This technique is modified from the batch method for purification of double-stranded (ds) DNA developed by H. Gamper, N. Lehman, J. Piette, and J. Hearst (1985) Purification of Circular DNA Using Benzoylated Naphthoylated DEAE-Cellulose, DNA 4: 157-164. I suggest that relevant sections of this paper be read before using the technique for the first time.
1. Calculate how much BND-cellulose you will need to bind the "single-stranded" component of the DNA you are working with. To make this calculation, assume that 5% of your DNA is single-stranded. The actual amount is less, but it's better to err on the safe side. My data show that 0.2 ml of packed BND-cellulose (Serva or Sigma; this corresponds to about 100 mg of the Sigma product) will bind at least 10 µg of ssDNA. This observation is consistent with the binding data of Gamper et al.
Note that RNA competes with ssDNA for binding to BND-cellulose. If your DNA prep is not free of RNA, you must calculate how much extra BND-cellulose will be required to bind the total of DNA + RNA. You may find that additional RNase treatment (note: prolonged incubations at 37°C are not recommended for preservation of replicating intermediates) or CsCl centrifugation (which is compatible with preservation of replicating intermediates) would be worth while. The simplest method for removing RNA is to include some RNase in addition to restriction endonuclease during the restriction digestion which usually precedes BND-cellulose fractionation.
2. If prepared BND-cellulose is not available, then use the following technique to prepare some. The amounts used in the following protocol should provide you with far more prepared BND-cellulose than required for a single fractionation. The excess should be stored at 4°C.
a) Weigh out 2 gm of BND-cellulose. Place in 15 ml conical centrifuge tube.
b) Add 10 ml of 5 M NaCl and suspend the BND-cellulose, making sure that all particles are wet.
c) Centrifuge by accelerating to 2000 rpm, then slowing with brake.
d) Remove supernatant, resuspend pellet in 10 ml of 5 M NaCl, and repeat centrifugation.
e) Repeat 3 more times.
f) Wash pellet 1 time in 3 pellet volumes of water to reduce the salt concentration.
g) Resuspend the pellet in 10 ml of 1.0 M NET (1.0 M NaCl, 10 mM Tris, pH 8.0, 1 mM EDTA) and centrifuge (2 times).
h) Resuspend in a final volume of 10 ml of 1.0 M NET. Store in refrigerator until used.
3. My early observations suggested that Serva BND-cellulose gave slightly better yields, but that Sigma BND-cellulose had greater affinity for ssDNA. In addition, the Sigma product was more granular (less fibrous) and had a slightly lighter, more yellowish color than the Serva material. More recent observations suggest that these properties may change, depending on the lot number of the BND-cellulose. We have found that one particular recent lot of Serva BND-cellulose (lot #22088) did not provide satisfactory enrichment of replicating DNA. In contrast, an earlier lot (control #21025) provided excellent enrichment. Sigma BND-cellulose lot #22H7040 also provides excellent enrichment. Other lots of BND-cellulose from both manufacturers should be purchased in small amounts and tested before large amounts are ordered.
4. If the required amount of BND-cellulose is 500 µl packed volume or less, use the microcentrifuge method described below. For larger volumes, use the syringe column method described by Gamper et al. or use larger centrifuge tubes (centrifugation at 2000 rpm for 60 sec is usually sufficient to pellet the BND-cellulose in larger tubes). Microfuge method: for each DNA sample to be fractionated, prepare a 1500 µl Eppendorf tube with the correct amount of packed BND-cellulose in 1.0 M NET. Do this by centrifuging each slurry of BND-cellulose in a microfuge for 20 sec. Remove the supernatant (this may be difficult to do completely, in this and subsequent steps, because the upper portion of the slanted pellet tends to collapse as supernatant is removed around it, but do the best you can) and add more slurry if necessary to get proper final volume of packed material, or resuspend in fresh 1.0 M NET (use a small spatula, with rapid twisting motion, to aid in resuspension) and remove slurry if necessary to get proper final volume of packed material. Wash the BND-cellulose in the tubes by resuspending in 1 ml of 1.0 M NET and centrifuging, 3 times. Remove final supernatant.
5. The DNA should be in 1.0 M NET, in a volume about twice the packed BND-cellulose volume. The DNA can be transferred into the correct volume of 1.0 M NET by EtOH precipitation followed by dissoving the pellet in 0.8 times the correct volume of TE buffer. Once the pellet is completely dissolved, add 0.2 times the correct volume of 5 M NaCl and mix. Apply the sample to the packed BND-cellulose, resuspend the BND-cellulose in the sample (with the aid of a spatula as above), allow about 30 sec for completion of adsorption, then centrifuge 20 sec (as above). Save supernatant as separate fraction.
6. Wash the BND-cellulose 5 times with volumes of 1.0 M NET that are equal to or twice the volume of packed BND-cellulose (depending on whether your primary goal is to maximize the yield of replicating DNA or to reduce contamination of the caffeine wash fraction with nonreplicating DNA). Save the supernatants as separate fractions.
7. Wash the BND-cellulose 6 times with volumes of 1.0 M NET, 1.8% caffeine, that are equal to or twice the packed BND-cellulose volume (depending on whether your primary goal is to maximize the purity of replicating DNA or to maximize the yield of replicating DNA). Save the supernatants as separate fractions.
8. At this point, the fractionation results may be assayed by agarose gel electrophoretic analysis of samples of appropriate volume from each fraction. Use sufficient volume to be able to detect a signal, but remember that the high salt in each sample may cause electrophoretic artifacts if too large a sample volume is used.
9. Pool fractions according to the results of the gel assay in step 8, or, if you don't wish to run a gel assay at this point, simply pool the first 5 salt wash fractions (discard the 6th) and pool all 6 caffeine wash fractions. Centrifuge at 10,000 rpm in nonsiliconized glass tubes or Eppendorf tubes for 10 min to pellet particles of BND-cellulose (important). Transfer supernatants to fresh tubes. Any glass tubes used at this point should be siliconized. Siliconized Corex 15 ml tubes are good if the pooled volumes are in the 1-5 ml range.
10. Add an equal volume of isopropanol to each tube. Mix gently but thoroughly. Leave in -20°C freezer for 30 min or longer.
11. Centrifuge 10,000 rpm for 90 min.
12. Remove supernatant. Air dry. Dissolve "pellets", which may be invisible or may be a thin turbid film spread over a wide area, in 150 µl of TE. Transfer to Eppendorf tube, then rinse the Corex tube with another 100 µl of TE, which should be pooled with the original solution in the Eppendorf tube. Larger volumes of TE may be used at this point, if necessary to get good solution.
13. Add 1/9 volume of 3 M potassium acetate, mix, and then add 2.2 volumes of ethanol. Place in -20°C freezer for 30 min to overnight.
14. Centrifuge in microfuge for 30 min. Remove supernatants. Rinse pellets in 70% ethanol. Air dry. Dissolve pellets in about 300 µl TE. Repeat steps 13 and 14, but dissolve final pellets in volumes of TE buffer appropriate for subsequent manipulations. Typically, the caffeine wash DNA will be dissolved in about 20 µl TE and the salt wash DNA will be dissolved in about 400 µl TE. Freeze samples after they are dissolved.
15. To know what volume of the DNA in your caffeine and salt wash samples to load onto your first dimension gels for 2D gel analysis, it is necessary to measure the concentrations of your salt and caffeine wash DNAs. The salt wash DNAs should be concentrated and pure; therefore, it should be easy to determine their concentration by measuring their absorbance at 240, 260, and 280 nm. The value of OD260 should exceed that of OD240 (and OD260, though this is never a problem) by at least 30%. If it does not, then your salt wash DNA is not sufficiently pure for concentration determination by OD measurement. You will need to measure the concentration by comparison with a standard in gel electrophoresis, as described next for caffeine wash DNA. To measure the concentration of your caffeine wash DNA, take a small portion (about 1/10) of the solution of caffeine wash DNA in TE buffer (from step 14 above) and run it on a gel along with a range (broad enough to cover the range of expected amounts of caffeine wash DNA) of known amounts (from OD measurements) of salt wash DNA in parallel lanes. By comparison of the caffeine wash lane with the known salt wash lanes, it should be possible to determine the concentration and total amount of caffeine wash DNA in your prep.
15. Note: our experiments suggest that optimal results are obtained with a single BND-cellulose fractionation. Although the caffeine wash fraction obtained from a single fractionation is contaminated with a small amount of double-stranded DNA, and although all detectable traces of this double-stranded DNA can be removed by a second fractionation, there is also considerable loss of replicating DNA during the second fractionation so the signal to noise ratio is not improved. Most of the nonreplicating DNA which contaminates the replicating DNA in the caffeine wash is not purely double-stranded DNA but rather double-stranded DNA with small gaps or with single-stranded tails (produced by shearing). There is also some purely single-stranded DNA. These purely or partially single-stranded nonreplicating molecules cannot be removed from the replicating molecules by additional BND-cellulose fractionations.
Return to Huberman lab home page
Go to directory of HTML versions of these protocols