105
038

 

11415513 1 -- “and i

I n47 3 aufg“::)y
4%ng 3’3:
'6 k 4...,“ “:2?
UR“; §V§f¢€ “Mn+__'
v-"vikug _-,.~9’

 

 

_.._‘_
""‘lfimz .. J
‘

This is to certify that the

thesis entitled
DEVEIDPMENT OF A GRGJING DROP FLIDRESCFNCE DETECTOR

presented by

Claude Bruce Coffin
has been accepted towards fulfillment

of the requirements for

Masters of Science—degree in Biochemistry

jute/l. Ream

4mm professor

Date W

07639

 

 

 

 

OVERDUE FINES:
25¢ per day per item
RETURNING LIBRARY MATERIALS:

Place in book return to remove
charge from circulatton records

 

DEVEIDPMENT OF A m DROP FLLWESCENCE DETECTOR

By

Claude Bruce Coffin

A THESIS

Suhnitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Dapartment of Biochemistry

1981

ABSTRACT

DEVEIDPM'INT OF A GRENING DIDP FLUORESCENCE DE'I‘ECIOR

By
Claude Bruce Coffin

A fluorescence detector designed for use with High Performance
Liquid Cinematography has been constructed. It measures the fluores-
cence of a growing drop, eliminating the chranatographic and spectro-
scopic limitations imposed by the use of a flow cell. ,

A chranatographic method has been develoPed to analyze the
ocmposition of the cap structures of messenger'RNA. The method
employs an enzynatic digestion of the messenger RNA, concentration
on a strong anion exchange oolum, and subsequent separation by
reverse phase chromatography. Detecticn is acccmplished by fluores~
cence using the growing dmp detector.

IEDICA‘I‘ION

his thesis is dedicated to Pam, Chris, and Robbie.

ii

'lhe author is deeply indebted to Dr. Fritz Rottman for his
guidance and persmal interest throughout the course of this research.
'Ihe author also wishes to ezqaress his appreciation for the help and
moom'agemmt of Drs . Jack Holland and Ron Patterson .

'lhe author wishes to express his gratitude to the faculty of
the Department of Biochemistry, Michigan State University, for their
tutelage, and to the Department itself for the financial aid and
experience gained as a Graduate Teaching Assistant.

The author wishes to express his deep appreciatim to his wife and
family for their support and encomagement during the course of
his studies. 'Ihe author also wishes to express his gratitude to
Karen Friderici and Dr. John Nilson for the friendship, help, and

encouragement that they gave to him.

iii

TABIEOFWI'ENI‘S

Page

List of Tables ..................................................... v
List of Figures .................................................... vi
List of Abbreviations .............................................. vii
Introduction ....................................................... 1
lbthods and mterials .............................................. 13
Results ............................................................ 36
Discussion ......................................................... 56
Bibliography ....................................................... 65

iv

LIST OF TABLES

Page

I . Corrected respmse of fltores oence detector
with regard to omlposition of cap structures ................. 42

LIST OF FIGURES-—

Page
1. 5' terminal end of emeryotic mRNA ............................. 4
2. 7-methylguanosine of the cap structure ......................... 6
3. Physical layout of components of growing drop
flmrescence detector .......................................... 15
4. Optical system of growing drOp fluorescence detector ........... l7
5. Illustration of liquid and ..................................... 20
Dripper needle assembly ........................................ 22
7A. Schematic of timing circuit .................................... 24
B. Timing sequence of the detector ................................ 26
8. Schematic of amplification circuit ............................. 28
9. Schematic of timing circuit .................................... 30
10. Schematic of HPLC and detectors ................................ 34
11. Response as function of amomt of
2 , 5-diphenyloxazole injected ................................... 38
12 . Corrected fluorescence as fimctim of flow rate ................ 40
13. Chramatograms of cap standards
A. using absorbance detection .................................. 44
B. using fluorescence detection ................................ 46
14. Chramatograms of poly A(-) RNA
A. using absorbance detection .................................. 49
B. using fluorescence detection ................................ $1
15. Chromatograms of bovine pituitary poly A(-)RNA with cap standards
A. using absorbance detection ................................... 53
B. using fluorescence detection ................................. 55

geegsaég

91*
PP

LIST OF'ABBREVIATIONS

Ribonucleic acid

lbssenger Ribonucleic acid

Gas Chromatography

liquid Chramatography

High Performance Liquid Chromatography
Ultraviolet

Photomultiplier

Transistor-transistor logic

Inside diameter

Outside diameter

INTRODUCTION

'Ihe transfer of genetic information fram DNA to protein has long
been an area of intense biochemical interest. As the general processes
and mechanisms of transcription and translation have been established,
much attention has focused upon the control mechanisms involved in
these processes, resulting in a considerable body of knowledge
omoerning the regulaticn of gene expressim.

It has became evident that gene regulation in eukaryotes occurs
at both tl'e transcriptional level, involving the turning on of the
gene and the beginning of transcription, and the post-transcriptional
level where the immediate gene product, a high molecular weight RNA,
is processed into the much smaller, covalently mdified, mature
mRNA (1) . These post-transcriptional processing events include excision,
splicing, addition of polyadenylic acid to the 3' end of the mlecule,
and modification of the 5' end of the molecule to form cap structures.

The purpose of this investigation is to deve10p an analytical
method for the efficient analysis of the camposition of the 5'
terminal cap structures of eukaryotic mRNA.

It was first reported in 1974 that eukaryotic mRNA, unlike
prokaryotic mRNA, is methylated (2,3) but to a much lesser extent than
either ribosomal or transfer RNA. Hydrolysis of eukaryotic mRNA
yielded the four cannon nucleosides methylated at the 2' position

6

of the ribose ring, N -methyladenosine, and a nucleoside later

1

2
identified as 7-methylguanosine. The results of alkaline hydrolysis
indicated that most, but not all, of the methyl groups appeared in
alkali resistant oligonucleotides . The general structure of these
oligonucleotides was proposed by Rotmen et a1. (5) to be a 7-methyl-
guanosine at the 5' terminus of the mRNA linked in a 5'—5' manner to
one or two adjacent 2 '-O-methylnucleosides via a perphosphate bcnd.
These structures are oamonly called cap structures (Figure 1). Cap
structures have been found cm the mRNA of many eukaryotic species and
viruses that infect eukaryotes.

Pmperties of the cap structures

The cap compounds have some interesting attributes in camon due
to their unique structure. 7—methy1guanosine occurs at the 5' end of
almost all eukaryotic mRNA's md is the (lily place in the message that
it is fund. The methylatim of guanine at the seven position
introduces a positive charge (11 the imidazole ring (Figure 2) and
whai 7,9 disubstituted, as in 7-methylguanosine, lowers tl'e pKa of
the N-l nitrogen by two pH mits (6) . Mtlermore, the glycosidic
bard of 7-methylguanosine is less stable than that of guanosine (6) .
At neutral and alkaline pH's, 7-methylguanosine will ring-0pm
between the 8 and 9 positions of the imidazole ring forming
2-amino-4-hydroxy-5- (N-methyl) fonmamido-6-ribosylaminopyrimidine .
Finally, 7-methylguanosine has a fluorescence quantum efficiency
about two orders of magnitude greater than the non-methylated
molecule.

The phosphodiester linkage in the cap structure is very Lnusual

and provides the cap structures with several interesting features.

Figure l. The 5' terminal end of
eukaryotic messenger RNA. The structures m7GpppN
are referred to as cap 0's, m7Gppme as cap 1's,

and m7Gppmeme as cap 2's.

 

5'-TERMINAL END
0 CH3

II I+

HNA/I[N\
HZN’KN N

/CH2 :0:

OH OH

 

I I l
<9 <9 <9
-o-u—o-'p-o—-p-o
II
o o o

O
I
N

0
Z

 

l'

 

.mousuonuum moo on» mo ocflmocmamamsum5|n one

.m muzmwm

 

 

z

 

rally——
0

 

N wmzmzu

7
First, it links the tenminal 7-methylguanosine in a 5'-5' manner to

the penultimate nucleoside. The result of this 5'-5' linkage, the
presence of 2',3' cis hydroxyls at each end of the mRNA, has been
utilized effectively in various investigations including elimination
of the cap by periodate oxidation and ’5 -elimination (8,15,16,17),
radiochemical labeling of the cap by periodate oxidation and reductian
with al'I-borohydride (8,9), and selective retention of both ends of the
mRNA via affinity chromatography using a borate stationary phase (10,11) .
Biologically, the 5'-5' linkage may enhance the stability of the mnRNA
by preventing the action of 5' exonucleases (12). Secondly, the
length of the phosphodiester band is important to the conformation of
tie nucleosides in the cap structures. Studies using proton NMR
suggest that, at least for the cap structure m7GpppAmpA, the
7-methylguanosine can bend about the phosphodiester bond and
intercalate between the two adenosine moieties (12). This unusual
arrangement of the bases , providing hydrOphobic and ionic regions

at the end of the mINA, requires a long, flexible spacer such as

that provided by the three phosphates .

The presence of 2'-O-methyl groups in cap structures is
implicated in the canformation of the cap and the possibility of
base stacking (13) . It causes the phosphate backbone to elongate and
provide efficient space for the terminal 7-methylguanosine to

intercalate between the adjacent bases (13).

Biological role of the cap structures
The occurence of the unusual cap structures at the 5' end of

nearly all eukaryotic mRNA has stimulated numerous investigations into

8
tie possible biological role of these structures. Many of these

investigations have focused upon the possible role of cap structures
in protein biosynthesis.

The first report concerning the requirement for the methylation
of mRNA for efficient translaticrn described the comparison of
methylated vs unmethylated viral mRNA in an in vitro , wheat genm
cell-free translation system (14). It was reported that the methylated
mm stimulated portein synthesis much more effectively than the
otherwise identical unmethylated mRNA. Chemical reroval of the
7-methylguanosine by periodate oxidation and fl -elimination with
aniline was also shown to destroy the translational ability of
mRNA (8) . Subsequent studies using enzymatic renoval of the
7-methylguanosine (18,19) and more gentle conditions for the
,6 -eliminationn of the 7-methy1guanosine (15) have reinforced the
findings that the cap structure is required for efficient translation
of eukaryotic mRNA.

Cap analogues have also been found to affect the translational
ability of capped mmA' s. 7-methylguanylic acid and similar analogues
were fomd to inhibit the translation of a variety of capped mRNA's
(20,21,22) but have little effect an the translational ability of
normally mcanped mRNA (23) .

In additian to the role that the cap plays in protein synthesis,
the very early capping of HnRNA suggests that the cap might have
important roles in mRNA synthesis and mRNA stability.

Separation mefinods

The development of analytical methods for fine analysis of
molecules extracted fram biological samples is a difficult task due
to fine camplexity of the matrix. The generally accepted approach to
this problem is to selectively isolate the molecules in question. To
this end, a variety of separation techniques have been developed but
few have been so productively utilized as chrcrmatography in general and
liquid chromatography in particular.

Chrcmatography cansists of a two phase system in which separatian
occurs as a result of a differential affinity for the phases. Gas
chranatography, GC, matured earlier than liquid chromatography, LC,
and separations due to very minor structural differences have been
achieved by finis method. The most significant limitation of» GC
wifin respect to the separation of material with a biological origin
is that few biomolecules are volatile, a requirement for CC.
Canversely, liquid chromatography is well suited to fine separatian of
non-volatile, large, and ionic canpounds, such as finose found in a
biological sample.

The first sys tenatic development of a chromatographic separation
is attributed to Tsvet (24) when he separated various biochemicals
using column liquid chromatography in the early 1900's . Chrcmatography
developed at a slow and uneven pace through the first part of the
twentieth century. Although column LC was the earliest form, it
developed rather slowly by carparision to gas chromatography, paper
chromatography, and thin layer chranatography until the advent of
high performance liquid chromatography, HPLC .

HPLC is an evolutionary technique which builds upon the theory

10
and successes of Open column LC as well as fine other chromatographic

techniques. Central to the development has been the design of fine
column and column packing materials. These packing materials,
typically based npon a silica gel support, are characterized by a
smell particle size (3-20 micron), a very large surface area (200-800
mzlg), and a uniform, highly characterized stationary phase. These
paraneters have necessitated the development of precise, high pressure
pimping systems and in jectian valves , and low dead volume continuous
flow detectors. The one advance, however, primarily responsible for
the rapid growfin in I-IPLC is the developmt of bonded phase and
especially bonded reverse phase packing materials . Reverse phase
is a generic term encompassing a wide range of statiunary phases
typically cansisting of straight chain alkyl hydrocarbons covalently
bound to a porous silica gel (25) . A wide range of selectivities
are available wifin reverse phase material depending upon the
hydrocarbon chain length, banding chemistry, percent derivatizatien,
pore size of the silica gel, and abundance of residual silanols (26).
Incomplete understanding of the mechanians involved in retentian
on a reverse phase material has made selection of a material suitable
for a given analysis difficult. Fortunately, modification of the
mobile phase (27) and additionn of various modifiers, e.g. ion pairing
reagents (28) , will almost always offer sufficient selectivity to
effect the separation.

Chromatography of nucleic acids

 

The first reported use of column LC for the separation of nucleic

acids was by Cohn (29,30) over thirty years ago using Dowex anion

11
exchange resins. A decade later, Andersen reported the first use of

I-IPLC for the analysis of nucleic acids (31,32), again using Dowex
resins. Limitations of fine physical properties, particularly
canpressibility, of the gel materials led to the development of
solid core icrn exchange pellicular materials (33) in the late 1960's
and porous silica gel materials in the early 1970's (34) . These latter
materials have became standard for the analysis of nucleic acids (35) .
The use of ion exchange chromatography is not without difficulties
however and is subject to problem with stability and reproducibility.
Recently, reverse phase chromatography has been applied to the
analysis of rnucleic acids (9,36-42). The wide range of available
statiunary phases and the selectivity available through alteration and
modification of the mobile phase makes possible analysis of compounds
as widely divergent as positively charged bases (40) and highly
negatively charged cap structures (42) .

Fluorescence detection

 

The application of fluorescence detection to HPLC has proven
particularly successful for the trace analysis of many naturally
fluorescing (40) and derivatized campounds (43,45) . The desirability
of HPLC fluorescence detection results from the inherent sensitivity
(often two orders of magnitude more sensitive than UV absorbance
detectian) and selectivity of fine phenamena (44) .

In general, all systems designed for the measurement of fluoresw
cence contain certain components. These include : a) a UV light source,
b) a device to select the excitation wavelength, c) a sample to be

monitered, d) a device to select the emissian wavelength, e) a

12
photo-sensitive device to measure the flnoresced light, and f) .

electronics to convert fine data into a form convenient to the user.

In addition, application of a fluorometer to analytical HPLC detection
introduces the constraints that the illuminated sample must have a
small volume and the response time of the measuring system must be
rapid. Commercially available analytical fluorescence detectors meet
the above requirements in a variety of ways. They all have me
componnenet in cannon, however, a flow cell. The effluent from the
column is typically directed to a quartz flow cell with a volume of

10 pl to 100 V1. The design of the flow cell, especially with respect
to flow characteristics and clearance times , and material of construction
are critical to the performance of the detector. Significant improve-
mnents have been made by Martin et a1. (50) in fine design of a free
falling drop flnorescence detector in which a drop falls through a
light beam and also by Diebold et a1. (46) in thedevelopment of

a suspended drop laser flnorcmeter.

An HPLC fluorescence detector has been designed finat employs a
dripper needle suspending a growing drop. The use of a growing drop
rather than a drOp falling through a beam allows a much longer time
for the drOp to be illuminated. This enhances the signal to noise
ratio and results in a very low dead volnme. In this study, fine
sensitivity and selectivity of HPLC with fluorescence detectian is

utilized to deve10p a mnefinod to analyze the cap structures of
eukaryotic mRNA.

MEI‘I-DDSANDMATERIALS

Detector design; and construction
Optical System

The Optical system of the growing drop detector is represented
schematically in Figure 3. The light source consisted of either a low
pressure mercury-argon lamp (Oriel Corporation, Marlboro, Mass . , model
# C-l3-62) and power supply (Oriel model # C-73-l6) or a high pressure
mercury miniature arc lamp (Oriel model # 6281) and power supply
(Oriel model # 6240-3) . The light source is positioned either three
centimeters in front of the interference filter (low pressure mercury
lamp) or two meters in front of the interference filter (higln pressure
mercury arc lanp) . The difference in distance from lamp to filter
results from the presence of a colimating lens on the high pressure
arc lamp housing. The image of the lap or arc is focused immediately
belov fine dripper needle in both systems by a pair of UV grade fused
silica lenses (Oriel models A-ll-621-lO 0.5 " diameter, 25 mm focal
length and A-ll-651-30 1.5" diameter, 100 mm focal length) (Figure 4).
The excitation wavelength is selected by a 1.0" circular UV bandpass
interference filter (Ditric Optics Inc. , Hudson, Mass.) with a peak
percent transmission of 10 7.. at 254.0 nm and bandwidth of 7.0 nm. The
emission wavelength is selected by a 1.0" square three cavity band pass
interference filter (Ditric Optics Inc.) with a peak percent trans-

mission of 38 7.. at 370 mm and a bandwidth of 10.8 mm. The

13"

 

14

.uouoouoo oocooaocoaau mote e:_3:cr

ecu mo mucocooeoo on» no uoo>oH acoam>co one .m case“;

 

55.“. o.

 

2n—

 

 

 

mmSCoN “U

o o

mzmn
w2mn ”Hunuue

IOHOwau—mm

m McDOI

 

 

exam... 5.1.03

 

 

16

oouo ocfi3ouo ecu mo Eoummm HMOfluoo one

.Houoouoo oocoomouooam

.v wuoowh

mzm: o.

 

l7

 

v mmDOE

 

whommz
amen. .mo

18

photomultiplier tube (EMT Gencom, model 9789B) is of the end-on type
with eleven stages and a 1.0 on window. The photomultiplier is driven
by a high voltage power supply (Kepco, Flushing, N.Y. , model ABC) at
850-900 volts.

Liquid and

The liquid end of the detector is fabricated from a single block
of aluminum and is illustrated in Figure 5. The dripper needle
consists of 1/16" 0.D.x 0.010" I.D. stainless steel tubing (The Anspec
Conpany, Ann Arbor, Mi.) over which a piece of 1/8" 0.D. x l/l6" I.D.
teflon tubing (The Anspec Co.) has been fitted. The end of the dripper
needle is immediately above the focused image of the mercury lamp and
coaxial with the photomltiplier tube, see Figure-6. Opposite the
photomultiplier tube is a UV enhanced parabolic reflector (Melles Griot,
Irvine, Ca., model M00 009/028) positioned such that the drop is at

_ one focus.

Electronics

A block diagram of the electronics and a timing diagram are presented
(Figure 7) and also schematics for the amplification circuit (Figure 8)
and timing circuit (Figure 9).

The operation of the detector utilizes both the fluoresced and
reflected liglnt. As the illnminated drop grow on the dripper needle,
the current generated by the photomultiplier tube is amplified by a
high impedance, FET input Operational amplifier (Analog Devices,

Norwood, Pass. , model AD523J). The signal from this amplifier is
filtered (Analog Devices, model AD504J) and input into a modular

19

Figure 5. Illustration of liquid end.

fi'--1
I I
n__J
'i
l

A O

I
1

:n
I
I
_I

-l——

I

I

I
-r--

 

l

 

 

 

 

 

e 4
a.“ \Hl 1.: /
\
II» we
/ \
/TL.%
Ii TAIIP
TL
rm:

 

 

 

 

 

 

gm

 

 

 

 

 

 

 

 

ll

a tn .1 a a
Mall/L .1 a

nun: \ run 4&%

#1....th .Mrnh” .. a. #
. I\ _ r. #
.nl.m.nlt

 

 

21

Figure 6. The dripper needle of the

growing drop fluorescence detector.

FIGURE 6

STAINLESS
STEEL

 

 

 

 

TE FLON

 

_————— — _ —.—._
~_—————-—_-

 

LAMP
IMAGE

23

 

.uwoouao ocHEau on» mo owumEocom

.<h manuam

24

<5

  
   

mmEOEmE

mmmHZDOO

 

.EDOEO
02:22.

<5 mango

>I

25

.Houomuoo oocmomouooam mono

ocflzouo on» mo oocooomm oCHEHu one .mh masowm

26

 

 

9:4

 

 

 

 

 

 

mm mmDGE

mmE.

mom

mm SK

m0e<m<o200

.2205
Dme<Fmemuufi

oOmn

27

.uasouflo cofiucowwwaosm one no owumEooom

.m muooflm

FIGURE 8

"-‘TIMING CIRCUIT

 

 

7404

 

 

A0452

 

28

 

 

 

 

 

 

 

 

 

 

H

II

7|

 

 

 

 

PM

 

 

Hv—J l

 

‘5
O
LLI
o—t
== I
no
4.
r.
,_
N
no
f\
O
<

 

 

 

 

l l

 

 

 

VLIVL

 

 

 

l9I-t'l.

 

llllll

 

 

I7L 17L

 

IQLVL

 

 

 

 

 

l
///J \\ // \\\\

 

L9 LVL

 

 

 

 

IQLVL

 

 

 

 

 

 

I9 Wt

 

 

 

 

 

 

29

.uwoouflo ocflfifiu ecu mo oeuoEocom

.m musoflm

30

 

Om: - ‘Hll

mmszEwEl If) r — . —IllFl
Om: F mNPE. In—
J L _ +

.221.
AL aA

o MEDOE

 

 

 

 

 

 

 

 

 

 

 

m:

 

 

 

 

 

 

31

voltage to frequency convertor (Analog Devices, model AD452J) with a
maximum 100 KHz frequency response. The components for fine amplification
and voltage to frequency conversion are mounted in the photomnltiplier
enclosure and immediately adjacent to fine photomnltiplier. The
remainder of fine amplification circuit is housed with the timing circuit
in a separate enclosnm'e. All components unless otherwise specified are
standard 7400 series TIL. The signal from the voltage to frequency
convertor is divided by two and converted to a symmetrical square wave
by a flip-flop (SN7474) and used to clock a series of five 4 bit
binary counters (SN74161). The output of the last finree counters is
input to two hex menories (SN74174) . The data from these memories is
input onto a 12 bit digital to analog convertor (Analog Devices , model
7521JN) . The output of the digital to analog convertor is offset by
biasing an output amplifier (Motorola, model 741 GP) so that fine
signal ranges between 0-10 mV. V

The timing circuit obtains fine buffered signal from the photomulti-
pl ier and differentiates it (Analog Devices , model AD504JH) then
compares this signal to a reference potential. As the chap is slovly
growing, the amount of light passed to the photomultiplier increases
resulting in an increasing current to the first amplifier. As the
drOp falls, the positive voltage spike from the differentiator turns
the comparator on momentarily which triggers a monostable vibrator
(SN74123) . Qne output of this monostable, Q, triggers a front panel
LED and resets fine counters of fine amplifier circuit while fine comple-
mentary output, 0', triggers a second monostable (XRSSSCP) which stays
in the ON state for a period determined by a front panel potentiometer.

This monostable determines the integration time. As it turns off, it

32

triggers a third monostable which clocks the mnerories so that they
latch fine binary encoded output of the connters .

High performance liquid chromatoth
Chromatograph

The HPLC represented sdnematically in Figure 10 consisted of a
Milton Roy miniPump (Laboratory Data Control, Riviera Beach, Fla. , model
396-31), 0-5000 psig pressure guage (Ashcroft), Valco sample injection
valve. (Valco Instruments, Houston, T'x., model CV-6-UHPa-N60) with a
Bromlee guard column holder and anion exchange cartridge (The Anspec
Co.) in place of the sample loop, an ultraviolet absorbance detector
(Laboratory Data Control, model 1203, 254 nm) and a strip chart recorder
(Linear Instrument Co., Irvine, Ca. , model 255).

Column

The packing material used during this investigation was 10 micron
Whatman ODS-3 (Whaman Inc., Clifton, N.J.). This material was
pnmchased in bulk and packed downward using a 10 7; slurry of ODS-3
in 25 7; glycerol: 75 ‘7. methanol. The packing material was sonicated
for one minute finen quickly introduced into the slnnry reservoir and
packed into the analytical column at 10,000 psig using an HPLC column
packing apparatus (The Anspec Co.) utilizing a Haskell pneumatically
amplified air driven purp. The eluting solvent was HPLC grade mefinanol
(Burdick and Jackson, Muskegon, Mi.) . The eluting solvent was pumped
through the column for at least one hour at 10,000 psig and the bed
consolidated several times . The column blank was made from 316
Li-Chroma stainless steel (Handy and Harmon), 0.25 inch 0.D. by 2.1 mm

33

.oom: muouoouoo mo ucofimocmuum one one coonooumfiouco

ofloowa moonfiuownoo goes on» NO oeumEooom .OH ousowm

34

mew/:5

 

 

 

 

 

 

oOhOmhmc
mozwowmmODau

 

 

 

mOHOwwa
w oz<m m0wm<

 

 

 

 

2.23400

or mono:

 

 

mine

 

 

35
I.D. by 25 on, cleaned fioroughly and fitted with Valco column and

fittings (Valco Instruments). Frits consisting of a 0.125 inch
diameter fritted stainless steel pressed into a 0.25 inch diameter
Kel-F disc (The Anspec Co.) were used as bed snpports.

Pbbile phases

All solvents used were HPLC grade. The water used was double
distilled (second distillation in glass) and filtered through a
0.45 micron filter (Millipore Filter Corp., Bedford, Ma. , type HA).
PhOSphate buffer was made from analytical reagent grade phosphoric
acid (Mallinckrodt, St. Toms, lb.) and potassium hydroxide
(Mallinckrodt) and filtered in fine same manner as the water.

Digestion of RNA

The digestion was carried out according to the procedure outlined
by Albers et a1. (42) . Ten A26Ounits of poly A(-) RNA were digested
with 20 units of T2 RNase (Calbiochen) and 20 lag of P1 nuclease
(Yama Shovu Co.) in 400 l of 10 mM sodium acetate, pH 5, for three
hours at 37°C. The pH was adjusted to 9 with 10 mM ammonium hydroxide
and 5 units of dialyzed bacterial alkaline phosphatase (Worthington
Biochemical Corp.) was added and the reaction mixture incubated
for fortyfive minutes at 37'C. The solution was neutralized with
phosphoric acid and then spiked with a mixture of cap compounds
(P-L Biochemicals).

RESULTS

Linearity

The response of the growing drop detector as a function of amount
of 2 ,5-diphenyloxazole (Research Products Inter.) injected is displayed
in Figure 11. A 1 mg/ml stock solution of the 2,5-diphenyloxazole was
prepared and 1:10 serial dilutions made from it. A 100 1 sample loop
was used and injection volumes were limited to 10 to 50 1 volumes.
The mobile phase was 100 7. acetonitrile (Burdick and Johnson, UV) and
fine pump output was 0.27 ml/min. A 50 cm length of 1/16 inch 0.D. x
0.30 inch I.D. teflon tubing was used in place of the column to
minimize dead volume. The potential applied to the photomultiplier
was either 700 or 800 volts depending on the amount of solute injected.

A linear regression analysis yielded a y-intercept of 0.61 and a

correlation coefficient of 1.0.

Fluorescence as a function of flow rate

The corrected response of the detector, measured as the ratio of
fluorescent signal to absorbance signal, as a function of the drOp
period is presented in Figure 12. These data were obtained from the
isocratic separation of fine four cap 0 compounds (m7GpppN where N-A,U,
G, C) plus m7Gppme and m7GpppCmn (P-L Biochemicals). The mobile phase
was 50 mM potassium phosphate, pH 5.6, 1 g/l tetrabutyl amnion
hydroxide (Aldrich and 0.4 7. tetrahychofuran (Burdick and Jackson)

36

37

Figure 11. Response as a function of the

amount of 2,5-diphenyloxazole injected.

FLUORESCE NCE

100

50

 

O

O

38

FIGURE 11

 

2

2, 5-Dl PH ENYLOX AZOLE

3

(ng)

39

Figure 12. Corrected fluorescence as a
function of flow rate. The flow rate of the HPLC
was varied yielding drop periods from 1.2 seconds
to 10.7 seconds per drop. The corrected fluores-
cence was considerably different when the pen-
ultimate nucleoside was a purine compared to a

pyramidine.

40

 

 

FIGURE 12
2
.0
f PYRIMIDINES
o
1 o
PURINES
o
o
o 1 2 3 4 5 e 7 s 10
DROP PERIOD (sec/drop)

41
and the stationary phase was Whatmnan ODS-3 in a 2.1 mm I.D. by 25 cm

column. The absorbance detector and fluorescence detector were in
series with the absorbance detector first and the potential applied to
the photomultiplier was 880 V in all cases . The absorbance detector
(laboratory Data Control) was attenuated to 0.032 A254 full scale.

Differential response of fine caps

The corrected response of the fluorescence detector as a function
Of fine conposition of the cap structures is presented in Table I.
These data were Obtained from the isocratic separation of fine six
cap structures described in the section on fluorescence as a function

of flow rate .

Sgparation Of the cap compounds
Cap standards

Figure 13 depicts chromatograms of eight cap standards using the
A) absorbance and B) fluorescence detector respectively. The analytical
column was equilibrated with 50 mM potassium phosphate, 1 g/l tetra-
butyl anmonium hydroxide, and 0.4 7. tetrahydrofuran at a flow rate of
0.3 ml/min. The anion exchange column was switched out of the mobile
phase stream and flushed with 5 ml. Of water. A total Of 0.1 A254 of
cap standard mix was injected onto the anion exchange column and fine
column flushed with 5 ml Of water, removing all negatively charged
material. 'lhe cap mix was then injected onto the analytical column.
Upon elution Of m7GpppA the mobile phase was changed to 50 mM potassinm
phosphate, pH 5.6, 1 g/l tetrabutyl ammonium hydroxide, and 2.4 7.
tetrahydrofuran.

Table I. Corrected response of fluorescence detector as a function

of fine composition Of the cap structures.

 

 

 

 

 

DrOp period Average fluorescence Ratio average pyrimidine
absorbance average purine
4.7 sec/drOp pyrimidines 1.0
3.3
purines 0.30
8.7 sec/drop pyrimidines 1.4 2
.7
purines 0.52
9.0 sec/drop pyrimidines 1.7 3 3
purines 0.51 .
10.6 sec/ch'Op pyrimidines 2.0 3
.0

purines

0.66

43

Figure 13A. Chromatogram Of the cap
standards using absorbance detection. The mobile
phase was changed from A, 50 mM KH2P04 pH 5.6,

1.0 g/l TBA, 0.4% THF to B, 50 mM KH P04 pH 5.6,

2
1.0 g/l TBA, 2.4% THF.

FIGURE 13A

E<mEooo0~E

E<ooo ORE Illufllnuly

 

EOoooOnE

 

Doggone

 

 

Oooomqu

 

 

 

 

 

0.032

0.016

w02<mm0mm<

TIME (hr)

Figure 138. Chromatogram of cap
standards with fluorescence detection. The mobile
phase was changed from A, 50 mM KH2P04 pH 5.6,

1.0 g/l TBA, 0.4% THF to B, 50 mM KH2P04 ph 5.6,

1.0 g/l TBA, 2.4% THF.

FIGURE 13B

 

 

 

 

 

 

EOQQQGNE

Dana 05.:

 

 

 

 

 

l3

l2

 

 

wOZ w OmmeDJ n.

TIME (hr)

47
RNA blank

A diromatogran of poly A(-) RNA is presented in Figure 14.
A 3.0 A260 sanple of bovine pituitary poly A(-) RNA digested as
described above was loaded (nto the anim exchange column and
injectim mto the analytical oolum was accomplished as described in
the previous section an the cap standards.

RNA and cap stmdards

Figure 15 illustrates a separatim of cap standards injected
along with 3.0 A260 of INA. 0.1 A260 of cap standards plus 3.0 A260
of bovine pituitary poly A(-) RNA digested as described previome was
chranatographed in the same marmer as in the previous section.

Figure 14A. Chromatogram of poly A(-)
RNA with absorbance detection. The mobile phase
was changed from A, 50 mM KH2P04 pH 5.6, 1.0 g/l
TBA, 0.4% THF to B, 50 mM KHZPO4 pH 5.6, 1.0 g/l

TBA, 2.4% THF.

ABSORBANCE

FIGURE 14A
0.032

 

 

 

 

 

 

 

 

 

0.016

 

 

 

 

 

 

 

AB

 

 

 

TIME (hr)

50

Figure 143. Chromatogram of poly A(-)
RNA with fluorescence detection. The mobile phase

was changed from A, 50 mM KH PO4 pH 5.6, 1.0 g/l

2

TBA, 0.4% THF to B, 50 mM KH PO4 pH 5.6, 1.0 g/l

2
TBA, 2.4% THF.

FLUORESCENCE

 

FIGURE 148

AB

 

TIME (hr)

N—

w—

52

Figure 15A. Chromatogram of bovine
pituitary poly A(-) RNA with addition of cap
standards using absorbance detection. 3.0 A260

poly A(-) RNA was injected with 0.1 A of

260
cap standards. The mobile phase was changed

from A, 50 mM KH PO4 pH 5.6, 1.0 g/l TBA, 0.4%

2

THF to B, 50 mM KH PO4 pH 5.6, 1.0 g/l TBA, 2.4%

2
THF.

FIGURE 15A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r3

 

 

woz<mm0wm<

I: I\
H girl
IwLm
8095055
GQQQGBE
Danna}. .H
a
1'0
. L _
2 6
3 1..
0 0
0. 0.

TlME(hr)

54

Figure 153. Chromatogram of bovine
pituitary poly A(-) RNA with addition of cap
standards using fluorescence detection. 3.0 A260
poly A(-) RNA was injected with 0.1 A260 of
cap standards. The mobile phase was changed

from A, 50 mM KH PO pH 5.6, 1.0 g/l TBA, 0.4%

2 4

THF to B, 50 mM KH2P04 pH 5.6, 1.0 g/l TBA, 2.4%

THF .

FIGURE 158

 

 

 

 

 

 

 

 

 

 

 

 

 

E<wfiooo0~€ Hr:
E
E<aao0~E
<QQQO~E
EOQQQQNE
Sodom ONE
Dead 05.:
09:55:

r3

'2

r0

 

 

wozmowwm 0:4...

TlME(hr)

DISCIBS ION

Detector design
Optical

'Jhe desigi of the growing drOp fluorescence detector has been
strongly influenced by the absorbance and fluorescence characteristics
of 7-uethylguanosine (47), the mique canponent of mRNA cap structures.
The selection of a mercury lamp as the source of the excitation
radiaticn was based upon its strong anission line at 253.7 nm.
A 100 watt high pressure mercm:y arc lamp was initially selected
because of its high total output and a very small arc size but was
superceded by a 4.6 watt low pressure mermrry-argon lamp. The
low pressure lamp has a much higher output at 253.7 nm relative to
370 nm, the respective excitation and anission wavelengths of
7-methylguanosine. This change resulted in a significant
reduction in the background and an increase in sensitivity. One
disadvantage of the low pressure lamp is that it is not a point source
but rather has an illuninated lmgth of nearly one-half inch,
resulting in a line being focused on the drop rather than a spot.
'Jhe excitaticn wavelength is selected by a single interference
filter with the maximin percmt transmission at 254 nm. Multiple
interference filters, Optical filters, and canbinations of the two

were tried but invariably yielded lower signal to noise ratios .

5'6

57

(lice light of the proper wavelength has been selected, focused
an the drOp, and absorbance and fluorescence have taken place, the
fluoresced light must be gathered and directed to the photomultiplier.
The initial design of the detector used an integrating sphere with a
reflective surface to gather the fluores ced radiation . This is a
satisfactory arranganent except that ornly a small percentage of the
fluoresced light is mnitered by the photamltiplier. The current
design uses a relatively small cavity, a parabolic reflector, and
close proximity of the photanultiplier window to the chop. This
results in a significantly higher percentage of the fluoresced
radiatian reaching the photoniltiplier .

Finally, the use of a narrow bandpass interference filter
to select the emissian wavelength was chosen for its high rejection
of wavelengths outside the bandpass. It was found that rejection
of these wavelengths resulted in a higher signal to noise ratio
than did a wide bandpass absorbance filter, in spite of the loss
of a great deal of fluoresced radiation with the interference
filter.

Dripper needle

One of the most crucial aspects in the desigi of this
detector is the desigi of the dripper needle. Various configuratims
have been tried including single drippers with blunt and pointed
ends, dual needles with various orientations, and suspended drops
using two vertically aligned needles . The configuraticn that
proved to be the best with a variety of solvents is a piece of

1/16" 0.D. stainless steel tubing with a 90° polished blmt end

58
over which a piece of 1/8" 0.D. teflon tubing has been fitted.
The teflon tubing and the stainless steel tubing must be cut at
a true 90' and present a flat surface for suspension of the dmp.
This dripper needle design prevents creeping of the mobile phase
up the side of the needle because of the hydrophobic nature of the
teflon and yet allows easy positioning because of the stainless
steel. Althcmgh the drop size is quite large, on the order of
50 #1, even the earliest appearing peak required fourteen drops
to elute completely. Furthermore, there was no noticeable
degradation in resolution or efficiency between the two detectors .
This drOp size might becone a significant problen with nore
efficient or smaller columns but the problen can easily be
rectified by reducing the size of both the needle and the teflon

sleeve .

Electronics

 

Timing circuit
The timing circuit has been designed to be as simple as possible

yet make the detector free from effects due to variations in the
flow rate and dr0p residence time. The method chosen initiates
the timing sequence by the falling of one drap, integrating the
pulse train generated by the voltage to frequency convertor for
a given period (a front panel adjustment) and then clocking the
memories. Using this system, the integration has stOpped and the
accumulated count is stored in merory before the drop has fallen.
This allows for some variation in the ch'Op period, an interval
which is more variable than the integration period established by

59'
the timer. The detector will respond to small changes in the flow
rate due to the variation in the growth rate of the (trap but these
effects are minor. Sore care must be exercised in selecting the
integration period. especially if a change in the mobile phase
is made since a change in surface tension will result in a change in
the dr0p period. If the drop period becones shorter than the
integration time, the counters will be reset before the menories
have been clocked. This is easily prevented by setting the
integration tinne based upon the unobile phase with the shortest
drOp period.

A nunber of signals have been utilized to indicated the
falling of the drOp including unnitering the transmission or
reflection of the excitation wavelength with a photocell, detecting
the drop in conductance as the drop falls from two needles, and
monitering the change in the signal plus background from the photo-
multiplier tube, the systen currently in use. Although this
systen does limit the dynamic range of the detector, it has proven
quite versatile, working with a range of solvents from water to
acetonitrile, and the results are very reproducible. In order to
use the background signal to trigger the timing circuit, i the peak
background signal must be between four and six volts. The gain
must be adjusted so that a sufficient background does exist to
trigger the timing circuit in the absence of fluorescence from
the drop, but the gain must not be so high that the amplifiers
are saturated when the strongest fluorescent peak is monitered.

If this detector were to be used for more general applications ,

the use of a photocell that mnitersthe reflection of the excitation

60
wavelength would be the preferred method.

Amplificaiton circuit

The amplification circuit makes use of digital data trans-
mission, integration, and storage and holds some distinct advantages
over a completely analog circuit.

'nne original circuit design was completely analog. It used
a similar current to voltage convertor and amplifier circuit
but employed a sample and hold circuit to sample the potential of
an analog integrator and hold that potential until the next drop
fell. 'nnis circuit had significant problem with the pickup of
noise on the input to the current to voltage convertor and with
cheap in the sample and hold circuit.

The subsequent circuit replaced the sample and hold circuit
with analog to digital and digital to analog convertors . This
was a substantial improvement over the previous design but the
input to the current to voltage convertor was still susceptible to
noise and problens with the integrator began to appear. Both of
these problems are eliminated irn the current design.

Presently the system is designed with the analog amplification
and the voltage to frequency convertor circuits adjacent to the
photomltiplier in the photomultiplier housing. This minimizes
the problem of noise pickup between the photomultiplier and the
first amplifier and also allows transmission of the signal from
the photomultiplier in digital form which is inherently more
immune to noise than analog transmission. The serially transmitted

data is converted into parallel digital data by counting the pulses

61
from the voltage to frequency convertor. This is particularly
convenient because the integration period is easily varied up
to the limit of the counters. Five 4 bit counters allow for a drOp
period of ten seconds with the amplifiers saturated. The output
from tl'e last three counters is stored in the 2 hex memories
upon the signal from the timing circuit and the digital to analog
convertor outputs the updated analog signal .

Qnrcmatography

The chromatographic separation of the cap structures is
similar to the previous method used (42) with respect to the diges-
tion of the mINA and the fragments generated. Specifically, the
cap 2 structures are degraded to cap 1 structures and the analysis
consists of separating the cap 0 structures from the cap 1
structures. The composition of the cap 2 structures must be
inferred from the analysis of the 2 '-O-methy1nucleosides . The
chromatographic method utilizing the growing dr0p flLorescence
detector is less powerful as implemented than the method of
Albers et a1 . since it does not detect the 2 ' -0-methy1nucleosides .
It does however use a much simpler gradient systen erploying a
one step gradient rather than a series of complex gradients .
Furthermore, the method has the potential for analysing the
2 ' -O-methylnucleosides using conventional methods of detection.

The one feature of the want method that has allowed the
great simplification in the mobile phase is the use of an anion
exchange colum as a selective enrichment device. This column

is used in place of a sample loop and selectively retains the

62

strongly charged cap structures. Flushing this anion exchange
column with water will selectively elute all the non-negatively
charged conpounds , including all nucleosides , simplifying the matrix
considerably. The cap compounds are quantitatively eluted off
the anion exchange column by the mobile phase designed to separate
the various cap 0's and cap 1's on the analytical column. The
use of this enrichment column provides three important benefits.
It a) simplifies the matrix sufficiently that the cap structures
may be analysed without interference, b) allows the injection of
large volures without volume overloading the analytical column,
and c) allows the use of microbore columns .

The analytical column used for the separation of the caps
was selected after much entperimentation. The column bore was
selected based upon mass sensitivity and packing efficiency.
Scott et al. (48,49) have established the power of microbore HPLC
columns for separations requiring exceptionally high efficiency
or mass sensitivity. Since the cap analysis is carried out with
a very limited amount of sample, the increase in mnass sensitivity
with microbore columns is particularly attractive. Theoretically
as the bore of the column is reduced, the mass sensitivity will
increase as the ratio of the cross-sectional areas, up to the
point that the column overloads. This was demonstrated empirically
(data not shown) by a four fold increase in sensitivity by decreasing
the column bore from 4.6 mm to 2.1 mm. The limitations to reducing
the bore of the column are that extra column volunes becone very
significant and the smaller the bore, the harder the column is

to pack efficiently. The column used in the separations presented

63
has a bore of 2.1 mm and is 25 on long, the minimum length that
will yield the desired separation.

'flne stationary phase selected for the analytical column is
10 micron Whatmen ODS-3. This material was selected after an
evaluation of many reverse phase packings based upon its high
surface coverage, high permeability, ease of packing, selectivity,
and compatability with aqueous mobile phases. The compatability
with an aqueous mobile phase was a very important consideration.
Although the exact nature of the surface of the stationary phase
has not been adequately described, the description of a "molecular
fur" is frequently used. The conformation of this "molecular fur"
is a function of a variety of factors including the composition
of the mobile phase. Presumably as the mobile phase changes,
the conformation of the stationary phase will change to the most
tho-odynamically stable conformation. This is seen enpirically
as a change in the efficiency of the column. It was found that,
of the packings evaluated, the Mnatman ODS-3 was the least
affected by the polar mobile phases.

The mobile phase consisted of a low ionic strength buffer,
an ion pairing reagent, and an organic modifier. The separation
was dependent on all three components. The phosphate buffer was
chosen based upon the separation on the analytical column as well
as its ability to quantitatively elute the caps from the anion
exchange column. The ion pairing conpound, tetrabutyl amoninm
hydroxide, was in sufficient concentration that the separation
was not sensitive to minor changes in its concentration. The

use of tetrahydrofuran as the organic modifier was due to the

64
enhanced selectivity it offered with respect to m7GpppGn and m7GpppA
as compared to methanol and acetonitrile. It was found that a very
low concentration of tetrahydrofuran would allow the separation of
these compounds but neither methanol or acetonitrile would.

Igprovements in the cap analysis

There are several improvenents that could be made to this
method that would enhance it. The first, and most promising, is
the developnent of efficient microbore columns. The use of a column
packing system capable of packing at 25,000 'psi should allow
packing efficient columns with a 1.0 mm bore. This is especially
important with the fluorescence detector since the sensitivity,
corrected for absorbance , is strongly dependent on the drop period
and consequently the volumetric flow rate . The second improvenent
would be to collect the nucleosides washed off the enrichment column
on a reverse phase column and separate them as was done by Albers et al.
This requires the use of conventional detection systems. Finally,
significant improvenent can be made in the amount of sample required
with the fluorescence detector by either using a post-column
derivatization system to enhance fine quantum efficiency of fine
7-methylguanosine or use of a laser as a light source. Both
methods have the potential of increasing the sensitivity at least

two orders of magnitude .

BIBLIOGRAPHY

(l) Darnell, J. E., W. R. Jelinek, and G. R. I‘blloy, Science _l_8_1_,
1215 (1973).

(2) Perry, R. P., and D. R. Kelly, Cell 1, 37 (1974).

(3) Desrosiers, R., K. Friderici, and F. Rottman, Proc. Nat. Acad.
Sci. U.S.A. 21, 3971 (1974).

(4) Kaehler, M. , Dissertation Thesis of Doctoral Degree, Michigan
State university, 6.

(5) IZottman, I. R., A. J. Shatkin, and R. P. Perry, Cell 3, 197
1974).

(6) lawley, P. 1)., and P. Brookes, Biochem. J. §_9_, 127 (1963).
(7) {11aninst results.

(8) Mnthukristmnan, S., G. W. Both, Y. Eurvishi, and A. J. Shatkin,
I‘hture 225, 33 (1975).

(9) Breter, H., L. T. Manek, c. M. Hellman, and R. E. Rhoads,
Anal. Biochem. g, 102 (1979).

(10) Rosenberg, M. , Nucleic Acids Research 1, 653 (1974) .

(ll) Furvichi, Y., A. J. Shatkin, E. Stavnezer, and J. M. Bishop,
Nature 257, 618 (1975).

(12) Adams, J. M. , "The Ribonucleic Acids:, Springer-Verlang, New
York (1977).

(13) Kim, C. H., and R. H. Sarma, Journal of the American anenical
Sociegy .192, 1571 (1978).

(14) Both, G. W., A. K. Ba‘nerjek, and A. J. Shatkin, Proc. Nat. Acad.
Sci. U.S.A. 2;, 1189 (1975).

(15) Md, R. E.,and C. Lane, Nucleic Acids Res. _5_, 3237 (1978).

(16) Desrosiers, R., K. Friderici, and F. Rottman, Biochem. _l_4_,
4367 (1975).

65

(17)

(18)

(19)

(20)

(21)
(22)

(23)
(24)
(25)
(26)

(27)

(28)
(29)
(30)
(31)
(32)

(33)
(34)

(35)

(36)

(37)

66
Wei, C., A. Gerslnowitz, and B. Mass, Cell 4, 379 (1975).
Zan-Komlczenska, M. , M. Bretner, H. Sierakowska, E. Szczesna,
W. Filipowicz, and A. J. Shatkin, Nicleic Acids Research 4,
3065 (1977).

Miami-ism, 3., s. Szybiak, A. B. Legocki, and w. Filipowicz,
Eur. J. Biochem. 93, 69 (1978).

Hickey, E. 0., L. A. Weber, and c. Baglioni, Proc. Nat. Acad.
Sci U.S.A. 1;, 19 (1976).

Suzuki, H., FEES letters _7_2_, 309 (1976).

Sasavage, N. L. , K. Friderici, and F. Rattan, Nucleic Acids
Research 9, 3613 (1979).

Smnith, R. E., and J. M. Clark, Biochemistry 18, 1366 (1979).

 

Tsvet, M., Proc. Warsaw Soc. Nat. Sci. Bio. Sec. 14, no. 6 (1903).
Horvath, C., Trends in Analytical Chemistgy l, 6 (1981) Review.

Goldberg, A. , "Comparison of Reversed Phase Packings';', The
DuPont Co. (1981).

Kirkland, J. J., and L. R. Snyder, Introduction to Mndern
liquid Chromatography" , Second Edition, Wiley and Sons , New York.

Bidlingneyer, B. A., J. Chrom. Soc. _1_8_, 525 (1980).
Conn, w. E., m $93, 377 (1949).

Connw. E., J. Amer. ChemSoc. 7_1_, 2275 (1950).
Anderson, N. 0., Anal. Biochem. g, 269 (1962).

Anderson, N. 3., J. C. Green, N. L. Barber, and R. c. Ladd,
Anal. Biochem. 6, 153 (1963).

Pbrvath, C. C., and S. R. Lipsky, Anal. Chem. _3_9_, 1422 (1967).

Party, R. A., J. A. Schmit, and R. C. Williams, J. Chrom.
Science 1.1, 358 (1973) .

"Analysis of Nucleic Acid Constituents", Whatman, Inc. , Clifton,
N.J.

Krstulovic, A. M., P. R. Brown, and D. M. Rosie, Anal. Chem. £_n_9_,
2237 (1977).

Hartwick, R., A. Jeffries, A. M. Krstulovic, and P. R. Brown,
:1; Chron. Sci. 16, 427 (1978).

 

(38)

(39)
(40)
(41)

(42)

(43)
(44)

(45)
(46)
(47)

(48)
(49)
(50)

67

Walseth, T. P., G. oef£,'M. c. Macs, and N. D. Goldberg, Anal.
Biochem. 1(17_, 240 (1980) .

Hartwick, R. A., and P. R. Bram, J. Clnron. 4, 381 (1976).
Brown, P. R., and A. M. Krstulovic, Anal. Biochen _9_9_, l (1979).

Wakizaka, A., K. Kurosaka, and E. Okuhara, J. Chrom. 162,
319 (1979).

Albers, R. J., C. B. Coffin, and F. M. Rottman, Anal. Biochem.,
(in press).

Lindroth, P., and K. Mapper, Anal. Chem. 51, 166711979).

Flatmark, T. 8., W. Jacobsen, and J. Haavik, Anal. Biochem _l_0_7_,
71 (1980).

Yui, Y., Y. Itokawa, and C. Kawai, Anal. Biochem1Q_8_, 11 (1980).
Diebold, G. J., and R. N. Zare, Science 196, 1439 (1977).

Lang, M., F. Pochon, and A. Michelson, Biochen Bigplys. Acta.
16_9_, 338 (1968).

Scott, R. P., and P. Kucera, J. Chrom. 169, 51 (1979).
Scott, R. P., P. Kucera, and M.anroe, J. Chrom. 186, 475 (1979).

Martin, F., Maine, J., Sweeley, C., and J. Holland, Clinical
Chem. 12, 1434 (1976).

   

M'IIIIIIIIIIIIJIIILIIII]IIIIJIJIIIIIIIIIIIIEs