTERT in the wt and HMGB1 KO MEFs (Fig. 7a, left). Correspondingly, identical levels of mTERT protein were detected in the wt and HMGB1 KO MEFs by western blotting/antibody detection of total cellular proteins (Fig. 7a, right). On the other hand, qRT-PCR as well as semiq-RT-PCR (agarose gel) using specific primers for mTR revealed that the HMGB1 KO MEFs possessed up to 3-fold diminished levels of the mTR relative to the wt MEFs (Fig. 7b). Similar conclusion was also reached from the Northern blot analysis revealing significantly decreased levels of the telomerase RNA component mTR in the HMGB1 KO MEFs relative to the wt (HMGB1+/+) MEFs (Fig. 7c). In summary, our results demonstrated that lower telomerase activity in HMGB1 KO MEFs may be related to downregulation of the mTR levels (see “Discussion” section).

KO MEFs relative to the parental cells by qRT-PCR (Fig. 8b) or Northern blot analysis (Fig. 8c). In conclusion, results of this paper revealed that knockout of the HMGB1 gene in mouse embryonic fibroblasts results in increased chromosome instability, enhanced incidence of TIFs, telomere length shortening, and decreased telomerase activity, suggesting that HMGB1 protein is required for telomere maintenance.

HMGB2 protein inhibits cellular activity of telomerase HMGB1 is a member of HMGB family proteins involving closely related HMGB2, and also HMGB3 and recently discovered HMGB4 (Štros et al. 2007). Many of the in vitro-binding properties of HMGB1 and HMGB2 are very similar (Štros et al. 2007), albeit their biological roles may be distinct as revealed by different phenotypes of HMGB1 and HMGB2 knockout mice (Calogero et al. 1999; Ronfani et al. 2001). Therefore, we have studied whether HMGB2 deficiency in MEFs could also result in decreased telomerase activity as it was the case of the HMGB1 KO MEFs (Fig. 4a). Surprisingly, knockout of the HMGB2 gene in MEFs brought about >3-fold enhancement of telomerase activity (Fig. 8a) relative to the parental cells. Although we have not yet addressed mechanism of modulation of telomerase activity by HMGB2, we have consistently determined up to ~3-fold elevated levels of mTR in the HMGB2

Discussion HMGB proteins are subject of intensive research due to their involvement in numerous biologically important processes such as DNA replication, recombination, repair, transcription, tumor migration/metastasis, and also cellular signaling (Štros 2010). Experiments presented in this paper revealed HMGB1 involvement in telomere maintenance. A moderate telomere shortening (TRF and Q-FISH analyses) observed in HMGB1-knockout MEFs, as well as increased DNA damage response at telomeres, is most likely a consequence of decreased telomerase activity. Telomere shortening has been observed also in our parallel study in AtHMGB1knockout mutants of Arabidopsis thaliana (Procházková Schrumpfová et al. 2011). The occurrence of substantially longer telomeres (detected by Q-FISH) in mouse HMGB1knockout cells may be due to occasional alternative lengthening of telomeres by telomerase-independent recombination mechanism but we have no experimental evidence directly supporting this hypothesis. Various factors have previously been reported to influence telomerase activity in the cells, including intracellular location and phosphorylation of telomerase (Collins 2008). However, no distinct phosphorylation (Štros, unpublished data) or different nuclear cytoplasmic distribution of

Chromosoma

Fig. 7 Downregulation of telomerase RNA component mTR in HMGB1–knockout MEFs. a Left levels of mTERT in HMGB1+/+ and HMGB1−/− MEFs were determined by real-time qRT-PCR using GAPDH as a normalizer. a Right mTERT protein was analyzed in cellular lysates from the HMGB1+/+ and HMGB1−/− MEFs by Western blotting and antibody detection using α-TERT and β-actin antibodies (loading control). b Levels of mTR in HMGB1+/+ and HMGB1−/− MEFs were determined by qRT-PCR or semi-qRT-PCR (1.5 % agarose gels stained with ethidium bromide) using GAPDH as a normalizer. Telomerase activity is expressed relative to the activity in the HMGB1+/+ MEFs which is regarded 100 %. Results were calculated from at least three independent experiments, each in quadruplicates. c Northern blot analysis of mTR expression. a Total RNA from the wildtype (HMGB1+/+) and HMGB1−/− MEFs (SV40-T) was used for Northern blotting as detailed in “Materials and methods” section. The RNA blots were hybridized with 32P-labeled DNA probe against mTR (full length). Loading RNA controls were 18 S RNA (ethidium bromide) and actin (detected by hybridization with ~90 bp probe specific for actin)

telomerase was found in parental or HMGB1−/− MEFs (ESM Fig. S3). In addition, despite previously reported alternate splice variants in mTERT mRNA in MEFs (Sýkorová and Fajkus 2009), we have not observed any differences in alternative splicing of mTERT in the parental or HMGB1−/− MEFs (Dobšáková and Fajkus, unpublished data; Sýkorová and Fajkus 2009). Enhancement of telomerase activity in MEFs upon (plasmid directed) overexpression of HMGB1 provided further evidence for HMGB1 involvement in modulation of cellular

activity of telomerase. The inability of purified HMGB1 to enhance telomerase activity indicates that the observed higher telomerase activity in HMGB1-containing MEFs is not a consequence of a direct binding of the protein to active telomerase complex. The same conclusion seems to apply to HMGB2. Although HMGB2 could interact with telomerase in vitro, addition of purified HMGB2 protein to cellular lysate from MEFs could not affect telomerase activity in vitro (not shown). It is possible that HMGB1 (and also HMGB2) could modulate telomerase activity only in the appropriate cellular compartment and time window, possibly via an (unknown) factor or contribution of several factors involved in modulation of telomerase activity and/or assembly of functional telomerase (Collins 2008). For example, the ability of HMGB1 to interact with both mTERT and mTR could serve the purpose of promotion of assembly of the telomerase subunits into active telomerase complex, similarly to the functioning of other chaperones involved in telomerase assembly (Collins 2008). In this respect, HMGB1 could act either directly by interaction with the core telomerase subunits and/or indirectly via promotion of synthesis/folding of factor(s) participating in formation (assembly) of active telomerase (Collins 2008). The latter would be compatible with DNA microarrays revealing diminished expression of numerous heat shock proteins (chaperones) in HMGB1−/−MEFs relative to the parental cells (Krynetskaia et al. 2008), including Hsp90 mRNA as revealed by qRT-PCR (data not shown). Hsp90 protein has previously been shown to be an important factor in activation of telomerase (Holt et al. 1999; Woo et al. 2009). Our preliminary TRAP experiments also revealed partial inhibition of telomerase activity in human breast cancer (MCF-7) and ovarian carcinoma (A2780) cells upon knockdown of HMGB1, demonstrating the ability of HMGB1 to modulate telomerase activity in human cells [data not shown; notice that mouse and human HMGB1 proteins exhibit 86 % similarity in their amino acid sequences (Štros et al. 2007)]. In addition, in vitro pull-down experiments demonstrated HMGB1 binding to human TERT (ESM Fig. S5). Similar binding experiments with truncated forms of hTERT revealed that the presence of the RID2 domain (involved in TR binding) was required for efficient binding of TERT to HMGB1 (ESM Fig. S5). These results support the specificity of HMGB1–TERT interactions in vitro. Although we were unable to detect unambiguously by western blotting/antibody detection telomerase that would coimmunoprecipitate with HMGB1 upon addition of αHMGB1 antibody to cellular lysates from MEFs (ESM Fig. S5), the TRAP assays revealed consistently weak telomerase activity co-immunoprecipiting with HMGB1 (unlike the IP experiment with control antibody; ESM Fig. S4). The apparent discrepancy between the western blotting/antibody detection and the TRAP assay is likely related to the

Chromosoma

Fig. 8 HMGB2 inhibits telomerase activity in MEFs. a Telomerase activity assays (TRAP) of cellular extracts from HMGB2+/+ and HMGB2−/− MEFs immortalized by the 3T3 protocol were performed and quantified in a dual-color real-time PCR mode or by the conventional TRAP assay and gel electrophoresis (b). Arrowhead indicates internal TRAP standard of 56-bp. Presence (+) or absence (−) of HMGB2 gene is indicated. b HMGB2 knockout results in elevated levels of mTR. Levels of mTR in HMGB2+/+ and HMGB2−/− MEFs were determined by qRT-PCR or semi-qRT-PCR (1.5 % agarose gels stained with ethidium bromide) using β-2-microglobulin as a

normalizer. Telomerase activity is expressed relative to the activity in HMGB2+/+ MEFs which is regarded 100 %. Results were calculated from at least three independent experiments, each in quadruplicates. c Northern blot analysis of mTR expression. a Total RNA from the wildtype (HMGB2+/+) and HMGB2−/− MEFs (3T3) was used for Northern blotting as detailed in “Materials and methods” section. The RNA blots were hybridized with 32P-labeled DNA probe against mTR (full length). Loading RNA controls were 18S RNA (ethidium bromide) and actin (detected by hybridization with ~90 bp probe specific for actin)

extreme sensitivity of the TRAP assay which is capable to detect trace levels of endogenous telomerase associated with HMGB1. Only moderate association between HMGB1 and telomerase suggests that the putative HMGB1–TERT interactions in the cell may be weak and/or transient. HMGB1 is the most mobile chromatin-associated protein in the nucleus, and interactions of HMGB1 with proteins can increase residence times of both HMGB1 and its binding partners on chromatin (Agresti and Bianchi 2003; Agresti et al. 2005). Whether HMGB1 binding to telomerase can affect its binding to telomeric regions in chromatin remains to be established. No report is available in the literature whether HMGB1 associates with telomeres, albeit several moderately DNA sequence-specific HMG proteins such as HMG-5 from Caenorhabditis elegans (Im and Lee 2003), HMG1L10 and HMG-A1 (Dejardin and Kingston 2009) were reported to bind telomeric dsDNA. Although HMGB1 binds dsDNA with very little sequence preference, the protein can recognize and bind different DNA structures (such as bent, kinked, or unwound DNA) with very high affinity and specificity [reviewed by (Štros 2010)]. As telomeric repeats do not show any notable DNA curvature, kink, or other structural features which could enhance HMGB1 binding (Fajkus et al. 1995), it seems unlikely that HMGB1 would exhibit significantly higher affinity to telomeric sequences as compared to other dsDNA sequences. Telomerase activation by HMGB1 observed in this work cannot be caused by numerous downstream steps participating in regulation of telomerase activity in vivo, such as structural transition of telomeric DNA (t-loop, quadruplex), accessibility of telomere-G-overhangs, nucleoprotein

structure of telomeres, recruitment of telomerase to telomeres since these steps are not reflected by the TRAP assay in vitro. This, of course, does not exclude a role of HMGB1 in these processes. We can speculate that the telomere displacement loop (d-loop) may represent a potential binding site to HMGB1 considering its certain similarity to four-way junctions, the reported high-affinity binding sites for HMGB1 (Štros 2010). Apart from the above outlined mechanisms of possible involvement of HMGB1 in telomere maintenance, one of the most likely explanations for the decreased telomerase activity in the HMGB1 KO cells are significantly lower levels of mTR relative to the parental cells. Low levels of mTR may originate either from downregulation of mTR transcription, or from the reduced stability of mTR in HMGB1 KO cells. Here, we can speculate that diminished levels of mTR would then result in decreased amounts of assembled (active) telomerase enzyme. This conclusion is compatible with recent reports indicating that transcriptional regulation of mTR, rather than mTERT, is a limiting factor for telomerase activity in vivo [reviewed by Cairney and Keith (2008)]. Measurement of telomerase activity in cellular lysates from HMGB1 and HMGB2 KO MEFs provided evidence that the two closely related proteins of the HMGB family could differentially modulate (with opposite effects) telomerase activity in the cell. This finding is very intriguing given the extreme similarity between HMGB1 and HMGB2 (>80 % identical at the amino acid level) and indistinguishable biochemical properties in vitro (Agresti and Bianchi 2003; Štros 2010). However, it is possible that the biological

Chromosoma

roles of HMGB1 and HMGB2 proteins do not overlap completely as revealed by different phenotypes of HMGB1 and HMGB2 KO mice, suggesting that HMGB1 and HMGB2 are not perfectly equivalent in their functions (Calogero et al. 1999; Ronfani et al. 2001). While HMGB1 and HMGB2 proteins are widely expressed during embryogenesis, HMGB2 expression (unlike HMGB1 expression) is restricted in adult mice mainly to lymphoid organs and testes (Ronfani et al. 2001). However, HMGB2 (like HMGB1) is also expressed in all human or mouse immortalized cells (Ronfani et al. 2001). Given the relatively constant expression of HMGB1 in the cell [with changes in expression occurring mainly under certain conditions such as in cancer (increase) or when the cells are committed to differentiation (decrease; Agresti and Bianchi 2003; Štros 2010)], the final pool of HMGB1/2 proteins could be modulated by variation in HMGB2 levels (Ronfani et al. 2001). Whether differences in the expression of HMGB1/2 proteins in the cell could fine tune cellular activity of telomerase requires further investigation. The observed down- or upregulation of TR levels in the HMGB1 or HMGB2 KO MEFs, respectively, may suggest that HMGB1 and HMGB2 proteins could differentially modulate telomerase activity by affecting the TR levels leading to distinct amounts of assembled (active) telomerase enzyme. It would be also of interest to find out whether the other members of the HMGB family, HMGB3 and HMGB4 (Štros 2010), could also modulate cellular activity of telomerase, and to determine their possible functional interplay. The observed loss of telomere signals at the sites of chromosomal fusions (via nonhomologous end-joining pathway, NHEJ) is likely related to lower telomerase activity (de Lange 2005). The loss of telomeric repeats can also occur due to problems associated with DNA replication and repair, as supported here by enhanced occurrence of TIFs in HMGB1 KO cells (Fig. 3). Involvement of HMGB1 in NHEJ (Nagaki et al. 1998; Štros 1998; Štros et al. 2000) as well as in other major DNA repair pathways: nucleotide excision repair, base excision repair, and mismatch repair [(Liu et al. 2010) and references therein] has been demonstrated. Thus, it is possible that the observed genomic instability (Giavara et al. 2005) and telomere dysfunctions (this paper) in HMGB1 KO MEFs may be related, in addition to changes in telomerase activity, to the inability of the HMGB1-deficient cells to repair efficiently double-stranded breaks and/or other sites of DNA damage. Discovery of chromosome abnormalities and telomere dysfunctions in HMGB1 knockout mouse embryonic fibroblasts suggests an important role of HMGB1 protein in maintaining telomere integrity in mammalian cells. Our finding on distinct roles of HMGB1 and HMGB2 proteins in modulation of cellular activity of telomerase is highly relevant to cancer biology given the reported overexpression of HMGB1 and HMGB2 proteins in most tumors, and the

ability of HMGB1 to promote tumor growth and metastasis (Taguchi et al. 2000). Acknowledgments Plasmids pBluescript-mTR Native/short(noT7), HA-TERT-pcDNA3.1-Zeo and mHMGB2-Flag-pcDNA3 were kindly obtained from Scott J. Garforth (Einstein College of Medicine, NY, USA) and Lea A. Harrington (Department of Medical Biophysics, University of Toronto), respectively. Wild-type and HMGB1−/− mouse embryonic fibroblast cell lines (SV40-T) were kindly provided by Marco E. Bianchi (San Raffaele Research Institute, Milan). We also thank Thomas R. Cech (University of Colorado, Boulder) for providing plasmids phTERT-HA2 and phTR, and Joachim Lingner (UPLIN, Laussane) for providing plasmids encoding hTERT and its truncated forms. Antibodies to γH2AX and POT1 were kindly provided by Emilie Lukášová and Eva Bártová (Institute of Biophysics, Brno). We are also grateful to Emilie Lukášová and Pavel Matula (Faculty of Informatics, Masaryk University, Brno) for help with confocal microscopy and colocalization analysis using Acquiarium software (http://cbia.fi.muni.cz/acquiarium.html), respectively. This research was supported by grants to M.Š. from the Grant Agency of the Czech Republic (P301/10/0590 and P305/12/2475). JF was supported by the Grant Agency of the Academy of Sciences of the Czech Republic (IAA500040801) and by the project Central European Institute of Technology (CEITEC; CZ.1.05/1.1.00/02.0068) financed from European Regional Development Fund.

References Agresti A, Bianchi ME (2003) HMGB proteins and gene expression. Curr Opin Genet Dev 13:170–178 Agresti A, Scaffidi P, Riva A, Caiolfa VR, Bianchi ME (2005) GR and HMGB1 interact only within chromatin and influence each other’s residence time. Mol Cell 18:109–121 Cairney CJ, Keith WN (2008) Telomerase redefined: integrated regulation of hTR and hTERT for telomere maintenance and telomerase activity. Biochimie 90:13–23 Calogero S, Grassi F, Aguzzi A, Voigtlander T, Ferrier P, Ferrari S, Bianchi ME (1999) The lack of chromosomal protein Hmg1 does not disrupt cell growth but causes lethal hypoglycaemia in newborn mice. Nat Genet 22:276–280 Collins K (2008) Physiological assembly and activity of human telomerase complexes. Mech Ageing Dev 129:91–98 de Lange T (2005) Shelterin: the protein complex that shapes and safeguards human telomeres. Genes Dev 19:2100–2110 Dejardin J, Kingston RE (2009) Purification of proteins associated with specific genomic loci. Cell 136:175–186 Fajkus J, Kovařík A, Kralovics R, Bezděk M (1995) Organization of telomeric and subtelomeric chromatin in the higher plant Nicotiana tabacum. Mol Gen Genet 247:633–638 Fajkus J, Šimičková M, Malaska J (2002) Tiptoeing to chromosome tips: facts, promises and perils of today’s human telomere biology. Philos Trans R Soc Lond B Biol Sci 357:545–562 Fajkus J, Koppová K, Kunická Z (2003) Dual-color real-time telomeric repeat amplification protocol. Biotechniques 35:912–914 Falciola L, Spada F, Calogero S, Langst G, Voit R, Grummt I, Bianchi ME (1997) High mobility group 1 protein is not stably associated with the chromosomes of somatic cells. J Cell Biol 137:19–26 Fernandez-Capetillo O, Lee A, Nussenzweig M, Nussenzweig A (2004) H2AX: the histone guardian of the genome. DNA Repair (Amst) 3:959–967 Foddis R, De Rienzo A, Broccoli D, Bocchetta M, Stekala E, Rizzo P, Tosolini A, Grobelny JV, Jhanwar SC, Pass HI, Testa JR, Carbone

Chromosoma M (2002) SV40 infection induces telomerase activity in human mesothelial cells. Oncogene 21:1434–1442 Garforth SJ, Wu YY, Prasad VR (2006) Structural features of mouse telomerase RNA are responsible for the lower activity of mouse telomerase versus human telomerase. Biochem J 397:399–406 Giavara S, Kosmidou E, Hande MP, Bianchi ME, Morgan A, d’Adda di Fagagna F, Jackson SP (2005) Yeast Nhp6A/B and mammalian Hmgb1 facilitate the maintenance of genome stability. Curr Biol 15:68–72 Hemann MT, Greider CW (2000) Wild-derived inbred mouse strains have short telomeres. Nucleic Acids Res 28:4474–4478 Herbert BS, Hochreiter AE, Wright WE, Shay JW (2006) Nonradioactive detection of telomerase activity using the telomeric repeat amplification protocol. Nat Protoc 1:1583–1590 Holt SE, Aisner DL, Baur J, Tesmer VM, Dy M, Ouellette M, Trager JB, Morin GB, Toft DO, Shay JW, Wright WE, White MA (1999) Functional requirement of p23 and Hsp90 in telomerase complexes. Genes Dev 13:817–826 Im SH, Lee J (2003) Identification of HMG-5 as a double-stranded telomeric DNA-binding protein in the nematode Caenorhabditis elegans. FEBS Lett 554:455–461 Jaouen S, de Koning L, Gaillard C, Muselikova-Polanska E, Štros M, Strauss F (2005) Determinants of specific binding of HMGB1 protein to hemicatenated DNA loops. J Mol Biol 353:822–837 Krynetskaia N, Xie H, Vucetic S, Obradovic Z, Krynetskiy E (2008) High mobility group protein B1 is an activator of apoptotic response to antimetabolite drugs. Mol Pharmacol 73:260–269 Liu Y, Prasad R, Wilson SH (2010) HMGB1: roles in base excision repair and related function. Biochim Biophys Acta 1799:119–130 Malaska J, Skleničková M, Krejčí K, Fajkusová L, Bajer M, Hrstková H, Fajkus J (2000) Telomerase activity and expression and telomere analysis in situ in the course of treatment of childhood leukemias. Blood Cells Mol Dis 26:534–539 Murnane JP (2006) Telomeres and chromosome instability. DNA Repair (Amst) 5:1082–1092 Nagaki S, Yamamoto M, Yumoto Y, Shirakawa H, Yoshida M, Teraoka H (1998) Non-histone chromosomal proteins HMG1 and 2 enhance ligation reaction of DNA double-strand breaks. Biochem Biophys Res Commun 246:137–141 Paull TT, Rogakou EP, Yamazaki V, Kirchgessner CU, Gellert M, Bonner WM (2000) A critical role for histone H2AX in recruitment of repair factors to nuclear foci after DNA damage. Curr Biol 10:886–895

Poon SS, Lansdorp PM (2001) Measurements of telomere length on individual chromosomes by image cytometry. Methods Cell Biol 64:69–96 Procházková Schrumpfová P, Fojtová M, Mokros P, Grasser KD, Fajkus J (2011) Role of HMGB proteins in chromatin dynamics and telomere maintenance in Arabidopsis thaliana. Curr Protein Pept Sci 12:105–111 Ronfani L, Ferraguti M, Croci L, Ovitt CE, Scholer HR, Consalez GG, Bianchi ME (2001) Reduced fertility and spermatogenesis defects in mice lacking chromosomal protein Hmgb2. Development 128:1265–1273 Štros M (1998) DNA bending by the chromosomal protein HMG1 and its high mobility group box domains. Effect of flanking sequences. J Biol Chem 273:10355–10361 Štros M (2010) HMGB proteins: interactions with DNA and chromatin. Biochim Biophys Acta 1799:101–113 Štros M, Cherny D, Jovin TM (2000) HMG1 protein stimulates DNA end joining by promoting association of DNA molecules via their ends. Eur J Biochem 267:4088–4097 Štros M, Ozaki T, Bačíková A, Kageyama H, Nakagawara A (2002) HMGB1 and HMGB2 cell-specifically down-regulate the p53and p73-dependent sequence-specific transactivation from the human Bax gene promoter. J Biol Chem 277:7157–7164 Štros M, Bačíková A, Polanská E, Štokrová J, Strauss F (2007) HMGB1 interacts with human topoisomerase IIalpha and stimulates its catalytic activity. Nucleic Acids Res 35:5001–5013 Sýkorová E, Fajkus J (2009) Structure–function relationships in telomerase genes. Biol Cell 101:375–392, 371 p following 392 Taguchi A, Blood DC, del Toro G, Canet A, Lee DC, Qu W, Tanji N, Lu Y, Lalla E, Fu C, Hofmann MA, Kislinger T, Ingram M, Lu A, Tanaka H, Hori O, Ogawa S, Stern DM, Schmidt AM (2000) Blockade of RAGE-amphoterin signalling suppresses tumour growth and metastases. Nature 405:354–360 Takai H, Smogorzewska A, de Lange T (2003) DNA damage foci at dysfunctional telomeres. Curr Biol 13:1549–1556 Thomas JO, Travers AA (2001) HMG1 and 2, and related ‘architectural’ DNA-binding proteins. Trends Biochem Sci 26:167–174 Wang S, Zhu J (2004) The hTERT gene is embedded in a nucleaseresistant chromatin domain. J Biol Chem 279:55401–55410 Woo SH, An S, Lee HC, Jin HO, Seo SK, Yoo DH, Lee KH, Rhee CH, Choi EJ, Hong SI, Park IC (2009) A truncated form of p23 downregulates telomerase activity via disruption of Hsp90 function. J Biol Chem 284:30871–30880