Cell. Vol. 5, 301 -310, July 1975, Copyright®1975 by MIT
`
`Nucleic Acid Hybridization Using DNA
`Covalently Coupled to Cellulose
`
`Barbara E. Noyes and
`George R. Stark
`Department of Biochemistry
`Stanford University School of Medicine
`Stanford, California 94305
`
`Summary
`
`We describe a method for linking RNA and DNA
`covalently to finely divided cellulose through a di-
`azotized aryl amine, which reacts primarily with
`guanine and uracil (thymine) residues of single
`strands. The high efficiency of coupling and high
`capacity of the cellulose for nucleic acid make pos-
`sible a product with as much as 67 pg of nucleic
`acid per mg of cellulose. The product is especially
`suitable for hybridization experiments where very
`low backgrounds are important, and it is stable in
`99% lormamide at 80°C so that hybridized nucleic
`acid can be recovered easily. Full
`length linear
`Simian Virus 40 (SV40) DNA, produced by cleavage
`of SV40(l) DNA with S1 nuclease, can be coupled
`to diazo cellulose with an efficiency of 80-90%, and
`is effective in hybridization experiments with SV40
`DNA, complementary RNA synthesized in vitro
`from SV40(l) DNA with E. coli RNA polymerase,
`and the SV40-specific fraction of total RNA from
`SV40-infected and transformed cells. In these ex-
`periments an excess of cellulose-bound DNA was
`used, and the efficiency of hybridization was about
`90% when ribonuclease treatment of the hybrids
`was omitted.
`
`Introduction
`
`DNA-DNA or DNA-RNA hybrids formed in solution
`can be detected with a nuclease specific for single
`strands or isolated by chromatography on hydroxy-
`apatite. (For example, with SV40 DNA, see Sam-
`brook, Sharp, and Keller, 1972; Hansen, Pheiffer,
`and Hough,
`1974.)
`DNA-DNA reannealing
`competes with DNA-RNA hybridization in solution,
`especially if the DNA is in excess. Although this dif-
`ficulty can be eliminated by immobilizing the DNA
`on nitrocellulose filters (Gillespie and Spiegelman,
`1965) or in agar (Bolton and McCarthy, 1962; Han-
`sen et al., 1974), variable loss of DNA from the filter
`and high background levels in agar complicate the
`results. (For examples of the use of filters with SV40
`DNA, see Westphal and Dulbecco, 1968; Haas,
`Vogt, and Dulbecco, 1972; Holzel and Sokol, 1974).
`Loss from the filter can be particularly significant
`in experiments designed to quantitate low levels of
`specific RNA within a larger heterogeneous pool,
`since any RNA which hybridizes to DNA in solution
`is not detected.
`
`To circumvent these problems, we have devel-
`oped a new method for hybridization using DNA co-
`valently linked to cellulose. Shih and Martin (1974)
`coupled SV40 DNA to cellulose powder through the
`terminal phosphate groups using a water—soluble
`carbodiimide according to Gilham (1971). Poonian,
`Schlabach, and Weissbach (1971) and Arndt-Jovin
`et al. (1975) have coupled DNA to agarose activated
`with CNBr. These preparations have been used suc-
`cessfully in affinity chromatography, but their appli-
`cation to sensitive analytical hybridization may be
`limited because of the large amount of support ma-
`terlal used. Residual positive charges on the
`agarose, generated as a consequence of CNBr acti-
`vation, might also contribute to high background.
`The coupling procedure we describe is a modifica-
`tion of the one developed by Gurvich, Kuzovleva,
`and Tumanova (1961) for linking proteins covalently
`to finely divided aminobenzyloxymethyl cellulose.
`The high capacity of the cellulose and the facility
`with which the reaction can be performed are major
`advantages of the method. The DNA-cellulose can
`be used analytically to detect low levels of a specific
`nucleotide sequence within a larger heterogeneous
`pool as in filter hybridization techniques, or it can
`be used preparatively.
`
`Results
`
`Covalent Linkage of DNA to Cellulose
`Single stranded DNA can be linked covalently to
`reprecipitated,
`finely divided m—aminobenzyloxy-
`methyl cellulose after the primary aryl amino groups
`have been diazotized as described by Miles and
`Hales (1968). In the first experiment of Table 1, only
`8% of the input SV40 DNA was coupled to the diazo
`cellulose in borate buffer under conditions used
`successfully by Miles and Hales (1968) for the cou-
`pling of proteins. The DNA in this experiment had
`been denatured at pH 12 immediately before the
`diazo cellulose was added, but was extensively re-
`natured during coupling at high DNA concentration
`at 4°C (pH 8). However, when the DNA is kept dena-
`tured by doing the coupling in 70% or 80% dimethyl-
`sulfoxide (DMSO), more than 80% of the input can
`be added to the cellulose routinely. No DNA re-
`mains stably bound to amino cellulose which has
`not been diazotized. Experiments 2-4 of Table 1
`show that the efficiency of the reaction depends on
`the concentration of diazo cellulose, but is relatively
`independent of the concentration of DNA. in experi-
`ments using from 20-800 pg of SV40 DNA at con-
`centrations from 100-300 pg/ml, about 90% of the
`DNA can be linked covalently, provided that the cel-
`lulose concentration in the reaction mixture is at
`least 8-10 mg (dry weight)/ml. The percentage of
`DNA coupled decreases to 40-50% when the cellu-
`
`Page 1 of 10
`Page 1 of 10
`
`BD EXHIBIT 1007
`BD EXHIBIT 1007
`
`

`
`Cell
`302
`
`Table 1. Coupling of DNA to Diazo Cellulose
`
`DNA
`(pg/ml)
`
`245
`
`Buffer
`
`Borate
`70% DMSO
`70% DMSO
`
`Experiment
`1
`
`2
`
`3
`
`SV40 DNA
`Sonicated
`(6-73)
`
`S1 linears
`(165)
`
`S1 linears
`(165)
`
`99
`22
`43
`2.2
`0.5
`4
`
`4
`S1 linears
`(168)
`
`70% DMSO
`80% DMSO
`
`90
`280
`
`8
`11
`
`80% DMSO
`
`100
`
`80% DMSO
`
`200
`
`2.2
`2.2
`
`45
`53
`
`Dlazo Cellulose
`(mg/ml)
`
`% DNA
`Coupled
`
`Temperature
`(°C)
`
`12
`9.5
`10
`
`(NH2—cellu|cse)
`
`8
`87
`0
`
`80
`88
`
`4
`
`4
`
`25
`
`25
`4
`
`32P-labeled SV40(l) DNA, sonicated to an average size of 6-78 or digested with S1 nuclease to full length linear size (163), was suspended
`in 0.2 M borate buffer (pH 8), and diluted with additional buffer or DMSO. DNA in borate alone was denatured at pH 12, placed on ice,
`and readjusted to pH 8 immediately before addition to cellulose. DNA solutions were added to diazotized cellulose in small tubes, and
`the suspensions were stirred continually for 48 hr. The percentage of DNA coupled was determined from HP stably bound to the cellulose
`following extensive washing as described in the text.
`
`lose concentration is reduced to 2 mg/ml, and the
`coupling etiiciency is very poor if the cellulose con-
`centration is reduced further. When E. coli DNA or
`
`salmon sperm DNA (200 pg/ml) are coupled to cel-
`lulose (10 mg/ml), about 60% of the DNA remains
`stably bound. This decrease in the percent of DNA
`which couples compared to SV40 DNA is not under-
`stood. E. coli tFlNA (2 mg/ml) coupled to diazo ce|-
`lulose (2 mg/ml) in 70% DMSO with 47% efficiency.
`At cellulose concentrations above 8-10 mg/ml,
`the reaction is complete within 24 hr. At lower con-
`centrations, 48 hr are required to achieve maximum
`coupling. The time course and extent of the cou-
`pling reaction are about the same at 4°C or 25°C.
`The amount of DNA stably bound to the cellulose
`is determined after thorough washing with 80%
`DMSO and 0.1 X standard saline citrate (SSC, 0.15
`M NaCl—0.015 M sodium citrate). When DNA dena-
`tured before addition of cellulose is coupled without
`DMSO, as much as 50% of the DNA initially bound
`can be removed by washing with 80% DMSO or 99%
`formamide. However, less than 10% of bound DNA
`is lost during such washing after reaction in DMSO,
`and most of this material is released in the low salt
`washes. The percentage of polynucleotide found
`linked covalently to the cellulose was the same
`whether determined using radioactively labeled
`DNA, by spectral analysis of acid hydrolysates of
`the DNA—ce|lulose, or by spectral analysis of the
`DNA which did not couple.
`
`SV40 DNA
`
`% Coupled
`87
`
`Table 2. Coupling of Nucleotide Homopolymers to Diazo Cellulose
`Polymer Concentration
`(Am/ml)
`5 65
`'
`55
`5-15
`P°'V(U)
`20
`6.00
`PoIy(d'D
`<5
`e_25
`Pu|y(C)
`4,
`535
`POMG)
`
`poly(A)
`M5
`<5
`Polymers were precipitated with ethanol and resuspended in 80%
`DMSO in 0.2 M borate buffer (pH 8). The molecular weights of
`all polymers were greater than 1 X 105, except poly(G) which was
`3 X 104, Concentrations of the polymers were determined before
`the DMSO was added. Coupling was carried out at 4°C for 46
`hr in a total volume of 0.2 ml at a cellulose concentration of 10
`mg/ml.
`
`The linkage between nucleic acids and the diazo
`cellulose was
`investigated
`using
`nucleotide
`homopolymers. The data of Table 2 suggest that
`coupling occurs best through guanine and uracil
`residues. Poly(dT) couples less well than poly(G)
`and po|y(U), whereas poly(A) and poly(G) do not
`react with the diazo cellulose appreciably under the
`conditions tested. Although a low level of reaction
`with poly(A) and poly(G) cannot be excluded, the
`results do indicate that guanine and uracil (thymine)
`residues are probably the major sites of reaction.
`
`Page 2 of 10
`Page 2 of 10
`
`
`
`

`
`Page 3 of 10
`
`

`
`it was necessary to characterize the
`further,
`preparation used. Of 100 arbitrary units of cRNA,
`17 were RNAase resistant. When incubated with
`SV40 DNA—ceIIulose under standard conditions for
`
`24 hr, 75 units hybridized and 25 did not. Only 4
`T‘?
`I
`I
`I
`i
`
`o-
`
`..
`
`—I.o-
`
`E’
`‘C0:
`.5‘D
`‘I
`.0>
`
`I 4 U
`
`ED
`
`‘O
`_l
`
`J1
`0
`
`I
`
`I
`L0
`
`I
`
`I
`2.0
`
`lngl
`Log cRNA input
`Figure 2. Hybridization of SV40 DNA—Cellulose with Increasing
`Amounts of cFtNA
`SV40 DNA coupled to cellulose (50 ng DNA, 22 pg DNA/mg) was
`hybridized with increasing amounts of cRNA in 50% formamide
`buffer for 48 hr at 37°C. The samples were washed. treated with
`RNAase (20 pg/ml,
`1 hr,
`room temperature) in 2 X SSC and
`washed again. The cFtNA hybridized was determined after elution
`with 99% formamide and 0.1% SDS as described in the text. Each
`point represents the average 01 duplicate determinations. Similar
`results were obtained when the RNAase treatment was omitted.
`
`80 Tjr 1%‘ Wjr
`
`I
`
`§
`
`60
`
`40
`
`20
`
`83
`
`;:
`
`E5>I <Z
`
`n:U
`
`‘E0)
`‘.2
`(D
`o.
`
`0
`
`20
`
`40
`
`Hours
`
`60
`
`so
`
`Figures. Kinetics of Hybridization of SV40 DNA—Cellulose with cRNA
`SV40 DNA coupled to cellulose (500 ng DNA, 22 pig DNA/mg) was
`incubated with 2.5 ng cRNA in 50% formamide buffer at 37°C.
`The samples were washed, and the percentage of cRNA hybridized
`was determined after elution of the cRNA with 99% formamide
`and 0.1% SDS as described in the text. Each point represents the
`average of duplicate determinations. Similar results were obtained
`with samples treated with RNAase.
`
`Cell
`304
`
`in-
`ized selectively with SV40 DNA-cellulose,
`dependent of the amount of DNA coupled per mg
`cellulose over an 11 fold range. Treatment with
`RNAase A caused a 15% decrease in the percent-
`age of cRNA hybridized. This result
`is expected
`since sedimentation of the cRNA in 99% formamide
`
`and 0.1% sodium dodecyl sulfate (SDS) at 35°C in-
`dicated that about 12% of the preparation is larger
`than full
`length linear SV40 DNA. The relatively
`small effect of RNAase on the percentage of cRNA
`hybridized indicates that
`long regions of single
`stranded DNA must be available for hybridization.
`An alternative method of examining the accessi-
`bility of the DNA is to hybridize increasing amounts
`of cFlNA to a fixed amount of DNA—cellulose, as
`shown in Figure 2. At least2.5 ng of cFiNA (probably
`more at saturation) can be bound to 50 ng of DNA.
`Since most of the cRNA is complementary to only
`one of the two strands of SV40 DNA, at least one
`DNA strand in ten is accessible for hybridization
`with a long sequence of cRNA.
`
`15
`
`Efficiency of Hybridization of SV40
`DNA—Cellulose with cRNA
`The reaction between SV40 DNA—Cellulose and
`cRNA was investigated more thoroughly to deter-
`mine optimal standard conditions for assays of cell
`extracts containing SV40-specific FlNA. At DNA
`concentrations between 1 and 120 pig/ml, 70-80%
`of the input cRNA hybridized reproducibly with the
`SV40 DNA—cel|u|ose within 24 hr. Figure 3 shows
`kinetic data for a DNA concentration of 2.5 pg/ml
`and a cRNA concentration of 12 ng/ml. To evaluate
`the efficiency of hybridization reactions with cRNA
`
`Table 3. Hybridization of SV40 cRNA with DNA Celluloses
`% of Input cRNA
`Hybridized
`
`DNA—Cellulose
`pg DNA/mg Cellulose —FtNAase
`+FtNAase
`SV40
`6
`74.0
`60.0
`76.7
`60.3
`59.9
`49.3
`70.4
`49.9
`75.7
`63.9
`74.7
`66.9
`71.5
`55.9
`70.8
`55,7
`
`48
`
`67
`
`
`
`Salmon Sperm 10
`
`0.3
`0.5
`
`0.2
`0.3
`
`DNA—Cellulose containing 0.6 pg of DNA was incubated with SV40
`cRNA (3.6 ng) in 50% formamide for 18 hr under standard condi-
`tions. The total amount of cellulose in each reaction mixture was
`normalized to 0.1 mg by adding carrier cellulose. Treatment with
`RNAase A (20 ,ug/ml) was for
`1 hr at
`room temperature in
`2 x SSC.
`
`Page 4 of 10
`Page 4 of 10
`
`

`
`Hybridization with Immobilized Nucleic Acids
`305
`
`units of the cRNA which failed to hybridize in the
`first experiment hybridized to fresh SV40 DNA-
`cellulose in a second attempt. Therefore, 21 units
`of the original preparation failed to hybridize in two
`attempts. This material was completely RNAase
`sensitive, indicating that it was not double stranded;
`it probably represents transcripts of small amounts
`of cellular DNA contaminating the SV40 DNA
`preparation used as template. In addition to the 21
`units of material not complementary to SV40 DNA,
`the cRNA contains 62 units of single stranded and
`17 units of double stranded SV40 RNA. Since 75
`units did hybridize in the first experiment, we con-
`clude that
`the efficiency of hybridization was
`greater than 90%, and that in this case most of the
`double-stranded RNA did hybridize.
`
`SV40-Specific RNA in Infected and Transformed
`Cells
`To explore the utility of the technique for detecting
`low levels of SV40-specific RNA within the total RNA
`pool of eucaryotic cells, RNA was prepared from
`monkey cells (MA-134) productively infected by
`SV40 and from a hamster line transformed by SV40
`(C13/SV28). MA-134 cells were harvested 72 hr
`after high multiplicity infection with SV40 and after
`labeling with 3H—uridine for 20 min (71%—72 hr) or
`24 hr (48-72 hr). RNA was extracted from the super-
`natant solution following lysis of the cells with SDS
`
`in the presence of NaCl and removal of precipitated
`material by centrifugation (Hirt, 1967). Although
`RNA prepared in this way probably represents only
`about 50% of the total cellular RNA (Aloni, 1972),
`this procedure was chosen in order to compare our
`results with those of others. As indicated in Table
`4, 10% of the RNA from cells pulse-labeled for 20
`min with 3H-uridine hybridizes specifically to SV40
`DNA-cellulose. This agrees well with values of 10-
`20% and 10% reported by Aloni (1972) and Acheson
`et al. (1971) for viraI—specific RNA prepared in the
`same way from SV40 and polyoma-infected cells
`and assayed using filter hybridization techniques.
`Table 4 also shows that about 0.7% of the RNA
`prepared from lytically infected cells labeled for 24
`hr hybridizes with SV40 DNA—ce|lulose. This is with-
`in the range 0.1-1 % reported by Khoury and Martin
`(1972) for the percentage of SV40—specific RNA in
`infected AGMK cells as measured by reassociation
`kinetics.
`RNA from C13/SV28 cells was extracted after la-
`
`beling with 3H-uridine for 24 hr. As shown in Table
`5, about 0.02% of the total labeled RNA hybridizes
`specifically to the SV40 DNA-cellulose. This value
`is 10 fold larger than the one reported by Sambrook
`et al. (1972) for SV40—specific RNA in SV3T3 cells
`(assayed by hydroxyapatite chromatography after
`hybridization in solution), but our result is within
`the range of the values 0.01—0.025% reported by
`
`Table 4. SV40—Speciiic RNA in Infected Monkey Cells
`Labefing
`Period
`20 min
`(71 2/3-72 hr)
`
`Experiment
`‘l
`2
`3
`4
`
`24 hr
`(48-72 hr)
`
`5
`
`6
`
`7
`
`8
`
`pg DNA
`10
`10
`10
`24
`
`24
`
`10
`10
`15
`15
`20
`
`15
`
`,ug RNA
`2.3
`1.9
`1.7
`1.9
`
`0.95
`
`68
`143
`88
`177
`34
`
`88
`
`% RNA Hybridized
`8.0, 11.1
`9.8,
`9.3
`12.5
`8.9
`
`10.4 :: 1.8
`
`7 7
`
`0.58,
`0.74,
`0.76,
`0.77,
`0.91,
`
`0.67
`0.65
`0.83
`0.73
`0.85
`
`0.73
`0.70,
`0.74 i 0.09
`
`RNA was extracted from MA-1 34 cells 72 hr after infection with SV40 and either 20 hr (71 2/3-72 hr) or 24 min (48-72 hr) after labeling
`with 3H—uridine. The specific activity of RNA from cells labeled for 20 min was 6.4 X 105 cpm/pg, and from cells labeled for 24 hr,
`2.6 X 104 cpm/lug. Hybridization with SV40 DNA—cellulose and E. coli DNA-cellulose was carried out in 50% fcrmamide at 37°C for 40
`hr as described in the text. RNAase treatment (20 ,ug/ml RNAase A and 1 pg/ml T1 RNAase) was for 1 hr at room temperature in 0.2
`ml of 2 x SSC. Except for different RNA concentrations, which reflect differences in specific activity, the conditions in Experiments 1-4
`were the same as those in Experiments 5-8. The percentage of input RNA hybridized was corrected for a background determined from
`hybridization with E. coli DNA-cellulose. For RNA labeled for 20 min, this background was 0.4% of the input, and for RNA labeled for
`24 hr it was 0.009%.
`in Experiment 8, unlabeled total cellular RNA from uninfected MA-134 cells was used in place of yeast RNA as
`cold carrier in the hybridization mixture.
`
`Page 5 of 10
`Page 5 of 10
`
`

`
`Cell
`306
`
`
`Table 5. SV40—Specific RNA in Transformed Cells
`
`Average
`RNA
`pg
`pg
`cpm
`% of Total Labeled
`
`Experiment
`Fraction
`DNA
`DNA
`RNA
`Hybridized
`Material Hybridized
`1
`I
`SV40
`8
`206
`3454
`4254
`E34
`730
`1878
`2069
`239
`297
`2316
`2648
`427
`628
`1463
`3592
`322
`276
`
`2
`
`3
`
`4
`
`I
`
`II
`
`ll
`
`E. coli
`
`SV40
`
`E. coli
`
`SV40
`
`E. coli
`
`SV40
`
`E. coli
`
`8
`
`10
`
`10
`
`8
`
`8
`
`10
`
`10
`
`206
`
`105
`
`105
`
`346
`
`346
`
`263
`
`263
`
`0.018
`
`0.019
`
`0.0021
`
`0.0027
`
`C13/SV28 cells labeled for 24 hr with 3H—urldine were disrupted with a Dounce homogenizer and divided into 3 fractions for extraction
`of RNA as described in the text. Fractions I (nuclear pellet) and II (postmitochondrial supernatant) contained 72% and 23% of the total
`labeled RNA, respectively. Fraction lll (pellet from centrifugation at 17,000 x g for 15 min) contained 5% of the total labeled RNA and
`was not assayed. Hybridizations were performed in 50% tormamide at 37°C for 40 hr, and samples were processed as described in
`the text. RNAase treatment (20 pg/ml RNAase A and 1
`,ug/ml T1 RNAase) was for 1 hr at room temperature in 0.2 ml of 2 x SSC.
`The specific activity of the RNA was 6.2 x 104 cpm/,ug. The percent of the total
`labeled RNA hybridized is the average of duplicate
`determinations, corrected for the average background values obtained with E. coli DNA—cellulose.
`
`Aloni, Winocour, and Sachs (1968) for SV3T3 cells
`labeled with 3H—uridine for 22 hr (assayed by filter
`hybridization).
`
`Discussion
`
`Covalent Attachment of Nucleic Acids lo
`Cellulose
`DNA and RNA can be immobilized by trapping them
`in agar or cellulose or on nitrocellulose filters, or
`by linking them covalently to a solid support
`through the single terminal phosphate groups or
`through multiple points of internal attachment as
`to agarose after activation with CNBr (see review
`by Gilham, 1974). The method we describe results
`in covalent attachment of single stranded RNA or
`DNA at multiple points to very finely divided cellu-
`lose. Based on the data of Table 2, the linkage is
`primarily through guanine and uracil (thymine) resi-
`dues, as expected from the work of Cavalieri and
`Bendich (1950) and Robins (1958), who found that
`the disubstituted purine bases guanine and Xan-
`thine did couple with diazotized aromatic amines in
`dilute alkali at carbon 8, whereas the monosubsti-
`tuted bases adenine and hypoxanthine did not
`react. Substitution of guanine at position 9 (as in
`
`nucleotides and nucleosides) reduces the nucleo-
`philicity of carbon 8 so that reaction occurs instead
`with the primary amino substituent, or a ring nitro-
`gen, or both (Robins, 1967; Shapiro, 1968). Pyrimi-
`dine bases probably react with diazotized aryl
`amines through electrophilic substitution at carbon
`5 (Acheson, 1967; Robins, 1967). The methyl group
`at this position in thymine may account for the re-
`duced reactivity of this base. Barry and O'Carra
`(1973) reported the coupling of NAD+ to sepharose
`through a diazotized aryl amine. Although the site
`of attachment to the NAD+ is not known, the au-
`thors postulate linkage through the adenine moiety.
`in these experiments, 1% or less of
`the NAD+
`present in the reaction mixture actually coupled to
`the support. Such low levels of reaction with poly(A)
`or poly(C) would not have been detected in the ex-
`periment presented in Table 2.
`react much
`Single-stranded
`polynucleotides
`more readily than the double stranded species.
`With denatured SV40 DNA,
`it is difficult to prevent
`renaturation in buffer at low temperature because
`the concentration of DNA is high. For more complex
`DNAs, renaturation at similar high concentration is
`not as rapid. For example, denatured salmon sperm
`DNA can be coupled in borate buffer at pH 8, al-
`
`Page 6 of 10
`Page 6 of 10
`
`

`
`Hybridization with immobilized Nucleic Acids
`307
`
`though the efficiency is only 15-20%. This result
`suggests that DMSO may enhance the efficiency of
`coupling not only by maintaining the DNA in dena-
`tured form, but also by increasing the nucleophil-
`icity of the purine and pyrimidine rings. E. coli tRNA
`also coupled to diazo cellulose in borate buffer at
`pH 8, but again the efficiency of coupling (5—7%)
`is less than for polymers in DMSO. Reaction in this
`case is probably through the guanine and uracil
`residues of single stranded regions. Duplex mole-
`cules with single stranded tails could probably be
`coupled selectively through these tails, and it may
`be possible to develop conditions for selectively
`coupling partially denatured (AT- or AU-rich) re-
`gions of a fully duplex structure.
`in designing a particular procedure for coupling,
`it is important to consider the intended use of the
`immobilized nucleic acid.
`it
`little nucleic acid is
`available,
`it may be preferable to couple at
`low
`DNA:cellulose ratios to increase the amount of DNA
`
`bound. Large amounts of support, however, may
`give somewhat higher backgrounds in hybridization
`experiments. Alternatively, one can couple large
`amounts of DNA to very small amounts of cellulose
`in order to reduce nonspecific binding.
`In the ex-
`periments presented here, from 6-67 pg of DNA
`were coupled per mg cellulose without any notice-
`able effect on the efficiency of hybridization with
`cRNA.
`
`Hybridization with Immobilized Nucleic Acids
`Hybridization assays can be performed in 0.6 M
`NaCI at 65°C or in 50% formamide at 37°C with
`equivalent results. Because backgrounds appeared
`to be slightly lower in formamide and because it
`is more convenient to work at 37°C, the lower tem-
`perature was chosen for routine assays. The sensi-
`tivity of the procedure depends upon reducing the
`background to a low level. In addition to including
`cold, carrier RNA in the hybridization mixture, it is
`important to wash the hybrid cellulose extensively.
`No differences in background levels were observed
`when the cellulose was washed with 0.1, 2, or
`4 X SSC, and 6 washes with ice cold 2 X SSC
`were sufficient to reduce backgrounds to accept-
`able levels when the input was less than 105 cpm
`of CRNA.
`In experiments with high levels of RNA
`extracted from cells (3 X 107 cpm), the cellulose
`was washed with hybridization buffer containing
`25% formamide following RNAase treatment, which
`reduced the background without disrupting the
`RNA—DNA hybrids. The washing procedure most
`suitable for a given application depends upon the
`nature of the hybrid formed. For example, when
`d(‘4C—pC)zoo(pT)2oo was hybridized with poly(G)—
`cellulose in 25% formamide at room temperature for
`5 hr in the absence of cold carrier, 93% of the input
`
`Page 7 of 10
`Page 7 of 10
`
`radioactivity bound to poly(G)—ce||u|ose, and 8%
`bound to control cellulose. Washing with 99% form-
`amide and 0.1% SDS at 40°C removed all of the
`
`radioactivity bound to the control cellulose, but only
`10% of that bound to the po|y(G)—ce||ulose. Material
`which had hybridized selectively to the poly(G)-
`cellulose was then eluted with 99% formamide, 0.1%
`SDS at 85°C.
`
`in the experiments presented, RNA was labeled
`with 3H—uridine, and hybridized RNA was eluted
`from the cellulose, precipitated in the presence of
`HCl, and collected on glass fiber filters for counting.
`When using 32P-labeled material where quenching
`clue to the cellulose is not significant,
`it is conve-
`nient to collect the hybrid cellulose on glass fiber
`filters which may be dried and counted directly after
`extensive washing.
`The major difference between filter hybridization
`and the procedure we describe is that the DNA is
`covalently attached to cellulose in the latter. Be-
`cause the finely divided cellulose has a high capac-
`ity for nucleic acid, very low backgrounds, equiva-
`lent to those obtained in filter hybridization assays.
`can
`be obtained.
`In
`addition,
`the
`coupling
`procedure is technically simple and avoids harsh
`treatment of the DNA. Both the hybridized and non-
`hybridized fractions of a sample can be recovered
`easily for further use or analysis. The method can
`be applied analytically to determine low levels of
`a specific nucleotide sequence within a larger het-
`erogeneous pool, or it can be applied preparatively.
`For example, by coupling SV40 cFlNA to the cellu-
`lose,
`it should be possible to effectively separate
`the two strands of SV40 DNA. The cellulose is not
`
`suitable for column chromatography because it is
`so finely divided, but it sediments rapidly to form
`a firm pellet and can be used very conveniently in
`a batch-wise fashion. The DNA cellulose can also
`
`be reused after thorough washing. For example, in
`experiments with CRNA, the efficiency of hybridiza-
`tion with recycled cellulose was the same as with
`freshly prepared cellulose. in addition to RNA—DNA
`hybridization, the method can be adapted for DNA-
`DNA hybridization, or hybridization using nucleo-
`tide homopolymers. it may be useful in gene selec-
`tion, and possibly for isolating DNA or RNA binding
`proteins and studying their interaction with nucleic
`acids.
`
`Burrell and Horowitz (1975) have found that large
`periodate-oxidized RNAs (163, 238) couple to Se-
`pharose-dihydrazide through their 3‘
`termini
`less
`efficiently than smaller oxidized RNAs (58). Proba-
`bly the larger molecules cannot penetrate the Se-
`pharose beads well.
`it may be possible to avoid
`such steric effects by coupling periodate oxidized
`RNAs
`to
`finely-divided
`cellulose
`rather
`than
`Sepharose.
`
`

`
`Cell
`308
`
`Experimental Procedures
`
`Enzymes and Reagents
`RNAase—free DNAase l, RNAase T1, and RNAase A were pur-
`chased from Worthington Biochemical Corp. RNAase A (1 mg/ml)
`was heat—treated at 90°C for 10 min to inactivate any DNAase.
`Proteinase K was obtained from EM Laboratories, and S1 nuclease
`from Aspergillus oryzae was prepared by Elizabeth Swyryd from
`Takadiastase powder (Sankyo Co., Ltd., Japan) as described by
`Sutton (1971). Preparations of E. coli RNA polymerase prepared
`from E. coli B according to Burgess (1969) were generously pro-
`vided by William Wickner and James Alwine. DMSO (Matheson
`Coleman and Bell) was redistilled under vacuum, and the fraction
`collected at 83°C at 17 mm Hg was placed over 4A molecular
`sieves, flushed with nitrogen, and stored at —20°C. Formamide
`(Matheson Coleman and Bell) was washed twice with equal vol-
`umes of ether and stored under nitrogen at —20°C. Cu (OH)2 was
`from K and K Labs. The potassium salts of poiy(A), poly (U), and
`poly (C) were from Calbiochem, and poly (dT) and poly (G) were
`obtained from Miles Laboratories. 5—3H—uridine (27.8 Ci/mM), 5,6-
`3H—uridine—5’triphosphate (36.84 Ci/mM), and H332PO4 (carrier
`free) were from New England Nuclear Co.
`
`Cell Lines and Virus
`MA-134 cells, an established line of Green Monkey kidney cells
`obtained from J. Pagano, were grown on 100 mm plastic dishes
`(Nunc) in Dulbecco's modified Eagle's medium (Gibco) supple-
`mented with
`10% calf
`serum (Microbiological Associates),
`100 pg/ml streptomycin sulfate (Pfizer), and 500 units/ml penicillin
`G (Squibb) in a CO; incubator at 37°C. C13/SV28 cells, an SV40
`transformed hamster cell
`line, were obtained from C. N. Wiblin
`(Wiblin and MacPherson, 1972) and maintained in suspension cul-
`ture at 37°C in the above medium. The virus used was SV40 wild
`type 830 (M. Herzberg, J. E. Mertz, P. Berg, J. R. Cameron, and
`R. W. Davis, manuscript in preparation) derived from strain SVS
`(Takamoto, Kirchslein, and Habel, 1966).
`
`Preparation of DNA
`SV40 (I) DNA was isolated from MA-134 cells infected at a multi-
`plicity of 0.1 plaque forming unit/cell. When 75-80% of the cells
`appeared rounded (10-11 days), 32PO4 (100 poi/plate) was added,
`and 24 hr later the DNA was collected as described by Hirt (1967),
`extracted twice with equal volumes of chloroformzisoamylalcohol
`(24:1, v/v), and concentrated by ethanol precipitation at —20°C.
`Resuspended DNA was incubated at room temperature and pH
`12 for 2 hr, readjusted to pH 8, and DNA(l) was isolated by equilib-
`rium density centrifugation in CsCl—ethidium bromide (Radloff,
`Bauer. and Vinograd, 1967). Full length linear SV4O DNA was gen-
`erated from purified DNA(l) by digestion with S1 nuclease as
`described by Beard, Morrow, and Berg (1973). The digestion was
`performed at 37°C for 15 min at a DNA concentration of 56 pg/ml.
`Under these conditions, less than 10% of DNA(l) remained, while
`more than 80% of the material cosedimented with DNA(ll) upon
`centrifugation in an alkaline sucrose gradient. E. coli DNA was
`isolated from E. coli B according to Marmur (1961). Salmon sperm
`DNA was purchased from Calbiochem. DNAs were sonicated for
`four 30 sec intervals in 10 mM Tris buffer (pH 7.5),
`1 mM EDTA,
`and 10 mM NaC| with an MSE sonicator operated at maximum
`amplitude. As judged by sedimentation in an alkaline sucrose
`gradient,
`this procedure generated SV40 DNA fragments with a
`mean sedimentation coefficient of 6—7S and E. coli DNA fragments
`of about 168.
`
`Preparations of RNA
`C13/SV28 cells were grown in the presence of 3H~uridine (4 ,uM
`uridine, 9 juLCi/ml,1 l) for 24 hr. The cells were collected by centrifu-
`gation, washed three times with ice cold Earle‘s saline, allowed
`to swell for 30 min at 0°C in 2.5 vol of hypotonic MCT buffer (10
`mM triethanolamine hydrochloride, pH 7.4, 2 mM Caclz, 5 mM
`
`Page 8 of 10
`Page 8 of 10
`
`MgCl2), and disrupted in a Dounce homogenizer. A crude nuclear
`pellet (fraction l) was obtained by centrifuging the homogenate
`for 2 min at 850 X g. The pellet was washed once with MCT buffer
`0.25 M in sucrose. This wash, together with the remainder of the
`cell homogenate, was centrifuged for 15 min at 17,000 X g to ob-
`tain a postmitochondrial supernatant fraction (Ii) and a pellet (Ill).
`Each of the three fractions was digested with proteinase K (200
`pg/ml) for 15 min at room temperature in 0.01 M Tris buffer (pH
`7.4), 0.5% SDS, 0.1 M NaCl, 0.02 M EDTA (Mach, Faust, and Vas-
`salli, 1973). SDS buffer (0.02 M Tris buffer, pH 7.4, 1% SDS, 0.2
`M NaCl, 0.04 M EDTA) was added to dilute each fraction to about
`40 times the volume of original packed cells, and the pH and SDS
`concentration were adjusted by addition of 0.1 vol of 1 M Tris
`(pH 9), and 0.1
`vol of 10% SDS. Each fraction was ex-
`tracted three times with equal volumes of chloroformzphenol
`(50:50, v/v) at room temperature (Leder, 1972), and precipitated
`by addition of 2.5 vol of ethanol at —20°C. RNA was collected by
`centrifugation, resuspended in 3 ml of MCT buffer, and digested
`with DNAase (50 pg/ml) at 0°C for 1 hr. The material was again
`extracted 2 times with chloroformzphenol, precipitated twice from
`ethanol at —20°C, redissolved in 99% formamide, 0.1% SDS. and
`stored at —20°C. The specific activity of the RNA was 2.6 X 106
`cpm/Am.
`Tritiated RNA from SV40-infected cells was prepared from con-
`fluent monolayers of MA-134 cells infected at a multiplicity of 20
`plaque forming units/cell. After 48 hr, 3H—uridine (4 pM uridine,
`10 jLCl/ml) was added to the cultures. At 72 hr after infection,
`DNA and proteins were precipitated by addition of NaC| and SDS
`as described by Hirt (1967), and the RNA was extracted from the
`supernatant solution as described above. RNA from infected cells
`pulse labeled for 20 min 72 hr after infection was prepared under
`the same conditions.
`Unlabeled total cellular RNA was extracted from MA-134 cells
`disrupted by Dounce homogenization as described for C13/SV28
`cells. Yeast RNA from Gal|ard—Schlesinger was extracted with
`phenol and reprecipitated from ethanol twice before use. E. coli
`tRNA (Schwarz/Mann) was used without further purification.
`RNA complementary to SV40 DNA (cRNA) was synthesized in
`vitro using E. coli RNA polymerase. The reaction mixture containing
`28 pg of SV40(l) DNA, 50 jug of polymerase, 10 mM MgC|2, 0.1
`mM EDTA, 0.1 mM dithiothreitol, 150 mM KCl, 0.3 mM ATP, CTP,
`and GTP, and 0.1 mM 3H—UTP in 1 ml of 40 mM Tris buffer (pH
`7.9), was incubated at 37°C for 2 hr (Burgess, 1969). RNAase-free
`DNAase (30 pg/ml) was added at 37°C for 30 min, and then EDTA
`was added to 0.04 M. The cRNA was extracted 3 times with phenol
`sat