A mm- m sagsmva TECERIQE‘B m
TEE mmmmmax 0? ma
2 gommmcmsm
mm 0% M5 amréoascm

T519929 50:9 fire Dog?“ 0‘ M. S.
,HECEIGAH STATE WWW?
ELe-e M. Pike .

1971

 

LIBRARY

Michigan State
University

 

ABSTRACT

A RAPID AND SENSITIVE TECHNIQUE FOR THE
DETERMINATION OF THE 2'-_(_)_-METHYLNUCLEOSIDE
RATIO OF AN RNA MOLECULE
BY
Lee M. Pike

A method is presented for the determination of the 2'-_Q-methyl-
nucleoside ratio in an RNA molecule. The RNA to be analyzed is
enzymatically digested to the nucleoside level with a combination of
pancreatic ribonuclease, snake venom phosphodie sterase, and
bacterial alkaline phosphatase.

Isolation of the 2'-2-methylnucleo sides is achieved on an
AG .lx4(borate) column. The ribonucleo sides are retained on the
column since a negatively charged borate complex forms across
the cis-Z', 3'-hydroxyl groups. Recovery of the 2'-Q_-methylnucleo-
sides is good but deoxyribonucleoside contamination is possible
if the RNA sample is not free of DNA.

The isolated 2'-Q—methylnucleo sides are trimethylsilylated
with BSA to form volatile derivatives for a gas-liquid chromato-
graphic separation. The separation is achieved on the slightly polar
liquid phase OV-l? under isothermal conditions. Peak areas are
determined by weighing. The derivatives are stable at room temper-
ature for a minimum of three hours and their response to the hydrogen

flame detector of. the GLC is reproducible.

A RAPID AND SENSITIVE TECHNIQUE FOR THE
DETERMINATION OF THE Z'-Q-METHYLNUCLEOSIDE
RATIO OF AN RNA MOLECULE

BY

Lee M. Pike

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Master of Science

Department of Biochemistry

19 71

Dedicated to my wife, Mary and to our son, Lee, Jr.

ii

I thank Dr. Fritz Rottman for his guidance in my re search
and graduate training. My thanks also go to Joseph Abbate, Brian
Dunlap, and the other members of the Rottman group for their '
assistance and beneficial discussions in the course of this work.

My thanks are extended to Dr. William Wells for his assist-
ance with gas chromatographic techniques and to Roger Lane and
Ray Hammond for their help in the interpretation of mass spectra.
I also thank Dr. Allan Morris who served on my guidance committee.

I especially thank my wife for her love, faith, and encourage-
ment during my graduate training. She is also thanked for the
typing of this manuscript.

The financial assistance from the .National Institute of Health
and the Department of Biochemistry, Michigan State University

is appreciated.

iii

TABLE OF CONTENTS

Page
LIST OF TABLES......................................vi
LIST OF FIGURES......................................Vii
LIST OFABBREVIATIONS..............................viii

HTEMTURE REVIEW...0..O.COCOOOOOOOOOOOOOOOOOOOO0.1

Distribution of 2'-9_-methylnucleosides . . . . . . . . . . . . . . 1
Correlation of 2'-2-methylnucleo sides with rRNA

processing..................................2
Correlation of Z'-9_-methylnucleosides with

nuclear RNA metabolism.....................3
Quantitation of 2'- 2-methylnucleo side content

ofan RNA molecule..........................4
GLC ofnucleosides................................7

INTRODUCI‘IONOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOIO
MATERIALSOOOOOOO0.0.0.000...OOOOOOOOOOOOOOOO0......11

METHODS...‘.0...OOOOOO-OOOOOOOOOOOOOOOOOOOOOOOOOOOOO 13

Purification of Russell's viper venom phospho-
diesterase..................................l3
Hydrolysis ofRNA................................13
Isolation of 2'-0-methylnucleosides.................l3
Nucleoside derivatization..........................14
GLC analysis.....................................15
Mass spectral analysis............................16

EXPERIMENTALOOOOOOOOOOO00.0.0.0...0.0.0.000000000017

1. RNA hydrolysis and isolation of 2'-Q-methyl-
nuCIeosideBOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO00.17

Enzymatic RNA hYdrOIYSi-sooooooooooo0000000000000017
Behavior of nucleo sides on AG 1x4 (borate) column. . . 18

iv

Page

Isolation and characterization of Z'-_Q-methyl-
nucleo side fraction wheat germ rRNA. . . . . . . . . . 22
Pyridine and water hydrolysis of RNA. . . . . . . . . . . . . . . 25

II. Derivatization and GLC of nucleo side 3. . . . . . . . . . . . . . 30

BSA derivatives.................................. .30
GLC behavior and retention times of
nucleoside derivatives. . . . . . . . . . . . . . . '. . . 31
Stability and reproducibility of derivatives. . . . . .33
Nucleoside ratio determination of synthetic
copolymers.................................39
Alternative derivatization techniques. . . . . . . . . . . . . . . .42-
BSTFA derivatives of nucleo side 8. . . . . . . . . . . . . 42
Mass spectral analysis of A and Am
derivatives formed with BST FA. . . . . . . . . .46
Effect of methoxyamine-hydrochloride pretreat-
ment on the nucleoside derivatives formed
by BSA................................57
TSIM derivatives of nucleosides. . . . . . . . . . . . . . .58
Effect of methanolic acetic anhydride pretreat-
ment on the nucleoside derivative formed
by BSTFA.............................6l

DISCUSSIONO.OOOOOOOOOOOOOOOOOOO...OOOOOOOOOOO00......61

REFERENCESOO0......OOOOOOOOOOOOOOOOOOOOOOOOO0......73

LIST OF TAB LES

Table Page

1. The recovery of 2'-_Q_-methylnucleosides from
the AG 1X4 (borate) celumIIOOOOOOCOOCOOOCOO0.0.0.021

2. The retention of ribonucleosides on the AG
1x4 (borate) COlmnnOOOCOOOOOOOOOOOOOOOSOOOOOOOOOOOZ3

3. The isolation of 2'-_9_-methylnucleo sides from
wheat germ rRNA...............................24

4. The progress of destruction of adeno sine by
incubation with pyridine and water. . . . . . . . . . . . . . . . 50

5. The retention times of the BSA derivatives of the
2'-9_-methylnucleosides..........................55

6. The stability and reproducibility of the TMS
derivatives of the 2'-_Q-methylnucleo side 8. . . . . . . . . .58

7. The relative molar response values for the 2'-Q_-
methylnuCIeOSideSOOOOOOOCOOOOOOOOOOOOOOOOOCOOOOOhl

8. The relative molar response values for the '
nucleosides in synthetic copolymers. . . . . . . . . . . . . . . ’42

9. The nucleo side ratios of synthetic copolymers. . . . . . Hi

10. The retention times of the BSTFA derivatives
of the ribonucleo sides and the 2'-(_)_-methylnucleo-

BideSOOOOO0....OOOOOOOOOOOOOOOOOOOOOOO00.0.00...1+6

11. The effect of pretreatment with methoxyamine-
hydrochloride on the retention times of the BSA
derivatives of the 2'-_9_-methy1nucleo side a. . . . . . . . . . 61

Figure

1.

2.

9.

LIST OF FIGURES

Page

An elution profile of the 2'-_9_-methylnucleo-
sides from the AG 1x4 (borate) column. . . . . . . . . . . . . 20

A gas-liquid chromatogram of the TMS
derivatives of the 2'-9_-methylnucleo sides
on OV-l7 COlumnOOOOOOOOOOOOOOOOOOOO00.0.000000055

The mass spectrum of TMS4-adenosine. . . . . . . . . . . . 1+9
The mass spectrum of TMSS-2'-_(_)_-methy1adeno sine. 51
The structure of TMS4-adeno sine. . . . . . . . . . . . . . . . . 53
The structure of TMS3-2'-9_-methyladeno sine. . . . . . . 53

The structure of the TMS -methoxime deri-
vative 0f Z'-2-methchyti<fine. o o o o o o o o o o o o o o o o o o o 0 60

The mass spectrum of TMSz-5'-acetyl-
2'-_0_-methY13den08i.neooooooooooooooeooooooooooooo 61"

The structure of TMS -5'-acetyl-Z'-2-
memYIadeDOSi-nee0.0.2000000000000000000000000000067

vii

A260

BSA
BST FA
C

Cm
DEAE-cellulo se
eV

G

Gm
GLC
HMDS
mRNA
OV-l
OV-17
RNA ’

rRNA

LIST OF AB BREVIATIONS

Adeno sine

2' -_Q-methyladeno sine

Absorbance at 260 nanometers

N, O-bis (trimethylsilyl) acetami de
N, O-bis (-trimethylsilyl, trifluoroacetamide‘ '6'
Cytidine - ‘

Z ' -_Q-methylcytidine
Diethylaminoethylcellulo se
Electron volt

Guano sine

2 ' -Q-methylguano sine

Ga 8- liquid chromato g raphy
Hexamethyldi sili zane
Messenger RNA

Methyl silicon gum
Methyl-phenyl silicon gum
Ribonucleic acid

Ribo somal RNA

viii

sRNA

TMCS

TMS

tRNA

T SIM

Um

RNA soluble in 2M NaCl
Trimethylchlorosilane
Trimethylsilyl
Transfer RNA
Trimethylsilylimidazole
Uridine

2 ' ~2-methyluridine

LIT ERAT URE REVIEW

Ribonucleic acid is a polymer containing principally the purine
and pyrimidine nucleosides adenosine, guanosine, cytidine, and
uridine. Methylated modifications of these four nucleosides occurring
naturally are known as ”trace" or "minor" nucleosides, since they
represent only a small fraction of the total nucleo sides present.
Thirteen different ring-methylated nucleosides and five nucleo sides
methylated on the 2'-hydroxyl group of the ribose moiety have been
identified (Starr and Sells, 1969). The ribose methylated nucleo-
sides are of interest to us with respect to the influence of a blocked
2'-hydroxy1 group on the structure, function, and metabolism of
RNA molecules (Bobst et al, 1969; Price and Rottman, 1970; Dunlap
et al, 1971).

. Distribution of 2 ' -_Q- methylnucleo side 5

 

The occurrence of 2'-_(l-methylnucleosides in RNA was first
demonstrated by Smith and Dunn (1959). They were isolated in low
yield from the alkaline stable oligonucleotides in RNA from such
varied sources as wheat germ, tobacco and beet leaves, and rat
liver. Biswas and Meyers (1960) reported the presence of Cm in
a preparation of cytidine obtained from acid hydrolysis of RNA in
I the blue-green algae Anacv stis nidulans. Bacterial, yeast, and
several mammalian RNA's were shown by Hall (1964) to contain -

significant levels of the 2'-_Q_-methylnucleosides.

l

Lane and his co-workers have developed techniques (described
below) for detecting the presence and distribution of the 2'-_Q-Ine thyl-
nulceosides. Their studies have included characterization of the
2'-9_-methylation patterns of RNA from wheat germ (Singh and Lane,

1964a), yeast (Gray and Lane, 1967), and Esherishia coli (Lane and

 

Tamaoki, 1969).

The 2'-_9_-methylnucleosides are ubiquitous; they have been 5-
found in all organisms assayed to date, and in nearly all classes of I
RNA. The exceptions are viral and bacterial mRNA (Moore, 1966),
and SS rRNA from bacteria (Brownlee et a1, 1967) and mammals
(Forget and Weissman, 196 8). Eukaryotic mRNA has not been
analyzed for the presence of 2'-.(_)_-methylnucleosides.i

Correlation of 2'-__Q_-methLlnucleo sides with rRNA processing

Studies on high molecular weight rRNA isolated from eukaryotic
cells indicate that the major site of methylation in these molecules
is the 2'-hydroxyl group of ribose (Wagner et a1, 1967; Iwanami and
Brown, 1968; Lane and Tamaoki, 1969). The kinetic studies of
Greenberg and Penman (1966) and Zimmerman and Holler (1967)
showed that preribosomal 45S RNA is methylated as it is synthesized
in the nucleolus. This methylation has been reported to be a pre-
requisite for maturation and processing of 45S RNA into 185 and
283 RNA (Vaughan et a1, 1967). In HeLa cells, the processing of the

455 RNA into 188 and 28S RNA has been demonstrated to involve

no significant loss of Z'-_Q-methyl groups (Wagner et al, 196 7;

Weinberg et al, 1967). Conservation of 2'-Q-methy1 groups is also
high in the maturation process of 455 RNA in Novikoff hepatoma
ascites cells with a loss of only 15-17% (Choi and Busch, 1970).
These re sults are consistent with a processing mechanism for 45S
RNA involving specific loss of unmethylated or undermethylated
regions of the molecule.

Correlation of Z'-.Q-methylnucleo sides with nuclear RNA metabolism

 

Another group of RNA molecules displays a relationship between
its 2'-_(_)_-methylation patterns and metabolism procedures. The
nucleus contains a group of six to nine RNA molecules of discrete
sizes ranging from 4-75 which were reported independently in rat
liver by Dingman and Peacock (1968) and in HeLa cells by Weinberg
and Penman (1968). They were.) subsequently shown to be present
in cells of several organisms and many vertebrate tissue culture
lines (Rein and Penman, 1969). Many of these RNA molecules were
shown to be highly methylated since they were heavily labeled in
14C--methylmethionine labeling experiments (Weinberg and Penman,
1968; Zapisek et al, 1969; Moriyama et al, 1969).

The distribution of these methyl groups has been examined
in the nuclear RNA from two sources. Three fractions of Chinese
hamster cell nuclear RNA have been shown by Zapisek et al, (1969)
to have a large fraction of their methyl label in 2'-_Q_-methyl groups.

Ro-Choi et al (19 70) have isolated the nuclear 4. SS RNAm from

Novikoff hepatoma ascites cells and shown that five of the eighty-

five nucleotides it contains are Z'-_g-methylated, including the
3‘-terminal nucleoside (El-Khatib et a1, 1970).

Metabolically these nuclear RNA molecules are very stable
(Weinberg and Penman, 1969; Moriyama et al, 1969). Several of
the molecules are as stable as rRNA, while others have half lives
of one day or more. This stability distinguishes these small mole-
cular weight nuclear RNA molecules from the bulk of nuclear RNA
which has a very rapid turnover rate (Soeiro et a1, 1968; Aronson
and Wilt, 1969). Whether or not there exists some relationship
between metabolic stability and the degree of 2'41 -methylation is
unknown, but Novikoff hepatoma ascites cell nuclear 4. SS RNAl,
one of the least stable of these nuclear RNA molecules (Ro-Choi,
1971), has been shown to contain no Z'-_Q-methyl groups among its
102 nucleotides (Ro-Choi et al, 1971).

Quantitation of 2'-.Q-methylnucleo side content of an RNA molecule

 

Two techniques arose independently for the early quantitative
analysis of the 2'-_Q—methylnucleoside levels in RNA. Singh and
Lane ( 1964 a and b) utilized the pr0perty of alkaline stability of
oligonucleotides containing blocked 2'-hydroxyls. The stable
oligonucleotides were separated from the mononucleotides by salt
fractionation on a DEAE-cellulose column. Two dimensional paper
’ chromatographic separation of the sixteen dinucleo side monopho sphates
allowed estimation of the abundance of 2'-9_-methylnucleosides based

on the percentage of the initial ultraviolet absorbing material recovered

from each spot on the chromatogram. Alternately, the di- and tri-
nucleotide fraction could be enzymatically digested to the nucleoside
level. Paper chromatography in several steps could then be used
to separate the eight resulting nucleosides; they could be quantitated
by recovery of ultraviolet absorbing material as described above.

If an attempt is made to increase the sensitivity of the above
assay by labeling with l4C-methyl groups, an exaggerated estimate
of sugar methylation may be obtained. RNA may contain other
dinucleotide sequences which display alkaline stable properties
although they are methylated only on their bases (eg. N6,N6-
diniethyladenylyl-(3', 5' )-N6, N6-di1nethyladeno sine-2'(3')-pho sphate;
Nichols and Lane, 1966a). There is the additional problem of coincident
base methylation of some of the 2'-_(_)_-methylnucleo sides (eg. N4-
methyl-2'-_Q-methyl- cytidylyl-(3', 5')-cytidine; Nichols and Lane,
19666).

In the second technique, Hall (1964) made use of the inability
of vii-methylnucleo sides to form complexes with borate; ribonucleo-
sides will form such complexes across the (cis)2' ,3 '-hydroxyl groups.
RNA was hydrolyzed to its constituent nucleic sides enzymatically and
the nucleosides were fractionated on a partition column of diatomaceous
earth with a solvent system of l-butanol, ammonium hydroxide and
sodium borate. Paper chromatographic separation of the Z'-p_-

methylnucleo sides allowed estimation of their abundance based on the

percentage of the intial ultraviolet absorbing material recovered

for each Z'-_Q-methylnuc1eo side.

These two techniques have been the only available methods
for measuring the amount of each 2'-_Q-methylnucleo side in an RNA
molecule. The enrichment or isolation step is necessary since the
2'-_Q-methy1nucleosides represent a small fraction of the total nucleo-
sides in RNA and would otherwise be masked. However, the paper
chromatographic separations used in both of the above techniques
are time consuming (approximately one day for each dimension or
system used). In addition, they require sufficient quantities of the
nucleosides for complete resolution and exact measurement of
ultraviolet absorbance after elution from the chromatogram (ca
20 nmole or O. 25 A260 of each nucleoside per determination).

Abbate and Rottman (19 71) recently developed a more rapid and
sensitive technique for quantitation of the total percentage of nucleo-
sides which are Z'-_Q-methylated. It is based on a report (Baskin
and Dekker, 1967) that perchloric acid hydrolysis of 14‘C-—methyl
labeled RNA released only sugar g-methyls as 14C--methanol; labeled

base methyls were not released. The 14

C-methanol was isolated by
repeated co-distillation with 50% methanol to determine the amount
of Z'-2—methylation.

The method of Abbate and Rottman (1971) has several advantages.
The perch loric acid hydrolyzate is directly analyzed for the released
methanol by gas chromatography, reducing the risk of loss by handling.

Since the methanol originating from the 2'-_Q_-methy1 group is not

diluted with added methanol, the 14C-label is not necessary. Virtually
any RNA sample or mixture of nucleosides can be analyzed. However,
the data obtained reflects only the total abundance of sugar hydroxy-
methylation, regardless of which nucleosides are methylated.

GLC of nucleosides

 

The use of gas-liquid chromatography (GLC) for analysis of
nucleic acids is a relatively new technique. Base ratios can be
determined using a mixture of bases obtained by perchloric acid
hydrolysis of RNA (Hashizurne and Sasaki, 1968), but quantitation
of 2'- 2-methy1nucleo sides in a RNA molecule would require analysis
of a mixture of intact nucleo side 5. Since nucleosides have a low ’
volatilities, it is necessary to render them less polar and more
volatile by blocking the amino or hydroxyl groups on the base in
order to elute them from a GLC column at temperatures which will
not result in their decomposition.

Miles and Fales (1962) used'various combinations of the acetyl,
methyl, or isopropylidene derivative for this purpose, but they
experienced difficulties with multiple derivatives and trailing of
cytidine and guanosine. Hancock and Coleman (1965) used trimethyl-
silyl (TMS)derivatives with more success. Subsequent evaluation of
the TMS and other silyl derivatives of the nucleo sides by Sasaki and
.Hashizurne (1966), Gehrke et a1 (1967), and Hancock (1968) has led A
to the conclusion that the TMS derivatives are most suitable since

the :problem of multiple derivatives is minimal.

Hancock and Coleman (1965) and Sasaki and Ha shizume (1966)
used the procedure developed by Sweeley et a1 (1963) for silylation
of sugars to form TMS derivatives of nucleosides. The dry standard
nucleosides were dissolved in a mixture of pyridine, hexamethyl-
disilizane (HMDS) and trixnethylchlorosilane (TMCS); silylation
time ranged from 5 to 120 minutes with or without heating to enhance
the reaction. The resulting mixture of derivatives was injected
onto a GLC column consisting of a diatomaceous earth support coated
with a liquid phase of l to 5% methyl silicone guzm (OV-l or S E-30).
Peaks were obtained for all four nucleosides, but they were asym-
metrical with broad shoulders leading adeno sine and guanosine and
all four trailed badly.

Gehrke et al (1967) achieved a measure of success when they
added N, O-bis (trimethylsilyl) acetamide (BSA) to some nucleosides
dissolved in acetonitril for silylation; they did not include cytidine,
which has been difficult to derivatize. Jacobson et a1 (1968) used
only BSA for a very successful derivatization of both standard nucleo-
side mixtures and nucleo sides obtained by enzymatic hydrolysis of
RNA. BSA was added directly to the ly0philized incubation mixture
and heated at 120° for 120 minutes. The TMS derivatives were
resolved on a GLC column packed with a methyl-phenyl silicone
gum (4% OV-l7 on Gas Chrom Q) with a programmed temperature
increase of two degrees per minute ~from 1600 to 240°. Single

peaks were obtained for all derivatives with the exception of TMS-

guano sine which had two peaks. Theirs was the first successful
attempt at a gas chromatographic quantitation of the nucleoside

ratios in RNA.

INTRODUCTION

The purpose of this research was the deve10pment of a more
rapid and sensitive assay for quantitation of 2'-_Q-methy1nucleoside
ratios than the elaborate and involved paper chromatographic systems
currently in use. Coupled with knowledge of the total abundance of
2'-_0_-methylnucleo sides in an RNA molecule obtained by the method
of Abbate and Rottman (1971), the 2'-_Q-methylnucleoside ratio would
permit calculation of the abundance of each 2'-0-methy1nucleoside
in that RNA molecule. The two techniques would thus fingerprint the
2'-_Q—methylation patterns in RNA molecules. These fingerprints
could be used to compare a variety of RNA molecules for inherent

differences or changes in the 2'-2-m¢hy1ation pattern.

10

MATERIALS

Adenosine, guanosine, cytidine, and uridine were purchased
from Sigma Chemical Company. The 2'-_Q-methyladenosine (Am)
was prepared from adenosine treated with diazomethane in an aqueous
solution of 1, Z-dirnethoxymethane by the method of Broom and Robins
(1965) as modified by Rottman and Heinlein (1968). The 2'-_Q-methy1-
cytidine (Cm) was prepared by a similar chemical methylation of
cytidine (Rottman and Johnson, 1969). The 2'-2-methyluridine (Um)
was prepared by dearnination of Cm with potassium nitrite in acetic
acid as described by Martin et al (1968). The alkaline stable di-
nucleosides monopho sphates and 2'-2—methy1guanosine (Gm) were
isolated from wheat germ rRNA using the techniques of Singh and
Lane (1964b). Commercial wheat germ was purchased locally.

Snake venom pho sphodie sterases (Russell's viper and Crotalus

adamenteus) and bacterial alkaline phosphatase (BAPSF) were

 

purchased from Worthington Biochemical Corporation. Bovine
pancreatic ribonuclease A was purchased from Sigma Chemical
Company. Bio-Rad AG 1x4 (Cl'), 100/200 me sh and Dowex-SO,
W-X8, 100/200 me sh were purchased from Bio-Rad Laboratories. '
N, O-Bis (trimethylsilyl) acetamide (BSA), N, O-bis (trimethylsilyl)
trifluoracetamide (BSTFA) and trimethylsilylimadazole (TSIM)
were purchased from Regis Chemical Company. The 2% OV-l7
(methyl, phenyl silicon gum) on 80/100 mesh Gas Chrom Z was

11

 

12

the generous gift of W. W. Wells. The 1% OV-l (methyl silicon gum)
on 100/200 me sh Gas Chrom Z was purchased from Applied Science

Laboratories. All other chemicals were reagent grade.

METHODS

Purification of Russell's viper venom pho sphodiesterase.

 

Russell's viper venom pho sphodiesterase had been purified
to remove contaminating phosphatase activity. ‘ This was achieved on
a cation exchange re sin (Dowex-SO, W-X8, 100/200 mesh) column
washed with sodium hydroxide and equilibrated with sodium acetate

(Keller, 1964).
Hydrolysis of RNA

 

Digestion of RNA was effected with the following reaction
mixture per milligram of RNA: alkaline phosphatase, 15 ug; ribo-
nuclease A, 20 ug; purified Russell's viper venom pho sphodie sterase,

45 ug of crude Crotalus adamenteus venom pho sphodie sterase, 45

 

 

ug; in a total volume of 0. 3 ml which was 0. 1M in ammonium carbonate
buffer, pH 8. 8 and 2mM in magnesium acetate. A 24 hour digestion

at 37°C achieved hydrolysis to the nucleoside level. Completeness

of the reaction was checked by electrophoresis on Whatman #1 for

90 minutes at 400v in 0. 05M Tris-acetate, pH 7. 5.

Incubation of a pyridine and water mixture ( 1:1, vzv) with RNA,
nucleotides, or adeno sine.was done in sealed glass tubes. The tubes
were heated for various times a 108° and the contents were analyzed
by paper chromatography after removal of pyridine.

Isolation of 2'-_Q_-methy1nucleo side 3
Bio-Rad AGlx4 (Cl‘) was converted to the borate form by

washing with 1. ON sodium hydroxide until the silver nitrate test

13

14

was negative for Cl', washing with H20, and finally washing with
0. 7M boric acid. An 8. 0 x 1. 6 cm coluInn was used for an incubation
mixture of 10 mg of RNA. Prior to use the column was washed with
500 ml 0. 7M boric acid and finally equilibrated with 5.00 ml 0. 3 M
ammonium borate, pH 6. 0.

The pH of the nucleo side sample was adjusted to 6 with 1. ON
HCl prior to addition to the column. The 2'-_9_-methy1nuc1eo side
fraction was eluted in 450 ml 0. 3M ammonium borate, pH 6. 0 at a
flow rate of 1. 0 ml per minute. The sample was lyophilized and the
boric acid was removed as the methyl ester by repeated evaporation
from methanol (Balle et al, 1966).

Nucleoside derivatization

 

Derivatizations of the nucleosides for GLC were performed
i115 dram vials sealed with a teflon-lined screw cap. Trimethyl—
silylation was done with BSA, BSTFA, or TSIM. BSA was added in
a 100 mole excess and the solution was heated for 120 minutes at
125° (Jacobson et al, 1968). BSTFA was added in 100 mole excess
and the solution was heated for 90 minutes at 90°. TSIM was added
in 100 mole excess and the solution remained at room tempe rature
for 1-2 hours before injection.

Occasionally, the nucleo sides were treated with methyoxyarnine-
hydrochloride or methanolic acetic anhydride as described below
prior to silylation. iMethoxyamine-hydrochloride was added in 40

mole excess with dry pyridine as a solvent. The solution was heated

 

15

for 3 hours at 75°, evaporated by a stream of dry nitrogen, and the
nucleosides were silylated with BSA as above. Acetylation was done
in a solution of 0.1 m1 acetic anhydride and 0. 4 m1 absolute methanol.
The solution was treated for 2 hours at 800 or remained at room
temperature from 5 to 96 hours. Following evaporation of the solvent
by a stream of dry nitrogen, the nucleosides were silylated with
BSTFA as above.

GLC analy sis

 

GLC analyses were conducted on a Hewlett-Packard FandM
402 'High Efficiency Gas Chromatograph equipped with a flame ionization
detector. The instrument was connected to a 1 mV Sargent Model
SRG recorder. The colurrm consisted of U- shaped glass tubing
(6 ft x in i. d.) packed with 1% OV-l on silanized Gas Chrom Q or
2% OV-17 on silanized Gas Chrom Z. The columns were conditioned
prior to use as follows: a) nitrogen flow for 5 minutes at room
temperature; b) heated to 3500 for 1 hour with no carrier flow;
c) cooled to room temperature; and d) column heated to 250° for
24-48 hours with flowing nitrogen. The oven temperature was kept
40-50° higher than normal operating temperature when the column
was not in use.

Chromatographic conditions for OV-l? columns were: isothermal
column temperature of l95--205O C; injection heater, 200°; detector
heater, 210°; and carrier gas, nitrogen, 45 cc per minute. Conditions

for OV-l columns were: isothermal column temperature of 160-

180°; injection heater, 170°; detector heater, 180°; and carrier
gas, nitrogen, 45 cc per minute.

A 1-7 ul aliquot containing 0. 03-1. 0 ug (0. 12-3. 8 nmole) of
each nucleo side was injected into the gas chromatograph. Peaks
were compared by cutting out the peaks, weighing them, and correct-
ing for differences in attenuation and chart speed of the recorder.

Mass Spectral Analysis

 

Low resolution mass spectra were recorded on an LKB-9000
combined gas-liquid Chromatograph-mass spectrometer using an
ionizing energy of 70 eV. The ion source was a 290° and the molecular
separator was at 230°. The trap current was 60 uamps. The GLC
column consisted of coiled glass tubing (4 ft x 3 m i. d. ) packed
with 3% SE-30 on silanized Supelcoport ( 100 120 me sh). Helium
carrier gas was maintained at 35 cc periminute. The column had a
linear program rate of 2° per minute from an initial temperature
of 220° to a final temperature of 250°. The data was processed by

an on-line digital computer system (Sweeley et al, 1970).

EXPERIMENTAL

1. RNA hydrolysis and isolation of 2'-_.Q-methylnuc1eo side 5.

 

Enzymatic RNA hydrolysis

 

Enzymatic digestions were attempted on several types of
RNA by the procedures described in the methods section. Pancreatic
ribonuclease was added to endonucleolytically cleave the RNA into
short pieces to create more sites for the exonucleolytic action of
snake venom pho sphodie sterase. Bacterial alkaline phosphatase
was added to hydrolyze the phosphate from the resulting nucleotides
producing nucleosides for subsequent analysis. The digestions were
done at alkaline pH to create optimal conditions for the alkaline
phosphatase and the snake venom pho sphodie sterase.

The products of the enzymatic digestions were assayed by a
90 minute electrophoresis at pH 7. 5 in 0. 05M Tris-acetate buffer
at 400V. Standards used to check for completeness of reaction
were: a) a mixture of the four normal ribonucleosides; b) a mixture
of the four normal 5'-ribonucleotide s; and c) a mixture of wheat
germ alkaline stable dinucleo side monopho sphates. The uncharged
nucleo sides remain at the origin, while 5'-mononudeotides and
dinucleoside monophosphates migrate towards the positive pole.

The reaction was judged incomplete if ultraviolet absorbing
material was present in any region of the electrophoreogram other
than the nucleoside region; a 3% contamination was detectable.

17

18

By this criterion, Crotalus adamenteus venom pho sphodiesterase

 

and purified Russell's viper venom pho sphodie sterase were equally
successful in accomplishing the digestion. However, Russell's viper
venom phosphodie sterase may be preferred since it has been purified
to eliminate nucleo side deaminase activity.

Behavior of nucleosides on AG 1x4 (borate) column

 

Isolation of the 2'-9_—methylnucleo side fraction is necessary
to avoid its being masked by the ribonucleosides as discussed in the
Literature Review. The technique developed in this research utilizes
the formation of a borate-ribonucleo side complex, but separation
is effected on a simple AG 1x4 (borate) column rather than the elaborate
partition column described by Hall (1964). The separation depends
on a borate complex forming with the ribonucleosides and their
consequent retention on the resin. The 2'-9_-methylnuc1eo sides
pass through the colunm with the purines being held up slightly on
the basis of a partial charge or polarity of the base. Furukawa et
a1 (1965) used a similar Dowex-l column to isolate 2'-0-methy1ated
intermediates in a chemical synthesis of 2'-_(_)_-methyluridine. There
has not been a report of the use of this column with RNA hydrolyzate s.
Deoxyribonucleosides represent a potential source of contamination
in this technique. Since they lack the 2'-hydroxy1, no borate complex
is formed and they Chromatograph similarly to the 2'-_0-methylnuc1eo-

sides on the AG 1x4 (borate) column. All RNA molecules investigated

must therefore be ascertained to be free of DNA or be treated extensively

19

with deoxyribonuclease during their purification.

Recovery of standard 2'-_Q_-methy1nucleo sides from the AG 1x4
(borate) column was determined to insure that no loss occurred
during this isolation. In addition, the ability of the column to retain
ribonucleosides was checked by applying a mixture of U, A, C, and
G (in an undetermined ratio) to the column and monitering the effluent
for the appearance of material which absorbs at 260 nm.

A profile of the first 200 m1 of the elution of a mixture of Um,
Am, and Cm from the column as described in Methods is shown in
Figure l. The fir st peak contains Um and Cm and their elution is
complete in 200 ml. The broad second peak contains Am, but its
recovery is not complete until about 400 ml have passed through
the column. Elution of Gm would begin under the Am peak and
would continue through another 50 ml volume after the Am elution
is complete.

The recovery of each 2'-9_-methylnuc1eoside was tested in
duplicate and the results are reported in Table 1. Amounts were
determined by absorbance at 260 nm at pH 6 (neutral). The results
were good for Um, Am, and Cm with recovery of over 95% of the
approximately 20 A260 units used. Further experiments were
limited by the minimal amounts of Gm available. A slight error
in determining the absorbance in this low range would have a larger
effect than in the previous cases, however. Larger amounts of

Gm will be available in the future since it is being chemically

20

 

0.50-

O. 35"”

 

 

 

 

 

 

Volume (m1)

Figure 1. An elution profile of the 2'-2-methy1nucleosides
from an AG 1x4 (borate) column.

A mixture of Um (0. 97mnole), Am (0. 50 umole), and Cm
(l. 61 umole) were applied in 0. 5 m1 H20 at pH 6 to an AG 1x4
(borate) column which has been prepared as described in Methods.
Elution was effected with 400 m1 0. 3M ammonium borate, pH 6. 0;
a profile of the fir st 200 ml of that elution is shown above. Peak
I contains Um and cm; peak II contains Am.

Nucleo side Amount Amount Pe rcent

 

Applied Recovered Recovered
(umole) (umole)
Um-Z 2. 02 2. 04 101
Am-l 1. 31 l. 26 96. 2
Am-z 1. 31 1. 29 98. 4
Cm-l Z. 85 2. 78 97. 5
Cm-Z 3. 12 3. 02 96. 7

Table 1. The recovery of 2'-0-methylnucleo sides from the AG
1x4 (borate) column.

Two samples of each 2'-0-methy1nuc1eosides were applied to
separate AG 1x4 (borate) columns in 1. 0 ml H20 at pH 6, followed
by a rinse of 0. 5 ml H O. The nucleosides were eluted with 0.3M
ammonium borate, pH 6. 0, at a flow rate of 1 ml per minute.
Fractions were analyzed with a Gilford UV Spectrophotometer for
the recovery of 2'-0-methylnuc1eoside 8. All recoveries were
complex within the first 450 m1 of column effluent as discussed in
the text.

21

 

22

synthesized in our laboratory. If the recovery cannot be increased.
data will be obtained to calculate a recovery factor to normalize
the Gm content of the column effluent of an RNA hydrolyzate prior
to further analysis.

The retention of a mixture of ribonucleosides was tested in
triplicate under the same conditions by which the 2'-_Q_-methy1nucleo-
sides were completely eluted. As shown in Table 2, 210 A260
units of an approximately equimolar solution of ribonucleosides
were still retained after 500 m1 of the buffer had passed through the
column. The capacity of the column was seen to be closely related
to its size since the application of a larger amount of ribonucleo-

. sides resulted in an earlier elution.

Isolation and characterization of 2'-Q_;methjlnuc1eoside fraction
from wheat germ rRNA

 

The 2'-0-methy1nucleosides were recovered from the Crotalus

adamenteus venom pho sphodie sterase digestion of wheat germ rRNA

 

using the AG 1x4 (borate) column. The digestion mixture was separated
into three portions of approximately 200 A260 units each. The
fractions were applied to separate columns and the 2'-g-methy1-
nucleosides were eluted under the previously described conditions

to effect their separation from ribonucleosides. The amount recovered
_ is reported in Table 3, both as A260 units and as percentage of

260 nm absorbing material applied to the column. There is some

variation, but the amount recovered is in the range of the l. 7 percent

Nucleo side Amount Amount Per cent Re cove ry

 

Applied Recovered Recovered Volume

(A260) (A260) (ml)
UACG-l 210 0 0 480
UACG-Z 250 0 0 420
UACG- 3 210 0 0 510

Table 2. The retention of ribonucleo sides on the AG 1x4 (borate)
column. .

Three aliquots of an approximately equimolar mixture of U,
A, C, and G were applied to separate AG 1x4 (borate) columns in
1. 0 ml H O at pH 6. The column was washed with 0. 3M ammonium
borate, p 6. 0, at a flow rate of 1 ml per minute. Fractions were
analyzed with a Gilford UV Spectrophotometer for the appearance
of eluted material. The indicated recovery volumes were collected
prior to the elution of any detectable ribonucleo side.

23

Crotalus adamenteus Amount Amount Per cent

 

 

Venom Biosphodiesterase 'Applied Recovered Recovered
Digestion (A260) (A260)
1 210 5. 2 2. 5
2 210 3. 5 1. 7
3 200 5. 5 2. 8

Table 3. The isolation of 2'-_0_-methylnucleo sides from wheat
germ rRNA.

A single incubation mixture containing 28 mg of wheat germ
rRNA was digested with Crotalus adamenteus venom pho sphodiesterase
as described in Methods. Corripletbness of the reaction was shown
by electrophoresis as discussed above. The mixture was adjusted
to pH 6 by addition of 1. ON HCl and the A260 was recorded. The
mixture was divided into three portions and each portion was applied
to a separate AG 1x4 (borate) column. The 2'-2—methylnucleoside
fraction was eluted with 0. 3M ammonium borate, pH 6. 0. Recoveries
are given as amounts of abosrbing material recovered and as percent
of the amount applied.

24

 

value reported for the 2'-_Q-methy1nucleo side content of wheat germ
ribo somal RNA (Abbate and Rottman, 1971).

The 2'-_Q-methy1nucleoside fraction was eluted from the
AG 1x4 (borate) column in 450 ml 0. 3 M ammonium borate; this
is equivalent to 8. 3 grams of boric acid. Removal of the borate
was accomplished by lyophilizing the column effluent and evaporating
several times from methanol. Ammonia is volatile and the boric
acid was removed as the methyl ester (Balle et a1, 1966).

Following removal of the boric acid by evaporation from
methanol as the methyl ester, the 2'-9_-methy1nucleoside fraction
designated 1 was checked by paper chromatography in 4-1-2
solvent (ethyl acetate, n-propanol, water; 4:1:2; v:v:v; upper
phase). .This solvent re solves normal nucleosides into two
spots-eone containing U and A and the other containing C and G.
The 2'-9_-methylnucleosides and deoxyribonucleosides are simi-
larly resolved. Since the six spots are cleanly separated, con-
tamination of the 2'-_(_)_-methy1nuc1eo side fraction with ribonucleo sides
or deoxyribonucleo sides would be indicated by the presence of these
spots on the chromatogram.

After spotting the sample and appropriate standards, the
chromatogram was developed for six hours. There was some diffi-
culty with salt, but the chromatogram showed that while there was

no ribonucleoside contamination, small amounts of deoxyribonucleo-

 

26

sides were present. These results indicate that while there is
no contamination with ribonucleo side 5, extra care must be taken
to remove DNA during the RNA preparations to avoid contamination

with deoxyribonucleo side 3.
Pyridine and water hydrolysis of RNA

Enzymatic hydrolysis of RNA is effective, but the nucleo-
side fraction is contaminated with protein at the conclusion of the
hydrolysis. The possibility of using a chemical hydrolysis of
RNA was investigated with the objective of eliminating any potential
difficulty imposed by the presence of protein in the isolation or
subsequent resolution of the 2'-Q_-methy1nucleoside fraction.

Alkaline hydrolysis of RNA is not satisfactory in isolation
of 2'-Q_-methy1nucleosides since the phosphate forms a cyclic inter-
mediate across the cis-2',3'-hydroxy1s as part of the initial cleavage.
Hydrolysis of RNA by ammonia under pressure, fir st reported by
Levene and Jacobs (1910) completely deaminates cytidine (Gullard
and Hobday, 1940) and probably proceded by an alkaline hydrolysis
mechanism in any case. Mineral acid hydrolysis of RNA results
in cleavage of the 'glycosidic bond with purines while perchloric

acid hydrolysis results in complete destruction of all nucleosides

27

to the free base level (Hashizume and Sasaki, 1968). Catalytic hydro-
lysis of RNA with lanthanum was reported by Alleniand Bacher (1951)
but they report a 2-3% loss of 260 mu absorbing material per hour.

Bredereck et a1 (1941) refluxed yeast RNA in a pyridine-water
mixture for 108 hours and reported hydrolysis of the RNA with
recovery of the four ribonucleosides. They used the procedure as
a source of nucleo sides rather than as a method of RNA analysis.

If no undesirable side reactions occur, hydrolysis of RNA by incubation
in pyridine and water would be attractive since pyridine could be

easily removed when the hydrolysis was complete. After several
evaporations, an incubation mixture would contain only nucleosides

and trace amounts of phosphate. Since total recovery of the nucleo-
sides from a hydrolyzed RNA sample is essential if they are to

be used to determine nucleoside ratios, this research included a

study of the effect of the pyridine and water incubation on nucleo-
sides and nucleotides.

The pyridine-water hydrolysis of RNA was investigated with
respect to recovery of nucleo sides and formation of by-products.
The initial study was a time analysis of the hydrolysis of yea st
sRNA in order to establish the length of time necessary to achieve
degradation to the mononucleo side level. Sealed vials containing
20 mg of yeast sRNA in a total volume of 1. 0 ml (pyridine, water;
1:1; v:v) were heated at 108° and samples were taken after one,

three, four, and six days. After several evaporations with water

28

to remove pyridine, the samples were spotted on Whatman #1.

A mixture of the four ribonucleo sides, a mixture of the four ribo-
nucleotides, a mixture of alkaline stable wheat germ dinucleo side
monopho sphates, and a sample of unhydrolyzed yeast sRNA were
included as standards. The chromatogram was developed for 18
hours by descending chromatography with Heppel's solvent (n-propanol,
ammonia, water; 55:10:35; v:v:v). Heppel's solvent leaves long
oligonucleotides or unhydrolyzed RNA at the origin. Short oligo-
nucleotides (< 10) and mononucleotides are resolved by size with the
shorter chains having the larger Rf values. Nucleosides move
even faster than the mononucleotides.

After one day of incubation the yeast sRNA had been hydrolyzed
into a mixture of nucleosides, nucleotides, and short oligonucleotides.
Some material still chromatographed in the dinucleo side monopho sphate
region after three days, but hydrolysis to the nucleoside level was
complete after four days. Comparison of the three, four, and six
day samples indicated an increasing amount of fluorescent material
with time. This fluorescent material chromatographed in the region
of the nucleo side s, but it has an unique Rf.

Since it was known to take at least four days to hydrolyze yeast
sRNA by this approach, the fate of several nucleotides during an
incubation under similar conditions was studied. Recovery of
nucleoside as measured by absorbance at 260 nm initially seemed

very good; the values were in the range of 95%. However, analysis

of the products of the incubation indicated deamination and cleavage

of the glycosidic bond were occurring after only 40 hours of incubation.
In addition, the fluorescent material seen in the yeast sRNA hydrolyzate
began to appear upon further incubation.

The fate of adenosine during a pyridine-water incubation was
followed in a time course analysis. As shown in Table 4, adenosine
was either hydrolyzed to the free base or dearninated (or both) to
a significant extent in the period of time necessary to hydrolyze the
yeast sRNA sample. Values for adenosine are maximum values
since they may contain unresolved hypoxanthine. At least 16% of the
adenosine has been destroyed in the fir st 38 hours of incubation.
This value increases to a minimum of 19% in the four days which
were required to hydrolyze yeast sRNA. The data is not sufficient
to determine a rate of destruction but a definite decrease in amount
of adenosine with time is seen.

It is of interest that no fluorescent material was produced in
the adenosine incubation, even at the longest incubation period.

The fluorescent compounds may occur as a result of some reaction
of phosphate ion with the nucleo sides or with pyridine since the
fluorescent material appeared following hydrolysis of the mono-
phosphates (eg. AMP) as well as the hydrolysis of RNA.

Extensive destruction of the nucleo side is occurring in the
time periods necessary for complete hydrolysis of yeast sRNA by

pyridine-water reflux. The appearance of cleavage and demination

Time of Adeno sine Adenine Ino sine

 

Incubation (Percent) (Percent) (Percent)
( hours) - . .. -
38 83. 7 8. 1 8. l
67 81. 8 9. 1 9. 1
84 83. 3 8. 6 8. 0
96 80. 7 7. 8 11. 4
108 76. 0 ll. 5 12. 5
120 77. 8 9. 7 12. 4
133 77. 4 8. 4 14. l
144 75. 1 9. 3 l5. 6

Table 4. The progress of destruction of adeno sine by incubation
with pyridine and water.

Vials containing 4. 1 umoles of adeno sine were incubated as
described in Methods and samples were taken at various time 8.
After several evaporations from water on a flash evaporator to
remove pyridine, the samples were spotted on Whatman #40 (acid
washed). Adenosine and its deamination product, inosine, were
included as standards, as well as adenine and hypoxanthine, the
respective products of hydrolysis of their glycosidic bonds. The
chromatogram was developed for 18 hours by descending chromato-
graphy with 86% n-butanol as solvent. After chromatography, the
standards were arrayed in the following order by increasing Rf
value: inosine, adeno sine, hypoxanthine and adenine. The spots
corresponding to adeno sine, adenine, and ino sine were cut out and
eluted in l. 0 ml of 0. IN HCl. Since hypoxanthine nearly cochromato-
graphed with adenosine, it was not possible to quantitate this base.
The A260 for each spot was measured and the values were converted
to moles using the respective extinction coefficients.

30

31

products of adenosine when the nucleoside alone is incubated is
enough to invalidate the approach. Production of a fluorescent
by-product when the nucleotide or intact RNA is incubated renders
the procedure completely unacceptable for quantitation of nucleo-

side ratios.

11. Derivatization and GLC of nucleo sides.

The mixture of 2'-_(_)_-methylnucleo sides obtained from the
AG 1x4 (borate) column must be resolved and quantitated for a ratio
determination. For my purposes, GLC was investigated as a potential
tool for this resolution due to its sensitivity and convenience.
Jacobson et a1 (1968) reported a lower limit of 0. 1 ug (ca 0. 4 nmole)
of each nucleoside in their GLC determination of nucleoside ratios,
which is much less than the amount required for paper chromato-
graphic determinations (ca 20 mnole). Several analyses may be
performed in a relatively short time. A single sample may be in-
jected several times, giving data which is statistically more signi-
ficant than the single A260 reading available after paper chromato-
graphy.

BSA derivative 3

 

The 2'-_0_—methy1nucleosides must be derivatized to increase
’ their volatility prior to a GLC analysis. The BSA silylation pro-
cedure of Jacobson et a1 (1968) was used. The TMS-derivatives

formed from the 2'-_Q_-methylnuc1eo sides must have well resolved

 

32

peaks for the determination of a ratio. Two column liquid phases
were investigated for this separation. They were OV-l, the non-
polar methyl silicon gum, and the slightly polar methyl-phenyl
silicon gum OV—l7 used by Jacobson et a1 (1968). The resolution
of nucleosides achieved by Jacobson et a1 depended on increasing the
column temperature from 160° to 240° by temperature programming
to elute the derivatives from a 4% liquid phase. Lower percentages
of liquid phases were used in this work in an attempt to elute the
peaks isothermally, thus decreasing the time of analysis and creating
more reproducible conditions.
GLC behavior and retention times of nucleo side derivatives.

Treating 2'-_Q-methylnucleo sides with BSA as described in the
methods section results in a single derivative for Um and Am as
shown by the presence of a single symmetrical peak when they are
chromatographed on either OV-l or OV-l7. Cm and Gm form multiple
derivative 3, indicated by a trailing peak for Cm on both phases and
two peaks for Gm on OV-l7. The trailing peak for the ratio deter-
mination; the two peaks formed by the Gm derivatives may also be
treated together since the first peak is much larger than the second
and they are not well resolved.

Retention time is determined by the time of elution of the
crest of the GLC peak (major peak for Cm or Gm). Table 5 contains
the retention times of the TMS derivatives for the 2'-_(_)_-.methy1nucleo-

sides as resolved on OV-17 or OV-l under the conditions described

Nucleo side Retention Time

 

ov.17 ov-1

192° ' 111°

Um 6. 4 6. 1
Am 14. 2 12. 9
on}. 21. 6 16. 5
Gm 29.4 ---

Table 5. The retention times of the BSA derivatives of the
Z'-£—methylnucleo sides.

The TMS derivatives were prepared separately for each 2'-
_Q_-methylnucleo side with BSA as described in Methods. Samples were
injected onto a 1% OV-l column or a 2% OV-l7 column of the GLC.
Retention times are given in minutes at the indicated column temperatures.

33

34

in the methods section. It is seen that there are more than 7 minutes
between the closest two peaks when OV-17 was used. Separation
was not successful on OV-l since the Am and Cm derivatives are not
resolved sufficiently to give discrete peaks.

Figure 2 is a compo site drawing of the GLC tracing of a mixture
of Um, Am, and Cm which was derivatized with BSA and injected
on OV—l7 and the GLC tracing of Gm which has been treated similarly.
It was not possible to obtain a GLC tracing of a mixture of the four
2'-0—methy1nucleosides at this time since the Gm is in short supply
as discussed earlier. The composite was drawn to make the excellent
resolution of the four peaks more obvious.
Stability and reproducibility of derivatives.

Another aspect of the GLC nucleoside ratio determination is
the reproducible formation of derivatives which are stable over a
reasonable period of time. The reproducibility and stability of a

compound are traditionally studied using its absolute molar response.

 

1The absolute molar response is calculated as the ratio of the
response of a known amount of a compOund to the response of an
equimolar amount of an internal standard, an added compound which
is known to give a stable and uniform response. The reproducibility
of derivative formation is established by obtaining a constant value
for the absoluet molar response of a compound in a series of derivati-
zations. The stability of a derivative is similarly established by
. obtaining a constant absolute molar response value from a series
of injections of one sample over the time interval of interest.

 

Figure 2. A gas-liquid chromatogram of the TMS derivatives
of the 2'-_g-methylnuc1eosides on an OV-l? column.

The TMS derivatives were formed for a mixture of Um, Am,
and Cm and chromatographed on a 2% OV-17 GLC column as described
in Methods. A sample of Gm was treated similarly and chromato-
graphed under identical conditions on the same column. The chromato-
grams were then drawn as the compo site to indicate relative retentions
and separations as discussed in the text. Peak I is TMS-Um, peak
II is TMS-Am; peak 111 is TMS-Cm; and peak IV is TMS-Gm. The
attenuation was increased by a factor of 4 at the point indicated by
the arrow.

35

36

 

 

 

 

 

 

 

 

Time (min)

 

 

asuods 93 .10199190

37

The absolute molar response is not necessary for a ratio determination
since the relative amounts of the nucleosides will suffice. In this
case, a stable relative molar response for the nucleoside derivatives
may be used to establish their stability and reproducibility. 2 The
relative molar response values may also be used to convert experi-
mentally determined peak weight ratios into molar ratios for the
nucleo side ratio determination.

The stability and reproducibility of the TMS derivatives of the
2'-_0_-methy1nucleosides were checked by comparing their relative
molar response values over a period of time and from one derivati-
zation sample to another. Table 6 gives the values obtained for the
molar response of Um with respect of Am and the molar response
of Cm with respect to Am in that experiment. The values for both
responses are grouped for easy comparison.

The relative molar response values are internally consistent
within one sample over a time span of two and one-half hours, and
externally consistent from one sample to another. There was not
sufficient quantity of Gm at this time to test the stability and re-

producibility of its response as the other nucleosides were tested.

 

2 O O O
The relative molar response 18 the relative response of one

p compound in a mixture with respect to an equimolar amount of
another which is cho sed as standard. The reproducibility and
stability of derivative formation are established by obtaining constant
values for the relative molar response values in the same manner
as with the absolute molar response.

Table 6. The stability andreproducibility of the TMS derivatives
of the 2'-_0_-methylnucleo side 3.

Three samples of a mixture of Am (0. 351 umole), Um (0. 328
umole), and Cm (0. 634 umole) were treated with BSA at two day
intervals as described in Methods. They were maintained under
anhydrous conditions at room temperature and an aliquot was
periodically injected onto a 2% OV-l7 column of the GLC. The peaks
were cut out, weighed, and the weights were corrected for differences
in attenuation and chart speed. These corrected peak weights were
used to calculate a peak weight ratio of Um with respect to Am and
Cm with respect to Am. The relative molar responses were deter-
mined by correcting the peak weight ratio for the actual molar
ratios present in the mixture of nucleosides.

38

Time
(min)

Zero

30

60

90

120

150

Zero

30

60

90

120

150

Relative molar response of UmzAm

 

I II III
0. 74 ---- 0. 81
0. 78 0. 80 -..--
0.73 0. 79 0.80
--.... 0. 81 0. 81
--..- 0. 77 0. 79
0. 84 ---- ----

Relative molar response of Cm:Am

 

I II 111
0.14 ---- 0.02
0.17 0.16 -..--
0.21 0.15 0.17
....-- 0.18 0.18
..-..- 0.14 0.22
0.17 ---- -..--

39

40

Preliminary results indicate that the peak size remains faifly constant
upon repeated injection of aliquots of the same volume from one
sample.

Table 7 contains the mean relative molar response values for
TMS-Um, TMS-Am, and TMS-Cm with their standard deviations.
The relative molar response for Gm will be determined at a later
time. A comparison of the relative molar responses reveals that
TMS-Cm given a much poorer response than earlier TMS-AIn or
TMS-Um, which may partially account for the slightly higher I
variation in the TMS-Cm response indicated by the higher standard
deviation. The values are in a workable range, although the low
response of the Cm derivative rais es the minimum amount of 2'-Q-
methylnucleosides required to determine a ratio.

Nucleoside ratio determination of synthetic copolymers

Nucleoside ratios were determined for synthetic copolymers
prepared with polynucleotide phosphorylase in our laboratory (Rottman
and Johnson, 1969). The ratio of the nucleosides resulting from an
enzymatic digestion of each copolymer was obtained by GLC of their
TMS derivatives. A paper chromatographic ratio determination was
also done as a check on the accuracy of the GLC determination.

For the GLC determination, the relative molar response
values reported in Table 8 were obtained from mixture of known
amounts of standard nucleo sides as described earlier. The artifical

mixtures were prepared to approximate the ratios expected in the

 

Um Am Cm
Actual Molar Ratio 0. 94 l. 00 l. 81
Mean Peak Weight Ratio 0. 74 1. 00 0. 31
Relative Molar Response 0. 79 l. 00 0. 17
Standard Deviation O. 03 ---- 0. 02

Table 7. The relative molar response values for 2'-0-methy1.-

nucleo side 8.

The values presented above represent the mean of the 12

values obtained from the experiment discussed in Table 6.

41

 

Actual Molar Ratio 1 o. 0572 o. 0481
Mean Peak Weight Ratio 1 o. 0847 ’f o. 0574*
Relative Molar Response 1 l. 48 l. 19
Standard Deviation - 0. 04 0. 02

In: 7; * 11:5

Table 8. The relative molar response values for the
nucleosides in synthetic copolymers.

Mixtures of U and A or U and Am were prepared in the molar
ratios indicated above. TMS derivatives were prepared with BSA
and GLC was performed on an OV-l7 column. The relative molar
response of A and Am were determined with respect to U as described
in the footnote to Table 7.

42

43

copolymers. The corrected peak weight ratios were calculated
for the nucleo sides in enzymatically digested samples of poly (U,A)
and poly (U, Am) after their separation on GLC. The relative molar
response values discussed above were used to calculate the molar
ratios of the nucleosides from the peak weight ratios for these
synthetic copolymers. 3

The molar ratios of the nucleosides in the two copolymers
are given in Table 9. Part A contains the values obtained by GLC
and Part B contains the values obtained by paper chromatography.
The two techniques give values which are in good agreement.

Alternative derivati zation technique s

BSTFA derivatives of nucleo sides

There are some practical difficulties to be faced in using
BSA as a TMS donor for nucleosides. Both BSA and its by-product
mono (trimethylsilyl) acetamide, have longer retention times and
higher detector responses than is'de sirable, which restricts the
lower limit for elution times of peaks. The detector must be fre-
quently cleaned since silicon dioxide, one of the combustion pro-

ducts of the hydrogen flame, rapidly coats the detector resulting in

 

3For example, the poly (U,Am) polymer gave a corrected peak
weight ratio of 11. 8 for U with respect to Am. Since a mole of Am
gave 1.19 times as much response as a mole of U, the response
of U must be increased by a factor of 1. 19 to obtain the molar ratio.
The molar ratio is therefore 14. 1 moles of U per mole of Am.

A. Determined by GLC of TMS derivatives.

 

poly (U, A) poly (U, Am)

U:A ratio U:Am ratio
Mean Peak Weight Ratio 4. 04* ll. 8*
Molar Ratio 6. 0 14. 1
Standard Deviation 0. 2 0. 4

Tn: 10; *n- 3

B. Determined by Paper Chromatography

 

poly (U, A) poly (U, Am)
U:A ratio U:Am ratio
Molar Ratio 4‘ 6. 6 13. 9

*Based on duplicates

Table 9. Nucleoside ratios of synthetic copolymers.

Samples of poly (U,A) and poly (U, Am) were digested with
Russell's viper venom pho sphodiesterase as described in Methods.

For the GLC ratio determination, the incubation mixtures were
lyophilized and silylation was achieved with BSA as described in
Methods. GLC was done on a 2% OV-l7 column. Following the
measurement of the peak weights, the molar ratios were calculated
using the relative molar response values from Table 8.

For the paper chromatographic ratio determination, duplicate
samples of each enzymatically digested copolymer were spotted on
Whatman #1 (acid washed) with standard nucleo sides and the chromato-
gram was developed with n-butanol, ammonia, water (86:5:14; v:v:v)
for 16 hours. The spots representing the nucleosides derived from
the c0polyme rs were eluted in 2. 0 m1 0. IN HCl. The absorbance
at 260 nm was recorded, molar values were calculated using the
appropriate extinction coefficients, and the molar ratios were determined.

44

baseline noise and reduced response of the desired peaks. These
difficulties can be reduced by use of BSTFA, a BSA analogue, since
its retention time as well as that of its by-product, mono (trimethyl-
silyl) trifluoroacetamide, are greatly reduced. Since the BSTFA
combustion product, silicon tetrafluoride, passes through the de-
tector to' a greater extent than its BSA analogue, there is much

less coating action.

It was for these reasons that Hancock (1969) tried unsuccessfully
to use a mixture of BST FA and pyridine for the nucleo side silylation.
The use of pyridine in silylation of nucleo sides has led to unsatis-
factory derivatives with BSA Vand TSIM as well as with BSTFA in
our experience. Um and Am derivatives form trailing peaks and
Cm does not give a derivative which elutes at all under our usual
GLC conditions. The use of BSTFA was therefore investigated without
added solvent in a manner similar to the BSA silylation.

Initially, the ribonucleosides were used as model compounds.
Table 10 contains the retention times for the BST FA formed TMS
derivatives of the ribonucleo sides when chromatographed on OV-l .
The peaks are respectable and re solution is good. The 2'-_(_)_-methyl-
nucleo sides were then derivatized with BST FA; the retention times
of their TMS derivatives are also reported in Table 10. Um and
Am form satisfactory derivatives with retention time 5 similar to
those formed with BSA, but Cm forms two derivatives with widely

spaced retention times on both OV-l and OV-l7. Use of both TMS-

Nucleo side Retention Time

 

 

ov-17 ov-1
£0

U 7. o

A 14. 6

c 17. 5

G 30. 4

122° 1613

Um 5. 6 l3. 2

Am 12. 5 31. 6

Cm 19. 2 38. o 21. 3 35. 7

Table 10. The retention times of the BSTFA derivatives of the
ribonucleosides and the 2'-2-methy1nuc1eosides.

The TMS derivatives were prepared separately for each
ribonucleOside and 2‘-Q~methylnucleo side with BSTFA as described
in Methods. Samples were injected onto a 1% OV-l column or a
2% OV-l7 column of the GLC as indicated above. Retention times
, are given in minutes at the indicated colunm temperatures.

46

Cm peaks for quantitation was not practical since the first peak
interferred with the elution of TMS-Um and the second peak on
OV-l7 was very broad (it spanned several minutes) and it would
interfer with the elution of TMS-Gm. The first TMS-Cm peak on
OV-l disappeared upon addition of one microliter of water to the
reaction mixture. The parallel experiment was not attempted with
OV-l7, since the second peak is not desirable anyway.

BSTFA was not used for silylation of the nucleo sides since
Cm formed two unusable derivatives. Jacobson et a1 (1968) reported
a similar formation of two derivatives for C when trimethylsilylation
was done with BSA at higher temperatures. Since BSTFA is a more
potent TMS donor, a similar situation has arisen, even at the lower
temperature at which the derivatization was performed.
Mass spectral analysis of A and Am derivatives formed with BSTFA

Mass spectra were obtained for a group of trirnethylsilyl
nucleo side derivatives. The extent and position of hydrogen re-
placement by TMS groups was examined to characterize the derivatives.
The mass spectral analyses of BSTFA formed TMS-A and TMS-Am
are presented here to illustrate the method by which the derivatives
were characterized.

The molecular weight of adenosine is 267; the molecular
weight of the TMS derivative would therefore be 267+72n with n
representing the number of added TMS groups. A molecular ion

with a mass to charge ratio ( m/e) 33, 411, 483, 555, or 627 would

48

be expected for the substitution of 1 through 5 hydrogens respectively.
The m/e 555 ion on the mass spectrum of TMS-A in Figure 3
represents the molecular ion for TMS-A indicating the substitution

of four hydrogens by TMS groups. A molecular weight of 555 is
supported by the presence of ions at M-15 representing the loss of

a methyl group, M-73 representing the loss of TMS group, M-90
representing the loss of trimethylsilanol, and M-107 representing the
loss of the 5'carbon with its constituents (CHz-O-TMS).

The molecular weight of Am is 281, 14 more than the molecular
weight of A since a methyl group has been substituted for the hydrogen
of the Z'hydroxyl. Since the 2'hydroxy1 is no longer available as
a site for TMS substitution, the molecular weight would increase
216 mass units by the addition of only three TMS groups if the .Am 7
silylation parallels that of A. As seen in Figure 4, the expected
molecular ion m/e 497 is present on the TMS-Am mass spectrum.
There are also ions at M-15 representing the loss of a methyl group .
M-3l representing the loss of an OCH3 group, and M-103 representing
loss from the molecular ion of C-5' with its constituents.

There are four potential silylation sites on the adenosine mole-
cule; the three hydroxyl groups of the ribose moiety and the amino
group on the purine ring. The fragmentation patterns discussed
below indicate that each of these four sites is silylated. The structure
proposed for TMS4-A, shown in Figure 5, is consistent with the

structure pr0posed by Lawson et al (19%) for the structure of TMSS-

Figure 3. The mass spectrum of TMS4-adenosine.

49

on

 

0mm

00m

owe.

0N4

0mm

wx

com

owN

CNN 00"
.1- L’PW‘D lb L—InL —

 

 

Laa1aa-.1

 

ION .

19‘

low

low

 

r:02

Figure 3. The mass spectrum of TMS4-adenosine.

49

 

can

con

omv

ONQ

0mm

mx

oom

omN

CNN 00—

:FLIrLW-rtg F'ihdl‘Lqirtp

om
_

r

 

 

Laa1ga1.a

 

[FL

Iom

row

 

rco"

Figure 4. The mass spectrum of TMS3-2'-_(_)_-methy1adenosine,

51

m“

L P h hIFVL hr+b p PIP PLLL

 

 

 

0mm Gin oom own ONN own 0%” 00“ ow
41

‘PL! ,
111 4 44.41
ION

Iom

 

 

102

53

ocwm onvvafifloEnMn . N -
ummzH mo ongosflm van. .0 mufiwwh

mIUO OMS...

I I
I I
.0 £003:
2 ZJ
A _ z
Z \
mZFIZ

ogmoaovmuwmzfi
mo ongofinum ”AH. .m onfimfim

mZPO OMS...
I I
I I

o «xuomzh
z 2
A _ J
2 \z

m ZFIZ

54

5'-AMP. Since the fragmentation patterns of TMS3-Am indicate

that each of the three remaining silylation sites of Am also contain

a TMS group, the structure in Figure 6 is proposed for TMS3-Am.
Fragmentation of the molecular ion is consistent with the

stability of the purine ring as reported by others for the nucleoside

(Shaw et a1, 1970; Hecht et a1, 1969) or nucleotide (McCloskey et a1,

1968) and for TMS derivatives of the nucleotides (Lawson et al, 1971)

Simple cleavage of the glyco sidic bond leads to ion B, which represents

the base fragment for either TMS4-A or TMS -Am. This ion is

3
unms NHms NHTMS
N N
(’50 ”in m
u N
" '1 H {1 H
a C 0
"k2“ ‘ m/¢907 sale 30!

accompanied by ions C and D indicating the presence of one or two
hydro gens which have been abstracted from the ribose moiety.
Alternate fates of ion C are 1038 of a TMS methyl group giving
rise to an ion at Ink 192 or the addition of a ribose TMS group
resulting in the m/e 280 ion.

The ion at m/e 236 is produced by cleavage of the molecular
ion to include C-l', 0-4', and two hydrogens with the base fragment

Either E or F may represent the structure of this ion since one of

BTMS H+7TMS
l , ,
c I /c\l
/‘*\ /
HO H O H
E. F
m]: 93‘ tale 3?‘

the hydrogens is of dubious origin and both of the structures would

55

be stable. Production of the ion is seen to be a major pathway for
fragmentation of both molecules by the high relative intensity of
their respective ions.

A diagnostic ion, G, is produced when the base fragment in-
cludes both C-l', and C-Z'. A similarbut less frequent ion, H, is

produced when 0-4' is included with the base fragment containing

 

C-1' and C-Z'. The difference of 58 mass units between the TMS n
B _
g o B -;
H‘. CHI' , “1'
' ml: 39! , 3" 07”“; OTMS ml: 337 , K'OTMS 1
R
G mle OH, F: “C”: H 05/0 177) ‘300H,

group and the methyl group located on the Z'hydroxyl is reflected in
the mass of the ions produced. A methyl group from TMS and a
hydrogen may be lost from ion G, possible as methane, producing

an ion at m/e 306 for TMS -A and an m/e 248 ion for TMS3-Am.

4

When the ribose moiety retains the charge upon cleavage of the
glyco sidic bond, a hydrogen is typically lost from the ribose ion
leading to ion S. A series of related ions is produced by loss of
portions of this parent ion. A TMS methyl group can be lost from

TM socn, o 3
Q Q

TMSO R TMSO K

5 .T
"’9 3"” R: arms Ml: Mr, RvoTMS'
”/e 9‘70, 3’ OCH-i n/e I37, 8: 0W3

the intact ribose producing ions at m/e 334 for TMS4-A and

m/e 276 for TMS3-Arn. Loss of 05' with its constituents may

56

occur resulting an ion J.
Loss of the C-2' ether group from ion S produces ion K with

both TMS4-A and TMS3-Am. When the TMS ether group is lost

TMSOCH, o o O
Q H,c=<_>i Tmsocu,<__>t
+
TMSO TMSO OTMS' 01—“?
k L N
m]: 95“! rule 35'? m/e 959

from C-5' or C-3' or ion S, ions L or N respectively are produced

for'TMS4-A. The related loss of the TMS ether from ion S of TMS3-

Am usually occurs with a hydrogen yielding ion 0 or P, respectively.

. O 0
TMS OCH;
0 P
m]: 9 0° un/e 200

Loss of the C-2' ether group may be preferred since the relative

intensity of the m/e 200 ion in the mass spectrum of TMS3-Am.
Significant ions are also produced by the loss of two ether

groups from ion 5. Loss of the C-2' ether group yields ion R when

the C-5' TMS ether is lost or ion T when the C-2' TMS ether is lost.

0
0 Two C ",9 +
4.
1.6.4} HQ

Tms'o T ' R
R ml: '69 u
"/6 l6? Mle n9 , 3: am:

m/e IH , 8 = 0013
Loss of both the C-5' and C-3' TMS ether groups produces ion U.

Apparently the coordinate loss of two ether groups is more easily

achieved from C-5' and C-3' than from either of the other two

57

combinations since the relative intensity of the m/e 169 ion (U) is
larger than the relative intensity of the rule 111 ion ( R T) in the mass
spectrum of TMS3-Am. This hypothesis is supported by the un-
published mass spectrum of TMS3-3'-_Q_-methyladenosine for which
the relative intensity of the m/e 169 ion ( R+T) is larger than the
relative intensity of the m/e 111 ion (U).

The charge may also reside with the ribose moiety following
the cleavage of the molecular ion in vhich C-1‘ and 0-4' reside with
the base. When this occurs, the very stable ion W is produced in

TMs‘o-CH. ~CH'—‘-’CH=-'C"- 0!?
W

m/e 930 , R. OTMS‘

m/e 173, 3" 0"”:
good yield as shown by its high relative intensity. Distribution of
the charge over C-4', C-3' and C-2' in this four carbon remnant
of the ribose moiety probably accounts for its stability.

Characteristic ions are present on both spectra resulting

from fragmentation of the ribose ring leaving three carbon pieces
with their constituents. Ion X represents C-S', C-4' and C-3', an
m/e 217 ion for both TMS4-A and TMS3-Am. Ion Z represents

C-4', C-3' and C-Z', also an m/e 217 ion for TMS4-A, but an m/e

159 ion for TMS3-Am. Fragmentation which results in the production
+

M *-
TMSO-CuzzcurxcH-o-rms' cu,=co~rm:— CH-oR
x Z
mic 9'7 “I: an, Re 74»:

01/2 I57, K’CH3

 

58

of ion X seems to be preferred since this ion has a higher relative
intensity than does the Z ion in the mass spectrum of TMS3-Am.
This preference may be explained on the basis of the stability of the
Xions resulting from distribution of the charge over three carbons
as in the case of ion W.

An m/e 103 ion is produced if the charge resides with C-5'
and its constituents when it cleaves from the molecular ion or from n

ion S. Several ions are charact eristically produced on mass spectra

 

of TMS derivatives from anomolies of the trirnethylsililanol molecule. h
The TMS ion at mile 73 is produced in sufficient quantities to be the

base peak in these spectra. Other examples are the m/e 89 ion

(aTMS) and the m/e 147 ion (DMS:6-TMS) formed by loss of a

methyl group from the di-TMS ether.

Effect of methoxyamine-hydrochloride pretreatment on the nucleo-

side derivatives formed by BSA.

The response ratio for TMS-Cm with respect to TMS-Arm
displays more variation than the response ratio for TMS-Um with
respect to TMS-Am as seen in Table 7. Butts (1970) reported
improved gas chromatographic behavior of cytidine and deoxy-
cytidine upon treatment with methoxyamine hydrochloride in pyridine
to form the methoxime derivatives prior to silylation. He further
reported that the reaction to form methoximes is unique to cytidine
and its derivatives; other nucleo sides do not react. In an attempt

to avoid difficulty with the TMS-Cm response, the TMS-methoxime-

59

Cm derivative was formed. The structure shown in Figure 7 assumes
that methoxime-Cm is silylated at both available site‘s.

Table 11 contains retention times for 2'-_(_)-methy1nucleo sides
which were treated with methoxyamine hydrochloride prior to silylation
or which were silylated only. Methoxyamine hydrochloride treat-
ment does not alter the retention time of either the Um or Am derivat-
ives, but there is an obvious difference in the retention times of the
Cm derivatives. TMS-methoxime-Cm has a retention time of less
than one-third the retention time of TMS-Cm on either OV-l or OV-17.
The derivative formed is superior to TMS-Cm on both liquid phases
since it forms a single sharp peak with little trailing. TMS-methoxime-
Cm cannot be used to determine the 2'-9_-methylnucleoside ratio
in an RNA molecule however, since it is not well resolved from
TMS-Um on either liquid phase as shown in Table 11.

TSIM derivatives of nucleo side 5.

In addition to the TMS-methoxime-Cm derivative, two other
approaches were used in an attempt to form a more satisfactory
derivative of Cm. Both approaches were based on the assumption
that the available sugar hydroxyls of a nucleoside are rapidly silylated
and the subsequent partial silylation of the amino or keto groups
on the purine or pyrimidine rings is responsible for multiple
derivatization. The fir st approach was treatment of the nucleosides
with trirnethylsilylimidazole (TSIM), which should silylate hydroxyl

groups only (Pierce, 1970).

NOCH3

21‘

TMSOCH: O

 

H H
H H

TMSO O CH;

Figure 7. The structure of the TMS -methoxime
derivative of 2'-_Q-methylcytidine.

60

Nucleo side Retention Time s

 

 

OV-l OV-l7
182°
Am 11. 2
Am (methoxyamine treated) 11. 2
0
171° 186
Um “ 10. 8 14. 9
Um (methoxyamine treated) 11. 2 l4. 9
169° 200°
Cm ' 3. 18 22. 6
Cm (methoxyamine treated) 8. 8 7. 6
169° 186°
Um 10. 5 l4. 9

Cm (methoxyamine treated) 8. 8 15. 8

Table 11. The effect of pretreatment with methoxyamine-hydro-
chloride on the retention times of the BSA derivatives
of the 2'- g-methylnucleo sides

Samples of Um, Am, and Cm were treated with methoxyamine-
(hydrochloride in pyridine as described in Methods. These samples

and similar untreated samples were then silylated with BSA and G. C

was performed on a 1% OV-l column or on a 2% OV—l? column.

Retention times are given in minutes at the indicated column temperatures.

61

 

62

The TMS derivatives of Um and Am formed by treatment with
TSIM have retention times which are very similar to those formed
by treatment with BSA on both OV-l and OV-l7. The peaks are
symmetrical and display a minimum of trailing. However, the
silylation of Cm is less satisfactory than treatment with BSA since
a series of peaks was formed. The peaks were not well resolved
on either liquid phase. They overlapped on OV-l and a trailing peak
r esulted; they were spread out on OV-l7 resulting in a broad plateau
effect rather than well defined peaks. Since silylation of the 2'-£-
methylnucleosides with TSIM does not result in a satisfactory deri-
vative of Cm, the procedure was not used.
Effect of methanolic acetic anhydride pretreatment on the nucleo-

side derivative formed by BST FA.

 

The second approach to more clearly defining the sites of
trimethylsilylation of the nucleo sides was an initial treatment with
acetic anhydride in methanol. Selective N-acetylation on the presence
of hydroxyl groups is routinely achieved with sphingolipids by this
treatment (Gaver and Sweeley, 1966; Carter and Caver, 1967).
Acetylation of the amino group of the purine ring was therefore
attempted prior to silylation of the sugar hydroxyl group.

Two derivatives of Am were obtained by acetylation with acetic
anhydride in methanol followed by silylation with BSA. The first
peak eluted with a retention time similar to that of TMS3-Am (pre-

pared by treating Am with BST FA) indicating that no acetylation

63

has occurred. The other peak has a retention time about one-third
longer than TMS3-Am, suggesting that acetylation has occurred.
There was some indication that higher yields of acetylated Am were
obtained if room temperature acetylations were performed over
longer periods of time (two days or longer) or if short term acetyl-
ations were carried out at elevated temperatures (90°).

Mass spectral analysis of the two peaks confirmed the con- r
clusion that acetylation of Am produced the second peak. The mass

spectrum of the first peak eluted was identical to that of TMS3-Am, f

 

shown in Figure 4. The mass spectrum of the second peak is shown
in Figure 8. Since an acetyl group has a molecular weight of 43,

30 mass units less than a trimethylsilyl group, the number of acetyl
groups present on the molecule can be deduced from the molecular

weight change. The difference in molecular weight between TMS3-Am

and the compound in the second peak is 30 units, indicating the
presence of one acetyl group.
The location of this acetyl group can be determined by com-

parison of the fragmentation pattern of TMS -acetyl-Arn with the

2

-Am and TMS4-A discussed earlier.

The purine ring has a strong m/e 208 ion in both TMS3-Am and

TMS4-A which would be changed to an ion at m/e 178 if the acetyl-

fragmentation patterns of TMS3

ation has occurred on the amino group of carbon six. The 2'-Q_-
methylribose moiety has a m/e 290 ion in TMS3-.Am, the expected

58 mass units lower than the ribose moiety ion of m/e 348 in' TMS4-A.

Figure 8. The mass spectrum of TMSz-5'-acetyl
2’-2-methy1adeno sine.

64‘

o"

 

 

rx

 

 

00"

 

om

low

is

 

66

Acetylation of the 2'-_g-methylribose moiety would result in a further
reduction of 30 mass units giving an ion at m/e 260 for TMSz-acetyl-
Am. As seen on the mass spectrum in Figure 8, the m/e 208 ion
has a high relative intensity, indicating that the base remains un-
altered. The ion representing the ribose moiety has shifted from

m/e 290 to m/e 260, however, indicating that the acetylation has

occurred on the sugar.

Acetylation of the Slhydroxyl is indicated by loss of the ion

4.

at m/e 217 (TMS'O-CHz-‘CH'J-‘vCH-OTMS) which originates from C-5', *

 

C-4', and C-3' of TMS-Am as discussed earlier. The m/e 187
+

ion (CH3-COO-CmH-OTMS) replaces this ion as would be
expected for a C-5' acetylation. The m/e 103 ion (CHZOTMS)
representing C-5' and its constituents is also missing from the mass
spectrum of TMSZ-acetyl-Am. Its replacement ion at m/e 73
(CHz-OOC-CH3) cannot be distinguished from the ion produced by
TMS alone without high resolution mass spectroscopy.

The complete structure of TMSz-S'acetyl-Am is given in
Figure 9. Since the acetylation occurred at the 5' hydroxyl rather
than at the amino group on the purine ring of Am, this derivati;

zation procedure was not applied to Cm. Incomplete acetylation

was another difficulty to be faced if the investigation was to continue.

NHTMS
“213
9 K” "
CH3°C’OCH3 O

H H
H H
TMSO OCH;

Figure 9. The structure of TMS -5'-acetyl
2'-_Q-methyladenosine.

67

 

DISC USSION

A study of the distribution of 2'-_Q-methy1nucleosides in an
RNA molecule can be approached from several levels. The first
is quantity, a determination of the percent which the combined 2'-_9_-
methylnucleo sides represent of the total nucleosides (Abbate and
Rottman, 1971). These data are easily obtained and are informative.
Although differences may be more subtle, comparisons of RNA
molecules can be made on the basis of the percentage of the nucleo-
sides which are 2'-_(_)_-methylated.

'A more complete analysis, short of the complete sequence
of the RNA, would include characterization of the 16 alkaline stable
dinucleotide s. Lane and his co-workers have perfected this technique;
very complex contrasts and similarities become evident when com-
parisons are made on this basis (Lane and Lane, 1971). Unfortunately,
isolation and resolution of the 16 dinucleotides is tedious and involved
(Gray and Lane, 1967). Digestion of this fraction with snake venom
phosPhodiesterase and separation of the eight resulting nucleo sides
is nearly as involved and gives less information.

The re search presented here makes possible an RNA chara-
cterization which is intermediate to the two discussed above--with
(respect to sophistication of the information obtained as well as the
ease with which it is done. The objective was a more rapid technique

by which one could obtain the 2'-_(_)_-methy1nucleo side ratio of an RNA

68

 

69

molecule for comparative purposes.

Isolation of the 2'-_g—methylnucleo side fraction from an
enzymatic digest of RNA as described in this research utilizes the
borate-complex formed with ribonucleosides to retain them on an
AG-l column. Hall (1964) also used this principle, but he used a
diatomaceous earth partition column which is more difficult to work
with than the AG-l column. Complete recovery of the 2'-_(_)_-methyl-
nucleoside standards was achieved on the AG-l columa.

The amount of RNA which can be applied to the AG-l column
under the reported conditions is restricted. Although the amount
of 2'-9_-methylnucleosides obtained from 200 A260 units of digested
RNA will suffice for mo st determinations, it would be convenient
if larger amounts could be used in some cases (eg. with RNA mole-
cules containing very low percentages of the 2'-0-methylnucleo-
sides). An initial measurement of the total percentage of 2'-0—
methylation in an RNA sample (Abbate and Rottman, 1971) should
be done to establish the amount of RNA required for the ratio deter-
minations.

The AG-l column isolation requires extra caution to insure
that deoxyribonucleo sides do not contaminate the 2'-2-methy1nucleo-
side fraction. Extensive treatment of RNA samples with deoxyribo-
nuclease during their preparation should minimize the danger of
contamination. If contamination is suspected, the 2'-_Q-methyl-

nucleoside fraction may be analyzed for deoxyribonucleosides by

paper chromatography following the GLC determination. To this
end, the TMS groups may be hydrolyzed from the nucleosides with
water or methanol.

The above research describes the successful separation of the
TMS derivatives of the 2'-_0-methylnucleo sides on a 2% OV-l7 column
of the GLC. The use of GLC for separation of the isolated 2'-_Q_-
methylnucleo sides takes advantage of the sensitivity of the hydrogen
flame ionization detector for quantitation. Model studies by Jacobson
et a1 (1968) showed that ratio determinations of ribonucleosides
can be done with as little as 0. 2 ug of each nucleoside onto the GLC
column. An even lower limit is indicated from the studies done
with standard nucleo sides in determining the copolymer ratios re-
ported above. As little as 0. 02 ug of Am may be quantitated success-
fully. The other nucleo sides would require larger amounts of material
since they have relative molar response ratios below one with respect
to Am. Cm is the limiting nucleo side with respect to detection; a
minimum injection of 0. 1 ug would be required under the conditions
presented.

Since the minimum volume of BSA which can be handled conven—
iently is in excess of the volume which must be injected, a single
ratio determination requires more material. The extra volume
permits multiple injection of a sample, a process which results in
statistically reliable data. In addition, the unused portion of a sample

may be saved for subsequent checking of an analysis by evaporating

71

the BSA and storing the nucleoside mixture.

Both Hall and Lane made use of paper chromatographic separ-
ations. Quantitation of unlabeled nucleosides was accomplished by
eluting the spots corresponding to each nucleo side (or dinucleotide)
and measuring the ultraviolet absorbance of the material recovered
from the chromatogram. The sensitivity of this approach is limited
by the amount of material necessary to obtain an accurate absorbance
reading (ca 7 ug of each nucleoside). The material eluted from the
chromatogram gives a single absorbance reading and there is no
recourse for checking its validity (i. e. the spots may not be completely
eluted). 1

The sensitivity of paper chromatographic determinations may
be greatly increased through the use of radiochemically labeled
RNA. Lane and Tamoaki (1969) have made successful use of this
technique with their characterization of l"ircumethyl labeled rRNA
molecules from L-cells.

Unfortunately, the use of labeled RNA may often require a
sacrifice in convenience. Labeled precursors must be obtained,
a process which may involve several complicated synthetic steps
or a great deal of expense. Specific samples must be prepared to
contain the label, when the isolation steps are difficult this process
Imay become tedious.

In several instances labeled RNA may be difficult if not imposs-

ible to obtain. The brine shrimp cyst is almost completely impervious

to RNA precursors (McClean and Warner, 1971). Since the oocyte

of the South African clawed toad, Xenopus laevis, undergoes oogene sis

 

eight weeks prior to ovulation, the labeled precursors must be
administered at this time to obtain labeled rRNA (Brown and Littna,
1964). RNA obtained from the organs of higher eukaryotes (eg.
mammals) cannot be labeled without complicated experimental designs.

There are some difficulties to be faced with the GLC analysis
of the 2'-_(_)_-methylnucleo side derivatives. From the low relative
molar response of Cm, it is concluded that only a small portion of
the Cm is being derivatized. Multiple derivatives form under more
severe conditions than those reported, which no derivative forms
a t all if the temperature if lower.

The formation of the methoxime derivative of Cm prior to its
trimethylsilylation with BSA results in a derivative which gives a
good response with a single, symmetrical peak on the OV—l7 GLC
column. This derivative can be used very successfully for quantitation
with synthetic copolymers which do not contain Um or U. Since the
retention times of TMS-Um and TMS-methoxime-Cm nearly coincide,
the derivative cannot be used with natural RNA molecules.

The technique is designed for the calculation of the 2'-_Q-methyl-
nucleoside ratio since the percent of the nucleosides which are 2'-9_-
methylated can be obtained rapidly and accurately with small amounts
of RNA as described by Abbate and Rottman (1971). An attempt to

obtain absolute amounts of the 2'-_Q-methylnucleosides would complicate

 

73

the procedure since an internal standard would have to be added and
since the handling of the 2'-_Q-rnethylnucleo side fraction would have

to be quantitative.

REFERENCES

Abbate,J. and F. Rottman, 1971, Anal. Biochem. In press.

Aronson,A.I. and F. H. Wilt, 1969, Proc. Nat. Acad. Sci., U. S.
62 186.

Bacher, J. E. and F. W. Allen, 1951, J. Biol. Chem. 188 59.

Balle, G., P.A. Cerutti, and B. Witkop, 1966, J. Am. Chem. Soc.
_8_8 3946.

Baskin, F. and C.A. Dekker, 1967, J. Biol. Chem. 242 5447.
Biswas, B. B. and J. Meyers, 1960, Nature 186 238.

Bobst, A. M., F. Rottman, and P. A. Cerutti, 1969, J. Am. Chem.
Soc. 91 4603.

Bredereck, H., A. Martini, and F. Richter, 1941, Ber. dtsch.
cheIn. Ges. _7_4 694.

Broom, A. D. and R. K. Robins, 1965, J. Am. Chem. Soc. _8__7_ll45.
Brown, D. D. and E. Littna, 1964, J. Mol. Biol. 8669.

Brownlee, G.G., F. Sanger, and B.G. Barrell, 1967, Nature 215
735.

Butts, W. C., 1970, J. Chrom. Sci. _8_474.

Carter, H.E. and R. C. Gaver, 1967, J. Lip. Res. _8_ 391.
Choi, Y. C. and H. Busch, 1970, J. Biol. Chem. 2421954.
Dingman, C.W. and A. C. Peacock, 1968, Biochemistry '_7_659.

Dunlap, B. E., K. H. Friderici, and F. Rottman, 1971, Biochemistry
10 2581

El-Khatib, S. M., T. S. Ro-Choi, Y.C. Choi, and H. Busch, 1970
J. Biol. Chem. 245 3416.

Forget, B.G. and S. M. Weissman, 1967, Science 158 1695.

74

75

Furukawa, Kobayashi, Kanai, and Honjo, 1965, Chem. Pharm.
Bull. (Tokyo)_l_l_’:_ 1273.

Gaver, R. C. and C. C. Sweeley, 1966, J. Am. Chem. Soc. fl 3643.

Gehrke, C.W., D. L. Stalling and C.D. Ruyle, 1967, Biochem.
Biophys. Res. Comrnun. _Z__§ 869.

Gray, M.W. and B.G. Lane, 1967, Biochim. Biophys. Acta_134_243.
Greenberg, H. and S. Penman, 1966, J. Mol. Biol. 31 527.
Gullard, J. M. and G.I. Hobday, 1940, J. Chem. Soc. 746.

Hall, R. H., 1964, Biochemistry _3_ 876.

Hancock, R. L., 1968, J. Gas Chrom. 6431.

Hancock, R. L., 1969, J. Chrom. Sci.l 366.

Hancock, R. L.,and D. L. Coleman, 1965, Anal. Biochem. _l_0_ 365.
Hashizurne, T. and Y. Sasaki, 1968, Anal. Biochem.__2_4 232.

Hecht, S. M., A. S. Gupta, and N.J. Leonard, 1969, Anal. Biochem.
_3_(_) 249.

Iwanami, Y. and G. M. Brown, 1968, Arch. Biochem. Biophys.
126 8.

Jacobson, 1M., J. F. O'Brien, and C. Hedgcoth, 1968, Anal. Biochem.
25 363.

Keller, E. B., 1964, Biochem. Biophys. Res. Commun.lz 412.
Lane, B.G. and T. Tamaoki, 1969, Biochim. Biophys. Acta 179 332.

Lawson, A. M., R. N. Stillwell, M. M. Tacker, J. Tsuboyama, and
J. A, McCloskey, 1971, J. Am. Chem. Soc. _9_3_ 1014.

Levene, P.A. and W.A. Jacobs, 1910, Ber. dtsch. chem. Ges.
_4_3_ 3150.

Martin, D. M. G., C. B. Reese, and G. F. Stephenson, 1968, Bio-
chemistry 1 1406.

76

McClean, D. K. and A. H. Warner, 1971, Dev. Biol._2_l_ 88.

McCloskey, J. A., A. M. Lawson, K. Tsuboyama, P. M. Krueger,
and R. N. Stillwell, 1968, J. Am. Chem. Soc. 224182.

Miles, H. T. and H. M. Fales, 1962, Anal. Chem. 31 860.
Moore, P.B., 1966, J. Mol. Biol. _l_8_ 38.

Moriyama, Y., J. L. Hodnett, A.W. Prestyako, and H. Busch,
1969, J. Mol. Biol. 31 335.

Nichols, J. L. and B.G. Lane, 1966a, Biochim. Biophys. Acta.
119 649.

Pierce, A. E. , 1970, 'Silylation of Organic Compounds, " Pierce
Chemical Company; Rockford, Illinois.

Price, A. R. and F. Rottman, 1970, Biochemistry-3 4524.
Rein, A. and S. Penman, 1969, Biochim. Bi0phys. Acta. 190 l.
Ro-Choi, T. S. , 1971, personal communication.

Ro-Choi, T. S., Y. Moriyama, Y.C. Choi, and H. Busch, 1970,
J. Biol. Chem. 245 1970.

Ro-Choi, T.S., R. Reddy, T.Takano, D. Henning, and H. Busch,
1971, Fed. Proc. 22 1101.

Rottman, F. and K. Heinlein, 1968, Biochemistry l 2634.
Rottman, F. and K. L. Johnson, 1969, Biochemistry, 8 4354.
Sasaki, Y. and T. Hashizume, 1966, Anal. Biochem. 16 1.

Shaw, S. J., D. M. Desiderio, K. Tsubayama, and J.A. McCloskey,
1970, J. Am. Chem. Soc. 22 2510.

Singh, H. and B. G. Lane, 1964a, Can. J. Biochem. 43: 87.
Singh, H. and B.G. Lane, 1964b, Can. J. Biochem. :15 1011.
smith, J. D. and .D. B. Dunn, 1959, Biochim. Biophys. Acta: 73.

Soeiro, R., M.H. Vaughan, J. R. Warner, and J.E Darnell, Jr.,
1968, J. Cell Biol. 39 112.

77

Starr, J. L. and B.H. Sells, 1969, Physiol. Rev. 432 623.

Sweeley, C.C., R. Bentley, M. Makita, and W.W. Wells, 1963,
J. Am. Chem. Soc. _8_5 2497.

Sweeley, C. C. , ‘B. D. Ray, W1. Wood, and J. F. Holland, 1970,
Anal. Chem. 4; 1505.

Vaughan, Jr., M. H., R. Soeiro, J. R. Warner, and J.E. Darnell, Jr.,
1967, Proc. Nat. Acad. Sci., U. S. 28 1527.

Wagner, E. K., S. Penman and V.M. Ingram, 1967, J. Mol. Biol.
29 3‘71.

Weinberg, R. A., U. Loening, M. Willems, and S. Penman, 1967,
Proc. Nat. Acad. Sci., U. S. 5_8_ 1088.

Weinberg, R.A. .nd 5. Penman, 1968, J. Mol. 13101. _3_§_ 289.

Weinberg, R. A. and S. Penman, 1969, Biochim. Biophys. Acta
190 10.

Zapisek, W. F., A. G. Saponara, and M. D. Enger, 1969, Biochemistry
8 1170.

Zimmerman, E. F. and B.W. Holler, 1967, J. Mol‘. __2_3 149.

MICHIGAN STATE UNIVERSITY LIBRARIES

II | I IIIIIIIIHIIH

3 119303174 5296