Note: Descriptions are shown in the official language in which they were submitted.
2009855
2558-348/B16
ELECTROPHORETIC SIEVING IN GEL-FREE MEDIA
WITH DISSOLVED POLYMERS
This invention relates to electrophoretic separa-
tions, and to separations of species in a sample based on
molecular size.
Molecular sieve electrophoresis is a powerful
method for separating macromolecular solutes both among -
themselves with high resolution on the basis of molecular
size and from solutes of lesser molecular size. The gel
media in which these separations take place however require
careful preparation and special handling techniques, with
problems in reproducibility and stability.
Capillary free zone electrophoresis, on the other
hand, is also of interest for certain types of separations,
since it permits the use of high voltages which provide the
advantage of relatively high speed. The small size of the
capillary further permits the separation of extremely small
samples in a buffer solution without the use of complex
media such as a gel or paper, and with essentially no band
broadening. Capillary free zone electrophoresis is
particularly useful in the separation of small peptides and
proteins. Separation occurs on the basis of the charge/mass
ratio, however, and for this reason certain separations are
very difficult to achieve by this method, notably those
involving high molecular weight polynucleotides and many
SDS-treated proteins.
Gel media may be placed in capillaries for molec-
ular sieve separations, but the preparation and use of such
gels is particularly problematic, since they undergo phys-
ical and chemical changes with each use and thus forpractical purposes can only be used once. This is detrimen-
tal to the reproducibility of the separations and to the
~.
, :
~s~ .i ,,j,,, ,,",",,,~ ,",.
2 Z009855
efficiency of the technigue. In addition, it raises serious
problems for those capillarieg which are incorporated into
cartridges designed for automated instrumentation.
It has now been discovered that sample ions, and
particularly biomolecules, may be separated from each other
on the basis of molecular size by electrophoresis through an
aqueous solution of a non-crosslinked polymer of a selected
molecular weight (or molecular weight range) and
concentration. The molecular weight of the polymer will be
selected as described below to correspond to the molecular
weight range of the sample ions in a manner which will
inhibit the migration of the sample ions through the
solution to ~arying degrees. Macromolecular sample ions and
other biological species may thus be separated from each
other and from sample ions of lesser size without the use of
a gel. The terms "macromolecule" and "macromolecular" are
used herein to refer to species having molecular weights of
at least about 10,000.
The polymers used herein are generally non-
crosslinked polymers. Branched or linear polymers may be
used, linear polymers being preferred for many applications.
In addition, the polymers may be neutral or charged, neutral
being preferred in applications where charge interaction
between the sample ions and the polymer is sought to be
avoided.
Cellulose derivatives have been used in capillary
electrophoresis for suppressing electroendosmosis and other
types of bulk flow by increasing the viscosity of the buffer
solution, and for preserving the capillary as well. The
quantities used for this purpose are small, however, with no
substantial tendency to detain the sample ions during their
migration or to affect their separation. The present
j 35 invention resides in the discovery that dissolved. linear
polymers in general produce a molecular sieving effect when
used in certain amounts, these amounts being generally
.,
, :, , : -, : , . -, ,:: . :: : - :.
20098S5
higher than the amounts of the cellulose derivatives used
for suppressing bulk flow.
Aqueous media with dissolved polymers in
accordance with this invention may be used for biomolecular
separations in general, although they are of particular
utility in separations performed in capillary columns with
high voltage, i.e., high performance electrophoresis. The
use of polymers in this manner permits the separation of
species which vary in molecular weight with insufficient or
no variation in charge/mass ratio, and lends itself to easy
preparation of the separation media and high
reproducibility.
As in the known use of cellulose derivatives
referred to above, the dissolved polymers further serve to
suppress bulk flow due to their inherent increase in
viscosity. Examples of bulk flow occurring spontaneously
are electroendosmosis, hydrokinetic flows (due to
hydrostatic heads), and convection. At these polymer
level~, however, the decrease in sample ion mobility caused
by the presence of the polymers varies both with the size
and concentration of the polymer and the size of the sample
ion, a feature which does not occur at the low levels at
which the cellulose derivatives have been used or
suppressing bulk flow.
Other features and advantages of the invention
will be apparent from the description which follows.
The polyme~rs used in connection t~he present inven-
tion must be water-soluble and, as stated above, are
'~ preferably linear.
~¦ Selection of the polymer is optimally geared
, toward the particular sample ions being separated. The
j 35 molecular weight of the polymer is of primary in~erest in
making this selection. In general, polymers varying widely
in molecular weight may be used. Resolution of the sample
.,.
' - ~:,
4 Z009855
ions will generally improve, however, as the polymer
molecular weight approaches the range of the molecular
weights of the sample ions. The best results are obtained
with polymers having an average molecular weiyht which is
between the lowest and highest molecular weights of the
sample ions, and in particular witll polymers whose molecular
weight range covers (i.e., is at least coextensive with) the
molecular weight range of the sample ions. In preferred
embodiments, the polymer has an average molecular weight
lOwhich is from about 10,000 to about 2,000,000, and within
about 0.1 to about 200 times, more preferably from about 0.2
to about 20 times, and most preferably from about 0.5 to
about 2 times the average molecular weight of the sample
ions.
15Within these parameters, the particular type of
polymer may vary widely. For aqueous systems, examples of
linear polymers which may be used are water-soluble
cellulose derivatives and fully water-soluble polyalkylene
glycols. Specific examples of such cellulose derivatives
are sodium carboxymethyl cellulose, sodium carboxymethyl 2-
hydroxyethyl cellulose, 2-hydroxyethyl cellulose, 2-
hydroxypropyl cellulose, methyl cellulose, hydroxypropyl -
methyl cellulose, hydroxyethyl methyl cellulose,
hydroxybutyl methyl cellulose, and hydroxyethyl ethyl
cellulose. Preferred cellulose derivatives are those with a
highly hydrophilic character, and consequently high water
solubility and minimal affinity to the sample ions. Methyl
cellulose is particularly preferred. Celluloses are
generally characterized in terms of the viscosity of aqueous
solutions in which theyiare dissolved at specified
concentrations and temperature. With this in mind, and
depending on the size of the sample ions sought to be
separated, the cellulose derivative may vary widely in terms
of this viscosity characterization. For example, cellulose
¦ 35 derivatives may be used which are characterized as producing
viscosities ranging from about 15 centipoise to about 17,000
centipoise when dissolved in water at 2 weight percent
....
, - .,- . ~ : ,: ; , .. - .. ,.. ... : . ~ ,,
` . . .. . . .... , ` . .....
2009855
measured at 25C, although in the context of this invention
they would be used at other concentrations. Polymers such
as these are useful in separating polynucleotides with
chains ranging from about 10 to about 10,000 base pairs.
Preferred cellulose derivatives for use in the present
invention are those which would have viscosities of from
about 1,000 to about 10,000 centipoise if prepared as 2%
aqueous solutions measured at 25C. It is to be understood
that these viscosity characterizations are intended merely
as an indication of the molecular weight of the polymer, and
not of the actual viscosity when used in the context of the
present invention.
Preferred polyalkylene glycols are polyethylene
glycols having average molecular weights of at least about
' 15 10,000. Particularly preferred are polyethylene glycols
', with average molecular weights of at least about 20,000,
most preferably at least about 30,000. As an example,
mixtures of sample ions ranging in molecular weight from
about 10,000 to about 100,000 may be separated with
polyethylene glycols ranging in molecular weight from about
10,000 to about 100,000.
Examples of branched polymers which may be used in
accordance with the invention are soluble starches and
starch derivatives. A specific example is hydroxypropyl
starch.
Mixtures of polymers in which varying molecular
weights are purposely combined may also be used. This will
be particularly useful in separating sample mixtures which
have a wide range of molecular weights, thus providing
separation over the entire range.
The use of charged polymers is an option which can
provide a further separation parameter to the system. This
will vary the interaction between the species and the
polymer, and may thus be of use depending on the particular
mixture Of species present in the sample, and the type Of
separation sought among these species.
,.'
Z009855
The quantity of polymer to be dissolved in the
resolving solution may vary widely, and will be any quantity
which extends the retention time of the sample ions to such
varying degrees that effective separation on the basis of
molecular size is achieved. Clearly, this will vary with
various parameters of the system, including for example the
column configuration and length, the presence and effect of
other factors influencing the separation such as charge and
electrophoretic mobility, the molecular structure, intrinsic
viscosity and interactive character of the polymer itself,
and the range of and differences between the molecular
weights of the sample ions. The degree to which the
retention times for the sample ions should be extended for
best results will vary with the sample composition and the
polymer being used. For separations of macromolecular
species, increases in retention time of at least about 25%,
preferably at least about 35%, and most preferably at least
about 50%, will provide the best results. For polyalkylene
glycols, particularly polyethylene glycols, concentrations
of at least about 2% by weight, preferably at least about
3%, and most preferably from about 3% to about 30% by weight
will give the best results. For cellulose derivatives,
preferred concentrations are at least about 0.1% by weight,
with about 0.1% to about 30% by weight more preferred, and
about 0.1% to about 10% by weight particularly preferred.
To conduct the separations in accordance with the
present invention, operating conditions and procedures used
in conventional electrophoretic separations, including
appropriately selected buffer systems, may be used. The
invention is of p~rticul!ar utility in high performance
electrophoresis as performed in capillaries. Preferred
capillaries are those having internal diameters of less than
about 200 microns, more preferably less than about 100
microns, and most preferably about 25 microns to about 50
microns. Voltages of at least about 1000 volts.~re
preferred, witll at least about 3000 volts particularly
preferred.
2009855
The following examples are offered strictly for
purposes of illustration, and are intended neither to define
nor to limit the invention in any manner.
In each of these examples, electrophoresis was
performed on an HPE-100 high performance electrophoresis
instrument, a product of Bio-Rad Laboratories, Inc.,
Hercules, California. Capillary tubes of 20 cm length by
25 ~ inner diameter, and 50 cm length by 50 ~ inner diam-
eter, coated with linear polyacrylamide as described in
10 Hjerten, U.S. Patent No. 4,680,201, issued July 14, 1987,
were used. A conductivity bridge Model 31 from Yellow
Springs Instrument Co., Yellow Springs, Ohio, was used, and
detection was performed on-line in the capillary itself, by
UV absorption. The hydroxypropyl methyl cellulose was
obtained from Sigma Chemical Co., St. Louis, Missouri, in
powder form, specified in terms of its viscosity when
prepared as an aqueous solution at a concentration of 2% by
weight and measured at 25C. The specified viscosities were
15, 50, 100 and 4000 centipoise, and the powdered polymer
will be referred to herein for convenience as "15-
centipoise," "50-centipoise," "100-centipoise," and "4000-
centipoise hydroxypropyl methyl cellulose." The 4000-
centipoise hydroxypropyl methyl cellulose was estimated to
have an average molecular weight of approximately 900,000.
The methyl cellulose was also obtained from Sigma Chemical
Co., similarly specified as producing a viscosity of 4000
centipoise in a 2% aqueous solution at 25C, and will be
referred to herein in a manner similar to the hydroxypropyl
methyl cellulose. The myoglobin was Type III from horse
30 heart, the albumin was Fraction V bovine serum albumin. The -
myoglobin, albumin and the substance P were also obtained
from Sigma Chemical Co. The DNA fragments used in Example 4
were a mixture used as Low Range Size Standards supplied by
Bio-Rad Laboratories, Hercules, California, and included
fragments containing 88, 222, 249, 279, 303, 634~ 800, 1434
and 1746 base pairs. -~
~: :
8 Z009855
EXAMPLE 1
This example demonstrates the effect of polyethyl-
ene glycol (PEG) as a solute in an electrophoretic medium
used for separating myoglobin and substance P. The first
part of this example serves as a control test performed in
the absence of the PEG, while the second shows the effect
which the PEG has on the component separation.
A. Electrophoresis without sieve-Pr motina polvmer.
A sample solution was prepared by dissolving
substance P and myoglobin in 10 mM pH 2.5 phosphate buffer
to achieve concentrations of 100 ~g/mL of substance P and 50
~g/mL of myoglobin. A 3-~L sample of the solution was
loaded electrophoretically on a coated 20 cm x 25 ~
capillary cartridge filled with the buffer solution.
Electrophoresis was performed on the loaded sample by
applying a potential of 8000 V across the capillary, with
detection at 200 nm with a sensitivity range of 0.02 AUF.
The myoglobin eluted at a retention time of 2.8
minutes, and the substance P at 3.3 minutes. Note that the
myoglobin migrated through the capillary faster than the
substance P.
B. Electrophoresis in presence of PEG.
The experiment of Part A was repeated, the only
difference being that the buffer solution in the capillary
further contained PEG with an average molecular weight of
approximately 35,000 at a concentration of 5 weight percent.
The sample components eluted in a reversed elution
order and with lengtheneid retention times. The substance P
eluted first at 4.9 minutes, followed by the myoglobin at
5.9 minutes.
EXAMPLE 2
This example illustrates the use of PEG in the
electrophoretic separation of the monomer, dimer and trimer
9 20~9855
of albumin. The first part of this example i5 a control
test, while the second includes the use of PEG.
A. Electrophoresis without sieve-promotinq polymer.
A sample was prepared by dissolving albumin as a
mixture of the monomer, dimer and trimer in 10 mM pH 2.5
phosphate buffer to a total albumin concentration of
100 ~g/mL. The sample was loaded, run and detected on the
same column using the same conditions as in Example lA. The
result was a single sharp peak at a retention time of 3.5
minutes.
B. Electrophoresis in presence of_PEG.
The experiment of Part A was repeated with the
lS inclusion of PEG with an average molecular weight of
approximately 35,000 at a concentration of 5 weight percent
in the column solution. Monomer, dimer and trimer formed
separate peaks at retention times of 5.6 minutes, 6.8
minutes and 7.7 minutes, respectively.
EXAMPLE 3
This example illustrates several conditions where
water-soluble polymers were in the buffer solution but
incomplete or no separation occurred due to insufficient
amvunts of polymer or due to polymer chains of insufficient
length.
A. Albumin with 2~ PEG of molecular weiqht 35.000.
The experiment of Example 2B was repeated, using
2% of the PEG rather than 5%. The monomer was detected at a
retention time of 3.7 minutes, and the dimer at 4.0 minutes,
with the trimer peak not distinguishable. The monomer and
;; dimer peaks overlapped.
Z0~9855
B. Albumin with 5~ PEG of molecular we~ght 6C00.
The experiment of Example 2B was repeated, using
PEG of an average molecular weight of 6000 rather than
35,000. All of the albumin components passed the detector
as a single peak at a retention time of 4.3 minutes, with no
separation a~ong the three.
C. Albumin and myoalobin/ _bstance P with 0.2~ 4000-
centi~oise hvdroxyDropy~ yl_cell_lose.
A solution was prepared by dissolving 4000-
centipoise hydroxypropyl methyl cellulose in 0.1 M pH 2.5
phosphate buffer at 0.25 weight percent. The experiment of
Example 2B (albumin sample) was then repeated, followed by
the experiment of Example lB (sample containing myoglobin
and substance P), using this hydroxypropyl methyl cellulose-
containing buffer solution in place of the PEG-containing
buffer solution in the capillary in each case.
The monomer, dimer and trimer components of
albumin passed the detector together as a single peak at a
retention time of 3.1 minutes. In the myoglobin/substance P
run, the myoglobin displayed a retention time of 3.1 minutes
~ and the substance P a retention time of 3.7 minutes, which
3 are essentially the same and in the same order as when no
polymer was present in the buffer solution (EY~ample lA).
This result suggests that the sieve passages
around the 0.25% hydroxypropyl methyl cellulose were too --
large to have any effect on the protein, although they might
well create a sieving effect with DNA.
.j .
D- 10~ Ethylene alycol and 10/_glycerin.
As the monomer of PEG, ethylene glycol has similar
hydrophilicity characteristics. This experiment
demonstrates that, like PEG, ethylene glycol and its analog
glycerin (1,2,3-propanetriol) both increase the solution
viscosity, but neither produce the sieving effec~
attributable to PEG. Viscosity increases in themselves are
therefore not responsible for the sieving effect.
.
;~
.q
. ~ " ~ .~: .. ~ . , ~ .. . ~ i .... . .. ... .... ..
2009855
11
Separate solutions were prepared, one containing
ethylene glycol and the other glycerin, both at 10% by
weight in 0.1 M pH 2.5 aqueous phosphate buffer. A third
solution, containing 0.05% of the 4000-centipoise hydroxy-
propyl methyl cellulose solution in the same buffer, wasprepared for comparison. The quantity of the latter is the
amount used in the prior art for suppressing bulk flow.
Using each of the three solutions, a sample con-
taining the albumin and fragments of substance P ranging
from 4 to 11 amino acids in size was subjected to electro-
phoresis using the same operating column and conditions as
in Example 2B. In the comparison run, the substance P
fragments separated into individual peaks, but the albumin
components passed the detector as a sinyle peak at a reten-
tion time of 2.5 minutes, showing no separation betweenmonomer, dimer and trimer. In the ethylene glycol run, all
peaks passed the detector in the same order as in the con-
trol run, again with no separation of the albumin components
I into separate peaks. The albumin retention time was 3.4
i 20 minutes. In the glycerin run as well, all peaks passed the
detector in the same order as in the control run, again with
no separation of the albumin components into separate peaks.
The albumin retention time was 3.5 minutes.
EXAMPLE 4
This example illustrates the separation of DNA
fragments of differing lengths, using the Low Range DNA Size
Standards of Bio-Rad Laboratories. The first part of this
example is a control test, while the second includes the use
30 of hydroxypropyl methyl cellulose at a concentration high
enough to cause a sieving effect.
: .
A. Electroph_r sis without sie~ romo in polymer.
A 3-~L sample of the mixture was electropho-
3S retically loaded onto a 50 cm x 50 ~ capillary filled with a
buffer solution made up of 0.089 M Tris-boric acid, 0.002 M
. ''
,
~ ''i'; ' ': : ': ' ~ : :' ': ' ~ ~
20C~9855
12
ethylenediamine tetraacetic acid, and 0.1% sodium dodecyl
sulfate, at a pH of 8Ø
Detection was done at 260 nm with a sensitivity
range of 0.005 AUF. The DNA fragments in the sample passed
5 the detector in three overlapping peaks at retention time~
of about 5-7 minutes, indicating poor if any separation of
the fragments.
8. Electrophoresis with hvdro~syprQpvl methyl cellulose in
10sieve-promotina amount.
The experiment of Part A of this example was
repeated, the sole difference being that 0.5% of the 4000-
centipoise hydroxypropyl methyl cellulose was additionally
included in the buffer solution. The result this time was
15 that the nine sizes of DNA fragments were well separated,
with the retention times listed in Table I:
TABLE I
ELECTROPHORESIS OF DNA FRAGMENTS
Retention
Fraament size Time
88 base pairs 16.2 minutes
222 18.0
249 18.4
279 18.7
303 19.4
634 22.1
800 22.8
30 1434 24.0
1746 24.4
EXAMPLE 5
This example illustrates the separation of a
35 mixture of at least fifteen lengths of DNA fragments, the
lengths differing by 123 base pairs, beginning with 246 base
:: :
" . , ,, , . ,.. , ,,.. j . ~ , , ~, ... , :. - .
Z009855
13
pairs. Methyl cellulose was used to obtain the sieving
effect.
A 3-~L sample of the mixture was electropho-
retically loaded onto a 50 cm x 50 ~ capillary filled with a
buffer solution made up of 0.089 M Tris-boric acid, 0.002 M
ethylenediamine tetraacetic acid, and 0.5% of the 4000-
centipoise methyl cellulose solution at a pH of 8Ø
Detection was done at 260 nm with a sensitivity
range of 0.005 AUF. The fragments passed the detector as
separate peaks, with the retention times listed in Table II:
TABLE II
ELECTROPHORESIS OF DNA FRAGMENTS
Retention
Fraament size Time
246 base pairs 20.3 minutes
369 22.2
492 23.8
20615 25.3
738 26.~ -
861 27.2
984 27.7
1107 28.2
251230 28.5
1353 28.7
1476 28.9
1599 29.0
1722 29.2
301845 29.3
,, :.
EXAMPLE 6
This example demonstrates the lack of effect of 1%
hydroxypropyl methyl cellulose on ths mobility of small
ions, using a range of amounts of the polymer including
amounts which are sieve-promoting in the preceding examples.
The solute in this case is sodium chloride.
'.,
,
'.'
Z009855
14
The polymers used in this group of tests were 15-
centipoise, 50-centipoise, 100-centipoise, and 4000
centipoise hydroxypropyl methyl celluloses (HMC's). Each
type of HMC was dissolved at 1% by weight in both water and
a 20 mM aqueous sodium chloride solution. The
conductivities of the resulting solutions were then measured
and compared as indications of the effect of the polymer on
the mobility of the sodium and chloride ions. The results
were as listed in Table III below.
TABLE III
CONDUCTIVITY OF NaCl SOLUTIONS
conductivity conductivity
15Type of HMCin waterin 20mM ~laCl
used at 1% at 1%
~entipolse)(~ hos) _ ~Imhos~
---* 2.85 3900
140 4000
69 3900
25 100 154 3950
4000 34 3950
~ Control: no polymer present.
3 30
3 The lack of variability of the numbers in the right column
indicates that the mobility of the sodium and chloride ions
~, is unchanged by the presence of the polymer.
~ 35 The foregoing is offered primarily for purposes of
`. illustration. It will be readily apparent to those skilled
in the art that modifications and variations in the pro-
cedures, materials, quantities and operating conditions
described above may be made without departing from the
spirit and scope of the invention.
, ~".,,~ ",
. ~ , -" ',
