Page designed and maintained by Octavian Henegariu (Email: or ).
	WARNING: The information provided in these pages is copyrighted and is intended for individual use only. No parts of this work (text, tables or pictures) may be commercialized, published or otherwise reproduced without the written consent of the author. Ref: Nature Biotechnology 2000, Vol 18, p345-348. Get article in PDF format here

---

2. DNA labeling of and purification

2.1. Enzymatic DNA labeling

Two procedures were used to label DNA, nick translation and PCR labeling

Legend for Fig. 1 and Fig. 2 is included in the main text.

2.1.1. Nick translation

Uses the simultaneous activity of two enzymes: (1) DNase I, which in the presence of Mg++ ions becomes a single stranded endonuclease (Fig. 1a), and creates random nicks in the two strands of any DNA molecule. (2) E. coli polymerase I, which through it's 5'-3' exonuclease activity removes nucleotides "in front" of itself, while the 5'-3' polymerase activity adds nucleotides to all the available 3' ends created by the DNase (Fig. 1b, red bars). This exonuclease/polymerase activity, moves (or "translates") any single stranded nick in the 5'-3' direction. When nicks on opposite strands meet, the DNA molecule breaks.

Prepare (and store at -20 C) aliquots of 10x NT buffer (500 mM Tris, pH 7.5, 100 mM MgCl2, 10 mM DTT, 0.5 mg/ml BSA) and 10x beta-mercaptoethanol (100 mM). Always prepare fresh 10x DNase solution [mix 1.0-1.3ul DNase (stock, 3mg/ml) in 1000 ul water, keep on ice for a few minutes, and use immediately].
Prepare stock solutions of nucleotides: 1mM each d(ACG)TP and 5mM dTTP. When a labeled nucleotide (in this case labeled dUTP) is used, it replaces 1/3 to 1/8 of the dTTP in the reaction, depending on the fluor/haptene conjugated to the dUTP.

2.1.2. PCR

The principle of the PCR reaction is illustrated in Fig. 2. Two single stranded DNA primers (18-30 bp long), one forward and one reverse (in other words, with their 3' ends pointing toward each other - yellow arrows, Fig. 2) are synthesized. The primers usually match a known DNA sequence and are used to amplify the fragment in-between.
After adding the primers, the Taq polymerase (or other thermostable polymerase), the buffer and the DNA template, the reaction mix is denatured by heating 30-60 seconds at 94 C (denaturing step). Then, the temperature is dropped to 50-60 C for 30-60 seconds, to allow the primers to anneal (Fig. 2a) to their target sequences (annealing step). Then the temperature is raised to 68-72 C (optimal temperature for Taq polymerase) for 0.5-4 minutes, to allow the enzyme to synthesize the new DNA strands (extension step, Fig 2b). These temperature steps are repeated again (usually 30 cycles), allowing exponential amplification of the DNA molecule between the two primers (Fig. 2c). If part of the dTTP in the reaction is replaced by labeled dUTP, PCR can be used to label the newly synthesized DNA molecules with fluorescent dyes or haptenes.

The 10x PCR buffer includes: 500 mM KCl, 100 mM Tris, pH 8.4, 15-20 mM MgCl2.

2.1.3. Labeling reactions using commercial or custom-made fluorescent-dUTP

When using commercially-labeled dUTP, the dTTP in the reactions is reduced to about 2/3 (130 um in PCR and 35 um in nick translation), whereas the dUTP is 1/3-1/8. Reactions work well, even if the overall dTTP+dUTP amount is somewhat variable.

The same variability of labeled dUTP in the reaction is seen with the custom-made nucleotides. In this case, though, the dTTP is reduced to about 1/3 (70 um in PCR and 17 um in nick translation). The following volumes of 1 mM custom-made nucleotide solutions are added to each reaction (in 100 ul PCR or nick translation):
2 ul (20 um) DEAC, CB, Cy3.5, Cy5.5
3 ul (30 um) R6G, TAMRA, TxR
5 ul (50 um) OG, A488, Cy3, Cy5
and 6-7 ul (60-70 um) AMCA, FITC, BIO, DIG, DNP.
Replacing so much dTTP is possible, because in the custom-synthesized fluorescent nucleotides 50% allylamine-dUTP is non-conjugated. The unconjugated allylamine dUTP readily replaces dTTP in the reaction. Using variable volumes of fluorescent dUTP in labeling reactions results in a variable dye:dTTP ratio, depending on the dye used. Nevertheless, labeling results are good, indicating that a precise dTTP/dUTP analog ratio does not appear to be necessary.

When using custom labeled dUTP:

• PCR labeling reaction requires additional Bovine Serum Albumin (BSA) at 0.4 mg/ml final concentration and additional magnesium (up to 4-5 mM final concentration). It is likely that BSA (other proteins could probably serve the same function, which is likely the trapping of oxidation products/free radicals originating from the non-reacted dye, allylamine, or impurities in these reagents). Magnesium is probably partially chelated by the free amine groups of the dye, and needs to be increased to 4-5mM compared to the initial 1.5mM provided by the PCR buffer only. Various Tris, glycine and ethanolamine concentrations did not influence efficiency of DNA labeling.
  
• Nick translation worked as usual (the NT buffer already provides 10mM magnesium ions in the reaction), but seemed somewhat improved by an increased BSA concentration (up to 0.2 mg/ml).

PCR labeling protocols (with commercial and custom fluorescent nucleotides).
Please note: the 10x PCR buffer used includes 15mM MgCl (providing by itself 1.5 mM Mg ions in the final reaction)

Degenerate oligonucleotide priming-PCR (DOP-PCR) labeling is a useful technique, in which virtually any template DNA can be amplfified using degenerate primers. I used two types of primers with similar results: one primer has the degenerate nucleotides a few bases from the 3' end (6MW, Telenius), whereas the other primer has the degenerate bases at the 3' end (Primer A, S. Bohlander). Both primers worked equally well in my hands. For more info regarding the primers, please consult papers published by the two authors mentioned.
1. Primer 6MW: 5' ccg act cga gnn nnn nat gtg g
2. Primer A: 5' tgg tag ctc ttg atc ann nnn nn
[3. Primer B: 5' aga gtt ggt agc tct tga tc (this primer can be added to the PCR reaction after a few cycles with primer A, to re-amplify all products started by primer A. Primer B also has a rare restriction site, but is not essential for the reaction]

PCR program for DOP-PCR amplification and labeling (includes a few cycles with very low annealing temperatures, to allow the primer(s) to bind randomly to most available DNA sequences in the template DNA):

2.1.4. General observations regarding DNA labeling

Different dyes are used in different amounts in the labeling reactions, because their bulkiness and/or electrical charge probably allows the polymerase to incorporate them only occasionally. The same labeled nucleotide may not be incorporated when it appears two-three times in a row, and some nucleotides, as is the case for the Cy3.5- and Cy5.5-dUTP inhibit the reaction even when added in small amounts. It is possible that they "block' the activity of the polymerase, or that the polymerase has to stop after incorporating only one such modified nucleotide into the DNA. The DNA "labeled" with these two nucleotides yielded extremely weak or no FISH signals.

Other dyes, like DEAC and the rhodamine derivatives, inhibit PCR amplification when added at the same concentration as FITC, biotin, digoxigenin. However, by performing parallel labeling reactions which use increasing or decreasing amounts of labeled nucleotide, a point of compromise can be identified, where DNA synthesis is not fully inhibited, and the DNA is sufficiently labeled to yield results. On the other hand, biotin-, digoxigenin- and FITC-dUTP can be added to the labeling reactions in quite high amounts, replacing almost half the amount of dTTP, and do not inhibit DNA amplification too significantly. Of course, very high amounts of these dyes are not needed, as labeling takes place when roughly 1/5-1/3 dTTP is replaced.

Yet other dyes, like the newly tested CMF is bulky and titration of the CMF-dUTP amount does not influence the overall amount of DNA produced in the reaction, indicating that this modified nucleotide is NOT used at all by the polymerase.

Fig. 3. Same DNA template (paint probe cocktail for M-FISH analysis) was labeled with a degenerate primer (DOP-PCR) in identical conditions, with (1) FITC-dUTP (7ul/100ul reaction), (2) R6G-dUTP (3ul/100ul reaction), (3) TxR-dUTP (2ul/100ul reaction), (4) BIO-dUTP (7ul/100ul reaction) and (5) DEAC (2ul/100ul reaction). Various dyes influence the level of DNA amplification differently. DEAC is more of an "inhibitor" than FITC or biotin. 7ul/100ul reaction of DEAC-dUTP or TxR-dUTP would completely inhibit DNA synthesis. The bright "spots" originate from the free fluor in the reaction, excited by the UV light when visualizing the ethidium-bromide stained gel. "M" is the size marker (1kb ladder).
Fig. 4. Simple method to verify DNA labeling. A 200 BP PCR fragment (1) was labeled with (2) BIO-dUTP, (3) DIG-dUTP, (4) FITC-dUTP and (5) TRITC-dUTP. TRITC is tetramethylrhodamine, similar to TAMRA. Because of the labeled nucleotide incorporation, the PCR products runs at slightly different speed in the gel.

2.2. Post-labeling DNA processing and purification

After labeling the DNA needs to be processed and purified before it is used in FISH (or some other applications).

2.2.1. DNase treatment (for FISH and other hybridization protocols)

For FISH, the "golden rule" is that the labeled DNA fragments used in hybridization need to be between 200-500bp long, otherwise, a relatively high backgrounds starts to become visible on the slide, and the hybridization signal becomes more punctate. Whereas this is automatically taken care of in a nick translation, where the DNase amount is chosen so as to yield fragments shorter than 500 BP, in PCR, a partial DNase treatment is required after the reaction.

The 10x DNase digestion solution, is obtained by mixing 400 ul water, 4 ul 1M MgCl2 and 1-2 ul 3 mg/ml DNase stock solution (final DNase concentration in the reaction is about 1.5 ug/ml). The DNase solution can be added directly into the PCR vials. Reaction takes place at room temperature for 10-15 minutes, and is stopped by heating 2-3 minutes at 94 C. An example of an appropriate DNase treatment is shown below, where the DNA fragments (shown in Fig 5) are digested below the 500 BP mark (in Fig 7). The short fragments yield best FISH signals.

Strength of DNase solutions may vary, so it is important that any batch of DNase is tested, by digesting the same sample of a PCR product with the same amount of DNase for different period of times (2-20 minutes).

Fig. 5. Whole chromosome paint probes labeled by PCR. No DNase treatment was applied. Note the apparent length of the fragments, with the DNA smear higher than 12kb, the size of the longest fragment of the marker (= 1kb ladder, Gibco BRL). Fig. 6. Same PCR products, after 5 minutes DNase digestion. Fig. 7. Same PCR products after 15 minutes DNase digestion. FISH signals were optimal using the products from Fig. 7. The yellow arrow points to the 500 BP mark.

2.2.2. BSA removal

The main "problem" when using custom-made nucleotides, is that there is a high amount of protein (BSA) in the reaction. When large volumes of labeled DNA (50-500 ul PCR reactions) need to be ethanol precipitated and resuspended in 10-12 ul hybridization buffer (10% dextran sulfate, 50% formamide, 2x SSC), the relatively large amount of BSA in the pellet prevents the pellet (and the DNA) from resuspending in the buffer. Therefore, after the labeling reactions are completed, the BSA can be removed using one of the procedures mentioned below.
The high temperature of the denaturing step in the PCR cycle (usually at 92-95 C) denatures/precipitates most of the BSA, which can be seen as small white flakes floating in the PCR mix. The same phenomenon happens after heat-inactivating the nick translation reaction. Therefore, a first step in removing the BSA can be a simple 30-60 seconds centrifugation of the PCR or nick translation reactions at 14,000 rpm in a tabletop centrifuge. The protein will be pelleted at the bottom of the vial(s), and the reaction mix (PCR or nick translation) containing the labeled DNA will be transferred in a clean vial. This centrifugation is optional and DOES NOT remove all the BSA.

To remove the BSA, three different protocols were tested:

• (1). The simplest procedure is a 5 minutes phenol extraction: 1/2 volume phenol and 1/2 volume chloroform:isoamylalcohol (24:1) are added to the labeled DNA, followed by vortexing, 3-4 times at 20-30 seconds interval, and 1-2 minutes centrifugation at full speed in a microfuge. The upper phase is collected, ethanol-precipitated and resuspended in 12 ul hybridization buffer. Phenol extractions lead to more than 50% losses (Fig 8) for DNA labeled with digoxigenin and rhodamine derivatives (especially Texas Red, but also R6G, RGr and TAMRA). Such DNA could still be used for FISH, but signals were less intense. To avoid such DNA losses, alternative protein removal procedures were used.
• (2). Spin columns (10 ug DNA capacity) (Quiagen, Valencia, CA) could be used to remove BSA from 100-200 ul PCR reactions, with only slight DNA losses, which are normal part of the column purification process.
• (3). Alternatively, 1-2 ul proteinase K (2 mg/ml) is added to every 100 ul PCR labeling reaction and the vials incubated 30 minutes at 37-45 C. Reaction is stopped by adding the irreversible protease inhibitor Pefabloc SC (Boehringer Mannheim, Indianapolis, IN), at a final concentration of 2-4 MM. The DNA can be simply precipitated and used. No DNA losses occur, and the DNA can be easily resuspended in hybridization buffer after precipitation.

Currently in this laboratory, when DNA labeled with different fluors is used in the same hybridization experiment (such as M-FISH analysis), the DNA labeled with DIG and rhodamine derivatives is pipetted into one vial and subjected to proteinase K purification method. DNA labeled with all other fluors/haptenes is added into a second vial and subjected to phenol extraction. Then, the (BSA-free) DNA is all mixed together, ethanol precipitated and used.
The DNA could all be subjected to proteinase K digestion, which would make the protocol simpler. However, phenol extraction is a very short and efficient procedure which removes all protein and leaves the DNA clean, and it is worth using (by comparison, after proteinase K digestion, although the BSA is largely removed, the DNA solution still contains the proteinase K itself and oligopeptides originating from the BSA breakdown, so it is not completely "clean").

Fig. 8. and Fig. 9. In both images, the first lane corresponding to any fluor shows the original PCR product, whereas the second lane shows the same volume of PCR product after phenol extraction. Most DNA losses during phenol extraction occur when using rhodamine derivatives to label DNA (R6G, TxR in Fig 8). Some DIG labeled DNA is also lost (Fig 9), whereas BIO labeled DNA is less affected. FISH tests showed that BIO labeled DNA can be subjected to phenol extraction whereas DIG labeled DNA should not. The control DNA (Ctrl) is Cot-1 DNA. Length marker is 1 kb ladder (GIBCO).