`
`BIOCdEMISTRY
`
`138,267-284
`
`(1984)
`
`REVIEW
`
`Hybridization of Nucleic Acids Immobilized on Solid Supports
`
`JUDY MEINKOTH
`
`AND GEOFFREY WAHL
`
`The Salk Institute, Post Office Box 8.5800, San Diego, California 92138
`
`I. INTRODUCTION
`Two decades have elapsed since the devel-
`opment of methods
`for immobilizing DNA
`on nitrocellulose paper (4950) and
`for de-
`tecting the fixed nucleic acid with radioactive
`probes (18,25). Since this technology
`(which
`we will refer to as mixed phase hybridization)
`made it possible to process many samples si-
`multaneously,
`it led to a rapid demise of the
`cumbersome and tedious centrifugation
`(45)
`and column
`techniques
`(3,5,11) commonly
`used for the separation of single- and double-
`stranded nucleic acids. While the mixed phase
`hybridization
`technology was limited
`in its
`applications
`initially by the inability
`to obtain
`a diversity of gene-specific
`hybridization
`probes, molecular cloning
`techniques have
`now eliminated
`this problem. Within
`the last
`seven years, mixed phase hybridization
`is no
`longer a technology
`in search of a problem
`to
`solve; rather,
`it forms
`the cornerstone of the
`gene (and gene product) detection methods
`which have revolutionized our understanding
`of gene structure, genome organization, and
`the control of gene expression. The sensitivity
`(< 1 pg of complementary
`sequence), speed
`(~24 h), and convenience
`(simple machines
`and inexpensive materials are required) of the
`nucleic acid hybridization
`procedures have
`enabled them to be applied not only to basic
`research problems, but also to the diagnosis
`of heritable diseases and to the detection of a
`wide variety of microbial and viral pathogens.
`
`I&Y WORDS: DNA blot hybridization; RNA blot hy-
`bridization; nucleic acid probes.
`
`theme of
`Many variations on the central
`detection of immobilized
`nucleic acids have
`appeared in the past several years. To the nov-
`ice first encountering
`these techniques,
`the
`variations between procedures can be bewil-
`dering and result in the adoption of needlessly
`complicated methods. To those experienced
`in the art, the most expedient methods may
`not always be employed. Our aim in this re-
`view is to describe methods for detecting elec-
`trophoreticahy
`fractionated nucleic acids ei-
`ther after transfer to solid supports or directly
`in the gel matrix, and for detecting sequences
`in unfmctionated nucleic acids applied directly
`to solid supports. We will emphasize the most
`sensitive and rapid techniques, simplifying
`the
`procedures on the basis of published obser-
`vations or where our experience and that of
`many colleagues permit.
`
`PARAMETERS
`II. HYBRIDIZATION
`The kinetics of hybridization
`of RNA or
`DNA probes with DNA
`tethered
`to nitrocel-
`lulose (NC)’ or free in solution are very similar
`(25,49,50,72),
`suggesting
`that parameters
`which
`influence nucleic acid reannealing
`in
`solution will have similar effects in mixed
`phase systems. In this section, we discuss the
`‘Abbreviations used: NC, nitrocellukxe; SSC ( 1 X), 0.15
`M NaCl, 0.015 M Na citrate, pH 7.0; SSPE (IX), 0.18 M
`NaCl, 0.01 M NaPO,, pH 7.7,O.OOl M EDTA, MS, mul-
`tiple sclerosis; BSA, bovine serum albumin, Fraction V;
`SDS, sodium dcdecyl sulfate; DPT, diazophenylthioetheq
`EtBr, ethidium bromide; NT, nick translation; Pol I, E.
`coli polymerase I; DTT, dithiothreitol; PEG, polyethylene
`glycol.
`
`267
`
`0003-2697184 $3.00
`Copyright 8 1984 by Academic
`rights of reproduction
`All
`in any
`
`Inc.
`phu.
`form resewed.
`
`Exhibit 2018 Page 1
`
`Enzo Exhibit 2018
`Hologic, Inc. v. Enzo Life Sciences, Inc.
`Case IPR2016-00820
`
`
`
`268
`
`MEINKOTH
`
`AND
`
`WAHL
`
`rate and
`parameters which affect hybridization
`of these
`hybrid stability. An understanding
`parameters enables one to derive hybridization
`conditions which should yield the optimal sig-
`nal to noise ratios.
`
`A. Hybridization Rate
`for single-
`The rate of hybrid
`formation
`stranded probes in mixed phase hybridizations
`should
`follow
`first-order kinetics since the
`concentration of probe is in vast excess over
`that of target sequences. For complementary
`probes (e.g., those generated by nick trans-
`lation as described in Section TVA), the sit-
`uation
`is more complicated since probe rean-
`nealing in solution decreases the concentration
`of probe available for hybridization with the
`target. For illustration, we will consider
`the
`case for single-stranded probes.
`The time required
`for half of the probe to
`anneal with the tethered DNA
`is
`
`(ex-
`rate constant
`where k is the first-order
`pressed as liters mol nucleotide-’
`s-‘) for for-
`mation of a hybrid molecule and C is the
`probe concentration
`(mol nucleotides/liter).
`The hybridization
`rate constant
`(k) is a func-
`tion of probe strand
`length
`(L), molecular
`complexity
`(iV, the total number of base pairs
`in a nonrepeating
`sequence),
`temperature,
`ionic strength, viscosity, and pH (the latter
`four factors are included in the nucleation rate
`constant kb) (84)
`
`k =
`
`k’,Lo.5N-
`
`PI
`Ionic strength has little effect on the rate con-
`stant as long as it is kept above 0.4 M (e.g., k
`increases 1.5-fold between 0.4 and 1 .O M NaCl)
`and pH effects are small (i.e., <1.3-fold)
`in
`the range pH 5.0 to 9.15 when
`the salt con-
`centration
`is above 0.4 M NaCI
`(84). The
`maximum
`rate of hybridization
`in solution
`(at 1.0 M NaCI) has been determined empir-
`ically to occur at 25°C below
`the tm of the
`duplex
`(12) (tm is the temperature
`at which
`half of the hybrids are dissociated). This ob-
`
`[41
`
`c
`
`.
`
`to mixed phase hy-
`servation also pertains
`bridizations which utilize probes longer than
`approximately 150 nucleotides ( 17). However,
`the temperature
`giving
`the optimal hybrid-
`ization rate is depressed mrther in mixed phase
`hybridizations
`for probes shorter than 150 nu-
`cleotides (7,79).
`condi-
`hybridization
`Under
`the optimal
`tions, the Wetmur and Davidson
`relationship
`(84) pertains to nucleic acids of all sizes mea-
`sured [even oligonucleotides as short as 14 bp
`(SO)] and can be written
`k = 3.5 X 105Lo.5N-1.
`131
`Combining Eqs [I] and [3] yields an expres-
`sion for the time (in seconds) required
`to an-
`neal half of the probe to the target:
`Nln 2
`t “2 = 3.5 x 105. rp.5.
`This equation
`indicates that the hybridization
`time
`is minimized
`by using probes of low
`complexity at high concentrations.
`Thus, a
`probe with a complexity
`(and size) of 15 bp
`has a tl,2 of 250 s at 10 rig/ml (3 X low8 mol
`nucleotides/liter). On the other hand, the pre-
`diction
`from
`the equation
`that longer probes
`should decrease hybridization
`times
`is not
`likely to be true
`for some mixed phase hy-
`bridizations
`since some immobilized
`target
`sequences may not be accessible to long probe
`molecules (7,17).
`The rate of probe reannealing can be en-
`hanced in solution
`(83) and in mixed phase
`hybridizations using anionic dextran polymers
`(e.g., dextran sulfate 500) (78). In mixed phase
`hybridization,
`the effect of dextran sulfate is
`most pronounced
`for polynucleotides
`longer
`than about 250 nucleotides
`(G. Wahl, un-
`published observations) and has no eIIii
`on
`oligonucleotides 14 bases long (R. B. Wallace,
`personal communication).
`The increase in the
`apparent
`rate of hybridization
`is approxi-
`mately 3-fold for mixed phase hybridizations
`which utilize single-stranded probes and up
`to loo-fold
`for hybridizations utilizing nick-
`translated probes (78) (see below
`for a de-
`scription of nick translation.) This effect has
`been attributed
`to the accelerated rate of for-
`
`Exhibit 2018 Page 2
`
`
`
`NUCLEIC ACID HYLWDIZATION
`
`269
`
`or “hyperpoly-
`mation of probe “networks”
`mers” between
`the partially overlapping se-
`quences of the molecules generated by nick
`translation
`(78). It is important
`to emphasize
`that the ability
`to form such networks
`is crit-
`ically dependent on probe size since the over-
`laps in small probes are not likely to be able
`to initiate or maintain
`the formation of stable
`networks.
`
`6. Hybrid Stability
`is a
`The formation of nucleic acid hybrids
`reversible process and an understanding of the
`parameters which affect their stability enable
`one to derive the optimal conditions
`for dis-
`criminating
`between perfect and
`imperfect
`hybrids. The melting
`temperature
`(T,)
`is af-
`fected by ionic strength (M, in mol/liter), base
`composition
`(% G + C), the
`length of the
`shortest chain in the duplex (n), and the con-
`centration of helix destabilizing agents such
`as formamide
`(e.g., see (43,66)). The following
`equation has been derived from analyzing the
`influence of these factors on hybrid stability:
`T, = 81S”C + 16.6 log M + 0.41(% G + C)
`- 500/n - 0.61(% formamide).
`[5]
`This equation pertains
`to probes longer than
`approximately
`50 nucleotides. Hybrids be-
`tween oligonucleotides
`(14-20 bp) and
`im-
`mobilized DNA show decreased stability (79)
`and an empirical formula has been determined
`to define the optimal conditions
`for their hy-
`bridization
`(81). The
`temperature
`at which
`50% of these short duplexes dissociate (Td)
`when
`the hybridization
`is performed under
`standard conditions
`(e.g., 0.9 M NaCl)
`is:
`T,(Y) = 4(G + C) + 2(A + T)
`[6]
`where G, C, A, and T indicate the number of
`the corresponding nucleotides in the oligomer.
`A temperature 5 ’ below the T, is used to detect
`hybridization
`between perfectly matched
`molecules (74,79). Further considerations
`for
`hybridization with oligonucleotides are given
`below (see Section IVC).
`formed between
`The stability of duplexes
`strands with mismatched bases is decreased
`according
`to the number and location of the
`
`for
`mismatches and is especially pronounced
`short (e.g., 14 bp) oligonucleotides. For hybrids
`longer than 150 bp, the T, of a DNA duplex
`decreases by 1 “C with every 1% of base pairs
`which are mismatched
`(6). For hybrids shorter
`than 20 bp, the T, decreases by approximately
`5°C for every mismatched base pair (79-8 1).
`In order
`to minimize
`the hybridization
`of
`probe to related but nonidentical
`sequences,
`hybridization
`reactions must be performed
`under the most stringent conditions possible.
`From
`the discussion above, hybridization
`stringency can be altered by adjusting the salt
`and/or
`formamide
`concentrations and/or by
`changing the temperature. The stringency can
`be adjusted either during
`the hybridization
`step, or in the posthybridization washes (see
`Section VII). It is often convenient
`to perform
`the hybridization at low stringency and wash
`at increasing stringencies, analyzing the results
`after each wash. This enables the detection of
`related sequences and the monitoring
`of the
`effectiveness of the washes in removing
`these
`sequences. This strategy also enables one to
`obtain an estimate of sequence relatedness.
`
`III. DETECTION OF NUCLEIC ACIDS
`IMMOBILIZED ON SOLID SUPPORTS
`OR
`IN AGAROSE GELS
`Techniques are now available
`for immo-
`bilizing both DNA and RNA on solid supports
`consisting of NC
`(1,2,25,7 1,75), diazotized
`cellulose
`(1,2,67), Ecteola cellulose
`(65),
`DEAE-cellulose
`(29), and nylon [e.g., see (8)].
`We review below the blotting applications best
`served by these solid supports, some new ap-
`plications
`for immobilized
`nucleic acids, and
`some of the parameters affecting
`transfer ef-
`ficiency.
`
`A. Uses of Different Solid Supports
`The mechanism of binding of nucleic acids
`to NC is unknown, but has been assumed to
`be noncovalent. This assumption
`led to the
`belief
`that immobilized
`nucleic acids might
`be eluted from nitroccllulose during stringent
`washing procedures and provided one incen-
`tive for developing diazotized paper supports
`which bind nucleic acids covalently
`[e.g., see
`
`Exhibit 2018 Page 3
`
`
`
`270
`
`MEINKOTH
`
`AND WAHL
`
`indepen-
`in numerous
`(1,2,48,67)]. However,
`dent experiments with RNA or DNA bound
`to NC,
`it has been possible to remove
`the
`majority of the hybridized probe molecules
`(see Section VII) without eluting significant
`quantities of the bound nucleic acids. The hy-
`bridization signal does not change significantly
`after at least six probe
`removal washes (J.
`Meinkoth
`and G. Wahl, unpublished obser-
`vation). Due to the relatively
`low cost, con-
`venience of use, and reusability, NC is clearly
`the solid support of choice for most mixed
`phase hybridization
`applications. Since frag-
`ments smaller than about 200-300 bp bind
`poorly to NC but are covalently bound
`to di-
`azotized papers (62,73),
`the
`latter are rec-
`ommended
`for experiments
`requiring
`the
`binding of small fragments.
`A novel
`technique has recently been de-
`scribed for the transfer of RNA
`fractionated
`in agarose gels to Ecteola paper
`(65). The
`binding of RNA to Ecteola paper is noncova-
`lent and can be reversed, allowing
`for its sub-
`sequent translation
`in vitro. Alternatively,
`the
`RNA can be translated
`in situ on the paper
`and the protein products
`immobilized at their
`site of synthesis. This procedure
`is more rapid
`and sensitive than others which require RNA
`elution
`from gels prior
`to translation. RNA
`immobilized on NC has also been reported to
`be a substrate for in vitro translation and for
`reverse transcriptase
`(9). Alternatively, RNA
`may be transiently bound
`to commercially
`available DEAE membranes and eluted
`in a
`biologically active
`form using strong dena-
`turants (e.g., 6 M guanidine
`thiocyanate)
`(29).
`
`for Detecting Nucleic Acids
`6. Techniques
`after Electrophoresis
`through Gels
`
`(7 1) for “blotting”
`technique
`Southern’s
`electrophoretically
`fractionated DNA
`from an
`agarose gel to NC by passive diffusion estab-
`lished that transfers were faithful
`replicas of
`the high-resolution gel patterns. Passive trans-
`fer techniques consist of four steps: electro-
`phoresis, transfer
`[with
`the attendant DNA
`fragmentation
`(optional),
`denaturation,
`and
`neutralization
`steps], fixation onto NC, and
`
`sequences
`detection of specific immobilized
`by hybridization.
`Commonly
`encountered
`problems
`in DNA electrophoresis have been
`the subject of a recent paper and will not be
`discussed here (70). Early reports
`indicated
`that efficient
`transfer of large DNA was dif-
`ficult to achieve, leading to the development
`of chemical
`(78) or uv light treatments
`(41)
`to fragment
`the DNA prior to transfer to en-
`sure equal
`transfer efficiencies of all DNA
`molecules. The chemical
`fragmentation
`(de-
`purination) method has been modified by de-
`creasing the depurination
`time (in 0.25 M HCl)
`to 7.5-10 min. Longer
`treatments generate
`smaller fragments which do not bind well to
`NC and result in lower hybridization
`signals.
`Solarization of uv filters over time reduces uv
`transmission
`to the gel and results in irrepro-
`ducible
`fragmentation.
`Fragmentation
`of
`DNA
`to average single-strand
`lengths of ap-
`proximately
`1000 bp enables
`the efficient
`transfer and retention of DNA molecules re-
`gardless of their original size or conformation
`(e.g., supercoil, linear, etc.). Fragmented DNA
`also transfers very rapidly; only l-2 h are re-
`quired to transfer >90% of the DNA
`from a
`1 S-cm-thick gel. A good rule of thumb
`is that
`transfer can be stopped when the gel thickness
`decreases to approximately
`1 mm since the
`gel concentration
`at this point
`is sufficiently
`high to prevent
`further
`transfer. Procedures
`for electroeluting DNA
`from agarose gels are
`therefore unnecessary since they require as
`much time as passive diffusion of fragmented
`DNA. However,
`in order
`to resolve small
`DNA
`fragments (i.e., ~0.5 kb), it is necessary
`to use polyacrylamide
`or composite acryl-
`amide-agarose gels (4,62). Electroelution may
`be advantageous for the transfer of small DNA
`molecules since DNA diffuses slowly from ac-
`rylamide gels and diazotized paper, the solid
`support of choice, has a limited
`time over
`which
`it can bind DNA
`(e.g., see (4) for ex-
`amples). Denaturation
`of small DNA
`fmg-
`ments has been reported to promote
`their re-
`tention on nitrocellulose
`(75). The use of large
`volumes oftransfer buffer (e.g., 20X SSC, 20X
`SSPE, or 1 M NH40Ac)
`is unnecessary since
`efficient transfer can be obtained from a simple
`
`Exhibit 2018 Page 4
`
`
`
`NUCLEIC ACID HYBFUDIZATION
`
`271
`
`apparatus consisting of a tray with two to four
`pieces of buffer-saturated blotting paper (e.g.,
`Whatman 3MM or S&S No. 470 WH) or a
`sponge.
`While most protocols specify a 2-hr baking
`step to fix nucleic acids to NC, we have found
`that it is only necessary to bake for as long as
`is required
`to completely dry the paper. This
`is usually
`lo-20 min at 80°C
`in a vacuum
`oven evacuated
`to 30 mm Hg. Baking
`times
`in excess of 24 h can produce brittle NC which
`shatters upon
`further manipulations. Nylon
`supports have been developed which remain
`pliable and easy to handle after baking. How-
`ever, the signals obtained
`from genomic DNA
`transferred
`to some nylon supports by the
`Southern
`technique can be as much as 25-
`fold
`less than
`those obtained
`for the same
`DNA
`transferred
`to NC
`(J. Meinkoth,
`un-
`published observations) and the background
`can be much higher. A new durable support,
`Genetran
`(manufactured
`by Plasco in Wo-
`bum, Mass.), has given signals equivalent
`to
`NC with
`low noise levels in preliminary ex-
`periments.
`the original
`of
`A recent modification
`Southern transfer protocol is the “bidirectional
`blot” which allows two identical blots to be
`produced from a single gel (69). This technique
`is valuable for restriction mapping studies (i.e.,
`enabling one to hybridize with
`two different
`probes simultaneously) and for checking
`the
`specificity of multiple cloned probes. DNA
`fragmentation
`is required
`for the transfer
`to
`be efficient since the gel diminishes
`in thick-
`ness very rapidly using this procedure (transfer
`from a 0.7% gel is complete after 1 h). Small
`DNA
`fragments can also be transferred bi-
`directionally
`from polyacrylamide
`gels (69).
`With 1 O-20 pg of DNA, bidirectional blotting
`can be used to detect unique sequence genes
`in mammalian genomic DNA
`(G. Wahl, un-
`published observations).
`An alternative to the conventional Southern
`transfer technique
`is to hybridize
`radioactive
`probes directly with DNA
`in dried agarose
`gels (586876).
`The sensitivity
`is the same as
`that of standard DNA blots with nick-trans-
`lated probes and is at least 5-fold greater with
`
`oligonucleotide probes (79) (see Section IVC).
`Hybridization
`in the presence of 50% form-
`amide
`for extended periods
`(e.g., >36 h at
`42’C) can solubilize the gel. DNA
`fragments
`l-50 kb long are quantitatively
`retained
`in
`the dry gel, but
`there is some loss of small
`fragments (~3 12 bp) during drying and/or hy-
`bridization. Dried gels can be rehybridized
`with different probes. Elution of hybridized
`DNA has been reported
`(58), but in our ex-
`perience it is often incomplete
`(G. Wahl, un-
`published observations).
`( 1,2)
`technique
`The
`“Northern
`transfer”
`was designed to transfer RNA
`from agarose
`gels to solid supports. The sensitivity
`is suf-
`ficient
`to detect low abundance
`(~0.0 1% of
`total mRNA) eukaryotic RNA molecules. A
`number of improvements have been made in
`the original procedure
`to make it more sen-
`sitive and easier to perform. The denaturant
`originally used, methylmercuric hydroxide (I),
`has been replaced by less toxic ones such as
`glyoxal(44,75)
`or formaldehyde
`(59) and NC
`has replaced diazotized paper as the solid sup-
`port of choice (75). Fragmentation of the RNA
`is not required prior to transfer when glyoxal
`or formaldehyde are used as denaturants. Un-
`der these conditions,
`it is possible to quanti-
`tatively
`transfer RNA molecules 10 kb or
`larger (55; J. Meinkoth,
`unpublished
`obser-
`vations). The sensitivity of detection of form-
`aldehyde denatured RNA
`transferred
`to NC
`in greater than that obtained using other com-
`binations of denaturants and solid supports
`(85). Bidirectional
`transfer of RNA
`from a
`0.75% methyl mercury agarose gel to diazo-
`tized paper has been reported
`(69) and so it
`is likely that
`the same methodology
`can be
`applied
`to glyoxal or formaldehyde
`agarose
`gels for transferring RNA to nitrocellulose pa-
`per. In situ detection of RNA
`in agarose gels
`has also been achieved (76).
`
`for Detecting DNA and
`C. Techniques
`in Unfractionated Nucleic
`RNA Sequences
`Acid Samples
`
`to be able to measure
`It is often sufficient
`the amounts of DNA or RNA sequences with-
`
`Exhibit 2018 Page 5
`
`
`
`272
`
`MEINKOTH
`
`AND
`
`WAHL
`
`their sizes. Electrophoresis
`out determining
`and transfer are unnecessary
`in such cases.
`Procedures have been developed over the past
`several years to allow the quantitation
`of se-
`quences in unfractionated DNA or RNA ap
`plied
`to NC. The sensitivity of these “spot
`hybridization” methods (32) is often superior
`to that of the blotting
`techniques since both
`degraded and
`intact sequences which are
`bound
`to the solid support contribute
`to a
`single concentrated hybridization
`signal. The
`nucleic acid can be loaded onto the solid sup-
`port rapidly using multisample vacuum man-
`ifolds which apply the sample in narrow rect-
`angular slots or in round dots. The strongest
`signals are obtained using deproteinized
`nu-
`cleic acid samples since protein co-immobi-
`lization competes with
`the nucleic acid for
`binding sites on NC and also adds to the back-
`ground
`(9). Vanadyl
`ribonucleosides,
`ribo-
`nuclease inhibitors, significantly
`increase hy-
`bridization
`signals with RNA blots
`(9; S.
`Albanil and D. Richman, personal commu-
`nication).
`It
`is also essential
`to
`limit
`the
`amount of nucleic acid applied so that
`the
`paper is not saturated (e.g., approximately 0.5-
`1 &mm*)
`or else significant variability
`in
`hybridization
`signals is observed. With DNA
`“slot blots,”
`it is possible
`to detect unique
`sequence genes in 5 pg of mammalian
`cell
`DNA after a 2-hr autoradiographic
`exposure
`(at -70°C with an intensifying screen) when
`the hybridization mix contains 10% dextran
`sulfate and
`lo7 cpm/ml
`(approximately
`100
`rig/ml) of filtered nick-translated
`probe (see
`Sections IV and VII).
`While there is one commonly used method
`for applying DNA to NC [see (32) and Section
`VII], several variations have been reported for
`RNA (9,75,85). We and many colleagues have
`found
`the “cytodot”
`procedure
`(85)
`to be
`rapid, reproducible, and sufficiently sensitive
`to detect an mRNA present with an abun-
`dance of 0.1% of the total poly(A)+ RNA after
`hybridization with 0.3 to 1 .O X 1 O7 cpm/ml
`of nick-translated probe using a 24-h exposure
`(at - 70 “C with an intensifying screen). A new
`procedure,
`“Quick blot,” utilizes NaI to lyse
`the cells and to allow for the direct application
`
`of RNA to NC (9). In the presence of nonionic
`detergents such as Nonidet P-40, mRNA but
`neither DNA nor poly(A)- RNA adheres to
`the filter (9). Minor variations of the Quick
`Blot RNA procedure allow DNA to be applied
`exclusively (9). We have found
`that the times
`required
`to perform
`the cytodot and Quick
`Blot procedures are comparable, but the su-
`persaturated Nal solutions
`required by the
`Quick Blot procedure are difficult
`to manip
`ulate.
`
`IV. PROBING STRATEGIES AND
`PROBE PREPARATION
`Recombinant DNA
`techniques have made
`it possible to obtain
`large quantities of gene-
`specific probes
`in either plasmids
`[e.g., see
`(42)], or single- (46) or double-stranded phage
`vectors [e.g., (42)]. A variety of strategies have
`been devised for preparing probes from
`these
`vectors and
`for hybridizing
`them with
`im-
`mobilized nucleic acids. Some of the most
`commonly used and convenient methods are
`summarized below.
`
`A. Nick Translation of
`Double-Stranded
`DNA
`Double-stranded DNA (linear, supercoiled,
`nicked, or gapped circular) can be labeled to
`specific activities > 1 O* cpm/Hg with deoxynu-
`cleotide 5’-[32P]triphosphates by “nick
`trans-
`lation”
`(63). This method
`takes advantage of
`the ability of Escherichia co/i DNA polymerase
`I to combine
`the sequential addition of nu-
`cleotide residues to the 3’-hydroxyl
`terminus
`of a nick (generated by pancreatic DNase I)
`with the elimination
`of nucleotides
`from
`the
`adjacent 5’ phosphoryl
`terminus. Since the
`nicks are introduced at random sites in the
`duplex, this method generates a population of
`radioactive
`fragments which partially overlap
`each other. Under standard conditions
`(i.e.,
`saturating nucleotide
`triphosphate concentra-
`tion), the size of the fragments
`is determined
`by the DNase concentration,
`Fragments ap-
`proximately 500- 1500 nucleotides
`long pro-
`duce optimal signal to noise ratios. This is
`presumed
`to result from
`their ability
`to hy-
`
`Exhibit 2018 Page 6
`
`
`
`NUCLEIC ACID HYBRIDIZATION
`
`273
`
`bridize with each other in overlapping com-
`plementary
`regions
`to form
`“networks”
`or
`“hyperpolymers,”
`a process accelerated by
`dextran sulfate 500 (78). Longer probes have
`been correlated with higher backgrounds,
`especially when using dextran
`sulfate
`(G.
`Wahl, unpublished
`observations). Shorter
`probes give less of an enhancement with dex-
`tran sulfate, presumably because the short re-
`gions of overlap cannot either form or sustain
`stable hypetpolymers. The variability
`in probe
`sizes generated by nick translation
`reactions
`performed at different
`times may explain
`in-
`consistencies in the magnitude of the dextran
`sulfate effect (78). Conditions given in Section
`VII have been adopted
`in order to minimize
`such variability.
`There are two general strategies for using
`nick-translated probes. First, a nick-translated
`DNA
`(e.g., a plasmid or phage containing an
`insert, or the purified
`insert fragment) can be
`hybridized directly with the immobilized
`tar-
`get DNA
`(Fig. 1A). Second, the immobilized
`target sequences can be detected by a two-
`stage “sandwich hybridization”
`( 19-22): in the
`first stage, the target is hybridized with un-
`labeled plasmid sequences or single-stranded
`phage DNA which
`is attached
`to sequences
`complementary
`to the target;
`in the second
`stage, labeled sequences complementary
`to the
`vector are hybridized with
`the vector
`tails
`which protrude
`from the hybridized
`target se-
`quences (e.g., see Fig. 1C). The latter approach
`has the advantage that it requires the synthesis
`of only one labeled probe
`for detecting all
`target sequences. A potential advantage is that
`the signals produced
`from short and long in-
`serts should be the same since it is the labeled
`vector sequences which produce
`the signal.
`This type of approach should prove to be ben-
`eficial for gene quantitation
`experiments since
`signal strength variations due to differences in
`probe sizes are eliminated.
`
`6. Single-Stranded Probes
`
`signal generated by nick-
`The hybridization
`translated probes is compromised by the abil-
`ity of the complementary
`strands to hybridize
`
`A) Nek
`
`translallon
`
`R) Single strand MlS/prtmer
`
`extensnn
`
`C) Sswle strand Ml3tsandwch
`
`D) RNA probes generated
`cloned DNA
`
`by transcription
`
`of
`
`E) Synthetic ollgonucleotldes
`
`FIG. 1. Strategies for probe preparation and target se-
`quence detection.
`
`to hybrid-
`in solution prior
`with each other
`with
`the immobilized
`nucleic acid.
`ization
`With regard to preformed hyperpolymers,
`the
`unhybridized single-stranded
`tails may be too
`short to anneal stably with
`the immobilized
`nucleic acid, and in cases such as hybridization
`with thin sections of tissues, large nucleic acid
`aggregates may not be able to dil%e
`through
`
`Exhibit 2018 Page 7
`
`
`
`274
`
`MEINKOTH
`
`AND
`
`WAHL
`
`the tissue section (7,17). Consistent with these
`ideas is the observation
`that small (i.e., ap-
`proximately
`100 bp) single-stranded probes
`give higher signals in in situ hybridization with
`tissue sections than self-complementary probes
`of identical sequence ( 17). We and others (e.g.,
`M. Akam, personal communication)
`have ob-
`served that
`the signals generated by single-
`stranded probes c”an equal and sometimes ex-
`ceed those generated by nick-translated mol-
`ecules.
`the single-stranded
`Vectors based upon
`phage M 13 have allowed several novel probing
`strategies to be developed
`[e.g., see (30,46)].
`Single-stranded Ml 3 phage DNA containing
`a specific insert can be hybridized with a small
`( 15 bp) oligonucleotide complementary
`to the
`viral (+ strand) sequences upstream of the
`insert. Extension of the primer away from the
`insert using the Klenow
`fragment of E. co/i
`DNA polymerase
`I (which
`lacks the 5’ - 3’
`exonuclease activity)
`in the presence of ra-
`dioactive nucleotides generates probes of high
`specific activity (30). Since few extended mol-
`ecules traverse the insert, a partially duplex
`probe is generated which can hybridize with
`target sequences complementary
`to the insert.
`It is important
`to note that
`the radioactive
`probe consists of M 13 sequences and the spec-
`ificity of the probe is dictated by the insert.
`Thus,
`this procedure should give greatly
`in-
`creased sensitivity over nick translations
`in
`cases where the insert may be very short. The
`sensitivity of this approach could be increased
`further by combining
`in one hybridization
`re-
`action separate Ml 3 clones which contain
`short stretches complementary
`to different
`re-
`gions of the same gene.
`A variation of the Ml 3 primer extension
`strategy is to employ “sandwich” hybridization
`to generate a universal probe
`for sequences
`cloned in M 13 (see Section IIIA and Fig. 1 C;
`D. Redfield, S. Albanil, D. Richman, and G.
`Wahl, unpublished
`observations; G. Guild,
`unpublished observations; M. G. Rosenfeld,
`unpublished observations). This approach has
`the advantage
`that the single-stranded DNA
`can saturate the target sites and can then be
`
`detected using nick-translated
`which can form hyperpolymers.
`
`molecules
`
`C. Oligonucleotide Probes
`
`(14-20 bp) de-
`Synthetic oligonucleotides
`rived from a known amino acid sequence can
`be prepared, obviating
`the requirement
`for
`cloning mRNA or genomic DNA sequences
`to obtain a gene-specific probe
`[e.g., see (79)
`for a review]. Hybridization
`conditions have
`been derived
`[see Section II and (74,79-81)]
`which allow one to detect the hybridization
`of a nonadecamer with its complement
`in a
`Southern blot of human DNA. By contrast,
`a nonadecamer with a single mismatch near
`the middle of the sequence can be prevented
`from hybridizing by using stringent hybrid-
`ization conditions or can be eluted by appro-
`priate washing conditions
`[e.g., see (79) and
`Section VI for the application of oligonucle-
`otide probes in medical diagnostics]. Due to
`the large signal differences obtained between
`perfect and imperfect hybrids with oligonu-
`cleotides, it has been possible to use complex
`mixtures of oligonucleotides as probes to de-
`tect a sequence represented by only one mem-
`ber of the mixture
`[see (79)]. Thus, one can
`utilize
`the oligonucleotide
`probe approach
`without
`knowing
`the exact nucleotide
`se-
`quence corresponding
`to an amino acid se-
`quence by synthesizing a mixture which con-
`tains the correct sequence as one component.
`Oligonucleotide
`probes hybridize about 5-
`fold more effectively with DNA
`fragments
`in
`agarose gels than with
`the same fragments
`transferred to NC (79; R. B. Wallace, personal
`communication).
`This is probably due to de-
`creased hybrid stability caused by some bases
`which are prevented
`from
`forming hydrogen
`bonds because they are tethered
`to the solid
`support. Prehybridization with solutions con-
`taining homochromatography
`mix 1 (3 1) has
`been observed to lower background hybrid-
`ization with oligonucleotide
`probes (Wallace,
`personal communication).
`It should be noted
`that oligonucleotides generally hybridize to the
`bulk of mammalian DNA
`(even under strin-
`
`Exhibit 2018 Page 8
`
`
`
`NUCLEIC ACID HYBRIDIZATION
`
`275
`
`gent conditions) and create significant back-
`ground
`[e.g., see (1611. Restriction
`fragments
`within
`the background hybridization
`“smear”
`could be obscured.
`
`D. RNA Probes Generated by in Vitro
`Transcription
`
`in hybridization
`The signal to noise ratio
`reactions with RNA probes
`is potentially
`greater than
`that with DNA probes
`for two
`reasons: (a) RNA-DNA
`duplexes are more
`stable than DNA-DNA
`duplexes
`(13), and
`consequently
`they can be formed and washed
`under more stringent conditions:
`(2) RNase
`can be used to remove nonspecifically bound
`RNA whil