Interstitial telomeric repeats-associated DNA breaks

ABSTRACT

During a cell's lifespan, DNA break formation is a common event, associated with many processes, from replication to apoptosis. Most of DNA breaks are readily repaired, but some are meant to persist in time, such as the chromosome ends, protected by telomeres. Besides them, eukaryotic genomes comprise shorter stretches of interstitial telomeric repeats. We assumed that the latter may also be associated with the formation of DNA breaks meant to persist in time. In zebrafish and mouse embryos, cells containing numerous breakage foci were identified. These breaks were not associated with apoptosis or replication, nor did they seem to activate DNA damage response machinery. Unlike short-living, accidental sparse breaks, the ones we found seem to be closely associated, forming discrete break foci. A PCR-based method was developed, allowing specific amplification of DNA regions located between inverted telomeric repeats associated with breaks. The cloning and sequencing of such DNA fragments were found to denote some specificity in their distribution for different tissue types and development stages.

KEYWORDS:

• Chromatin
• development
• DNA breaks
• ETUNEL
• interstitial telomeric repeats
• mouse
• zebrafish

Introduction

Maintenance of genome integrity is a prerequisite for the stable inheritance of genetic information over multiple cell generations. Accumulation of DNA lesions has impacts on aging, as well as on the development of cancer and other pathologies.^1,2 Among others, telomere damage drastically affects many events, from the expression of many genes to major changes of nuclear landscape.³ Cellular mechanisms involved in genome maintenance include the stabilization of chromosome ends by telomeres and a system of DNA-damage response (DDR) pathways that promptly detect double-stranded breaks.⁴ Telomeres perform a wide variety of functions, including maintenance of chromosome ends, their elongation by telomerase, and the regulation of adjacent genes.^5,6 Telomeric DNA consists of repetitive guanine-enriched sequences (TTAGGG in vertebrates). DNA damage signaling pathways, induced by lesions located at telomeres, often differ from the ones activated if chromosomes are damaged sporadically at internal sites.⁷ Along with telomeric repeats at chromosome ends, interstitial telomeric or telomere-like repeats of variable length are present virtually in all eukaryotic genomes.⁸ The opinions concerning the fact that they may be⁹ or may not be¹⁰ hot spots of DNA breakage are rather controversial.

Besides telomeres, there are other DNA strand breaks that normally should not induce the DDR, such as stalled replication forks at the borders of early and late-replicating chromatin domains or recombination sites.^5,8,10 Beyond that, repetitive elements in general and telomeric repeats in particular, including the interstitial ones as well, may require the recruitment of additional factors for efficient replication and for stabilization, especially in case of long tracks of repeats.¹¹ Given that telomeric repeats play a key role in DNA end protection,^5-7 we assumed that if, apart from chromosome ends, some stable DNA breaks arise at specific stages during cell cycle, or are associated with certain developmental events, they may be protected from immediate repair by telomeric repeats. Thus, non-random distribution of interstitial telomeric repeats relative to sites of specific DNA breakage can be hypothesized. In this work, we show that when stable DNA breaks can be detected in some cell types, such as liver or neural cells, these breakage sites appear to be distinct from the known ones, such as provisory breaks related to replication or transcription, they are not linked to apoptosis; also, that these breaks, at least in part, are associated with internal telomeric repeats.

Results

Tissue-specific detection of DNA breaks in situ in Danio rerio and mouse embryos

It is known that programmed cell death plays an important role in embryogenesis and is specifically induced at certain developmental stages.¹² Nevertheless, some DNA breaks not as massive as the ones occurring in apoptosis, but rather stable in time, may appear during normal cell differentiation and chromatin remodeling. To identify them, we developed an enhanced TUNEL procedure (ETUNEL), which utilizes rabbit anti-FITC antibodies, with subsequent detection with goat-anti rabbit FITC-labeled secondary antibodies for signal amplification. The increased sensitivity of the method allowed the identification of a cell subpopulation showing an intense non-uniform nuclear labeling, giving evidence that they contain foci of free DNA ends. These cells demonstrated non-random tissue specific distribution in vertebrates (Fig. 1). For zebrafish, the highest concentration of ETUNEL-positive cells was observed in the epithelia of both fins and body, and especially in the posterior lens (Fig. 1b). Less intense labeling was observed in ganglion cell layer nuclei (Fig. 1b). A rather heterogeneous population of ETUNEL-positive cells was detected in the forebrains and midbrains of 4dpf zebrafish embryos (Fig. 1a).

Figure 1. ETUNEL staining of 4 dpf zebrafish and 11 dpf mice embryos. There is an uneven distribution of the cells with numerous DNA breakage foci within the body, these ones being particularly abundant in epithelia, but also present in the midbrain and forebrain (a). A rather intense ETUNEL signal can be observed in posterior lens, and is also detectable in ganglion cell layer (b). The ETUNEL signal shows a nonuniform distribution within nuclei, often forming small foci, from several to hundreds per nucleus(c). In mouse 11 dpf embryo head ETUNEL-positive cells are concentrated especially near the ventricular zone of the brain, in forming embryonic eyes and epithelia (d). Epithelial (e) as well as neural cells (f) show nonuniform nuclear labeling, with defined breakage foci.

In most labeled cells, an intranuclear staining was distributed nonuniformly throughout the nuclei forming small, often coalescing foci 0.3–0.5 um in size (Fig. 1c). The number of such foci varied from several dozen to hundreds per nucleus. Overall staining intensity also varied depending on tissue of origin, being most prominent in skin and posterior lens (Fig. 1a, b). Rather similar picture has been observed in 11dpf mouse embryos (Fig. 1d), where skin epithelial and neural cells showed prominent and nonuniform labeling (Fig. 1d-f).

DNA breakage foci do not coincide with apoptosis-related DNA fragmentation, gammaH2AX repair foci, and seem to be not associated with replication

Most ETUNEL-positive cells display normal morphology without apparent signs of apoptotic death. Because some tissues contain a rather high amount of ETUNEL-positive cells, it is unlikely that all of them are committed to apoptosis. We performed double labeling of zebrafish embryo cryosections for ETUNEL, and apoptotic markers PARP and caspase 3 (Fig. 2a, b). ETUNEL-positive cells, which have well-defined clusters of breaks, show no apparent colocalization with PARP fragments (Fig. 2a), and do not express active caspase 3 (Fig. 2b). To explore the proliferative status of ETUNEL-positive cells, we performed double labeling of zebrafish tissues with ETUNEL and antibodies against PCNA. No apparent colocalization between the ETUNEL (Fig. 2c) and PCNA signal was found, so the observed breaks are generally not related to normal replication. The formation of DNA breaks commonly triggers DNA damage repair pathway activation, which usually comes out in H2AX histone isoform accumulation, phosphorylated at Ser139 (γH2AX) on chromatin around DSB sites.¹³ Zebrafish tissues showed virtually no γH2AX staining in both ETUNEL-positive and negative cells, with only a few γH2AX-positive cells being present, and no apparent colocalization of ETUNEL signal with γH2AX (Fig. 2d).

Figure 2. Double labeling of zebrafish 4dpf embryo cryosections for ETUNEL and apoptosis, proliferation and reparation markers. There is no evidence of colocalization between ETUNEL and PARP (a), caspase 3 (b), PCNA (c) and γH2AX (d) markers.

Chromatin status near DNA breakage foci

Tissue-specific distribution of the ETUNEL foci in zebrafish and mice, as well as their relative abundance, indicates that they are related to some of the free DNA ends, present in nuclei, which are unlikely to coincide with chromosome ends. Several processes, including chromatin remodeling, can also be associated with DNA break formation, being related either to transitions between hetero- and euchromatin or to open chromatin stabilization.^14-16 To verify whether the observed breakage foci correspond to sites of active chromatin remodeling, a co-staining with SNF2 was performed, showing that ETUNEL signals are not located at chromatin remodeling centers (Fig. 3a). To specify whether the observed foci of free DNA ends are assigned to euchromatin or heterochromatin regions, we analyzed the localization of the ETUNEL signal relative to euchromatin (H3K9-ac) and heterochromatin (H3K9–3me) markers (Fig. 3b, c). H3K9-ac and ETUNEL signals demonstrated a high degree of colocalization (Fig. 3b), while the H3K9-Me3 signal showed no significant overlapping with ETUNEL (Fig. 3c). As open chromatin domains at regions of intense transcription could also be associated with breaks,^17-19 a co-staining with RNA polymerase II was made. We found that the ETUNEL signal location generally doesn't correspond to sites of the most active transcription (Fig. 3d).

Figure 3. In 4dpf zebrafish embryo cells the distribution of ETUNEL signal (green) is different from the one of SNF2 (red) chromatin remodeling marker (a). There is substantial colocalization between ETUNEL (green) and euchromatin marker H3K9-ac (red) (b), and virtually no overlapping with histone H3K9-Me3 (red), specific to heterochromatic regions (c). ETUNEL sites (green) are not located at sites of intense transcription, marked by RNA polymerase II (red) (d).

Identification of interstitial telomeric repeats-associated DNA breaks (ITR-DB)

To identify these repeat-associated DNA breaks (both single- and double-stranded), we designed a PCR-based approach which includes a 2-step procedure. During the first step, all DNA ends are extended using terminal transferase reaction with dTTP. Next, a PCR reaction with a hybrid primer containing repetitive element sequence at the 3' end and polyA at the 5' end is used to amplify DNA stretches encompassed within oppositely directed repeat-associated breaks. We performed PCR with various primers specific to several types of repetitive sequences (telomeric repeats and 5 other types of repeats (Rep 1–5), which are oppositely directed and flank D.rerio transposons). Stretches of these transposon-flanking repeats are found approximately 130, 150, 660, 1220 and 5730 times per D.rerio genome, i.e. with similar frequencies as interstitial telomeric repeats (approximately 1100 times), and 3 of them are G-enriched (> = 50%). Analysis of the amplified fragments on agarose gel showed smear patterns for all tested primers, but we only observed a clear difference in fragment size distribution between terminal transferase-treated and control samples for telomere-specific primers (Figs. 4, 5). These results reflect non-random distribution of DNA breaks with preference toward ITR-containing regions.

Figure 4. Gel-electrophoresis of the amplified DNA fragments, selected by association of the free 3′ DNA ends with telomeric repeats on both sides, obtained by PCR with Term1-Term4 primer mixture. The distribution by size of the amplified DNA fragments from zebrafish hardroe (line 2,5), 4 dpf embryos (line 3,6) and adult fish (line 4,7) shows a general lowering of the molecular weight of the amplified fragments from germline to the embryonic stage and adult organism (a). There are also differences in the predominant fragment sizes of amplified fragments, comprised between DNA break-associated telomeric repeats for different zebrafish tissues, as seen for fin (line 1,7), gill (line 2,8), brain (line3,9), liver (line 4,10), muscle (line 5,11) and hardroe (line 6,12), also showing the highest molecular weight for germline and the lowest – for muscle (b). In both cases a specific signal appears after TdT treatment (a, line 2–4), (b, line 1–6), and there is almost no specific signal without TdT treatment (a, line 5–7), (b, line 7–12).

Figure 5. Gel-electrophoresis of the amplified DNA fragments, selected by possible association of the free 3′ DNA ends with other inverted repeats, originated from 5 types of D.rerio transposons, obtained by PCR with Rep1-Rep5 primers. The distribution of the amplified fragments appears to be constant and not dependent on fish age, neither it is associated with DNA breaks, as there is no apparent difference between samples, treated or not treated with TdT: 1,5 dpf embryos (lanes 1,6), 4 dpf embryos (lanes 2,7), 21 dpf zebrafish (lanes 3,8), adult female (lanes 4,9) and adult male (lanes 5,10).

Since we observed tissue-specific patterns of DNA break distribution by ETUNEL, we performed the same PCR procedure with template zebrafish DNA isolated from various organs, and from organisms at different developmental stages. Interestingly, agarose gel fractionation of PCR products revealed several unusual features (Fig. 4a, b). It was shown that the distribution of breakage sites varies depending on fish age, as shown by the germline 4 dpf embryo and adult fish, size dispersion of the amplified fragments gradually increasing from germline to embryo and finally to adult fish, shifting to lower molecular weight products (Fig. 4a). The control and TdT-treated samples displayed very different patterns of PCR fragment size distribution between different organs, varying much from cells in roe to muscle tissues in adult fish (Fig. 4b). These results prove that the genomic distribution of ITR-associated DNA breaks appears to be tissue-specific. In contrast, there is virtually no association of DNA breaks with 5 other types of zebrafish repeats (Rep1-Rep5) originated from transposons (Fig. 5).

To analyze the sequence features and genomic locations of break-associated regions, PCR fragments were cut out from gel, cloned into a pUC19 vector and sequenced. Our clone library contained 161 individual sequences, representing only a part of the amplified fragments. The sequences are available in the NCBI BioSample database (http://www.ncbi.nlm.nih.gov/biosample/) under accession number SAMN07285974. General features of the sequences in the clone library, including their length, position in the genome and identity percentage comparing to the annotated genome (assembly GRCz10) are given in Table 1. Sequence analysis revealed that clones were represented with different frequencies in the libraries, and could be aligned into homology groups, representing either the same sequences with polymorphism in the length of terminal telomeric repeats, or with additional polymorphism at internal sites (Table 1). Genomic localization of the sequences was rather similar for all of the analyzed fish tissues, notwithstanding apparent differences in fragment size distribution. The number of telomeric repeats next to the terminal stretches of poly-A varied from 3–4 to dozens, however no clear association with age or tissue type was found, regardless the differences in telomerase expression. However, it should be noted that even the enhanced TUNEL procedure did not allow us to detect individual break events, if they were not clustered together. Therefore, the presence of sparse breaks in all tissues cannot be excluded.²⁰ On the other hand, our ETUNEL procedure detects not only ITR-associated breaks, but all types of DNA breaks, although it is unlikely for the latter to persist in cell nuclei without being recognized by cell repair systems or telomerase, unless this recognition process could be trans-regulated and guided by break clustering.

Table 1. General features of the sequences, flanked by inverted telomeric repeats, associated with breaks.

name	length	group	seq in group	chromosome	start	end	length in the genome	identities
63_Skin_8m	449	I	3	chr2	6536371	6536783	412	92%
1_All_1y	114	I	3	chr2	6536509	6536604	95	83%
4b_All_1y	622	I	3	chr2	6537323	6537545	222	36%
2_All_1y	878	II	3	chr1	32166991	32167857	866	99%
5_All_1y	869	II	3	chr1	32166991	32167857	866	100%
4a_All_1y	927	II	3	chr1	32167490	32167817	327	35%
3b_All_1y	1033	III	2	chr2	41085127	41085899	772	75%
3a_All_1y	1005	III	2	chr2	41085631	41086277	646	64%
20_Brain_6m	473	IV	2	chr3	42146648	42146987	339	72%
14_Brain_6m	332	IV	2	chr3	42146679	42146987	308	93%
23_Gill_2y	232	IX	3	chr11	67269	67430	161	69%
27_Gill_2y	166	IX	6	chr11	67269	67373	104	63%
43_Hr_2y	224	IX	6	chr11	67269	67430	161	72%
44_Hr_2y	221	IX	6	chr11	67269	67430	161	73%
49_Fin_8m	229	IX	6	chr11	67269	67430	161	70%
24_Gill_2y	295	IX	3	chr22	42255951	42256165	214	73%
118_all_4d	449	V	14	chr16	4659882	4660070	188	42%
121_all_4d	452	V	14	chr16	4659882	4660070	188	42%
125_all_4d	448	V	14	chr16	4659882	4660070	188	42%
135_all_4d	450	V	14	chr16	4659882	4660070	188	42%
17_Brain_6m	449	V	14	chr16	4659882	4660070	188	42%
18_Brain_6m	452	V	14	chr16	4659882	4660070	188	42%
58_Hr_8m	453	V	14	chr16	4659882	4660070	188	42%
60_Hr_8m	453	V	14	chr16	4659882	4660070	188	42%
70_Gill_2y	448	V	14	chr16	4659882	4660070	188	42%
76_Gill_2y	450	V	14	chr16	4659882	4660070	188	42%
83_Gill_2y	449	V	14	chr16	4659882	4660070	188	42%
86_Gill_2y	450	V	14	chr16	4659882	4660070	188	42%
93_Hr_2y	450	V	14	chr16	4659882	4660070	188	42%
88a_Hr_2y	365	V	14	chr16	4659899	4660070	171	47%
19_Brain_6m	392	VI	3	chr1	42254595	42254970	375	96%
116_Skin_8m	1533	VI	3	chr7	62165117	62166623	1506	98%
36_Gill_2y	1400	VI	3	chr7	62269666	62271046	1380	99%
124_all_4d	703	VII	2	chr21	41154295	41154996	701	100%
21_Brain_6m	705	VII	2	chr21	41154295	41154998	703	100%
78_Gill_2y	1047	VIII	27	chr10	32330923	32331740	817	78%
104b_Fin_8m	245	VIII	27	chr18	36052648	36052803	155	63%
112_Hr_8m	383	VIII	27	chr18	36052648	36052803	155	40%
142_all_4d	256	VIII	27	chr18	36052648	36052803	155	61%
144_all_4d	254	VIII	27	chr18	36052648	36052803	155	61%
22_Gill_2y	288	VIII	27	chr18	36052648	36052803	155	54%
28_Gill_2y	257	VIII	27	chr18	36052648	36052803	155	60%
31b_Gill_2y	272	VIII	27	chr18	36052648	36052803	155	57%
35_Gill_2y	258	VIII	27	chr18	36052648	36052803	155	60%
37_Gill_2y	353	VIII	27	chr18	36052648	36052803	155	44%
40_Hr_2y	284	VIII	27	chr18	36052648	36052803	155	55%
42_Hr_2y	261	VIII	27	chr18	36052648	36052711	63	24%
45_Hr_2y	416	VIII	27	chr18	36052648	36052803	155	37%
48_Fin_8m	291	VIII	27	chr18	36052648	36052803	155	53%
51_Fin_8m	310	VIII	27	chr18	36052648	36052803	155	50%
55_Hr_8m	259	VIII	27	chr18	36052648	36052803	155	60%
57_Hr_8m	287	VIII	27	chr18	36052648	36052803	155	54%
59_Hr_8m	252	VIII	27	chr18	36052648	36052803	155	62%
62_Skin_8m	289	VIII	27	chr18	36052648	36052803	155	54%
74_Gill_2y	417	VIII	27	chr18	36052648	36052803	155	37%
82_Gill_2y	287	VIII	27	chr18	36052648	36052803	155	54%
85a_Gill_2y	265	VIII	27	chr18	36052648	36052803	155	58%
88b_Hr_2y	304	VIII	27	chr18	36052648	36052803	155	51%
90a_Hr_2y	256	VIII	27	chr18	36052648	36052803	155	61%
92_Hr_2y	226	VIII	27	chr18	36052648	36052718	70	31%
52_Fin_8m	352	VIII	27	chr12	50644186	50644260	74	21%
87_Gill_2y	344	VIII	27	chr12	50644186	50644251	65	19%
109_Hr_8m	1072	X	9	chr15	31769276	31770333	1057	99%
101_fin_8m	1070	X	9	chr15	31769279	31770333	1054	99%
113_Hr_8m	250	X	9	chr15	31769280	31769515	235	94%
26_Gill_2y	1062	X	7	chr15	31769282	31770333	1051	99%
115a_Hr_8m	1050	X	9	chr15	31769325	31770184	859	82%
104a_Fin_8m	949	X	9	chr15	31769469	31770333	864	91%
81a_Gill_2y	1139	X	9	chr15	31769535	31770329	794	70%
114_Hr_8m	573	X	7	chr15	31769864	31770333	469	82%
102_Fin_8m	627	XI	2	chr17	19518371	19518983	612	98%
29_Gill_2y	637	XI	2	chr17	19518371	19518983	612	96%
80b_gill_2y	1074	XII	2	chr13	5103713	5104540	827	77%
30a_Gill_2y	878	XII	2	chr13	5103788	5104536	748	85%
30b_Gill_2y	912	XIII	2	chr13	5102808	5103597	789	87%
80a_Gill_2y	1131	XIII	2	chr13	5102808	5103632	824	73%
137_all_4d	524	XIV	2	chr21	27739568	27740083	515	98%
32_Gill_2y	523	XIV	2	chr21	27739568	27740083	515	98%
89_Hr_2y	850	XIX	2	chr12	48315677	48316209	532	63%
91_Hr_2y	849	XIX	2	chr12	48315677	48316209	532	63%
33_Gill_2y	178	XV	6	chr10	14376420	14376591	171	96%
53_Hr_8m	171	XV	5	chr10	14376424	14376591	167	98%
90b_Hr_2y	172	XV	5	chr10	14376424	14376591	167	97%
115b_Hr_8m	183	XV	5	chr10	14376427	14376591	164	90%
133_all_4d	156	XV	6	chr18	19434080	19434230	150	96%
34_Gill_2y	243	XVI	8	chr24	27716	27955	239	98%
50_Fin_8m	213	XVI	8	chr24	27746	27955	209	98%
56_Hr_8m	213	XVI	8	chr24	27746	27955	209	98%
105_Hr_8m	652	XVI	8	chr24	27752	28403	651	100%
110_Hr_8m	656	XVI	8	chr24	27754	28404	650	99%
95_Hr_2y	654	XVI	8	chr24	27754	28404	650	99%
126_all_4d	532	XVI	7	chr24	27949	28404	455	86%
122_all_4d	737	XVI	7	chr24	28136	28404	268	36%
131_all_4d	192	XVI	4	chr17	40398860	40398965	105	55%
108_Hr_8m	357	XVII	2	chr9	13980369	13980720	351	98%
61_Skin_8m	274	XVII	2	chr9	13980475	13980720	245	89%
84a_Gill_2y	1122	XVIII	2	chr16	54148009	54148898	889	79%
117_Skin_8m	1110	XVIII	2	chr16	54148009	54149032	1023	92%
119_all_4d	451	XX	2	chr1	11806982	11807422	440	98%
148_all_d4	450	XX	2	chr1	11806982	11807422	440	98%
127_all_4d	508	XXI	3	chr4	15505881	15506389	508	100%
128_all_4d	511	XXI	3	chr4	15505881	15506389	508	99%
138_all_4d	507	XXI	3	chr4	15505881	15506389	508	100%
129_all_4d	511	XXII	3	chr9	54452453	54452826	373	73%
140_all_4d	509	XXII	3	chr9	54452453	54452826	373	73%
143_all_4d	510	XXII	3	chr9	54452453	54452826	373	73%
41_Hr_2y	363	*	1	chr24	27478	27821	343	94%
77_Gill_2y	577	*	1	chr18	40866	41110	244	42%
10b_Mus_6m	623	*	1	chr11	65310	65779	469	75%
96a_Hr_2y	337	*	1	chr11	67269	67449	180	53%
71_Gill_2y	959	*	1	chr10	4286638	4287379	741	77%
120_all_4d	550	*	1	chr23	4643596	4644126	530	96%
68_Gill_2y	582	*	1	chr18	5326771	5327324	553	95%
64_Gill_2y	342	*	1	chr4	6258878	6259207	329	96%
25_Gill_2y	417	*	1	chr3	7105715	7106123	408	98%
96b_Hr_2y	917	*	1	chr21	8044684	8045538	854	93%
46b_Hr_2y	1026	*	1	chr24	10045871	10046589	718	70%
141_all_4d	398	*	1	chr4	10520880	10521274	394	99%
72_Gill_2y	535	*	1	chr22	12017963	12018482	519	97%
136_all_4d	480	*	1	chr3	12047219	12047555	336	70%
38_Gill_2y	1432	*	1	chr6	13403695	13404823	1128	79%
100_Hr_2y	1288	*	1	chr1	13905917	13907203	1286	100%
146_all_4d	179	*	1	chr10	14376420	14376598	178	99%
16_Brain_6m	225	*	1	chr20	15125030	15125183	153	68%
11_Sc_6m	265	*	1	chr16	15141363	15141615	252	95%
99_Hr_2y	833	*	1	chr23	17587860	17588676	816	98%
67_Gill_2y	572	*	1	chr14	18903560	18904119	559	98%
75_Gill_2y	463	*	1	chr3	19648135	19648588	453	98%
132_all_4d	281	*	1	chr10	21295425	21295691	266	95%
69_Gill_2y	1205	*	1	chr1	21847111	21848262	1151	96%
39_Hr_2y	183	*	1	chr9	22588367	22588537	170	93%
12_Brain_6m	360	*	1	chr9	24945547	24945889	342	95%
97_Hr_2y	733	*	1	chr11	26547659	26548384	725	99%
130_all_4d	552	*	1	chr2	26636047	26636585	538	97%
79_Gill_2y	574	*	1	chr8	27049847	27050401	554	97%
139_all_4d	457	*	1	chr1	27567057	27567376	319	70%
47_Hr_2y	205	*	1	chr10	28217115	28217318	203	99%
111_Hr_8m	1098	*	1	chr25	30859335	30860427	1092	99%
149_all_d4	685	*	1	chr21	30971892	30972567	675	99%
73_Gill_2y	379	*	1	chr3	33748924	33749111	187	49%
107_Hr_8m	385	*	1	chr7	35499793	35500164	371	96%
123_all_4d	427	*	1	chr18	35539378	35539796	418	98%
145_all_4d	656	*	1	chr21	36020614	36021243	629	96%
15_Brain_6m	286	*	1	chr7	37660973	37661234	261	91%
85b_Gill_2y	563	*	1	chr21	37867240	37867617	377	67%
103b_Fin_8m	987	*	1	chr8	38270992	38271470	478	48%
103a_Fin_8m	765	*	1	chr8	38271776	38272534	758	99%
66_Gill_2y	1466	*	1	chr24	39651160	39651765	605	41%
106_Hr_8m	583	*	1	chr11	40974540	40975113	573	98%
98_Hr_2y	630	*	1	chr11	41271719	41272340	621	99%
54_Hr_8m	619	*	1	chr22	42254176	42254533	357	58%
134_all_4d	306	*	1	chr7	42306283	42306586	303	99%
8a_All_1y	988	*	1	chr5	42786443	42787016	573	58%
8b_All_1y	848	*	1	chr5	42787531	42787977	446	53%
147_all_4d	354	*	1	chr15	43039747	43040090	343	97%
6a_All_1y	681	*	1	chr1	44345640	44346002	362	53%
6b_All_1y	724	*	1	chr1	44347074	44347479	405	56%
94_Hr_2y	465	*	1	chr12	48289520	48289979	459	99%
7a_All_1y	980	*	1	chr1	49653737	49654198	461	47%
7b_All_1y	785	*	1	chr1	49655066	49655424	358	46%
65_Gill_2y	938	*	1	chr5	51332684	51333609	925	99%
13_Brain_6m	131	*	1	0	0	0	0	0%
9_Eye_6m	68	*	1	0	0	0	0	0%

Discussion

DNA breaks form constantly in living cells and can be caused by endogenous factors, normally accompanying such processes as replication, transcription, chromatin remodeling and recombination, or exogenous ones, such as ionizing radiation or oxidative agents. It is generally understood that if DNA breaks are left unrepaired, DNA damage response is immediately induced, with subsequent cell-cycle arrest, and the accumulation of DSB can lead to cell death, deregulation of cell growth,²¹ or result in genome instability.²²

Being a regulated process, multiple DNA breaks form after programmed cell death is induced.^23,24 Apoptosis-related formation of DNA breaks often accompanies differentiation.^25,26 Consequently, developmentally regulated apoptotic events are observed in lens and limb development,²⁷ recombination during immune cell development ²⁸ and neuron differentiation.² Nevertheless, occasionally, DNA breakage events are not associated with apoptosis, being linked to other processes, such as chromatin remodeling,²² and this interconnection may be reciprocal.^29-31

In the development of certain organisms, from ciliates to vertebrates, massive DNA fragmentation may accompany cell differentiation, and is related to somatic cell specialization, being coupled with deletion of some genome regions and amplification of the others.³² In these cases it usually accompanies terminal differentiation stages, and sometimes results in great differences between somatic and germline genomes.³³ However, in the majority of higher eukaryotes, the genome of somatic cells is rather stable, variations mainly affecting recombination during immune cell development or occurring in cancers, especially after genotoxic treatments.³⁴

The presence of tissue-specific, non-induced DNA breaks raises a question of whether these breaks occur at random or whether there is some sequence specificity for break formation. Some DNA sequences prone to breakage may correspond to recombination hotspots.^35-37 As has been shown by cloning and sequencing of DNA “forum” domains^37,38 or 50kb DNA fragments isolated from non-apoptotic human cells,³⁹ several types of repetitive sequences, which are prone to recombination, were found to associate with breakage sites. These include Alu repeats, the (TCAG)11 microsatellite, as well as CCAGCCTGG and AAAAAAAACAAAA motifs, the murine major satellite and the (TCCAA)n human satellite. Interstitial telomeric repeats (ITR) are another type of repetitive sequences in which numerous DNA lesions have been detected.^40-43

But, virtually in all of these cases, DNA fragmentation and subsequent healing is a rather rapid process, and, on the contrary, the accumulation of DNA breaks in differentiated cells is considered to be a consequence of their inability to efficiently repair DNA.⁴⁴

Our study demonstrates that besides transient DNA breaks, which sporadically arise and are promptly repaired during cell life or accompany cell death, there are foci of persistent DNA breaks which can be visualized in different cell types, such as neural or epithelial cells. The amount and distribution of free DNA end foci appears to be tissue-specific, so they may be associated with cell differentiation or may help to maintain specific chromatin architecture.

If, apart from chromosome ends, other stable DNA breaks exist in cells at some stages during a cell's lifecycle, or are associated with developmental events, it is conceivable that such breaks may be protected by telomeric repeats. Interstitial telomeric repeats, at least in part, may be linked to the protection of differentiation-related DNA breaks from repair or cell cycle checkpoint activation. Our findings suggest that certain interstitial telomeric repeats, as well as telomeres at chromosome ends, are associated with DNA breaks, limited to particular cell types during development. On the other hand, in certain cases, amplification of some internal chromosome regions could be hypothesized. Furthermore, if such amplification occurs, the generated short DNA molecules might also be capped by telomeric repeats, either added by telomerase, coming from internal genome regions or added by reverse transcription in a telomerase-independent manner.^45-47

There are different and controversial reports concerning whether the ITRs themselves are hot spots of breakage and recombination.^40,48-50 Some support this idea,⁵¹ while others extend it further, claiming that these potential breakage sites have a mutagenic impact on the adjacent regions.^40,48,50,52 Other reports show that ITRs are not hot spots of breakage or recombination, but rather are associated with chromosome regions of higher stability.^10,53 This can indicate that ITR are functionally different, their stability being dependent on length or genomic context. We found that some ITR are associated with stable DNA breaks organized in clusters and appear to be not a result of apoptotic events, recombination, replication or transcription by themselves, but might harbor mainly regulatory functions, from structural chromatin organization to involvement in epigenetic events. The fact that such repeats are present in rather distant species, such as fish and mice, can also point out that they might be involved in some evolutionary conservative regulatory mechanisms, and the variability in ITR number and location could reflect the plasticity of such processes.

Materials and methods

Cryosectioning

Adult zebrafish (line AB/TL) were dissected on ice, organs or whole embryos were placed in O.C.T. media (Fisher Scientific) and frozen in liquid nitrogen. Cryosections were made on CM3050 cryostat (Leica) within −18 to −25°C temperature range, with a slice thickness of 7 µm. Sections were transferred on polylysine coated slides and fixed in 4% paraformaldehyde in 1xPBS with 5 mM MgCl₂ (PBS*) for 10 minutes, afterwards they were washed 4 times in PBS* and permeabilized with 1% Triton X-100 in PBS* for 15 minutes, washed again and used for subsequent manipulations or stored in 20% glycerol, in PBS* at 0°C.

ETUNEL

Coverslips containing permeabilized zebrafish or mouse cryosections in 20% glycerol were washed 3 times for 10 min with PBS*, equilibrated in a TdT (Terminal deoxynucleotidyl transferase) buffer, 40 μl/slide, for 30 min. The slides were treated with TdT, 0,5 U/μl, and FITC-dUTP, 2,5 μM at 37°C for 1 hour.⁵⁴ The samples were washed 3 times for 10 minutes with 0,05% Tween 20 in PBS*, and blocked with 2% BSA in PBS for 2 hours at 37°C. To amplify the FITC signal, the slides were incubated with rabbit anti-FITC antibodies (Invitrogen) 1:100, in 0,05% Tween 20 in PBS*, and 1% BSA overnight at 4°C. Afterwards the samples were washed 5 times for 15 minutes with 0,05% Tween 20 in PBS* and incubated for 2 hours with FITC-labeled goat anti-rabbit antibody (IMTEK, Russia) at 37°C, at 1:1000 dilution. After washing 5 times for 15 minutes with 0,05% Tween 20 in PBS* coverslips were mounted in Moviol with DAPI.

Immunohistochemistry

Slides with fixed cryosections were washed 3 times for 10 minutes in PBS*. Afterwards, samples were washed in PBS*, 0,05% Tween 20 and then incubated in the same solution, supplemented with 2% BSA for 2 hours at room temperature. Then the slides were incubated with primary antibodies, diluted 1:100 to 1:1000 in 0,05% Tween 20 in PBS* and 1% BSA overnight at 4°C. The following primary antibodies and dilutions were used: mouse anti-FITC (Thermo Fisher Scientific) or rabbit anti-FITC (Abcam), mouse anti-PCNA (Abcam) at 1:1000, rabbit anti-γ-H2AX (Abcam) at 1:500, rabbit anti-caspase 3 at 1:500 (Life Technologies), rabbit anti-RNA polymerase II (Abcam) at 1:100, rabbit anti-PARP (Abcam) at 1:100, rabbit anti-SNF2 (Abcam) at 1:200, rabbit anti-Ac-H3K9 (Abcam) at 1:500, and rabbit anti-H3K9-Me3 (Abcam) at 1:200 dilution. After that, samples were washed 4 times (15 min each) in PBS supplemented with 0.05% Tween 20, and treated with goat anti-rabbit FITC or goat anti-mouse Alexa Fluor 555 conjugated secondary antibodies (Life Technologies) at 1:1000 dilution, or, alternatively, with goat anti-mouse FITC (Abcam) and goat anti-rabbit Alexa Fluor 555 (Thermo Fisher Scientific) conjugated secondary antibodies at 1:500 dilution. After washing 4 times for 15 minutes with 0,05% Tween 20 in PBS*, the slides were mounted in Moviol supplemented with DAPI to counterstain the DNA.

Fluorescence microscopy

Slides were examined and photographed with an Eclipse Ti-E inverted fluorescent microscope (Nikon, Japan) equipped with a × 60/NA = 1.4 lens and iXon cooled EMCCD camera (Andor), under the control of NIS-Elements 4.0 software. The entire section area was scanned automatically using a x20/NA = 0.45 lens and images were stitched using NIS-Elements. Serial optical sections were deconvolved using the AutoQuant blind deconvolution algorithm included in the NIS-Elements package.

PCR amplification of regions, flanked with inverted telomere repeats coupled with DNA breaks

High molecular weight genomic DNA from D.rerio was isolated according to The zebrafish book protocol.⁵⁵ For each sample, organs from 5 fish were pooled, or 10–15 whole embryos were used.

To amplify the regions flanked with telomere repeats coupled with breaks, isolated genomic DNA samples were preliminary treated with terminal transferase (Thermo scientific) in a TdT buffer, in the presence of 100 nM dTTP at 37°C for 1 hour to add poly-T tails to all 3′ DNA ends. Then the enzyme was inactivated by heating at 70°C for 10 minutes. Control samples were incubated with the same reaction mixture without enzyme addition.

The PCR reaction with a primer mixture, consisting of primers complimentary to the junction of telomeric repeats and a poly-T tail (Term1–4), was performed under the following conditions: 1x Taq buffer with (NH₄)₂SO₄, 3,5 mM MgCl₂, 0,5 mM dNTPs, 1 mM primer mixture (Table 2), and ∼100 fMol genomic DNA, 2U/50 µl Taq polymerase. The reaction was performed as follows: initial denaturation at 95°C for 5′, then 38 reaction cycles (95°C for 30," 56°C for 20," 72°C for 5′), and final elongation at 72°C for 10′.

Table 2. Primers, used to identify the correlation between DNA breaks and telomeric or other types of repeats.

Name	Sequences
Term1	AAAAAAAAAAAAAAAACCTAACCCTAACCC
Term2	AAAAAAAAAAAAAAAACTAACCCTAACCCTAA
Term3	AAAAAAAAAAAAAAAATAACCCTAACCCTAAC
Term4	AAAAAAAAAAAAAAAAAAACCCTAACCCTAAC
Rep1	AAAAAAAAAAAAAACAGTTTTGGGTGAAC
Rep2	AAAAAAAAAAAAAAAAAGCAGTGGCGCA
Rep3	AAAAAAAAAAAAAAAAAGTGCAAACTCCACA
Rep4	AAAAAAAAAAAAAAAAATCCATCAGACCACA
Rep5	AAAAAAAAAAAAAAAAATGTCAGGAGGAATG

To show whether the DNA-breaks could be found in association with other types of repetitive elements with a genome location similar to those of internal telomeric repeats, primers to the junction of poly-T tail and 5 other types of zebrafish repeats were used (Rep 1–5). PCR conditions were similar to those described for telomeric repeats, except the use of one primer (Rep 1–5 respectively) at 0,5 mM concentration instead of primer mixture. Annealing temperature was 58°C for Rep 2 and 56°C for Rep 1, Rep 3, Rep 4 and Rep5.

The amplification products were analyzed in 1% agarose gel, or cloned in pUC19 plasmid and sequenced.

Analysis of regions, containing telomeric repeat coupled with DNA breaks

For cloning, regions flanked with telomere repeats were amplified using telomeric primers (Term1–4 mix), phosphorylated using T4 Polynucleotide Kinase (Fermentas) according to the manufacturer protocol. PCR products were purified from 1% agarose gel using a GeneJet gel extraction kit (Fermentas). Blunt-ended products were obtained using T4 DNA polymerase. For cloning, dephosphorylated pUC19 vector cut on SmaI was used. Ligation was performed with T4 DNA ligase (Fermentas) according to the manufacturer protocol, using a 1:3 vector to insert ratio. Then the ligation mix was transformed into E. coli JM-109 strain by electroporation, the colonies were grown on LB with ampicillin and inserts were analyzed by electrophoresis. The plasmids with inserts were purified using plasmid MiniPrep kit (Fermentas) and sequenced.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are thankful for support from the Russian Scientific Foundation (project 14–24–00061) for animal care and cryosectioning equipment and service and to the development program of MSU (Plan of research work 5.13) for the equipment provided.

Funding

We are grateful to the Rostok group of companies for their support. This work was partially supported by the Russian Foundation of basic research (project 13–04–40199).