Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~ 094/07912 21 4&~62 PCT/US93/09174
A~ ~-lNl'l'Y PURIFICATION METHODS
INVOLVING AMINO ACID h~ llCS AS ELUTION REAGENTS
Background of the Invention
The invention relates to protein isolation and
purification tec-hniques.
There currently exists a variety of methods,
materials, and approaches for the separation of a
particular protein from the other components of a
lO biological sample. One general approach exploits the
non-specific affinity of a protein for a substrate. For
example, proteins may be separated based upon their
molecular charge using ion exchange chromatography,
whereby protein mixtures are applied to an oppositely
15 charged, chromatographic matrix, and the various proteins
bind to the matrix by reversible, electrostatic
interactions. The adsorbed proteins are eluted, in order
of least to most strongly bound, by increasing the ionic
strength or by varying the pH of the elution buffer.
Another general approach makes use of a protein's
physical characteristics as a means of separation. For
example, a protein may be separated based upon its size,
using gel filtration. By this method, protein mixtures
are applied to a gel-filtration column contAining a
25 chromatographic matrix of defined pore size. Proteins
are eluted, generally with an aqueous buffer, collected
as individual chromatographic fractions and analyzed.
Finally, a third general approach makes use of the
specific affinity of a protein for a purifying reagent.
30 A protein, for example, may be purified using an antibody
specific for that protein or conversely an antibody may
be purified using its specific antigen. Typically, the
antibody or antigen is bound to a column substrate and a
solution which includes the particular antigen or
35 antibody applied to the column, allowing immunocomplex
W094/n7912 Z 1 4 6 1~ 2 PCT/US93/09 ~
formation. Bound immunocomplex partners are then eluted
by destabilizing the antigen-antibody complex, e.g., by
exposure to buffers of very high ionic strength or high
or low pH. Alternatively, immunocomplex formation may be
5 exploited to purify the antigen or antibody by
immunoprecipitation. Antigen-antibody complexes may be
precipitated following ayyLeyation, or alternatively, one
of the binding partners may be covalently linked to a
solid particle (such as Sepharose or agarose) and
lO immunoaffinity complexes isolated by centrifugation. In
either method, the protein of interest is then released
from the complex, again, e.g., by exposure to buffers of
high ionic strength or high or low pH.
Of particular interest to molecular biologists are
15 isolation and purification methods for antibodies or
recombinant antibody fusion proteins. Structurally, an
individual antibody molecule includes two identical heavy
(H) rhA; n~ and two identical light (L) ch~; n~; each light
chain is disulfide bonded to a heavy chain, and the heavy
20 ~h~;n~ are disulfide bonded to each other to form the
basic dimeric structure of the molecule. Within each
chain, units made up of about llO amino acids fold up to
form compact domains, themselves held together by a
single internal disulfide bond. The L chain has two
25 domains, and the H ch~inc have four or five do-~;n~.
The first two N-terminal domains of the H r.h~; n~
interact with the two L chain domains, producing the "Fab
domain", a portion of the molecule which directs specific
antigen recognition and b; n~; ng. At the other end of the
3 0 molecule, the extreme C-terminal domains of the H ch~; nc
(termed the CH2 and CH3 domains) interact to produce the
"Fc domain", a portion of the molecule which directs a
number of immunoglobulin functions including binding to
cells, fixing complement, and traversing the placenta.
35 And, finally, lying between the Fab and Fc domains is a
~ 094/07912 2 el 4 6 1 6 2 PCT/US93/09174
small number of amino acids which make up the hinge
region, a flexible domain which facilitates the free
movement of the antigen binding portion of the molecule.
Summary of the Invention
It is an object of the invention to provide a
cheap and general method for eluting proteins adsorbed to
affinity chromatography columns.
As discussed generally above, affinity
chromatography is a highly effective method for purifying
10 proteins which exploits specific interactions between the
proteins to be purified and a solid phase immobilized
ligand. Typically the solid phase ligand has some unique
chemical character which results in the selective
adsorption of the protein of interest. Cont~;nA~t
15 proteins either do not bind the solid phase or can be
removed by washing the solid phase with appropriate
solutions. Historically, elution of an affinity column
has been accomplished by either of two methods: (i)
washing the matrix with a solution of specific ligand
20 resembling the immobilized ligand, or (ii) washing the
matrix with solutions of very high ionic strength, or
very high (>11) or very low (<3) pH. In general, method
(i) is more attractive than method (ii) because solutions
of very high ionic strength, while usually not
25 deleterious to proteins, are also usually not effective
at desorbing proteins which bind to the affinity matrix
tightly, and many proteins are labile to buffers
sufficiently acid or basic to elute the protein of
interest. However method (i) is not applicable if a
30 specific eluting ligand can not be found or if the use of
a specific eluting ligand is infeasible for practical
reasons, e.g. if the specific ligand is unstable or
expensive.
WO94/07912 ~ 6 2 PCT/US93/091~
Elution of proteins from immobilized ligands in
which the ligand itself is a protein can rarely be
achieved by the use of an eluting solution cont~; n i ng a
specific ligand. This is because the specific ligand
5 usually must be a protein or peptide fragment, and
elution is then infeasible for the practical reasons
mentioned above.
The instant invention describes a new method for
eluting affinity matrices based on the use of inexpensive
lO low molecular weight compounds which mimic the side
chains of specific amino acids. To desorb a protein
bound to a nonproteinaceous affinity ligand, the eluting
compounds are chosen to mimic the side ch~i nC of specific
amino acids participating in recognition of the ligand.
15 To desorb a protein bound to a proteinaceous affinity
ligand, the compounds may mimic the side chains of either
the protein to be purified or the immobilized ligand
protein.
Suitable side chain mimetic compounds are chosen
20 for their water solubility and compatibility with the
stability of the protein to be isolated. Because
protein-protein interactions are frequently stabilized by
hydrophobic interactions, and particularly by aromatic
hydrophobic interactions, the most favored elution
25 reagents will be chosen from the group comprising side
chain mimetics of histidine, tyrosine, tryptophan, and
phenylalanine. In particular examples, imidazole is an
appropriate choice for destabilizing interactions
promoted by binding to histidine residues. Hydroxylated
30 aromatic compounds such as phenols and phenolic compounds
whose solubility in water is promoted by other
substituents, such as aminophenols or
hydroxybenzenesulfonates, are appropriate mimetics for
tyrosine side ch~in~. Substituted indoles, such as
35 indole-3-sulfate or indole-3-acetic acid, or
~NO 94/07912 ~ ~ r ~ ~ PCI`/US93/09174
~ 5 ~
benzimidazole salts are appropriate mimetics for
tryptophan. Benzoate salts, phenylsulfonates, and water
soluble heterocycles, such as nicotinic acid salts or
nicotinamide, are appropriate mimetics for phenylalanine.
Useful amino acid mimetics may also be chosen from
those molecules which disrupt charged residue
interactions. Examples of compounds useful for
interfering with such charged residue interactions
include guanidine salts to mimic arginine, alkylamines to
10 mimic lysine, and alkanoic acids to mimic glutamate or
aspartate.
Finally, other compounds which are useful as
elution reagents in the invention include those which
interfere with aliphatic hydrophobic residue
15 interactions, for example, aliphatic alcohols, aliphatic
ethers of water soluble polyalkylene glycols, sulfate
esters of aliphatic alcohols and aliphatic amines.
Compounds with dual hydrophobic action, such as phenols
substituted with aliphatic groups, may also prove useful
20 in mimicking side chain residues.
Accordingly, in general, the invention features a
method of isolating a protein from a sample, involving
(i) providing a first molecule which is capable of
forming an affinity complex with the protein; (ii)
25 contacting the sample with the first molecule under
conditions which allow affinity complex formation; (iii)
isolating the complex; (iv) treating the complex with a
second molecule, the second molecule mimicking an amino
acid residue of either the protein or the first molecule
30 which is critical to the complex formation, so that the
second molecule disrupts the complex, causing the release
of the protein from the complex; and (v) isolating the
protein.
In a preferred embodiment, all or a part of the
35 second molecule mimics an amino acid side chain.
WO94/07912 2 1`4 ~ 1 6 2 PCT/US93/091~
In another preferred embodiment, the first
molecule is protein A and the protein to be isolated is
an antibody or an antibody fusion protein which includes
a protein A-binding domain. According to this
5 embodiment, the second molecule preferably mimics a
histidine residue and is, for example, imidazole.
In two other preferred embodiments, the first
molecule is an antibody and the protein to be isolated is
a recombinant protein; or the first molecule is an
lO antigenic protein and the protein to be isolated is an
antibody which specifically binds that antigenic protein.
By "mimicking an amino acid residue" is meant
being of the same or similar chemical composition to
either the amino acid residue itself or to a part of the
15 amino acid residue which is critical to that residue's
ability to interact with a proteinaceous or
nonproteinaceous affinity ligand (i.e., the "first
molecule"). Preferably, the portion of the amino acid
mimicked is that residue's side chain. Because such a
20 mimetic molecule is used as an elution reagent, the
molecule should preferably also be water soluble and
compatible with the stability of the protein to be
purified. Although any amino acid mimetic is useful in
the invention, those which disrupt hydrophobic protein-
25 protein interactions are preferred; these includeinteractions involving the amino acids histidine,
tryptophan, tyrosine, and phenyl~l~n;ne. Preferable
mimetics for these amino acids are described above.
By "antibody fusion protein" is meant a protein
30 which includes at least a portion of an immunoglobulin Fc
domain directly or indirectly covalently bonded to a non-
immunoglobulin polypeptide.
By "amino acid side chain" is meant that moiety
bound to the molecule's central carbon atom which
35 determines the amino acid's identity. Preferable side
~094/07912 21~ 6 ~ 6~ PCT/US93/09174
chains according to the invention include those
hydrophobic side ch~;nc which characterize histidine,
tryptophan, tyrosine, and phenylalanine.
By "protein A-binding domain" is meant that
5 portion of the immunoglobulin molecule which interacts
with the Staphylococcus aureus cell wall component termed
protein A. By crystallographic studies, this domain is
most likely positioned at the CH2/CH3 cleft.
Applicant has recognized that affinity
10 purification techn;ques may be modified such that amino
acid mimetics are used as elution reagents in the final
step of purification to release the protein of interest
from the affinity complex. Because this approach is
considerably gentler than more conventional elution
15 techn;ques (e.g., elution steps based on drastic changes
in solution pH), applicant's method facilitates the
isolation and purification of those proteins which are
destroyed (e.g., irreversibly denatured) or reduced in
activity by stAn~rd elution procedures. Moreover,
20 because the amino acid mimetics represent convenient and
inexpensive elution reagents, they may be utilized as
alternative elution reagents even for the purification of
more stable proteins.
Other features and advantages of the invention
25 will be apparent from the following description of the
preferred embodiments, and from the claims.
Brief Descri~tion of the Drawinqs
FIGURE 1 shows a schematic of the immunoglobulin
fusion protein CD62 Rg.
FIGURE 2A shows flow cytometric results of CD62 Rg
binding to the surface of neutrophils. FIGURE 2B shows
flow cytometric results of CD62 Rg binding to the surface
of H3630 cells. FIGURE 2C shows flow cytometric results
of CD62 Rg binding to the surface HSB2 cells. FIGURE 2D
WO 94/07912 ~ ~ PCT/US93/o9i~
2 ~;s ~.
shows flow cytometric results of CD62 Rg bin~;ng to the
surface of K562 cells.
Detailed Descri~tion
There now follows a description of a protein
5 isolation and purification procedure according to the
invention, and a description of its use in the isolation
of one particular immunoglobulin fusion protein.
Unusually gentle elution of the recombinant protein
facilitates purification without appreciable loss of
10 native bin~;ng reactivity. This example is provided for
the purpose of illustrating, not limiting, the invention.
Affinitv Purification of IaGl bv Elution with the
Histidine Mi~etic Imidazole
Human IgG1 was purified by affinity chromatography
15 followed by elution with the histidine mimetic i 1~AZO1e
as follows.
Human IgGl was loaded on protein A trisacryl beads
(Pierce, Rockford, IL), and washed with phosphate
buffered 5~1 in~. The beads were divided among several
20 small columns, and the columns were eluted with solutions
contAininq imidazole at 1, 2, 3, 4, or 5M conc~ntration,
ad~usted to a final pH of 6, 7, 8, or 9. The results are
given in Table 1 as the percent of ~ m elution
obtAine~ for any one pH.
TART,T~ 1
Fraction ~luteda
Imidasole
rMl ~ DH9 DH8 ~H7 pH6
1.00 0.59 + 0.03 0.54 + 0.02 0.58+ 0.06 0.49 + 0.03
2.00 0.69 + 0.13 0.80 + 0.04 0.80 + 0.03 0.76 + 0.01
3.00 0.78 + 0.26 0.83 + 0.04 0.93 + 0.00 0.86 + 0.01
4.00 0.94 + 0.14 1.00 + 0.05 1.00 + 0.00 0.91 + 0.04
5.00 1.00 + 0.08 0.96 + 0.01 0.93 + 0.07 1.00 + 0.03
a Mean + std. error
~VO 94/07912 ~ 1 ~ 61 6~ PC~r/US93/09174
Negligible amounts of IgG1 were retained by the
columns at the highest imidazole concentrations, at any
pH. In general the pH did not play a significant role in
mediating the elution power of imidazole, which was
somewhat unexpected, given that the pK of imidazole is
7.1, and so approximately 90% of the molecules would be
charged at pH 6, whereas approximately 90% would be
uncharged at pH 8.
Using this method, immunoglobulin fusion proteins,
as well as IgG1 alone, were purified from protein A
columns without loss of biological activity of the
protein moiety fused to the immunoglobulin constant
domain (in particular, see the purification of CD62 Rg
below). For several of the fusion proteins purified in
this manner, it was known that acidic elution conditions
destroyed the activity (i.e., the ligand binding
activity) known to reside in the portion of the protein
fused to the immunoglobulin domain. In general, the
purification of these immunoglobulin fusion proteins
involved an initial isolation of a crude preparation of a
fusion protein (including the hinge, CH2 and CH3 domains
of human IgGl joined to the extracellular domain of some
surface antigen) which had been prepared by transfection
of COS cells with the appropriate cDNA constructs. Media
supernatants were collected from transfected cells which
had been grown for a further 5 to 10 days, clarified by
centrifugation, and adsorbed to protein A trisacryl or
protein A agarose beads. The beads were collected,
washed thoroughly with phosphate buffered saline
containing 1% nonionic detergent (Nonidet P40 or Triton
X-lOo) followed by buffer alone, then eluted with 4 M
imidazole adjusted to pH 8 with acetic or hydrochloric
acids. The eluted fusion proteins were dialyzed against
buffer, or the imidazole was removed by two cycles of
W094/07912 æ ~ ~ 616 2 PCT/US93/091~
-- 10 --
centrifugal ultrafiltration (Centricon 30, Amicon Corp.,
Beverly, MA).
One particular example of such an antibody fusion
protein purification now follows.
Isolation of a Soluble CD62;Immunoglobulin Fusion Protein
CD62 protein chimeras were prepared by genetic
fusion of the first four N-terminal extracellular domains
of CD62 to the hinge domain of human IgGl as follows.
CD62 cDNA sequences encoding the lectin (L), epidermal
growth factor (EGF), and first two complement regulatory
protein repeat elements (CR) were amplified in polymerase
chain reactions using synthetic oligonucleotides designed
to allow fusion to the human IgG1 artificial splice donor
sequences described previously (Aruffo et al., Cell 61,
1303-1313, 1990) (Figure 1). The forward primer bore the
sequence GGC GCC GAA GCT TCC ATG GCC AAC TGC CAA ATA GCC
ATC TTG (SEQ ID NO:l), while the reverse primer bore the
sequence GGC CAG ATC TCC CTG CAC AGC TTT ACA CAC TGG GGC
TGG (SEQ ID NO:2); the sequence allowed the CD62 fragment
to be inserted as a HindIII to BgIII fragment into
HindIII- and BamHI-digested vector. To amplify the DNA,
20 PCR cycles were conducted, consisting of 30 s at 94C,
2 min at 45C, and 3 min at 72C, using the reaction
buffer recommended by the enzyme vendors (US Biochemical,
Cleveland, OH), and Mlul-digested DNA prepared from a
previously described endothelial cell expression library
(Bevilacqua et al., Science 243, 1160-1165, 1989). A
schematic of the resultant fusion protein, termed CD62
Rg, is shown in Figure 1.
The CD62 Rg expression plasmid was transfected
into COS cells using DEAE dextran as previously described
(Seed and Aruffo, Proc. Natl. Acad. Sci. USA 84:3365-
336~, 1987); typically, ten 100mm semiconfluent plates of
COS cells were transfected with each construct. Twelve
hours following transfection, cells were trypsinized,
~ 094/07912 ~ 6~ 6~ PCT/US93/09174
-- 11 --
seeded onto fresh 100 mm dishes, and allowed to grow for
7-10 days. On the fourth day, 5 ml of fresh media/10%
calf serum was added per dish. Supernatants were
harvested, centrifuged to remove nonadherent cells and
debris, pooled, and stored at 4C. Gel electrophoresis
of such supernatants demonstrated that the expression
plasmids encoded the recombinant globulins and that these
globulins appeared in soluble form in the supernatants of
the transfected COS cells.
Initial attempts to purify the CD62 Rg fusion
proteins by chromatography on protein A columns were
hampered by the lability of the fusion proteins to the
acidic buffers typically used to elute immunoglobulins.
To circumvent this problem, applicants eluted instead
with a solution of imidazole, reasoning that an excess of
this molecule would disrupt the interaction (as
postulated from crystal structure studies) between
protein A and the histidine residue contacts at the
CH2/CH3 cleft of the antibody fusion protein. 4M
20 imidazole proved to be a mild and effective eluant,
allowing retention of carbohydrate and tissue reactivity
(see below).
This imidazole purification procedure was carried
out as follows. Twelve hours following transfection, a
25 fraction of the COS cells transfected with each construct
were seeded onto flasks. Thirty-six hours post-
transfection, the cells were washed with phosphate-
buffered saline (PBS) and overlayed with cysteine-
methionine-free media for 30 min. [35S]-methionine and
t35S]-cysteine (TrAn~T~hel, ICN, Costa Mesa, CA) were
added to a final concentration of 150 ~Ci/ml, and the
cells were allowed to incorporate the label overnight.
The supernatants were harvested and incubated with 200~1
of protein A Trisacryl (Pierce, Rockford, IL) at 4C for
35 12 hr. The beads were collected by centrifugation and
W O 94/07912 21~ 616 2 PC~r/US93/091 ~
- 12 -
washed with PBS/1% Nonidet P-40. For analysis, the beads
were eluted with 200 ~1 of 1% sodium dodecyl sulfate.
Ten microliters of each eluate was loaded on a 6
discontinuous polyacrylamide gel with or without prior
exposure to mercaptoethanol. For preparative elution,
columns were washed with 5 bed volumes of 4M imidazole
(pH 8) (neutralized with acetic acid). Eluted proteins
were stored for short periods of time in imidazole at 4C
or 8C, or ~shAnged into PBS by centrifugal
ultrafiltration for longer term storage.
CD62 Rg Tissue Reactivity
To test the purified protein's ability to react
with cells and tissues in a manner characteristic of
CD62, the following binding assays were performed on
myeloid and tumor cell lines, i.e., cells normally bound
by native CD62.
Typically, 106 cells were incubated with undiluted
Rg supernatants for 30 min on ice in the presence of 10%
rabbit serum. Cells were washed once with PBS and
exposed to fluorescein-conjugated goat antibodies to
human IgG or IgM (Cappel, Malver, PA) at a concentration
of 1 to S ~g/ml for 30 min on ice, followed by fixation
in PBS contA;n;ng 4~ formaldehyde. Fluorescence profiles
were determined by stAn~Ard t~chn;ques with a FACScan
analyzer. Results are shown in Figure 2; solid lines
indicate reactivity with CD62 Rg, and dotted lines
indicate reactivity with control CD7 Rg protein.
Flow cytometry and fluorescence microscopy showed
that CD62 Rg reacted with a cell surface ligand on
freshly isolated human granulocytes, on the breast
carcinoma cell lines H3630 and H3396, and on the myeloid
cell lines HL60, THP-1, and U937. Cell surface
reactivity was not found with the leukemic T cell lines
HSB-2, Jurkat, or HPB-ALL, with K562 (erythroleukemia)
cells, HeLa cells, COS cells, RD (rhabdomyosarcoma)
~ 094/07912 PCT/US93/09174
~1 ~ 6~ ~
- 13 -
cells, H3606 and H3620 melanoma cells, or the L tk- and
NIH 3T3 murine fibroblast cells lines (Figure 2).
Control immunoglobulin fusion proteins CD7 Rg and CD8 Rg,
and native IgG, did not show appreciable reactivity under
- 5 these conditions (Figure 2). In many cases, the amount
of CD62 Rg bound to permeabilized cells greatly exceeded
the amount bound to unpermeabilized cells, suggesting
that substantial internal stores were present.
CD62 Rq Carbohydrate Reactivitv
Because glycolipids frequently express complex
carbohydrate determinants in lineage-restricted
developmental patterns, we investigated whether lipid
extracts of HL60 cells (a promyelocytic leukemia line)
would bind to CD62 Rg in either soluble or adsorbed form.
The upper and lower phases of a Folch partition of HL60
cells was subjected to thin layer chromatography on
silica gel plates, and the chromatograms were incubated
with radiolabeled CD62 or control fusion proteins,
washed, and subjected to fluorograph as follows.
Cells (1 x 108 to 5 x 108) were extracted by
homogenization with 20 vol of a 2:1 chloroform:methanol
solution. The crude extract was filtered through lipid-
free filter paper and subjected to repeated Folch
partitions as described (Hakomori and Siddiqui, Meth.
25 Enzymol. 32:345-367, 1974). Both upper and lower phases
were evaporated and subsequently dissolved in 200~1 of
methanol. Lipids from culture supernatants were
extracted (1:1 v/v) with butanol saturated with lM NaCl.
The butanol phase was dried by evaporation and the
residue resuspended in methanol.
Aluminum-backed silica gel HPTLC plates (5 cm x
7.5 cm) (E. Merck, Darmstadt) were used for
chromatography, and glycolipids were separated in
chloroform/methanol/water (120/70/14). After
chromatography, plates were dried, fixed by immersion in
WO94/07912 2 1 4 6 1 6 2 PCT/US93/091~
- 14 -
0.1% polisobutylmethacrylate in hexane (Magnani et al.,
Meth. Enzymol. 83:235-241, 1982), and incubated for 1 hr
at 22C in blocking solution (150mM NaCl, 3mM CaCl2, 2%
BSA). 35S-labeled Rg (1 x 105 to 2 x 105 cpm/ml), i.e.,
either CD62 Rg or control fusion protein ELAM-l Rg, was
added and allowed to incubate with the plates overnight.
The chromatograms were then washed twice for 30 min each
in 150mM NaCl, 3mM CaCl2, dried, sprayed with En3Hance,
and subjected to fluorography.
Glycolipids migrating either as a single band or,
in different solvent systems, as a closely spaced
doublet, were found to react strongly with CD62 Rg. No
reactivity was detected in ganglioside fractions under
these or more potently eluting conditions.
Parallel evaluation of the chromatographic pattern
of different purified glycolipids indicated that the HL60
lipids comigrated in three different solvent systems
[specifically, chloroform/methanol/water (120/70/14),
chloroform/methanol/water (73/21/4), and
chloroform/methanol/acetone/acetic acid/water
(10/2/4/2/1) (Ishizuka et al., J. Biol. Chem. 253:898-
907, 1978)] with commercial preparations of bovine brain
sulfatides (Sigma, St. Louis, M0; Matreya, Bellefonte,
PA), 3-sulfated galactosyl ceramides bearing heterogenous
fatty acyl substitution on the 2-amino position of the
sphingosine moiety.
Chromatography and analysis of the purified
glycolipids under the same conditions (i.e., two
micrograms (by dry mass) of each of the lipid s~n~rds:
either brain gangliosides (Sigma, St. Louis, M0; Matreya,
Bellefonte, PA), sulfatides (Sigma, St. Louis, M0;
Matreya, Bellefonte, PA), trisialyl ganglioside GTlb
(sigma, St. Louis, M0), galactosyl ceramides with
hydroxyl substitution (Sigma, St. Louis, M0), or
lysosulfatide (Sigma, St. Louis, M0; Matreya, Bellefonte,
~ 094/07912 2~ ~ 6~ ~ PcT/us93/09174
- 15 -
PA) reacted with CD62 Rg or control ELAM-1 Rg and
developed with chloroform/methanol/water 73/21/4)
confirmed that sulfatides reacted strongly with CD62, and
that the more polar form was recognized preferentially
under these conditions. Lysosulfatides, lacking the
fatty acyl substitution, were not recognized, nor were
galactosyl ceramides, lacking the sulfate residue, either
with or without hydroxyl substitution on the fatty acid
chain. Glycolipid bearing CD15 did not detectably react
with CD62 Rg under conditions allowing detection of
sulfatides. Neither CD7 Rg, CD8 Rg (Aruffo et al., Cell
61, 1303-1313, 1990), ELAM-l Rg (Walz et al., Science
250:1132-1135, 1990), intact IgGl, or a COS cell
preparation of a fragment of the IgGl corresponding to
the Fc fragment present in CD62 Rg reacted with
sulfatides.
Other Embodiments
Applicant has recognized that affinity
purification technigues may be modified such that amino
acid mimetics are used in the final step of purification
as elution reagents to release the protein of interest
from the affinity complex. Such an amino acid mimetic
elution step may be employed in any st~n~Ard affinity
purification procedure, e.g., to release a protein from a
column-bound complex or from a complex included in an
immunoprecipitate.
Because a number of amino acid mimetics (e.g.,
those described above) may be readily combined into a
cocktail, the general elution t~chn; que described herein
may be utilized for the purification of any protein,
including those proteins whose interaction contact points
or amino acid sequences are unknown. In one particular
example, most or all of the exemplary amino acid mimetics
listed above may be combined (solubility allowing) into a
WO94/07912 ~6~ PCT/US93/091
- 16 -
general purpose elution reagent which would disrupt any
of a number of affinity complex interactions based on
hydrophobic, aliphatic hydrophobic, and/or charged
interactions. Such a general purpose elution reagent
5 obviates the need to determine the amino acid sequence of r
specific affinity contact points or even of the protein
to be isolated.
Elucidation of such amino acid se~lences may,
however, be used to guide selection of those mimetics
10 which most efficiently elute the protein of interest.
For example, a generally useful affinity interaction
point would be that region of any given protein which
directs antibody complex formation; the identity of this
region could be determined either from the art or
15 experimentally (e.g., by successive deletion or point
mutation analysis of the protein followed by an assay of
the mutant protein's ability to bind antibody, e.g., by
immunoprecipitation or antibody column binding).
Examination of this region's amino acid sequence directs
20 the choice of mimetics which may be utilized as elution
reagents. ~or any given region of interaction, more than
one candidate amino acid contact point may exist.
Accordingly, more than one mimetic may be tested, alone
or in combination, to determine which mimetic best
25 disrupts the complex and directs the most efficient
protein elution.
~ 094/07912 2 1 4 6 1 6 2 PCT/US93/09174
8EO~ENCE LI8TING
(1) G~N~T. lN~O~ATION:
(i) APPLICANT: Seed, Brian
(ii) TITLE OF l-.v~.,lON: A~lNl~l~Y PURIFICATION
METHODS INVOLVING AMINO
ACID MIMETICS AS ELUTION
REAGENTS
(iii) NUMBER OF 8EQu~N~: 2
( iv) rO ~ PONDENCE ~nn~8
(A) ~nn~R~EE: Fish & Richardson
(B) ~TREET: 225 Franklin Street
(C) CITY: Boston
(D) 8TATE: Massachusetts
(E) COU~.~KY: U.S-A.
(F) ZIP: 02110-2804
~v) COMPUTER ~n~Rr~ FORM:
~A) MEDIUM TYPE: 3.5" Diskette, 1.44 Mb
~B) COMr~,~K: IBM PS/2 Model 50Z or 55SX
~C) OPERATING ~Y~,~M: MS-DOS (Version 5.0)
~D) 80FTWARB: WordPerfect (Version 5.1)
(v~) CURRENT APPLICATION DATA:
(A) APPLICATION NUNBER: 07/956,660
(B) FILING DATB: October 2, 1992
(C) CLAS8IFICATION:
(vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER:
(B) FILING DATB:
(Viii) A,,O~ Y/AGENT ~ OKMATION:
(A) NAMB: Clark, Paul T.
(B) REGI8TRATION NUMBER: 30,162
(C) REFERENCE/DG~l NUMBER: 00786/153001
~ix) T~T~rQMMUNICATION INFORMATION:
~A) TEL~nO..~: (617) 542-5070
(B) TELEFAX: (617) 542-8906
(C) TELEX: 200154
WO94/07912 PCT/US93/091~
21461~2
- 18 -
(2) lN~ ~TION FOR 8EQ~ENCE ID~ ~lCATION NUNBER: 1:
(i) 8EQU~N~ ~CTERI8TIC8:
(A) LENGTH: 42
(B) TYPE: nucleic acid
(C) 8TRaND~nN~ single
(D) TOPOLOGY: linear
(xi) SEQ~ENCE DESCRIPTION: SEQ ID NO: 1:
GGCGC~AG CTTCCATGGC CAACTGCCAA ATAGCCATCT TG 42
(2) lN~O~ ~TION FOR 8EQUENCB ID~L.ll~lCATION N~BER: 2:
(i) 8EQ~ENCE C~CTERI8TICS:
(A) LENGTH: 39
(B) TYPE: nucleic acid
(C) 8TRAND~nN~S: single
(D) TOPOLOGY: linear
(xi) 8EQUENCE DESCRIPTION: SEQ ID NO: 2:
GGC~.ATCT CCCTGCACAG CTTTACACAC TGGGGCTGG 39
What is claimed is:
