Volume 7 Number 4 1979
`
`Nucleic Acids Research
`
`Separation of o6go-RNA by reverse-pblse HPLC
`
`Gloria D.McFarland and Philip N.Borer
`
`Department of Chemistry, Univeni.ty of California, Irvine, CA 92717, USA
`
`Received 22 June 1979
`
`ABSTRACT
`
`A rapid and highly reproducible chromatographic technique
`has been developed for analysis and purification of complex
`mixtures of oligoribonucleotides. The method utilizes a column
`of microparticulate porous silica beads fully derivatized with
`octadecylsilyl groups. The column is eluted with gradients in
`acetonitrile/water/ammonium acetate pumped at pressures of
`1500-3000 psi. Most separations are completed in 5-15 min. with
`usually better than 11. reproducibility of absolute retention times
`and about 0.1% reproducibility of relative retention times. A
`single column accomplishes separations of mononucleosides,
`mononucleotide&, and larger oligomers through at least 20-mers.
`The absolute detection limit is ~1 pmole of base though most of
`the analytical separations described use ~1 nmole.
`In favorable
`circumstances it is possible to use the analytical columns to
`purify ~1 mg of an oligonucleotide in a single 10-30 min. elu(cid:173)
`tion.
`
`INTRODUCTION
`
`Most oligonucleotide syntheses have depended heavily on low(cid:173)
`pressure liquid chromatography by ion-exchange. 1- 5 Recent
`improvements in separation methods involve high-pressure ion(cid:173)
`exchange,6 RPC-5 (which combines ion-exchange and reverse-phase
`modes) , 7- 12 and reverse-phase chromatography on covalently bonded
`silica-octadecylsilane (ODS) columns where Khorana and his co(cid:173)
`workers have made considerable progress in separating derivatized
`and underivatized DNA oligomers. 13 All but the latter suffer
`from the disadvantage that the long elution times (30 min. to
`several days) severely limit their analytical usefulness. A
`major objection to RPC-5 is the current lack of a stable
`commercial supply and severe irreproducibility between batches. 8
`Additionally RPC-5 cannot be used to separate very short oligomers
`
`C Information Retrieval Limited 1 Falconberg Court London W1 V 5FG England
`
`1067
`
`1
`
`MTX1015
`
`

`

`Nucleic Acids Research
`
`because the non-covalently bound coating bleeds off the beads at
`the low salt concentrations required. This problem is not
`encountered with the bonded-phase ODS columns.
`Prepacked microparticulate silica-ODS columns such as those
`used in our work are available from several manufacturers. We
`have used four different ''Microbondapak Cl8" columns from Waters
`Corporation over a period of three years and find excellent
`reproducibility from column to column. A similarly formulated
`column from Altex Corporation provided different chromatographic
`patterns but would be suitable for separations like those de(cid:173)
`scribed here. A stable supply of these columns is assured
`because the manufacturers sell thousands each year due to their
`wide use for separations in organic chemistry.
`Reverse-phase separations are based on the exclusion of a
`solute containing lipophilic components from a polar mobile
`phase. When the mobile phase is made less polar partition no
`longer favors the stationary phase for the least lipophilic
`molecules; solutes elute from the column in order of decreasing
`polarity.
`In order to decrease separation time and band(cid:173)
`broadening it is often useful to perform gradient elution where
`the solvent mixture is made progressively less polar.
`In our
`situation the acetonitrile content is increased in a solvent
`that is otherwise composed of 1% ammonium acetate in water.
`Methanol is also used in place of acetonitrile by many chemists;
`we prefer acetonitrile for two reasons:
`(1) its vapor pressure
`is much lower which reduces problems with bubble formation by
`cavitation at high pumping rates, and (2) the viscosity of aceto(cid:173)
`nitrile-water mixtures is much lower than corresponding methanol(cid:173)
`water mixtures allowing columns to be operated at lower pressures
`thus decreasing wear on pumps and columns. Acetonitrle is cur(cid:173)
`rently 3 times more expensive than methanol but one gallon of
`acetonitrile suffices for 300 hours of continous chromatography
`using our system. Thus solvent cost is a rather small factor.
`
`MATERIALS AND METHODS
`HPLC System. A modular HPLC system was purchased from Waters
`Corporation (Milford, Massachusetts) including two model 6000A
`high pressure pumps, a model 660 solvent programmer, a model 440
`
`1068
`
`2
`
`

`

`Nucleic Acids Research
`
`uv absorbance monitor, and a model U6K sample injector. The
`outlet of a 30 em x 0. 39 em ID "Microbondapak Cl81' column was
`connected to the uv monitor and preceded by a guard column, cut
`from a length of a 2 mm ID "Corasil Cl8" column (also from Waters),
`which was connected to the injector. The guard column was cut
`to a length of 2 em with a tubing cutter being careful to prevent
`any loss of packing material. The guard columns had to be re(cid:173)
`placed at 6-8 week intervals when the normal operating pressure
`at 3 ml/min rose to ~3000 psi. A single 2 mm ID ~ 61 em Corasil
`Cl8 column can provide at least 20 guard columns. All column con(cid:173)
`nections were made with (0.009 in ID) stainless steel tubing and
`zero-dead-volume fittings (Parker-Hannifin, Irvine, California).
`The total void volume of the system was 1.24 ml from injector to
`uv detector.
`Solvents. Solvents had to be maintained at a very high
`degree of purity. Deionized water was further purified through
`a ''11illi-Q" (Millipore Corporation) system consisting of a pre(cid:173)
`filter, an activated charcoal filter, two ion-exchange cartridges,
`and a 0.22 micron final filter. To maintain high-purity water
`the system was allowed to recirculate at least 15 min before
`drawing off water; no more than 6~ should be removed before allow(cid:173)
`ing another 15 min of recirculation. Because gradient elution
`is frequently performed over a range of only 2-3% in acetonitrile
`concentration it is absolutely necessary to prepare solvents in
`a uniform manner.
`10% ammonium acetate (Mallinckrodt AR grade)
`was prepared by dissolving the entire contents of a ~ lb bottle
`in water, adjusting to pH= 5.9 with acetic acid and diluting to
`1.11 ~. then filtering through a 0.45 micron filter (Millipore(cid:173)
`HA) . Acetonitrile (Baker HPLC grade) was filtered through a
`Fluoropore (Millipore-FH) filter. Solutions are prepared by
`diluting appropriate amounts of CH3CN and 10% AmOAc to 9.00 t
`according to the chart
`on the following page. Solution a
`(or a') is supplied to the low concentration pump, called
`pump A; likewise b (or b') is supplied to the high side, called
`pump B. Most elutions were performed with solutions a and b,
`covering a range from 2% to 12% CH3CN. The model 660 programmer
`adjusts the pumping rates of pumps A and B to maintain a constant
`total flow rate and gives a readout of the portion of flow
`
`1069
`
`3
`
`

`

`Nucleic Acids Research
`
`Solution
`
`'%.CH3CN
`
`CH3CN(m1)
`
`10'1. AmOAc(mf.)
`
`a
`
`b
`
`a'
`
`b'
`
`2
`
`12
`
`1
`
`11
`
`180
`
`1080
`
`90
`
`990
`
`880
`
`790
`
`890
`
`800
`
`provided by pump B as "'1.B". Selection of a narrow range in
`CH3CN concentration allows greater accuracy in reproducing solvent
`composition and a 10'1. range between solutions a and b (or a' and
`b') allows easy computation of absolute CH3CN concentration at any
`time during the elution by the formula '1.B/10 + A • CH3CN con(cid:173)
`centration; A= 2% for solution a and 1% for a'.
`In the chrom(cid:173)
`atograms shown in Figs. 1-3 solvent strength increased linearly
`with time from t = o until the end of the gradient program; there(cid:173)
`after solvent composition and flow rate remained constant until
`all peaks were eluted.
`In certain separations (not shown) con-
`vex or concave gradients were adopted to optimize resolution of
`components. Flasks containing "'200 ml of pure CH3CN and "-11
`volumes of a and b were degassed in an ultrasonic bath at the
`beginning of each day. Solvent lines, pumps, and columns were
`stored in pure CH3CN and air bubbles carefully cleared from sol(cid:173)
`vent lines and pumps at the beginning of each day. Solvent switch(cid:173)
`over from pure CH3CN to solutions a and b was accomplished by
`gradient programs of 5 min at 3 mi./min.
`Sample Injection. Samples were always clarified by centri(cid:173)
`fugation if cloudiness was apparent. Syringes were purchased
`from Hamilton Corporation supplied with a 26S ga, 22° bent
`point needle designed to prevent damage to the U6K injector.
`Sample sizes smaller than 25 111 were tnjected without regard to
`chloride ion concentration but large samples and solvents con-
`
`1070
`
`4
`
`

`

`Nucleic Acids Research
`
`taining halide ions were avoided because of chemical attack on
`stainless steel. Following each nm a 3 m.l "bomb" of solution
`b (programmer set to 100%B) was run into the column to elute
`any strongly adsorbed species before returning to solvent
`conditions appropriate for the next run. The next sample can(cid:173)
`not be loaded onto the U6K injector until the bomb has cleared
`the sample loop and the column restored to initial conditions
`for the next nm.
`Isolation of Purifie'd Components. The analytical columns can
`be used for semi-preparative purifications in many instances.
`Acetonitrile and most of the water is removed from selected frac(cid:173)
`tions in a flash evaporator. Samples are redissolved in water
`such that the ammonium acetate concentration is 10% or less,
`frozen, and lyophilized to remove ammonium acetate. Dialysis
`or desalting on polyacrylamide gels is usually not necessary.
`Oligonucleotide Preparation. Nucleosides, nucleotides,
`dinucleoside monophosphates, poly U, and poly C were purchased
`from Sigma Chemical Corporation; poly A from Miles Biochemicals;
`alkaline phosphatase (calf intestine) from Boehringer-Mannheim.
`Primer-independent polynucleotide phosphorylase was purchased
`from P-L Biochemicals and made primer-dependent by limited
`trypsin digestion. 14
`Homologous series, oligo An• oligo en, and oligo Un, were
`prepared by alkaline hydrolysis at 60°C (3.3 mM poly X, 1.0 M KOH,
`100 pi total volume). Hydrolysis was stopped (at times indicated
`in the figure legends) by addition of 12 pi 10 M HCi04 and
`incubated at pH ~2 to eliminate 2'-3'-cyclic phosphates (60° for
`10 min) . Terminal phosphates were removed by incubation
`overnight at 37°C with alkaline phosphatase (300 pg/mi) after
`adding 6 pi 2M HEPES buffer and adjusting to pH=7. 5 with -v2pi
`10 M KOH. Solid KCi04 was removed by centrifugation. Oligo(cid:173)
`nucleotide preparations by primer-dependent polynucleotide
`2-5
`phosphorylase followed in methods described earlier.
`
`RESULTS AND DISCUSSION
`Dbt'ucleos'ide monophosphates. Several general features of
`reverse-phase HPLC separations of oligo-RNA are shown in Fig. 1.
`
`1071
`
`5
`
`

`

`Nucleic Acids Research
`
`A254
`
`.03
`
`.02
`
`.0 I
`
`e.a,7
`
`2 • •
`
`8
`
`1011
`12
`
`0~ u
`\....) "-Ju u
`.
`
`0
`
`5
`
`Time (min)
`
`10
`
`IS
`
`~· '
`
`\.
`
`16
`
`15
`
`Figure 1. 16 dinucleoside monophosphates separated by a 15 minute
`linear CH 3CN gradient, from 2% to 8% (using solvents a and b),
`at 3 ml/~n flow rate. The numbered peaks are identified in
`Table I.
`
`All 16 of the dinucleoside monophosphates were loaded onto the
`column c~o.5 nmole each) and separated by a 15 minute linear
`gradient covering a range from 2% to 8% CH3CN (Solvents a and b,
`gradient 0-+60%B). Elution parameters for this separation:
`peak identities, retencion times (tR), and retention times relative
`to ApA are collected in Table I along with parameters for a simi(cid:173)
`lar 15 minute 2% to 11% CH3CN program. The figure shows that
`most of the 16 components are clearly separated in this single
`15 minute run; further illustrations of the tremendous resolving
`power of this system are found in comparing the isomeric molecules,
`eg. ApC and CpA which elute 2.5 min. apart. It is usually true
`that e-rich oligomers elute first, followed by U-rich, G-rich,
`then A-rich oligomers in that order; CpC elutes first and ApA last.
`The arrangement of data in Table I follows this order, the XpC
`group first (X is any of the four bases), followed by the XpU,
`XpG, and XpA groups. Within each group it is seen that as X
`changed through C, U, G, and A there is a 1.5 to 3.5 min. inter(cid:173)
`val in tR. This interval is fairly constant for a specific 5'(cid:173)
`nucleotidyl base, eg. the difference in tR between 5'-G and
`
`1072
`
`6
`
`

`

`Nucleic Acids Research
`
`TABLE I
`
`Elution Parameters for 16 Dinucleoside Monophosphates
`Separated by Two Different 15 minute Linear Gradients, at
`
`3 ml/min Flow Rate, Using Solvents a and b
`
`0+60t B
`(2t to 8t CH 3CN)
`
`0+90t B
`(2t to llt CH 3CN)
`
`Number
`(Fig. 1)
`
`tR
`(min:sec)
`
`XpY
`
`tf Relative to
`A • 1. 00
`p
`
`tR
`(min:sec)
`
`tR Relative
`to ApA • 1.00
`
`CpC
`
`ope
`
`u
`~ GpC
`
`1
`
`4
`
`7
`
`3:42
`
`5:26
`
`6:50
`
`ApC
`
`12
`
`10:32
`
`CpU
`
`0 OpO
`llo
`~
`
`GpO
`
`2
`
`5
`
`9
`
`4:30
`
`6:45
`
`9:28
`
`0.236
`
`0.347
`
`0.436
`
`0.672
`
`0.287
`
`0.431
`
`0.604
`
`3:56
`
`5:37
`
`6:59
`
`9:55
`
`4:48
`
`6:45
`
`8:40
`
`11:42
`
`0.276
`
`0.394
`
`0.490
`
`0.696
`
`0.337
`
`0.474
`
`0.608
`
`ApO
`
`14
`
`12:30
`
`0.798
`
`CpG
`
`OpG
`
`l.'l
`llo
`~ GpG
`
`ApG
`
`CpA
`
`8 OpA
`
`~ GpA
`
`ApA
`
`3
`
`6
`
`10
`
`15
`
`8
`
`11
`
`13
`
`16
`
`5:00
`
`6:45
`
`9:50
`
`13:00
`
`8:01
`
`10:11
`
`12:10
`
`15~40
`
`0. 319
`
`0.431
`
`0.628
`
`0.830
`
`0.512
`
`0.650
`
`o. 777
`
`1.000
`
`5:10
`
`6:48
`
`9:04
`
`11:58
`
`7:50
`
`9:15
`
`11:16
`
`14:15
`
`0.821
`
`0.363
`
`0.501
`
`0.636
`
`0.840
`
`0.550
`
`0.649
`
`0.791
`
`1.000
`
`5'-A is 3:20±20 sec for each of the four groups. Such trends may
`be useful in adapting this analytical system to sequencing
`oligonucleotides.
`There is some variation in absolute retention times from day
`to day but retention times relative to tR • 1.000 for ApA are
`reproducible to ±0.020 or better. (±20 sec). Within runs on a
`
`1073
`
`7
`
`

`

`Nucleic Acids Research
`
`single day these relative tR values are usually reproducible to
`±0.005 (±5 sec). Since ApA usually elutes later than anything
`else it is a convenient reference for expressing relative retention.
`Nucleosides and nucleoside monophosphates. Table II shows
`the elution parameters for several monomeric constituents of RNA.
`The general trend of C, U, G, A in order of elution is again
`observed. We have exaudned the nucleosides (N), nucleoside-5 1 -
`phosphates (pN), nucleoside-3'-phosphates (Np), nucleoside-2':
`3'-cyclic phosphates (Np>), and nucleoside-3':5'-cyclic phosphates
`(cpN). We have not tabulated data on nucleoside-5'-diphosphates
`(ppN), triphosphates (pppN) and 2'-monophosphates (Np') but
`elution order follows this scheme: pppN, ppN, pN, Np, Np', Np>,
`N, cpN. We are usually unable to utilize the small difference in
`retention between pppN, ppN, and pN because of their close
`proximity to the void volume of the column; this is especially
`true for the pyrimidine nucleotides. The difference between the
`tR values of N and cpN are also usually small. Retention of
`the pN, Np, and Np' species are quite sensitive to the pH of
`the solvent so care should be taken to reproducibly prepare
`solvents with pH = 5.9.
`Homolgous series. Elution profiles of homooligonucleotide
`mixtures are presented in Figure 2. Panel a shows the separation
`of (Cp)nC oligomers, panel b (Up)n U oligomers, and (Ap)n A
`oligomers in panel c. The separations give excellent resolution
`of oligomers out through at least n = 10.
`In certain instances
`we have turned separation parameters to give resolution of n = 25
`from n = 26 oligomers and see no reason wny the resolution cannot
`be extended to longer oligomers. An oligomer with n • 10, eg.
`(Ap) 1oA, is a polyanion with 10 negative charges; one might sup(cid:173)
`pose that this would be such a highly polar molecule that it would
`not be retarded on a reverse-phase column.
`In fact, however,
`(Ap) 10A is retarded more than (Ap) 9A, so the separation ~;t be
`based upon the number of lipophilic substituents (bases); 4
`there
`is no need to resort to ion-exchange or expensive ion-pairing
`reagents.
`The amount of time required for a particular separation is
`directly related to the complexity of the mixture being analyzed.
`In the case of these homologous series, where up to twenty com-
`
`1074
`
`8
`
`

`

`Nucleic Acids Research
`
`TABLE II
`
`Elution p·arame·t·e·r·s· fo·r· Wucleosides and Nucleoside Monophosphatea.
`The following conditions were used: 16 minute linear gradient,
`from 0 to 90\ B, using solvents a and b (2\ to 11\ CH 3CN) at
`3 ml/min flow rate.
`tf Relative
`tR
`(min: sec)
`to pA • 1.00
`
`z
`llo
`
`pC
`
`pU
`
`pG
`
`pA
`
`Cp
`
`Up
`llo z Gp
`
`Ap
`
`Cp>
`" llo Up>
`z
`Gp>
`
`Ap>
`
`c
`z u
`G
`
`A
`
`cpC
`
`z cpU
`llo
`u
`cpG
`
`cpA
`
`1:28
`
`1:16
`
`1:32
`
`1:39
`
`1:28
`
`1:34
`
`2:04
`
`3:50
`
`1:22
`
`1:37
`
`3:03
`
`5:37
`
`1:55
`
`2:15
`
`4:19
`
`7:24
`
`2:31
`
`2:59
`
`4:07
`
`7:10
`
`0.122
`
`0.106
`
`0.128
`
`0.138
`
`0.122
`
`0.130
`
`0.161
`
`0.296
`
`0.113
`
`0.132
`
`0.249
`
`0.460
`
`0.152
`
`0.178
`
`0.341
`
`0.585
`
`0.206
`
`0.243
`
`0.335
`
`0.581
`
`1075
`
`9
`
`

`

`Nucleic Acids Research
`
`nao
`
`A~
`
`a
`
`2
`
`II
`
`4
`
`!
`
`2
`
`(C~2 C oligomers
`n
`120 sec hydrolysis
`60 min linear gradient
`1% to 4.3% CH3CN
`(solvents a' and b')
`
`.10
`
`0
`0
`
`A254
`
`n•O
`
`~ U oligomers
`n
`60 sec hydrolysis
`.I
`30 min linear gradient
`2% to 1% CH3CN
`(solvents a and b)
`
`0
`0
`A254
`
`03
`
`.. 0
`
`~n A oligomers
`180 sec hydrolysis
`30 min linear gradient02
`7% to 8.5% CH3CN
`(solvents a and b)
`
`Dl
`
`2
`
`II
`
`4
`
`4
`
`!
`
`60
`
`b
`
`c
`
`0
`0
`
`20
`TIME {min)
`
`30
`
`Figure 2. Elution profiles of homologous series. Each series was
`prepared by alkaline hydrolysis of the appropriate homopolymer.
`All elutions at 2 ml/min flow rate.
`
`1076
`
`10
`
`

`

`Nucleic Acids Research
`
`ponents have been separated in a single run (eg. (Cp)nC series),
`we have resorted to separations as long as 60 minutes in order to
`increase the resolving power. of the system, However, in sit-
`uations where only a few components are of interest, as it
`frequently occurs in practice, 10 to 15 minutes is sufficient;
`a 15 minute linear gradient from 1% to 3. 8% CH3CN at 3 ml/min
`flow rate will give baseline separation of C through (Cp) 4c
`in the (Cp)nC series.
`Block Copolymers. A coumon result of an oligonucleotide
`synthesis with primer dependent polynucleotide phosphorylase
`is a "block" addition product with a block of identical bases
`at the 3'-end and a primer molecule of defined sequence at the
`5'-end of the molecule; the primer may be as short as a dinucleo(cid:173)
`tide. Examples of the separation problems encountered are il(cid:173)
`lustrated in Figure 3 with molecules of the type, ApG(pX)n, where
`the primer is ApG. Separation of the ApG(pA)n (Figure 3a) and
`ApG(~U)n (Figure 3b) oligomers proceeds nicely with clear reso(cid:173)
`lution of products with n = 1-3 (the usual range of interest) in
`5-10 min. However, ApG(pC)n (Figure 3c) oligomers do not separate
`following the expected pattern, i.e. in order of increasing
`chain length.
`Instead, the reverse order is observed, in this
`case as well as in a number of other C righ oligomer series, under
`some separation conditions. The polyanion character of these
`molecules then must override the affinity of the lipophilic
`moities for the stationary phase under theae solvent conditions
`and the least charged oligomers are retarded most.
`Only at relatively low CH3CN concentrations do (Cp)nC
`oligomers elute in order of increasing chain length as in
`Figure 2. To obtain these results it is necessary to use solvents
`a' and b' which cover a range of CH3CN concentrations starting at
`1%; the use of solvents a and b (2% and 12% respectively) results
`in separations where the order of elution is reversed, as in
`Figure 3c, or even some in which no simple pattern of elution is
`discernible.
`Throughout all the separations described in this paper, pH
`has remained constant. However, it might prove useful in some
`difficult separations to lower the pH of the solvents to 3
`in order to positively charge adenosine and cytidine residues
`
`1077
`
`11
`
`

`

`Nucleic Acids Research
`
`.10
`
`a
`
`ApG(pA)n series
`10 min linear gradient
`6% to 8% CH3CN
`
`ApG(pU)n series
`10 min linear gradient
`6% to 7% CH3CN
`
`ApG(pC)n series
`10 min linear gradient
`6% to 7'7,. CH3CN
`
`0
`
`5
`
`TIME Cmlnl
`
`10
`
`15
`
`Figure 3. Block copolymers. Solvents a and b were used in
`eaCh at a flow rate of 3 ml/min.
`
`and thus change their elution properties. The use of pH above
`6 is not recommended because silica slowly dissolves in water
`above pH 6.
`The Microbondapak Cl8 column in combination with a single
`type of solvent, namely CH3CN/water solutions, over a very narrow
`concentration range, has proved to be a very versatile tool in
`analyzing mixtures of oligoribonucleotides. These mixtures can
`be the result of the use of biosynthetic enzymes such as poly(cid:173)
`nucleotide phosphorylase, RNA ligase, polynucleotide kinase and
`others in the synthesis of oligoribonucleotides of known sequence,
`for example, or they can be products of the activity of a variety
`
`1078
`
`12
`
`

`

`Nucleic Acids Research
`
`of nucleases on RNA, etc. The ability to analyze reaction mix(cid:173)
`tures in such short periods of time allows the study of the
`time course of these reactions in a quantitatively as well as
`qualitatively accurate manner. This could open the door to re(cid:173)
`action mechanism studies which have been limited by the cumber(cid:173)
`some and time consuming methods of product analysis previously
`available.
`There are indications that the system can be adapted to
`other uses, for example, the analysis of longer nucleic acids,
`which would be of value in the purification and sequence analysis
`of RNA and DNA. A further application should be in the separation
`of alkylated mono- and oligonucleotide products resulting from
`the action of activated carcinogens or mutagens.
`
`ACKNOWLEDGMENTS
`
`This work was supported in part by grants from the National
`Science Foundation (PCM76-04992 AOl), The National Institutes of
`Health (GU 24494), the Petroleum Research Fund (9196-G6), and
`the Research Corporation. We are grateful for helpful discussions
`with D. A. Brant, T. A. Posever 1 R. L. Boughton and applications
`personnel from Waters Corporation.
`
`REFERENCES
`1. Khorana, H. G. and Vizsolyi, J. P. (1961) :!· Am. Chem. Soc.,
`83 675.
`2. Borer, P. N., Kan, L. s. and Ts'o, P.O.P (1975)
`Biochemistry, 1ft• 4847; appendix in microfilm edition contains
`details of synt esrs-irid purification.
`3. Borer, P. N. (1972) Ph.D. Thesis, University of California,
`Berkeley.
`4. Martin, F. H. (1969) Ph.D. Thesis, Harvard University.
`5. Uhlenbeck, 0. C. (1969) Ph.D. Thesis, Harvard University.
`6.
`Leutzinger, E. E., Miller, P. S. and Ts'o, P.O.P. (1978)
`in Nucleic Acid Chemist~Improved and New Synthetic Pro(cid:173)
`cedures, MetliOcls and Tee gues • Townsencr;-L. B. , and Tips on,
`R. S. (eds.), Wiley, (1978).
`7. Walker, G. C., Uhlenbeck, O.C., Bedows, E., and Gumport,
`R. I. (1975) Proc. Nat. Acad. Sci. ~~t~ 72, 122.
`Shum. B. W. , iiiiO"Crother'S'";""D. M:""\1
`Nucleic Acids Res.
`.
`5, 2297.
`9. Hardies, S. C. and Wells, R. D. (1976) Proc. Nat. Acad. Sc1.
`USA, 73 3117.
`Egan.~ z. (1973) Biochim. Biophys. Acta, 299, 245.
`
`8.
`
`10.
`
`1079
`
`13
`
`

`

`Nucleic Acids Research
`
`11.
`
`12.
`
`13.
`
`14.
`
`15.
`
`. , Weeren, n. , Weiss, J. F. , Pearson, R. L. ,
`
`Pearson, R. L., WEiss, J. F., and Kelmers, A. D. (1971)
`Biochim. B·iobhys. Acta, 228 770.
`ttetmers, A.
`Stulberg, M. P., and Novelli, G. D. (1971) Meth. Enzymology,
`XX Part C, 9.
`Pritz, H. J., Belagafe, R., Brown, G. L., Fritz, R. H.,
`Jones, R. A., Lees, R. G., andiChorana, H. G. (1978)
`Biochemistry, 17 1257.
`Klee, c. B. anftinger, M. F. (1967) Biochem. Biophys. Res.
`COnin., 27, 356.
`Some exceptions have been observed, however,. with e-rich
`oligomers (see text under "Block COpolymers ) .
`
`1080
`
`14
`
`