Combinatorial Oligonucleotide FISH (COMBO-FISH): Computer Designed Probe Sets for Microscopy Research of Chromatin in Cell Nuclei

Abstract

Genome sequence databases of many species have been completed so that it is possible to apply an established technique of FISH (Fluorescence In Situ Hybridization) called COMBO-FISH (COMBinatorial Oligonucleotide FISH). It makes use of bioinformatic sequence database search for probe design. Oligonucleotides of typical lengths of 15–30 nucleotides are selected in such a way that they only co-localize at the given genome target. Typical probe sets of 20–40 stretches label about 50–250 kb specifically. The probes are either solely composed of purines or pyrimidines, respectively, for Hoogsteen-type binding, or of purines and pyrimidines together for Watson-Crick type binding. We present probe sets for tumor cell analysis. With an improved sequence database analysis and sequence search according to uniqueness, a novel family of probes repetitively binding to characteristic genome features like SINEs (Short Interspersed Nuclear Elements, e.g., ALU elements), LINEs (Long Interspersed Nuclear Elements, e.g., L1), or centromeres has been developed. All types of probes can be synthesized commercially as DNA or PNA probes, labelled by dye molecules, and specifically attached to the targets for microscopy research. With appropriate dyes labelled, cell nuclei can be subjected to super-resolution localization microscopy.

1. Introduction

Although first models assuming that chromatin in the interphase nucleus is well organized in distinct territories and domains, can be found in the late 19th and early 20th centuries [1, 2], experimental methods of visualization of genome architecture were missing until the 1970s/1980s [3]. With the breakthrough of three-dimensional (3D) light microscopy, especially 3D fluorescence confocal laser scanning microscopy also developments of specific labeling techniques like fluorescence in situ hybridization (FISH; for review see [4]) started their story of success in genome research and medical diagnostics.

FISH is based on the principle that a DNA probe either amplified in bacteria or by PCR represents the nucleic acid sequence of a given target DNA in the cell nucleus or of a metaphase chromosome [5]. Such a probe has to be thermally or chemically denatured into single DNA strands (if not synthesized by PCR) that can bind complementary to the single, that is, denatured DNA strands of given targets [6, 7]. The probes are labeled with fluorochromes. If these single-stranded probe molecules are added in excess to the denatured target strands, they specifically bind to their complementary target DNA so that a DNA–DNA hybrid with fluorescence labeling is formed [8]. Using various probes labeled with fluorochromes of different colors, multi-target visualization can be processed simultaneously [5].

With automated methods for artificial synthesis of high-purity DNA or PNA [9] oligonucleotides, probe sets of custom-made oligonucleotides have become available for many genomic target sites. Also, highly repetitive sequences in centromeres or telomeres were labeled [10]. By means of so-called “oligopaint” probes (fluorescently-labeled single-stranded DNA oligonucleotides) covering target sites by huge amounts of oligonucleotides, sequential, highly specific labeling of various targets from 5 kb up to a few Mb was performed. Oligopaint procedures depend on molecular biology techniques while COMBO-FISH probe design is fully based on computer database investigations. Further details of oligopaint can be found in References [11, 12, 13, 14].

Standard FISH probes as well as oligopainting probes work on probe-target base-pairing according to the Watson–Crick binding scheme, that is, the single-stranded probe complementarily binds to one target strand. This requires heat or chemical denaturation of the complete DNA in a cell nucleus [6, 7] which could impact the preservation of chromosome morphology and especially chromatin nano-structure [15]. In addition, FISH under vital conditions appears to be nearly impossible.

PNA oligonucleotide probes show a higher target affinity than DNA probes. This allows PNA probes sufficient access to DNA targets without additional heat denaturation, because native chromatin acts in an equilibrium state of single- and double-strand conformation [16]. Experimentally, this was only shown for repetitive DNA targets [17] but not in oligopainting experiments of complex targets.

In principle, COMBinatorial Oligonucleotide FISH (COMBO-FISH) potentially has several advantages over standard FISH and could also overcome all such drawbacks mentioned above. COMBO-FISH has been invented in the late 1990s [18] and experimentally realized in the early 2000s [19]. It only uses a few short oligonucleotides for specific labeling of a target. These COMBO-FISH probes can be synthesized as DNA or PNA sequences binding complementarily either as a Watson–Crick double-strand or as a Hoogsteen triple-strand [20, 21, 22] (Figure 1). A low number of probes labeling a target strand reduces synthesis costs and (triple)-strand binding without a strong denaturation step conserves chromatin morphology and organization with the nowadays advantage that the native chromatin structure can be analyzed on the nano-scale by super-resolution localization microscopy in 3D conserved cell nuclei [23, 24, 25].

Figure 1.

Snapshot of a molecular dynamics simulation showing Watson–Crick double-strand pairing and triplex structures of the two classical Hoogsteen pairs for parallel binding: C + *GC (left) and T*AT (right). Note: This figure was originally published in [20] and is reproduced with general permission of the publisher.

2. COMBO-FISH: principle and applications

In contrast to standard FISH where probes are usually cut and amplified by molecular biology techniques, COMBinatorial Oligonucleotide FISH (COMBO-FISH) follows a completely different strategy for probe design [19, 20, 26, 27]. If a genome of a species is sequenced and the sequence is cataloged in a database, oligonucleotide probes can be searched and a set of probes can be designed in silico. Also, their specificity is controlled by computer analysis searching for all possible binding sites of each probe [20, 28, 29].

Any given genome target that should be specifically labeled, can be selected in a DNA sequence database and the numbers of the beginning and end nucleotides are determined exactly. This cannot only be done for human but also for any other species with completely known DNA sequences that can be read in established DNA database archives like NCBI.

The process works as follows [20]: firstly, the beginning and end numbers of a given target (for instance, a gene) are selected. Then oligonucleotide stretches of 15–30 distinct bases are determined in such a way that just the combination of those stretches is singularly co-localizing at the given target site. All binding sites of each probe are determined. Finally, several probe combinations may be excluded from consideration: (a) these are those that have accessory binding sites at other locations in a genome than at the given target site or (b) those that co-localize at several loci in the same genome (Table 1). In addition to these principles governing the basic probe selection, further characteristic features (experimental and theoretical) can be taken into account to ensure a stable homogenous hybridization protocol. Among these, the most important ones are oligonucleotide length, binding energy [28], homo-purine/homo-pyrimidine sequences [30, 31], melting temperature [32], CG-content [32], etc. (see Table 2 as an example). Typically, such oligonucleotide stretches included in a probe set have a length of 15–30 nucleotides each.

Gene	Position on chromosome	Base position	B	R
NRAS	1p13.2	4773158.4824158	62	35
AKT3	1q44	1714365.2069716	55	37
MSH2	2p22.3-p22.1	5331083.5503008	80	49
GNLY	2p12-q11	17,295,832..17300298	113	33
RASSF1	3p21.3	587,156..598306	134	61
FHIT	3p14.2	3106362.3371407	84	50
PIM1	6p21.2	27,935,054..27940329	101	52
ABCB1 (MDR)	7q21.1	12,367,353..12576743	85	54
MET	7q31	41,489,038..41613011	90	26
CMYC	8q24.12-q24.1	191,836..197827	113	57
CDKN2A(p16)	9p21	21,935,774..21963322	101	53
PTEN	10q23.3	897,571..1000507	66	41
ATM	11q22.3	11,636,339..11779887	44	27
KRAS2	12p11.2	2,778,774..2824914	69	43
RB1	13q14	17452380.17630614	64	34
PNN	14q13	19,564,443..19572199	52	32
SNRPNup	15q12	3595436.3750373	40	21
IGF1R	15q25-q26	297,646..606403	101	59
FANCA	16q24.3	610,999..690100	49	32
D17S125	17p12-p11.2	7,147,833..7148031	105	30
ERBB2	17	1593840.1622888	159	77
LAMA3	18q11.2	2,933,849..3024059	120	59
AKT2	19q13.l-q13.2	13,007,022..13060073	81	43
MYBL2	20q13.1	7,348,623..7398038	77	39
PTPN1	20q13.1-q13.2	14,179,798..14253995	78	45
ZNF217	20q13.2-q13.3	17236471.17252494	69	39
PCNT2	21qtel	3053753.3175376	71	38
PDGFB (SIS)	22q13.1	18,834,148..18855844	114	60
TBX1	22q11.2	2892106.2918996	64	38

Table 1.

Examples of COMBO-FISH target sites (“Gene”), their positions according to the banding annotation (“Positions on chromosome”), start and end position in the human data base (“Base position”), possible number of probes (“B”) and the finally remaining specifically co-localizing probes (“R”).

Note: This table was originally published in [20] and is reproduced with general permission of the publisher.

Probe ID	Length (bp)	GC content (%)	Tm^a (°C)	Molar mass (g/mol)	Sequence
AMACR1	16	43.S	40.6	6110.2	AGG	AAG	AAG	GGG	AAA	A
AMACR2	16	43.8	34.4	6110.2	GGA	GGA	AAA	GAG	AAA	G
AMACR7	15	40.0	31.2	5781.0	AGA	AAG	AAA	AGA	GGG
AMACR9	17	35.3	30.6	6068.1	CTT	CTC	TTC	TTT	CTC	TT
AMACR10	17	58.8	45.3	6471.5	GAA	GAG	GAA	AGG	GAG	GG
AMACR11	16	37.5	37.9	5763.9	TCT	TCC	TTT	TCC	CTT	T
AMACR12	16	43.8	30.6	6110.2	GGA	GAG	AAG	AAA	GAA	G
AMACR13	17	76.5	57.9	6519.5	GGG	GGG	AAG	GGG	AGG	GA
AMACR14	15	66.7	43.3	5399.7	CCC	CTC	CCT	CTT	TCC
AMACR16	16	62.5	45.5	5703.9	TTC	CTC	CCT	CCC	CTC	T
AMACR17	15	40.0	31.7	5459.7	TTT	CTC	CTC	TTT	TCC
AMACR23	15	60.0	37.0	5829.0	GAG	AAG	AAG	AGG	GGG
AMACR26	15	53.3	37.0	5813.0	AAG	GAA	GGA	AGA	GGG
AMACR27	15	53.3	29.9	5813.0	GAA	GAG	AAG	GGA	GAG
AMACR28	16	37.5	34.6	6094.2	AGG	GAA	AGA	AGA	AAA	G
AMACR30	15	66.7	39.9	5845.0	GAA	GAG	GGG	GGA	GAG
AMACR31	15	53.3	34.0	5429.7	CCT	CCT	TTC	CTT	CTC
AMACR32	15	50.0	32.9	5733.9	CTC	CTC	TTT	CTC	CTC	T
AMACR33	15	33.3	28.2	5765.0	AAA	AGA	AGG	AAA	GAG
AMACR36	16	43.8	37.3	5748.9	TCC	CTT	TTC	TTC	TCC	T
AMACR37	17	41.2	34.1	6053.1	CTT	TCC	TCT	TCT	TTC	TC
AMACR38	17	41.2	36.9	6423.5	AGA	GAA	AGA	GGA	AAA	GG
AMACR39	17	52.9	42.2	6455.5	AGA	GGA	AGA	AAG	GGA	GG
AMACR40	16	37.5	34.6	5763.9	CTT	CTT	CCT	TCC	TTT	T
AMACR43	15	46.7	34.0	5797.0	AGA	GGA	GAA	AGG	GAA
AMACR45	15	40.0	27.1	5459.7	CTC	TCT	CCT	TCT	TTT
AMACR46	17	52.9	42.3	6023.1	TTT	CTC	CTC	CCC	TCT	CT
AMACR48	15	46.7	34.7	5444.7	TCT	TCT	TCC	TTT	CCC
AMACR50	15	40.0	34.9	5781.0	AAA	AGG	GAG	GAA	AAG

Table 2.

Example of a COMBO-FISH probe set for AMACR and some physical values considered in the selection of oligonucleotide stretches.

Note: This table was originally published under CC BY license in [32].

^a

Melting temperature (median value of the denaturation curve).

Several probe sets created according to this procedure (e.g., ABL, BCR, Her2neu, GRB7, AMACR, etc., see below) have been published or will be shown in the next chapter.

Finally, the computer-optimized probe set can be synthesized as DNA probes, PNA probes, SMART probes, or TINA (Twisted Intercalating Nuclear Acid) probes with high purity [30, 33, 34, 35]. PNA probes have a peptide backbone instead of a sugar–phosphate backbone of DNA. SMART probes also called molecular beacons consist of a stem-loop conformation quenching fluorescence by the closed loop until loop opening when probe and target are binding. TINA probes are oligonucleotides with additional anchoring molecules incorporated. Custom made oligonucleotide probes usually carry one dye molecule at one end or both ends each.

Depending on the base composition of the oligonucleotide probes, they can be designed in their 3′–5′ direction either for Watson–Crick binding (duplex forming probes result in complimentary probe-target double strands) or for Hoogsteen binding (triplex forming probes result in triple strands to homo-purine or homo-pyrimidine sequences as targets of the intact double strand) (Figure 1). COMBO-FISH probes targeting in Watson–Crick configuration are more flexible since they can be designed in a mixture of purine and pyrimidine bases. Hoogsteen binding probes use solely either purines or pyrimidines. It should mentioned that also exceptional (non-homo) Hoogsteen triples exist which can be incorporated into the probe design.

Due to the optical diffraction of a microscope lens, the point image of a fluorescence dye molecule spreads to an image of typically about 250 nm using a high numerical aperture lens. So the fluorescence of a probe combination within a target size less than typically about 250 kb merges into a homogeneous COMBO-FISH “spot”. Typical examples are shown in Figure 2.

Figure 2.

Example fluorescence microscopy images of COMBO-FISH labeling (arrows) of gene targets: (A) HER2/NEU gene labeling in a cell nucleus and (B) on metaphase (combination of 18 oligonucleotide probes); (C) TBX1 gene labeling in a cell nucleus and (D) in an early stage of mitosis (combination of 15 oligonucleotide probes); (E) FMR1 promotor region labeling on chromosome X of a male cell nucleus (combination of 20 oligonucleotide probes); AMACR gene labeling in a cell nucleus (combination of 29 oligonucleotide probes; see Table 2). Note: With the exception of the nucleus in (E), no counterstain was applied.

Detailed protocols for COMBO-FISH labeling of blood cells, fibroblasts, tumor culture cells, or tissue cells are described elsewhere [27, 29]. These protocols can be applied for duplex or triplex forming probe sets. COMBO-FISH for labeling of specific gene targets in cell nuclei has been described for several applications as for instance: The gene of the receptor tyrosine kinase 2 (HER2/NEU) [28, 34] (Figure 2A and B), the gene of the growth factor receptor-bound protein 7 (GRB7) [34], the breakpoint cluster region (BCR) on chromosome 22 [30], the ABL proto-oncogene 1 (ABL) on chromosome 9 [19, 30, 31], and T-box 1(TBX1) [33] (Figure 2C and D), the promotor region of the FMR1 gene [36] (Figure 2E) and the Alpha-Methylacyl-CoA Racemase coding gene (AMACR) on chromosome 5 [32] (Table 2; Figure 2F). Using the probe set for the ABL gene region and Spatially Modulated Illumination Microscopy [37], significant volume changes of the labeled regions were observed in cell nuclei of CML patients before and after medical treatment [35]. Beyond gene target labeling by probe sets of several different probe-sequences co-localizing at a given target only, also unique probes were found that were specifically labeling either interspersed genome regions by one copy each or centromeres by multi-copies (see chapter 4). Such an 18mer oligonucleotide PNA probe repetitively binding on centromere 9 was micro-injected into lymphocyte cell nuclei under in vivo conditions. After further incubation of the labeled cells, the sample was fixed and subjected to microscopy. The results indicated that the probe material was binding to the target region in the nuclei before fixation [33].

3. COMBO-FISH probe sets for genes and breakpoint regions in oncology

Cancer cells are mutations of former “normal” cells. One reason for cancer cell malfunctions are over-expressions of genes, which are mainly caused by two reasons, either one gene is over-expressed or one gene is amplified and additionally to the original gene the copies also express. The latter frequently occurring in solid tumors leads to measureable copy number increases which can be used as diagnostic parameter in tumor biology and medicine. In contrast to solid tumors, blood cell tumors show structural aberration-like translocations in the early stages which can then be followed by numerical aberrations of larger chromosome parts or even whole chromosomes. Translocation chromosomes are resulting by fusion of two parts of different chromosomes that were broken very exactly at breakpoints within a certain breakpoint region; for example, for CML (chronic myeloid leukemia) a famous hallmark is the Philadelphia chromosome with the translocation ABL-BCR t(9,22)(q34,q11), in which a fusion of a part of the abl-region on chromosome 9 with part of the bcr-region on chromosome 22 takes place (for review [38]). Since the breaks occur very exactly at certain breakpoints the fusion region can be transcribed into a functioning protein that does normally not exist in a cell and that is involved in CML-induction.

Although appropriate target sites for homo-purine or homo-pyrimidine probes only make about a few percent of the genome, some prominent tumor genes and breakpoint regions can be specifically labeled by uniquely co-localizing sets of COMBO-FISH probes. Others can only be labeled by a set overlapping on neighboring regions. Probe sets for Her2/neu, abl, and bcr have been published elsewhere [29, 30, 31, 34]. In the following we will show further examples.

It has been observed that in various types of cancer such as breast, ovarian, and squamous cell cancer an amplification of 20q13 occurs. When analyzing such cancers it has been found that often the region which encodes ZNF217 is amplified and an increased expression of the specific region of ZNF217 has been observed. In some cases also neighboring gene encoding sequences are also amplified. It has to be mentioned that the detection of the copy number of ZNF217 can be done by standard FISH, but the shortest available sequence to detect ZNF217 has a length of about 245 kb [39], which means—compared to the length of ZNF217 of about 16 kb—that further amplifications within the probe-binding sequence might not have been visualized using this method. To further reduce the length of the detected target, a COMBO-FISH probe set was designed (Table 3). As a result, we obtained 25 oligonucleotide probes within a range of 133 kb. When reducing the number of sequences to 21 by removing the first and the last three probes the length of the considered DNA region can be downscaled to 91 kb.

Probe number	Beginning nucleotide number
19,588	22,316,443	agggaggaaggagggaggaaggaagga
19,595	22,329,806	agagaaaagagagaaaa
19,600	22,332,226	agggaagaggaagagg
19,601	22,333,382	cctctctcttccctctccttcccctct
19,606	22,338,710	ttccttctttccttccttttctt
19,607	22,338,734	cttccttttctcccc
19,609	22,340,472	gagggaggagggagggaaaa
19,616	22,347,605	ggggaaagaaagaaa
19,623	22,352,916	tcttctcttttctttccttttcct
19,628	22,357,318	ctctttccccttctt
19,641	22,373,058	ggggagagaagaagg
19,643	22,373,175	ctttctcccttttccctcc
^*19650	22,376,231	ccttttctcccctcccctcccctcccct
+19,659	22,389,995	cttctttcctcctttt
19,662	22,396,940	ctttcctccctctctct
19,665	22,403,663	ctcctccttcctcccct
19,671	22,407,703	ccctttcccctcctccct
19,674	22,410,694	ggaggggaagaagaggg
19,675	22,411,085	aaggagaagagaaagagag
19,676	22,411,601	gggagggggaggaggggggagg
19,677	22,412,039	aggagaggggaaaag
19,685	22,421,078	aggaggaaagagaggg
19,710	22,436,400	ggggagggggaaaag
19,724	22,445,917	gagagagagagagagagagaagaaa
19,729	22,449,049	agaaagaaaagagaaagaagagaagag

Table 3.

List of COMBO-FISH probe targets for ZNF217 and its surroundings (target sequences are written from left to right in the 5′-XXXX-3′ direction).

+, probe completely on ZNF217; *, probe partly on ZNF217.

The gene TP53 (sometimes called p53) encodes the “tumor protein 53” (P53). Its purpose is to maintain genomic stability and to control cell growth. Moreover, it is important for the induction of apoptosis and the coordination of repair processes. Labeling of this gene can be obtained by five COMBO-FISH probes only (Table 4). The detection of 10 fluorochromes (dye molecules at both ends of each oligonucleotide probe) would require not only a sensitive microscope but also a background free preparation. Since in tumors P53 is inactivated (sometimes associated by a copy number loss of the gene, see, e.g., [40]), the protein MDM2 which can inactivate P53 when overexpressed can be investigated. In cells with an overexpression of MDM2 an extreme inactivation of the tumor suppressor protein can occur via binding of MDM2 to the transactivation domain of TP53. However, the treated gene, MDM2, does not contain sufficient homo-purine/homo-pyrimidine sequences so that a probe set has to be designed with an overlap on neighboring regions (Table 5).

Probe number	Beginning nucleotide number
7042	7,177,418	agaggagggggagaag
7046	7,181,933	aggaagaggaaggaga
7067	7,193,067	cttctttecctccct
7068	7,193,145	aaagaaggggaggga
7069	7,193,288	ttttctctctetctcctcccctctc

Table 4.

List of COMBO-FISH probe targets for TP53 (target sequences are written from left to right in the 5′-XXXX-3′ direction).

Probe number	Beginning nucleotide number
24,602	31,285,270	gaagagaagaaaggaga
24,603	31,287,071	aagagggaaggaaggg
24,605	31,288,864	cttttctcctccttct
24,608	31,292,189	aaggaagaggagaag
24,609	31,292,783	aagagagagaggggaggaaa
24,618	31,301,112	ctccctctccccctccctcttttccctcctt
24,626	31,312,583	tctctcttcctcttctt
24,648	31,336,603	aaggaggaaggagaggaaaa
24,651	31,339,431	aagagaagggaggggaa
24,656	31,341,897	gggaaggaggagaaggggggagg
24,657	31,342,918	cttctctctctctccccc
24,659	31,344,079	gggagaagggaagga
+24,663	31,349,552	aaaaggaagggagaaag
+24,682	31,375,469	gggaaaaggaagaag
24,690	31,388,934	ttttctcccccttcccccttct
24,692	31,390,072	aagagggaggggaaaag
24,694	31,390,823	gggaaggggagggaaggggaggggag
^*24695	31,390,850	ggaggaaagaagaaaaggaagggaaggggaggg
24,696	31,390,925	aggaagaggagaaggaaggaagaaaggaaagaaa
24,699	31,392,284	aaggaggaagagaaag
24,700	31,392,733	tttctcccttcttct
24,704	31,397,317	tttccttctccctttctct
24,706	31,398,270	ctttctttcctttcctctt
24,710	31,407,205	aagaggggaagggagag
24,724	31,422,982	aggaaaagaagaaaaga
24,726	31,425,374	aaagaaggagggaaa

Table 5.

List of COMBO-FISH probe targets for MDM2 and its surroundings (target sequences are written from left to right in the 5′-XXXX-3′ direction).

+, probe completely on MDM2; *, probe partly on MDM2.

CD44 is a receptor for hyaluronic acid, which plays an important role in cell migration, tumor growth and progression. Accumulating evidences have shown that the CD44 gene is abundantly expressed in cancer-initiating cells (CICs), and has thus been implicated as a CIC marker [41, 42] in several malignancies of hematopoietic and epithelial origin, including gastric cancer. Moreover, CD44 gene amplification was also found in gastric cancer. Table 6 shows the targets for an appropriate CD44 gene probe set.

Probe number	Beginning nucleotide number
24,971	35,103,408	gaaggggagaaggaggaaaggggaaggaaaggag
−24,972	35,108,828	gagggagggagagaaa
24,973	35,109,919	aagagggggaggggaa
24,975	35,112,662	agagagaggggagaggagagaaa
24,977	35,119,179	ctccctccttectcctc
24,978	35,120,540	ttttcctccctctccc
24,980	35,121,359	ctccttctccctttt
−24,981	35,122,451	tctctctttctctttctctctctctc
24,982	35,123,378	gggggggaagaggag
24,984	35,126,758	agggagagaaagaaa
24,985	35,129,005	tccttttcccttcct
24,986	35,132,494	tttcttcccctctct
−24,987	35,135,378	cctetctcctttette
−24,988	35,136,463	aggaaggagagaaagagag
24,991	35,139,280	gaaagggaaaaggaaag
−24,992	35,139,595	agagagagggagaaag
24,993	35,140,078	aaaaggggaggaaag
−24,994	35,141,641	tctctttccctctct
24,996	35,146,244	tcttcttttctcctttt
−24,998	35,147,116	tcttcctccctccctctctcc
24,999	35,147,168	tctccttcttcctcttt
25,004	35,151,365	ccttttctctctccc
25,013	35,169,216	aaaggggggaagaggg
25,014	35,173,890	tcttctctcctcccccc
25,019	35,181,667	ggggaaagagaagaa
−25,021	35,185,117	ccctctccctcctccc
25,022	35,185,516	aaggagaaagagaagg
25,026	35,192,486	agaaagaagaagaaaag
25,027	35,192,612	ccctctcccctccctctctccctcc
−25,028	35,192,661	cttcctttctcttct

Table 6.

List of COMBO-FISH probe targets for the CD44 gene (target sequences are written from left to right in the 5′-XXXX-3′ direction).

−, probes are excluded since they form secondary clusters.

Fusion proteins originate from reciprocal translocations. Sets of oligonucleotides were designed in such a way that translocations get cognizable. There are two breakpoint regions—one on every chromosome. Therefore, four sets of oligonucleotides are needed: One before and after the breakpoint on the two chromosomes. For microscopy labeling with different colors is necessary; the sets of the first chromosome need to be labeled for instance with a red dye and the ones on the second chromosome, for instance, with a green dye. After hybridizing there are the following possible results: (a) Two red spots and two green spots are close together for each color, but the red ones are clearly separated from the green ones. This is the normal case without translocation. (b) There are two parts where one red and one green spot are next to each other. In this case the translocation has occurred: Both chromosomes broke at the major breakpoint and the wrong ends were joined. With four colors more details are visible. For example, when one color is visible on two locations, another breakpoint has been observed.

The minimum requirement for detecting clusters of homo-purine/homo-pyrimidine sequences on DNA is 6 oligonucleotides within a range of 250 kb. It is not necessary that the oligonucleotides are all located on the breakpoint regions itself. To get bright and emphasized signals, sets with 30 oligonucleotids each were designed for the following examples: ABL - BCR t(9,22)(q34,q11); AML1 - ETO t(8;21)(q22;q22); MYC - IGH t(8,14)(q24,q32); PML - RARA t(15,17)(q22,q21);PLZF - RARA t(11,17)(q23,q21). Since these lists would extend the article to an inacceptable size, the lists will be available from the authors on request.

4. COMBO-FISH with repetitively binding, unique single probes, and applications of super-resolution localization microscopy

The following chapter will focus on further novel developments of COMBO-FISH using probe sets of only one uniquely binding oligonucleotide [29] that binds repetitively to a given target like a centromere so that the merging fluorescence leads to a microscopic signal. In Figure 3, examples for centromere 9 [33] and 17 [34] are shown.

Figure 3.

Example fluorescence microscopy images of COMBO-FISH labeling of centromere targets: (A) centromere 17 labeling in a lymphocyte cell nucleus with the repetitively binding probe (Alexa647-5′-cttctgtcttctttttata-3′) and (B) with the repetitively binding probe (Alexa488-5′-tataaaaagaagacagaag-3′). In (C) an overview image of centromere 9 labeling in lymphocyte cell nuclei with the repetitively binding probe (Alexa546-5′-aatcaacccgagtgcaat-3′) is shown indicating a hybridization efficiency better than 90%.

COMBO-FISH probes carrying one fluorochrome molecule at one end of each oligonucleotide are ideal nano-probes for Single Molecule Localization Microscopy [23, 24, 43] (SMLM) in order to analyze chromatin structure and architecture on the nano-scale in subchromosomal regions of cell nuclei [32, 43, 44, 45, 46, 47].

As an example, multiple copies of a repetitive probe for a tri-nucleotide expansion region were hybridized and analyzed quantitatively in cells with Fragile-X syndrome (FXS) or Martin-Bell-Syndrome. FXS belongs to the group of the so-called “trinucleotide repeat expansion disorders” consisting of the expansion of a trinucleotide frequency ((CGG)n-expansion) in the 5′ untranslated region of the Fragile-X Mental Retardation 1 gene (FMR1) on the X-chromosome. The enlargement of the CGG triplet-repeat results in a deactivation of the FMR1 gene and mental retardation of the patient. Multiplets of 6 trinucleotide units ((CGG)₆ or (CCG)₆ probes) were synthesized showing high specificity to the (CGG)-repeat expansion of the FMR1 gene; thereby a minimum of accessory binding sites were found due to the 6-times repetition. Considering the probe length of six trinucleotide units together with one dye molecule at one end, the results of SMLM indicated small chromatin loops for the expansion region rearranging chromatin on the nanoscale so that a deactivation could be explained by geometric reasons in the genome architecture [36].

Although established programs for the design of COMBO-FISH probes and probe sets were available [20, 26], novel so-called alignment-free investigations of k-mers, their frequencies and their positioning along the nucleotide sequence of a chromosome [48, 49] have found oligonucleotide probes that uniquely bind in a given repetition rate to chromatin sequences repetitively occurring as interspersed motives [29]. New generations of specific COMBO-FISH probes were elucidated against SINEs (Short Interspersed Nuclear Elements, e.g., ALU elements [32, 48, 50], Figure 4), LINEs (Long Interspersed Nuclear Elements, e.g., L1 [32]), or centromeres [44]. With these probes, first evaluations of the spatial organization of chromosome 9 were calculated (Figure 5) [32].

Figure 4.

(A) ALU-distribution along the genome: The intensity of the bars indicates the frequency within a 500 kb section of the given chromosome. Red: Position of the designed 17mer ALU probe. The sequence associated with the ALU probe appears in the entire genome at different frequency densities. Blue: Corresponding positions of the ALU consensus sequence. The number of emergence of the 17mer probe sequence was compared with the density of the ALU consensus sequence by using the program “Repeatmasker” [50]. Green: Distribution of a selected 17mer from the L1 element. Although this 17mer appears very often in the genome, the frequency density is significantly different from the selected ALU consensus 17mer. (B) Examples of SMLM images of cell nuclei after COMBO-FISH labeling with the 17mer ALU probe. Note: (A) was originally published under CC BY license in [48].

Figure 5.

(A) SMLM overlay image of Alu densities (green), a centromere 9 points cluster (red), and overlaying regions (blue). Note: Only one image plane (no projection) is shown, where one centromere 9 is located. A magnified point coordinate representation of the white box is shown below. (B)–(D) Estimates of chromosome 9 architecture by its centromere and genomic Alu. (B) Plot of the raw point matrix obtained from SMLM data of (A). A circular approximation of a chromosome 9 territory (black circle) modeled from the theoretical distribution of Alu elements (blue dots) around the chromosome 9 centromere (red dots). Lower image: Magnification of the region of interest in the upper image. (C) Idiogram of chromosome 9 showing the positional distribution of Alu probe binding sites (red), Alu consensus sequences (blue), and binding sites of a probe against genomic L1 elements (green). (D) The radial distribution of Alu signal points around a centromere 9 cluster centroid averaged over 38 centromere 9 clusters. Note: These figures were originally published under CC BY license in [32].

Using quantitative SMLM, the in such a way designed ALU COMBO-FISH probe has also been successfully applied in extending the standard methods of biological dosimetry, which aims at reconstructing or estimating from chromosome aberrations the dose from former radiation exposure [48, 50]. In addition, a novel improved preparation protocol circumvents any heat treatment for target denaturation so that mixed purine–pyrimidine probes can be used, that usually undergo Watson–Crick double-strand pairing (Figure 1). This so-called low-temperature protocol is the prerequisite to combine oligonucleotide-based COMBO-FISH and immunofluorescence staining by means of specific antibodies [29, 48].

5. Conclusion

COMBO-FISH offers a highly variable toolbox of labeling combinations and strategies for chromatin architecture and bio-medical research. Here we have introduced three strategies: (a) Design of a COMBO-FISH probe set which consists of several oligonucleotide probes that specifically co-localize at a given genome target as for instance, a tumor-relevant gene that could be involved in gene copy number changes or tumor-inducing translocations. (b) Design of a COMBO-FISH probe set which consists of one oligonucleotide probe which in many copies specifically co-localize at a given genome target as for instance a centromere. (c) Design of a COMBO-FISH probe set which consists of one oligonucleotide probe that uniquely occurs at several given repetitively occurring genome targets only as for instance SINEs or LINEs. The efficiency of the probe set can be further enhanced by incorporating structural [51, 52, 53] and dynamical parameters determined by molecular dynamics simulations (e.g., AMBER [54, 55, 56], GROMOS [57, 58], CHARMM [59, 60]) into the probe design, which we currently investigate.

In combination with super-resolution localization microscopy and novel tools of data evaluation and interpretation by geometric and topological algorithms [61, 62, 63, 64] COMBO-FISH probes offer new perspectives in understanding the reaction of chromatin as a system as a whole during gene expression, proliferation, or stress response [47].

Acknowledgments

The authors thank former and present team members, Wilhelm von Rosenberg, Laura Doll, Jens Rösler, Stefan Stein, Michael Stuhlmüller, Jin-Ho Lee, Florence Laure Djikimi Tchetgna, Matthias Krufczik, Aaron Sievers, and Jutta Schwarz-Finsterle at the Kirchhoff-Institute of Physics, Heidelberg University for images taken from their theses and publications.

Conflict of interest

The authors declare no conflict of interest.