Note: Descriptions are shown in the official language in which they were submitted.
- = -
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
USE OF NUCLEAR MAGNETIC RESONANCE
TO DF',SIGN l,IGAI~DS TO TARG~T RIOMO-,h'CUT,li.S
Technical Field of the Invention
The present invention pertains to a method for the use of two--lim~n.cional lSN/lH
NMR correlation spectral analysis to design ligands that bind to a target bi--molecule.
Back~Fround of the Invention
One of the most powerful tools for discovering new drug leads is random SCl~.,i,lg
of synthetic chemi~l and natural product ~l~t~b~es to discover compounds that bind to a
particular target molecule (i.e., the i~ntific~tion of ligands of that target). Using this method,
ligands may be irl~tifi~d by their ability to form a physical association with a target molecule
or by their ability to alter a function of a target molecule.
When physical binding is sought, a targe~ molecule is typically exposed to one or
more compounds ~.u~e~,lt;d of being ligands and assays are ~,.Çol,lled to r~termine if
complexes between the target molecule and one or more of those compounds are formed.
Such assays, as is well known in the art, test for gross changes in the target molecule (e.g.,
changes in size, charge, mobility) that in~ t~ complex formation.
Where functional changes are measured, assay conditions are est~hli~h~rt that allow
for measurement of a biological or chen ic~l event related to the target molecule (e.g., enzyme
catalyzed reaction, lcc~l~lo~ e(li~ enzyme activation). To identify an alteration, the
function of the target molecule is d~,Le~ ihled before and after exposure to the test compounds.
Existing physical and functional assays have been used successfully to identify new
drug leads for use in designing therapeutic compounds. There are, however, limitations
inherent to those assays that con~p~{".ise their accuracy, reliability and efficiency.
A major shortcoming of exi~ting assays relates to the problem of "false positives". In
a typical functional assay, a "false positive" is a compound that triggers the assay but which
compound is not effective in eliciting the desired physiological response. In a typical phys*al
assay~ a "false positive" is a compound that, for example, attaches itself to the target but in a
non-specific manner (e.g., non-specific binding). False positives are particularly prevalent
and problematic when screening higher concentrations of putative ligands because many
compounds have non-specific affects at those concentrations.
In a sirnilar fashion, existing assays are plagued by the problem of "false negatives",
which result when a compound gives a negative response in the assay but which compound is
actually a ligand for the target. False negatives typically occur in assays that use
concentrations of test compounds that are either too high (resulting in toxicity) or too low
relative to the binding or dissociation constant of the compound to the target.
Another major shortcoming of existing assays is the limited amount of information
provided by the assay itself. While the assay may correctly identify compounds that attach to
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
or elicit a response from the target m-)lecl~ ., those assays typicaUy do not provide any
i.lÇ ~ f;on about either specific binding sites on the target mnlecn1e~ or ~Llu~iluie activity
relationships I~L~ l the compound being tested and the target molecule. The inability to
provide any such inro~ Lion is particuIarly prohlem~tic where the scr~,~ g assay is being
5 used to identify leads for further study.
It has l ~,C~ y been suggested that X-ray crystallography can be used to identify the
bin-1ing sites of organic solvents on macrom- l~cllles However, this method cannot
clet~rmine the relative binding ~ffiniti~s at dirrclcl - sites on the target. It is only applicable to
very stable target proteins that do not dGualulc in the presence of high a ncentr~tions of
10 organic solvents. Moreover, this approach is not a sc~ g method for rapidly testing many
compounds that are ~hemic~lly diverse, but is limited to mapping the binding sites of only a
few organic solvents due to the long time needed to ~Ic~.. ;l-~. the individual crystal
structures.
Clompounds are screened to identify leads that can be used in tne design of new drugs
that alter the function of the target biomolecule. Those new drugs can be ~llu~iLuli~ analogs of
i~entified leads or can be conjugates of one or more such lead cc,lllyOullds. Because of the
problems inherent to eYi~tin~ scr~ellulg m~,thotl.~, those methods are often of little help in
~lPsignin~ new drugs.
There cnntinlles to be a need to provide new, rapid, effic~i~nt accurate and reliable
20 means of screening compounds to identify and design ligands that specifically bind to a
particular target.
Brief Sull~ of the lnvention
In its principal aspect, the present invention provides a process for the design and
25 i~len ~; fic,~t;on of compounds which bind to a given target biomolecule. That process
compri~es the steps of: a) identifying a first ligand to the target molecule using two-
dimensional l~N/~H NMR correlation spectroscopy; b) identifying a second ligand to the
target molecule using two--limt-n~ional 15N/lH NMR correlation spectroscopy; c) forrning a
ternary complex by binding the first and second ligands to the target molecule; d) ~1~l~....i.1i.~g
30 the three tlimpncinnal structure of the ternary complex and thus the spatial ~rient~tion of the
first and second ligands on the target molecule; e) linking the first and second ligands to form
the drug, wherein the spatial orientation of step (d) is m~int~ine~l
This aspect of the present invention uses the two-~limen~ional l5N/lH NMR
correlation spectroscopic screening process as set forth below to identify a first and
35 subsequent ligands that bind to the target molecule. A complex of the target molecule and two
or more ligands is formed and the three-dimensional structure of that complex is determined
preferably using NM~ spectroscopy or X-ray crystallography. That three-dimensional
~ ~ = - .
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
structure is use~ to ~1et~o.rmine the spatial u. ;~ ;on of the ligands re}ative to each other and to
the target molecule.
Based on the spatial nri~nS~tion, the ligands are linked together to form the drug. The
selection of an applupriate linking group is made by ~ g the spat;al ori~nt~tion of the
ligands to one another and to the target molecule based upon principles of bond angle and
bond length u~çulmaLion well known in the organic cl~ l art.
Thus, the molecular design aspect of the present invention comrri~es identifying a
first ligand moiety to the target molecule using two--limencional lSN/lH NMR correlation
spectroscopy; identifying subsequent ligand moieties to the target mn1~cl11e using two-
~limt~ncional lSN/lH NMR correlation spectroscopy; forming a complex of the first and
subsequent ligand moieties to the target molecule; ~~t~ g the three ~lim.on~ional structure
of the complex and, thus, the spatial ori~ntAti~ n of the first and ~.ubse~luenL ligand moieties on
the target molecule; and linking the first and subse~uGI~ ligand moieties to form the drug to
;lll~ the spatial nrit~nt~tinn of the ligand moi~o*~5
The i~ ntificAtion of subse~luçnt ligand moieties can be ywro~ ed in the ~hsçnce or
presence of the first ligand (e.g., the target molecllle can be bound to the first ligand before
being exposed to the test compounds for i~lentifir~tion of the second ligand).
The present invention further col~ pl~tes a drug ~lçeigne~ by the design process of
this invention.
(~hemi~sll compounds can be screened for binding to a given target biomolecule by a
process involving the steps of a) first g~,nG~ g a first two-dimensional lSN/lH NMR
correlation spectrum of a lSN-labeled target molecule; b) exposing the labeled target molecuLe
to one or a nli~lul~, of chemi~l compounds; c) next, generating a second two-flim~on~ional
lSN/lH NMR correlation spectrum of the labeled target molecule that has been exposed to
2~ one or a ll~l ue of compounds in step (b); and d) compAring said first and second two-
~lim~ n~ional lSN/lH NMR correlation spectra to determine differences be~ween said first and
said second spectra, the dirr~ ces identifying the presence of one or more compounds that
are ligands which have bound to the target molecule.
Where the process screens more than one compound in step (b), that is, a ~ ~G ofcompounds, and where a diLrGlGIlce between the first spectrum generated from the target
molecule alone and that generated from the target molecule in the presence of the mixture,
additional steps are performed to identify which specific compound or compounds contained
~ in the ll-L~Lulc iS binding to the target molecule. Those additional steps comprise the steps of
e) exposing the 15N-labeled target molecule individually to each compound of the mixture, f )
~5 generating a two-dimensional 1 5N/lH NMR correlation ~e1LI Ulll of the labeled target
molecule that has been individually exposed to each compound; and g) comparing each
spectrum generated in step f) to the first spectrum generated from the target molecule alone to
CA 02237336 1998-05-11
WO 97/18469 PCT~US96/18312
~le/c~ . . .i ne ~;liLrc;r~llces in any of those co~ >a,~;l spectra, the diLrt;~ ces idcntiry"lg the
~r,;,cnce of a compound that is a iigand which has bound to the target molecule.Because the chemicAl shift values of the particular lSN/lH signals in the two-
dimensional correlation ~e~;l ulll correspond to known specific locations of atomic groupings
5 in the target molecule (e.g., the N-H atoms of the arnide or peptide link of a particular amino
acid residue in a poly-peptide), the s(;~ g process allows not only for the for irl~-ntifir~ation
of which compound(s) bind to a particular target molecule, but also permit the dcti . .. ~i~.AIion
of the particular binding site of the ligand on the target molecule.
The dissociation constant, KD, for a given ligand and its target molecule can bedet~rrnined by this process, if desired, by performing the steps of a) gent~r~ting a first two-
dirnensional lSN/lH NMR correlation ~ ulll of a lSN-labeled target molecule; b)
exposing the labeled target molecule to various cnn~en~T~ti~nS of a ligand; c) g I~-A~ a
two-. li . . .~ ional IsN/lH NMR correlation ~e~ ulll at each conf~entratiQn of ligand in step
(b); d) CO~ rAI ing each ~e;LIum f~om step (c) to the first spectrum from step (a); and e)
cAl-~ulAting the dissociation constant bc~ ~,cn the target molecule and the ligand from those
~lirr~,~,nces according to the equation:
KD = ([P~O - X)([I~O - X3
x
An advantageous cArability of the screening method is its ability to ~l~terrnine the
dissociation constant of one ligand of the target molecule in the presence of a second molecule
already bound to the ligand. This is generally not possible with prior art methods which
employ "wet chemical" analytical methods of det~ormining binding of a ligand to a target
molecule ~ub~LLate.
The process of det~rminin~ the dissociation constant of a ligand can be pulru~ ed in
the presence of a second bound ligand. Accordingly, the 15N-labeled target molecule is
bound to that second ligand before exposing that target to the test compounds.
The ability of the scree~ g method to de~.lll,lle not only the existence of binding
between one ligand and the target molecule, but also the particular site of binding in the
presence of a second bound ligand, permits the capability to design a drug that comprises two
or more linked moieties made up of the lig~n(ls
In a ~c~re~lc;d embodiment of the present invention, the target molecule used in the
molecular design process is a polypeptide. The polypeptide target is preferably produced in
recombinant form from a host cell tran~r~,l,ned with an ~ ,ssion vector that contains a
polynucleotide that encodes the polypeptide, by culturing the transformed host cell in a
medium that contains an ~ssimilAhle source of lSN such that the recombinantly produced
polypeptide is labeled with lSN.
CA 02237336 l998-05-ll
WO 97118469 PCT/US96/18312
Brief ]~)eswl~,Lion of the Drawings
In the drawings which form a portion of the specific~tion
FIG. 1 shows a 15N/lH correlation ~e.iLlu,n of the DNA binding domain of
ulliru~ ly 15N-labeled human papillomavirus E2. The s~e.iL, u-n (80 complex points, 4
scans/fid) was acquired on a 0.5 mM sample of E2 in 20 mM phosphate (pH 6.5), 10 mM
dithiothreitol (D~I) and 10% d~lt~..;lll.l oxide (D20).
FIG. 2 shows 15N/lH correlation spectra of the DNA binding domain of uniformly
5N-labeled human papillomavirus E2 before (thin multiple contours) and after (thick single
o contours) addition of a final test compound. The final concentr~tion of compound was 1.0
mM. All other conditions are as stated in FIG. 1. Selected residues that show signific~nt
changes upon binding are inAi~fltecl
FIG. 3 shows lSN/lH-correlation spectra of the DNA binding domain of unirollllly15N-labeled human papillomavirus E2 before (thin multiple co~ u,~,) and after (thick single
15 contours) addition of a second test compound. The final concentration of compound was 1.0
mM. All other conditions are as stated in FIG. 1. Selected residues that show c; ~.... ric~
changes upon binding are indicated.
FIG. 4 shows 15N/lH correlation spectra of the catalytic domain of ulliLollllly ISN-
labeled stromelysin before (thin multiple contours) and after (thick single ~;ollLoL-l~,) addition
of a test compound. The final concentration of compound was 1.0 mM. The spectra (80
complex points, 8 scans/fid) were acquired on a 0.3 mM sample of SCD in 20 mM TRIS (pEI
7.0), 20 mM CaCl2 and 10% D2O. Selected residues that show ~ignific~nt changes upon
binding are indicated.
FIG. 5 shows 1SN/lH correlation spectra of the Ras-binding domain of uniforrnly
15N-labeled RAF peptide (residues 55-132) before (thin multiple contours) and after (thick
single contours) addition of a test co.,.~oulld. The final concentration of compound was 1.0
mM. The spectra (80 complex points, 8 scans/fid) were acquired on a 0.3 mM sample of the
RAF fragment in 20 nM phosphate (pH 7.0), lO mM DTT and 10% D20. Selected residues
that show ~ignifi~nt changes upon binding are inAic~teA
FIG. 6 shows lSN/lH correlation spectra of uniformly 15N-labeled FKBP before
{thin multiple contours) and after (thick single contours) addition of a test compound. The
final concentration of compound was l.0 mM. The spectra (80 complex points, 4 scans/fid)
was acquired on a 0.3 mM sample of FKBP in 50 mM phosphate (pH 6.5), lO0 mM NaCland 10% D20. Selected residues that show significant changes upon binding are in~ te~.
3~ FIG. 7 shows a first depiction of the NMR-derived structure of the DNA-binding
domain of E2. The two monomers of the symmetric dimer are oriented in a top-bottom
fashion, and the N- and C-termini of each monomer are inAi(~.ateA (N and C for one monomer,
N* and C* for the other). Shown in ribbons are the residues which exhibit .~ignifi{~nt
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
chPmical shift changes (ao(1H)>0.04 ppm; ao(l5N) >0.1 ppm) upon binding to a first test
compound. These residues co~ ond to the DNA-recognition helix of E2. Selected
residues are numbered for aid in vic~ 7~tinn.
FIG. 8 shows a second depiction of the NMR-derived structure of the DNA-binding
domain of E2. The two monomers of the ~y.~ . ;c dimer are ." ~~.~1~;1 in a top-bottom
fashion, and the N- and C~-termini of each monomer are in~lic~te~l (N and C for one mnnt~m~r,
N* and C* for the other). Shown in ribbons are the residues which exhibit eignific~nf
chemical shift changes (a~(lH)~0.04 ppm; ao(l5N) >0.1 ppm) upon binding to a second
test compound. These residues are located prirnarily in the dimer int~ ce region. Scle~ ed
residues are numbered for aid in vieu~li7~tinn.
FIG. 9 shows a depiction of the NMR-derived structure of the catalytic domain ofstromelysin. The N- and C-termini are intli~te-l Shown in ribbons are the residues which
exhibit cignif~ nt ch-qmic~l shift changes (ao(lH)>0.04 ppm; ~o(15N) ~0.1 ppm) upon
binding to a test compound. These either form part of the S 1 ' binding site or are spatially
ts ~ ~unal to this site. Selected residues are numbered for aid in vicll~li7~on.
FIG. 10 shows a ribbon plot of a ternary complex of first and second ligands bound
to the catalytic domain of stromelysin.
FIG. 11 shows the correlation between the NMR hin-lin~ data and a view of the
NMR-derived three-rlim~-ncinn~l structure of l?KBP.
FIG. 12 shows a ribbon plot of a ternary complex involving FKBP, a fragment
analog of ascomycin, and a b~n7~nili~le compound.
Detailed Description of the Invention
The present invention provides a rapid and ef~lcient method for ~l~ci~ninp ligands that
2~ bind to ~ d~eu~c targetmolecules.
T i~:~n~l~ are i(lentifitA by testing the binding of molecules to a target molecule (e.g.,
protein, nucleic acid, etc.) by following, with nuclear m~gnetic resonance (NMR)spectroscopy, the changes in chemic~l shifts of the target molecule upon the ~drlitinn of the
ligand compounds in the ~l~t~k~e
From an analysis of the cherni~zll shift changes of the target molecule as a function of
ligand concentration, the binding affinities of ligands for biomolecules are also detf.rrnin~-A.
The location of the binding site for each ligand is ~leterrnin~ from an analysis of the
chemical shifts of the biomolecule that change upon the addition of the ligand and from
nuclear Ov~rh~ r effects (NOEs) bt;lw~en the ligand and biomolecule.
Infc" IllaLion about the structure/activity relationships between ligands i~.ontifi~ by
such a process can then be used to design new drugs that serve as ligands to the target
molecule. By way of exarnple, where two or more ligands to a given target molecule are
nhfi~-l, a complex of those ligands and the target molecule is formed. The spatial
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
~" ;f~ ;nn of the ligands to each other as well as to the target molecule is derived from the
three-~im~n~ional ~L~u~ e. That spatial nriPnt~tion defines the ~ t~nf~e l~lwee~l the binding
sites of the two ligands and the ~ritont~ti~n of each ligand to those sites.
Using that spatial orientation data, the two or more ligands are then linked togeeher to
5 form a new ligand. T inking is accomplished in a manner that m~int~in~ the spatial orient~tion
of the ligands to one another and to the target molecule.
There are numerous advantages to the NMR-based discovery and design processes ofthe present invention. First, because a process of the present invention id~ntifi~ gands by
directly measuring binding to the target molecule, the problem of false positives is
10 ~ignifi~ntly recluced Because the presene process identifies specific binding sites to the
target molecule, the problem of false positives reslllting from the non-specific binding of
compounds to the target mc l~cllle at high co~ r, . ~ ;onc is kl i ~ t~ A
Second, the problem of false negatives is ~i~nific~ntly reduced because the present
process can identify compounds that specifically bind to the target molecule with a wide range
15 of dissociation constants. The dissociation or binding constant for compounds can actually
be ~leterrnin~d with the present process.
Other advantages of the present invention result from the variety and ~let~ fl data
provided about each ligand from the discovery and design processes.
Because the location of the bound ligand can be ~ir~ fl from an analysis of the
20 chemical shifts of the target molecule that change upon the addition of the ligand and from
nuclear Overh~ e.r effects (NOEs) between the ligand and biomolçcule, the binding of a
second ligand can be measured in the presence of a first ligand that is already bound to the
target~ The ability to simultaneously identify binding sites of dirr~ ligands allows a skilled
artisan to 1) define negative and positive cooperative binding between ligands and 2) design
25 new drugs by linking two or more ligands into a single compound while m~in~ining a proper
orientation of the ligands to one another and to their binding sites.
Further, if multiple binding sites exist, the relative affinity of individual binding
moieties for the dirÇeie"t Wnding sites can be measured from an analysis of the ch~mic~l shift
changes of the target molecule as a function of the added concentration of the ligand. By
30 simultaneously screening numerous structural analogs of a given compound, clet~ileA
structure/activity relationships about ligands is provided.
In part, the present invention provides a process of screening compounds to identify
ligands that bind to a specific target molecule. That process comprises the steps of: a)
generating a first two-~imen~ional lSN/lH NMR correlation spectrum of a 15N-labeled target
- 3~ molecule; b) exposing the labeled target molecule to one or more compounds; c) generating a
second two-dimensional 15N/lH NMR correlation spectrum of the labeled target molecule
that has been exposed to the compounds of step (b); and d) comparing the first and second
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
spectra to d~ e whether dirLr,.~ces in those two spectra exist, which differences in~1ins~t~
the presence of one or more ligands that have bound to the target molecule.
Where a process of the present invention screens more than one compound in step (b)
and where a dirr~.~,.lce l,~ ~n spectra is observed, additional steps are performed to identify
6 which specific compound is binding to the target ninlec~ s Those ~ isinll~l steps comprice
gen.or~ting a two-llim~ncional lSN/lH NMR correlation spectrum for each individual
compound and c~"..p,..;.-g each ~.~e~ um to the first spe~l-ul,. to determin~. whether
dirr~l~nces in any of those co~ ,d spectra exist, which ~liLrGlGIlces intiif~tr the presence of
a ligand that has bound to the target molecule.
Any 15N-labeled target molecule can be used in a process of the present invention.
Because of the .~ olL~Ice of proteins in medicinal çh~mictry, a ~lGrGll~ target molecule is a
polypeptide. The target m~lçcnl~. can be labeled with 15N using any means well known in
the art In a ~ ,r~ ,d çmho~ nt the target molecule is prepared in l~,COlllbill~l~ form using
transrc,l"~cd host cells. In an especially ~lGr~Gd embodiment, the target molecule is a
polypeptide. Any polypeptide that gives a high resolultion NMR ~e~ ulll and can be
partially or ullirollllly labeled with lSN can be used. The ~lG~dld~ion of ulliru~ ly 15N-
labeled exemplary polypeptide target molecules is set forth hGlcillarlGl in the Examples.
A L~ rGllGd means of plG~illg adequate q~l~ntiti~s of uni~ullllly lSN-labeled
polypeptides is to transform a host cell with an G~ ssion vector that contains apolynucleotide that encodes that polypeptide and culture the tran~roll--ed cell in a cultore
medium that contains ~ccimil~hle sources of lSN. Accimil~hle sources of 15N are well
known in the art. A pLGrGll~,d such source is 15NH4Cl.
Means for p~GJJdlillg G~ ion vectors that contain polynucleotides encoding specific
polypeptides are well l~nown in the art. In a similar manner, means for transforming host
cells with those vectors and means for culturing those transformed cells so that the
polypeptide is expressed are also well known in the art.
The screening process begins with the generation or acquisition of a two-~lim~c-onal
lSN/lH correlation spectrum of the labeled target molecul~. Means for generating two-
dimensional 1SN/lH correlation spectra are well known in the art [See, e.g., D. A. Egan~ et
al., Biochemistry, 32:8, pgs. 1920-1927 (1993); Bax. A.~ (~rzesiek~ S., Acc. Chem. Res.,
26:4, pgs. 131-138 (1993)].
The NMR spectra that are typically recorded in the screening procedure of the present
invention are two-riim~ncinnal 15N/1H heteronuclear single ~luanlul-l correlation (HSQC)
spectra. Because the lSN/~H signals corresponding to the backbone amides of the proteins
are usually well-resolved, the chernical shift changes for the individual arnides are readily
monitored.
In generating such spectra, the large water signal is suppressed by spoiling gradients.
To facilitate the acquisition of NMR data on a large number of compounds (e.g., a ~l~t~b~ce
CA 02237336 1998-05-11
WO 97/18469 PCTtUS96/18312
of synthetic or naturally occurring small organic compounds), a sample changer is employed.
Using the sample changer, a total of 60 samples can be run lm~t~nde-l Thus, using the
typical acquisition p~ f t''--~. (4 scans per free induction decay (fid), 100-120 HSQC spectra
can be acquired in a 24 hour period.
To facilitate procescing of the NMR data, con-l,ut~,l programs are used to ~ .r~,. and
o~ lic~lly process the multiple two--lim~nsion~l NMR data sets, including aroutine to
~ntom~ti~lly phase the two-.l;."e.-sional NMR data. The analysis of the data can be
facilitated by rO" ,~ .g the data so that the individual HSQC spectra are rapidly viewed and
compared to the HSQC ..~e~ w~ of the control sample c~ i.-g only the vehicle for the
added compound (DMSO), but no added compound. Detailed descriptions of means of
generating such two--lim~o-ncional 15N/lH correlation spectra are set forth hele"lar~ in the
Examples.
A r~ ,se~ /e two-~iim~.nsional lSNIlH NMR correlation ~7~JeCIlUlll of an lSN-
labeled target molecule (polypeptide) is shown in FIG. 1 (the DNA-binding domain of the E2
1 5 protein).
Following acquisition of the first ~ye~ ulll~ the labeled target molecule is exposed to
one or more test compounds. Where more than one test conlpowld is to be tested
~cimlllt~n~ously, it is ~,~GÇellcd to use a d~t~k~ce of compounds such as a plurality of small
molecules. Such molecules are typically dissolved in perdc;uL~ d dimethylsulfoxide . The
compounds in the ~t~kace can be purchased from vendors or created according to desired
needs.
Individual compounds can be select~cl inter alia on the basis of size (molecular weigh~
= 100-300) and molecular diversity. Compounds in the collection can have dirreiellt shapes
(e.g., flat aromatic rings(s), puckered aliphatic rings(s), straight and branched chain aliphatics
with single, double, or triple bonds) and diverse functional groups (e.g., carboxylic acids,
esters, ethers, ~mine,s, aldehydes, ketones, and various heterocyclic rings) for n~ ing the
possibility of discovering compounds that interact with widely diverse binding sites.
The NMR screening process utilizes ligand concentrations ranging from about 0.1 to
about 10.0 mM. At these concentrations, compounds which are acidic or basic can
cignific~ntly change the pH of buffered protein solutions. Chemucal shifts are sensitive to pH
changes as well as direct binding interactions, and "false positive" ch~,mic~l shift changes,
which are not the result of ligand binding but of changes in pH, can therefore be observed. It
~ is thus necessary to ensure that the pH of the buffered solution does not change upon addition
of the ligand. One means of controlling pH is set forth below.
Compounds are stored at 263~K as 1.0 and 0.1 M stock solutions in
dimethylsulfoxide (DMSO). This is nececc~ry because of the limited solubility of the ligands
in aqueous solution. It is not possible to directly adjust the pH of the DMSO solution. In
.
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
- 10-
addition, HCl and NaOH form insoluble salts in DMSO, so ~lt~rn~tive acids and bases must
be used. The following approach has been found to result in stable pH.
The 1.0 M stock solutions in DMSO are diluted 1:10 in 50 mM phosphate, pH 7Ø
The pH of that diluted aliquot solution is measured. If the pH of the aliquot is Im~h:~n~ed
5 (i.e., re",ail,s at 7.0), a working solution is made by ~ tin~ the DMSO stock solution 1:10
to make a 0.1 M solution and that solution is stored.
If the pH of the diluted aliquot is less than 7.0, ethanolamine is added to the 1.0 M
stock DMSO sntlltion, that stock solution is then diluted 1:10 with phosphate buffer to make
another aliquot, and the pH of the aliquot rechecked.
If the pH of the diluted aliquot is greater than 7.0, acetic acid is added to the 1.0 M
stock DMSO solution, that stock solution is then diluted 1:10 with phosphate buffer to make
anotner aliquot, and the pH of the aliquot ~ L rA
Ethanolarnine and acetic acid are soluble in DMSO, and the proper equivalents are
added to ensure that upon transfer to aqueous buffer, the pH is unchanged. Adjusting the pH
is an interactive procecs, repeated until the desired result is obL~ined.
Note that this procedure is performed on 1:10 dilutions of 1.0 M stock solutions (100
mM ligand) to ensure that no pH changes are observed at the lower cormçntr~tionc used in the
,filllGnLs (0.1 to 10 rnM) or in dirrGlGIlt/~,dker buffer ~,y ,L~n~s.
Following exposure of the 15N-labeled target molecule to one or more test
compounds, a second two--lirnencional 15N/lH NMR correlation spectrum is gen~r~t~
That second spectrum is generated in the same manner as set forth above. The first and
second spectra are then col,.~ ed to determine whether there are any differences beL~n the
two spectra. Differences in the two-flimPncion~l 15N/1H NMR correlation spectra that
in~ te the presence of a ligand correspond to 15N-labeled sites in the target molecule.
26 Those di~GrG-Ices are de~Gl"~ned using standard procedures well known in the art.
By way of example, FIGs. 2, 3, 4, 5 and 6 show comparisons of correlation spectra
before and after exposure of various target molecules to various test compounds. A ~let~
descnption of how these studies were performed can be found hereinafter in Examples 2 and
3.
Particular signals in a two-~imencional 15N/lH correlation spect-rum correspond to
specific nitrogen and proton atoms in the target molecule (e.g., particular amides of the amino
acid residues in the protein). By way of example, it can be seen from FIG. 2 that chemical
shifts in a two-dimensional 15N/lH correlation of the DNA-binding domain of E2 exposed to
a test compound occurred at residue positions 15 (I15), 21 (Y21), 22 (R22) and 23 (L23).
It can be seen from FIG. 2 that the binding of the ligand involved the isoleucine (Ile)
residue at position 15, the tyrosine (Tyr) residue at position 21, the arginine (Arg) residue at
position 22 and the leucine (Leu) residue at position 23. Thus, a process of the present
CA 02237336 1998-05-11
WO 97/18469 PCTtUS96tl8312
- 11 -
invention can also be used to identify the specific binding site bcL~ c~ll a ligand and target
molecule.
The region of the protein that is responsible for binding to the individual compounds
is ic~en*fie~ from the particular amide signals that change upon the addition of the
5 compounds. These signals are ~cign~d to the individual amide groups of the protein by
standard procedures using a variety of well-established heteron~ e~r multi-.li...~ nal NMR
To discover molecules that bind more tightly to the protein, molecnl~5 are selected for
testing based on the ~L~u~;Lul~/activity rel~tic)n~hiI s from the initial screen and/or structural
10 information on the initial leads when bound to the protein. By way of example, the initial
screening may result in the i~lentifi~tion of lig~ncls> all of which contain an ~ laLiC ring.
The second round of scr~ni-lg would then use other aromatic molecules as the test
compounds.
As set forth helci~larLc~ in Example 2, an initial s~ elullg assay for binding to the
15 catalytic domain of stromelysin i~l~ntifieA two biaryl compounds as lig~nrl.~ The second
round of screening thus used a series of biaryl derivatives as the test colllyou,lds.
The second set of test compounds are initially screened at a c~ CÇn I ~ ~11 ;on of 1 mM,
and binding constants are measured for those that show affinity. Best leads that bind to the
protein are then col}l~alcd to the results obtained in a functional assay. Those compounds
20 that are suitable leads are ~hemi~lly modified to produce analogs with the goal of discovering
a new ph~ eutical agent.
The present method also provides a process for de~Glll~li~lg the dissociation constant
between a target molecule and a ligand that binds to that target molecule. That process
comprises the steps of: a) generating a first two-~imt-n~innal lSN/lH NMR correlation
25 spectrum of a 15N-labeled target molecule; b) titrating the labeled target molecule with various
concentrations of a ligand; c) gGnGl~Ling a two--lim--n~iona} lSN/lH NMR correlation
spectrum at each concentration of ligand from step (b); d) comparing each spectrum from step
(c) both to the first spectrum from step (a) and to all other spectra from step (c) to 4uanLiry
difr~,~cnces in those spectra as a function of changes in ligand concentration; and e) calculating
30 the dissociation constant (KD) beLween the target molecule and the ligand from those
dirr~lellces.
Because of their importance in medicinal chemistry, a ~lGrGlred target molecule for
~ use in such a process is a polypeptide. In one L~IGrellGd embodiment, a process of
de~.lllilling the dissociation constant of a ligand can be ~Glfolllled in the presence of a second
35 ligand. In accordance with this embodiment, the 15N-labeled target mo}ecule is bound to that
second ligand before exposing that target to the test compounds.
Binding or dissociation constants are measured by following the lSN/lH chf mic~lshifts of the protein as a function of ligand concentration A known concentration ([P]O) of
CA 02237336 1998-05-11
WO 97/184C9 PCT/US96/18312
the target mo~ is mixed with a known cqnc en~tion ([L]o) of a previously iclçntif
ligand and the two--iim~ncional 15N/lH correlation spe~;L-u,ll was acquired. From this
spectrum, observed chemical shift values (~obs) are obtained. The process is r~eaL,d for
varying concentrations of the ligand to the point of s~tl-r~tiQn of the target molecule, when
5 possible, in which case the limiting chçmic~1 shift value for saturation ~~sa~ is measured.
In those ~itn~-on~ where saturation of the target molecule is achieved, the dissociation
collsL~L for the binding of a particular ligand to the target molçcnle is c~ ted using the
formula
KD = ([P]O - X) ([~]0 - X)
where [P]o is the totaI molar concentration of target mnl~nle; [L]o is the total molar
10 co,~e"L,~lion of ligand; and x is the molar concentration of the bound species. The value of x
is ri~ ~A from the equation:
X = ~obs ~ ~free
where of ree is the chernil~l shift of the free species; ~obs is the observed chemi~l shift; and
is the dirr~ ce belw~en the limitin~ ch~mi~:~l shift value for saturation (~sat) and the
chemic~l shift value of the target molecule free of ligand (~free)-
The ~ oçi~tir~n constant is then ~leterrnine~ by varying its value until a best fit to theobserved data is obtained using standard curve-fitting statistical methods. In the case whe
re
~sat is not directly known, both KD and ~sat are varied and subjected to the same curve-
fifflng procedure.
The use of the process described above to ~let~rmine the dissociation or binding20 affinity of various ligands to various target mflle~ules is set forth hereinafter in Examples 2
and 3.
Preferred target molecules, means for generating spectra, and means for comparing
spectra are the same as set forth above.
In its principal aspect, the present invention provides a process of ~ie~ignin~ new
2~ ligands that bind to a specific target molecule by linl~ng together two or more molecules that
bind to the target molecule.
The initial step in the design process is the j(lentific~tic-n of two or more ligands that
bind to the specific target molecule. The identification of such ligands is done using two-
~limt n~ional 15N/1H NMR correlation spectroscopy as set forth above.
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
- 13-
Once two or more ligands are iclentifi~d as binding to the target molecule at diL
sites, a complex bGtv~,~;n the target molecule and ligands is formed. Where there are two
n~lc, that complex is a ternary complex. Q..~ . y and other complexes are formedwhere there are three or more lig~n(1s
Complexes are formed by mixing the target molecule simlllt~ne,ously or sequentially
with the various ligands under ~ çs that allow those ligands to bind the target.
Means for d- t~ lhlg those conditions are well known in the art.
Once that complex is formed, its three-rlim~nsion~l structure is ~let~ h~ Any
means of deLG~ g three-din e.ncinn~l structure can be used. Such mfthorls are well
1 ç known in the art. Exemplary and p,~rG,l~d methods are NMR and X-ray crystallography.
The use of three-~lim~nsion~l double- and triple resonance NMR to ~l~l .. ;nP the three-
ion~ Llu~;Lulc of two ligands bound to the catalytic domain of stromelysin is set forth
in detail h~ ft~,~ in F.xampl~ 4.
An analysis of the three-(l; --~ )~ional SLluc~ulc reveals the spatial nrient~tion of the
ligands relative to each other as well as to the conrol,llaLion of the target molecule. First, the
spatial oriP.nt~tion of each ligand to the target molecule allows for illentific~tion of those
portions of the ligand directly involved in binding (i.e., those portions interacting with the
target binding site) and those portions of each ligand that project away from the binding site
and which portions can be used in subsequent linking procedures.
Second, the spatial rtrient~tion data is used to map the positions of each ligand relative
to each other. In other words, discrete ~list~n~e~s between the spatially oriented ligands can be
calculated.
Third, the spatial orientation data also defines the three-~iimPnsional relationships
amongst the ligands and the target. Thus, in addition to calculating the absolute ~iist~nç~s
between lig~n-l~, the angular orient~ti~ns of those ligands can also be det~ ~--i--e~
Knowledge of the spatial orientations of the ligands and target is then used to select
linkers to link two or more ligands together into a single entity that contains all of the lip~n~is
The design of the l~nkers is based on the distances and angular nriçnt~tion needed to ~ "~h~
each of the ligand portions of the single entity in proper orie~nt~t;l~n to the target.
The three--lim.on~ional conformation of suitable linkers is well known or readily
ascertainable by one of ordinary skill in the art. While it is theoretically possible to link two
or more ligands together over any range of ~list~nce and three--lim~-n~ional projection, in
practice certain limit~tinns of distance and projection are ~lGÇc;lled. In a L~l~r~l~d
embodiment, ligands are sepa,~Led by a distance of less than about 15 Angstroms (A), more
preferably less than about 10 ~ and, even more preferably less than about 5 A.
Once a suitable linker group is identified, the ligands are linked with that linker.
Means for linking ligands are well known in the art and depend upon the chemiçSIl structure of
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
- 14-
thc ligand and the linking group itself. T ig~nrls are linked to one another using those portions
of the ligand not directly involved in binding to the target molecule.
A deta~led descli~Lion of the design of a drug that inhibits the proteolytic activity of
stromelysin, which drug was cle~ignçtl using a process of the present invention is set forth
5 hc,.~,~7~ in Example 4.
The following Examples illustrate ~ Gd embotli~ of the present invention and
are not limiting of the ~perific~tion and claims in any way.
Example 1
]~ aL~.Lion Of Uniformly 15N-Labeled TargetMolecules
A. Stromelysin
Human stromelysin is a 447-arnino acid protein believed to be involved in proteolytic
degradation of cartilage. Cartilage proteolysis is believed to result in tlçgr,7r7.~tive loss of joint
cartilage and the resnlting ;~ 1 of joint function observed in both osteoarthritis and
rh~l-m~toirl arthritis. The protein possesses a series of r,70~inc incl7lrlin~ N-t~rn~in~l latent
and ~G~Lide domains, a C-terminal domain homologous with homopexin, and an internal
catalytic clom~in
Studies have shown that removal of the N-tr-rmin~l prosequence of approximately
eighty amino acids occurs to convert the proenzyme to the 45 kDa mature enzyme.
Furthr,rmore, studies have shown that the C-terminal homopexin homologous domain is not
required for proper folding of the catalytic domain or for interaction with an inhihitor (See,
e.g., A. I. Marcy, Biochemistrv~ 30: 6476-6483 (1991). Thus, the 81-256 amino acid
residue intt-rn~l segment of stromelysin was selected as the protein fragment for use in
identifying compounds which bind to and have the potential as acting as inhihitnr~ of
stromelysin.
To employ the method of the present invention, it was neces~. y to prepare the 81-
256 fragment (SEQ lD N0: 1) of stromelysin in which the peptide backbone was isotopically
enriched with and 15N. This was done by inserting a plasmid which coded for the
production of the protein fragment into an E. coli strain and ~rowing the g~nr-t~ ly-modified
bacterial strain in a limiting culture medium enriched with NH4Cl and C-glucose.The isotopically t-.nri(~h~1 protein fr~gm~nt was isolated from the culture medium,
purified, and subsequently used as the basis for ev~ ting the binding of test compounds.
The procedures for these processes are described below.
Human skin fibroblasts (ATCC No. CRL 1507) were grown and int11lcefl using the
procedure described by Clark et al., Archiv. Biochem. and Biophys., 241: 36-45 (1985).
Total RNA was isolated from 1 g of cells using a Promega RNAgentst~ Total RNA Isolation
System Kit (Cat.# Z5 1 10, Promega C~orp., 2800 Woods Hollow Road, Madison, WI 5371 1-
5399) following the manufacturer's instructions. A 1 ,ug por~on of the RNA was heat-
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
- 15-
den~Lul~,d at 80~C for five ~ uL~S and then ~,ubje~;lGd to reverse transcriptase PCR using a
GeneAmp~ RNA PCR kit (Cat.~ N808-0017, Applied Biosystems/Perkin-Elmer, 761 MainAvenue, Norwalk, CT 06859-0156) following the manufacturer's instructions.
Nested PCR was pe,ru~ ed using first primers (A) GAAATGAAGAGTC TTCAA
(SEQ ID NO:3) and (B) GCGTCCCAGGTTCTGGAG (SEQ ID NO:4) and thirty-five cycles
of 94~C, two minutes; 45 C, two mi~luLGs, and 72~C three min~ltes This was followed by
reamplification with internal ~lilllGl~7 (C) ATACCATGGCCTATCCAT TGGATGGAGC
(SEQ ID NO:5) and (D) ATAGGATCCTTAGGTCTCAGGGGA GTCAGG (SEQ ID NO:6)
using thirty cycles under the same conditions described ul~ ly above to generate a DNA
coding for amino acid residues 1 -256 of human stromelysin.
The PCR fragment was then cloned into PCR cloning vector pT7Blue~R) (Novagen,
Inc., 597 Science Drive, Madison, WI 53711) according to the m~mlf~ctllrer's instrnction~
The resllltin~ plasmid was cut with NcoI and BamHI and the stromelysin fr~m~nt was
subcloned into the Novagen expression vector pET3d (Novagen, Inc., 597 Science Drive,
Madison, WI 53711), again using the m~nnf~r,turer's instructions.
A mature stromelysin expression construct coding for amino acid residues 81-256
plus an inhi~*ng methionine was generated from the 1-256 expression construct by PCR
~mplifir~tion The resulting PCR fragment was first cloned into the Novagen pT7Blue(R)
vector and then subcloned into the Novagen pET3d vector, using the m~nnf~rtnrer's
instructions in the manner described above, to produce plasmid (pETST-83-256). This final
plasmid is if lentic~l to that described by Qi-Zhuang et al., Biochemistry. 31: 11231 - 11235
(1992) with the exception that the present codes for a peptide sequence beginning two amino
acids earlier, at position 81 in the sequence of human stromelysin.
Plasmid pETST-83-256 was transformed into E. coli strain BL21(DE3)/pLysS
2~; (Novagen, Inc., 597 Science Drive, Madison, VVI 53711) in accordance with the
manufacturer's instructions to generate an expression strain, BL21(DE3)/pLysS/pETST-255-
1.
A preculture l..e~ . was p~c~al~;d by dissolving 1.698 g of Na2HP4-7H2O~ 0.45 g
of KH2P04, 0.075 g NaCl, 0.150 g 1SNH4Cl, 0.300 13C-glucose, 300 ~lL of lM aqueous
MgSO4 solution and 15 ,uL of aqueous CaC12 solution in 150 mL of deionized water.
The resulting solution of preculture medium was sterilized and transferred to a sterile
500 mL baffle flask. Tmmt~ t--ly prior to inoculation of the preculture medium with the
bacterial strain, 150 ~L of a solution con~ining 34 mg/mL of chloramphenicol in 100%
ethanol and 1.5 mL of a solution cont~inin~ 20 mg/mL of ampicillin were added to the flask
contents.
The flask contents were then inoculated with 1 mL of glycerol stock of genetically-
modified E. Coli, strain BL21(DE3)/pLysS/pETST-255-1. The flask contents were shaken
(225 ~pm) at 37~C until an optical density of 0.65 was observed.
WO 97/18469 PCT/US96tl8312
- 16-
Af. .~ l;nnnntri~nt"~r~ waspreparedbydissolving 113.28gofNa2~IP4-7H20, 30 g of KH2PO4, 5 g NaCl and 10 mL of 1% DF-60 antifoam agent in
9604 mL of deionized water. This solution was placed in a New Brunswick Scientific
Micros r~,L~ ,nter (E~dison, NJ) and st~rili7r~l at 121~C for 40 minutes.
Tmm~ t-~.1ypriortoinoculationoftheÇ~.. -Il~.l-lnl;on~-,P.Ii-"--, thefollowingpre-
sterilized components were added to the f~.rm~nt~tion vessel contents: 100 mL of a 10%
aqueous solution of 15NH4Cl, 100 mL of a 10% aqueous solution of 13c-glllrose~ 20 rnL of
an aqueous lM solution of MgSO4, 1 mL of an aqueous lM CaC12 solution, 5 mL of an
aqueous solution of thiamin hydrochloride (10 mg/mL), 10 mL of a solution co~ ;.l;llp 34
mg/mL of chlor~mph- ni~ol in 100% ethanol and 1.9 g of ampicillin dissolved in the
chloramphenicol solution. The pH of the r~snlting solution was adjusted to pH 7.00 by the
addition of an aqueous solution of 4N H2SO4.
The preculture of E. Coli, strain BL21(DE3)/pLysS/pETST-255-1, from the shake-
flask scale procedure described above was added to the r~ .., cnnl ~ .. " ~i and cell growth
5 was ailowed to proceed until an optical density of 0.48 was achieved. During this process,
the f~rmr~ntpr contents were ~uLu~ llc~lly mslint~inr~cl at pH 7.0 by the ~d~ition of 4N H2S04
or 4N KOH as needed. The dissolved oxygen content of the r~..~ contents was
m~.int,.inr d above 55% air s?~hlr~.hion through a c~cad~ 1 loop which increased agitation speed
when the dissolved oxygen content ~ko~ed below 55%. Air was fed to the f~
conLGllLs at 7 standard liters per minute (SLPM) and the culture te. l lp~ ., was .~ 1 at
37'C throughout the process.
The cells were harvested by centrifugation at 17,000 x g for 10 minutes at 4''C and the
resulting cell pellets were collected and stored at -85~C. ~he wet cell yield was 3.5 g/L.
Analysis of the soluble and insoluble fractions of cell lysates by sodium dodecyl sulfate
26 polyacrylamide gel electrophoresis (SDS-PAGE) revealed that a~lu~-aL~ly 50% of the
5N-stromelysin was found in the soluble phase.
The isotopica71y-labeled stromelysin fragment prepared as ~leserihed above was
purified employing a modification of the technique described by Ye, et al., E3iochemistry. 31:
1 123~-1 1235 (1992).
The harvested cells were suspended in 20 mM Tris-HCI buffer (pH 8.0) sodium azide
solution containing 1 mM MgC12, 0.5 mM ZnC12, 25 units/mL of Benzonase~ enzyme, and
an inhibitor ~ Lul~; made up of ~(2-aminoethyl)-benzenesulfonyl flllon~le ("AEBSF"),
Leupeptin(~, Aprotinin(~9, and Pepstatin(~ (all at concentrations of 1 ~lglmL. AEBSF,
Leupeptin(~. A~ Linil~ , and Pe~:jL~ are available from ~Am~riczln Tn~-rn~ nal
Chemical, 17 SLl~Ll-l~ore Road, Natick, MA 01760.)
The resulting Il~ixLulc was gently stirred for one hour and then cooled to 4~C. The
cells were then sonically disrupted using a 50% duty cycle. The resulting lysate was
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/183i2
- 17-
centrifuged at 14,000 rpm for 30 .~ es and the pellet of in~o}llble fraction frozen at -80~C
for subsequent processing (see below).
Solid ~mmonillm sulfate was added to the supernatant to the point of 20% of
saturation and the reslllting solution loaded onto a 700 mL phenyl sepharose fast flow ("Q-
Sepharose FF") column (Pharmacia Biotech., 800 (~çntennizll Ave., P. O. Box 1327,
Piscataway, NJ 08855). Prior to loading, the sepharose column was equilibrated with 50
mM Tris-HCl buffer (pH 7.6 at 4~C), 5 mM CaC12, and 1 M (NH4)2SO4. The loaded
column was eluted with a linear gradient of decreasing co~ ns of aqueous
(NH4)2S04 (from 1 down to 0 M) and increasing concentr~tic)ns of aqueous CaC12 (from 5
o to 20 mM) in Tris-HCl buffer at pH 7.6.
The active fractions of eluate were collected and colue~ at~d in an Amicon stirred cell
(Amicon, Inc., 72 Cherry Hill Drive, Beverly, MA 01915). The conce~ a~d sample was
dialyzed overnight in the starting buffer used with the Q-Sepharose FF column, 50 mM Tris-
HCl (pH 8.2 at 4~C) with 10 mM CaCk.
The dialy_ed sample was then loaded on the Q-Sepharose FF column and eluted witha linear ~r~ nt conlplising the starting buffer and 200 mM NaCl. The purified soluble
fraction of the isotopically-labeled stromelysin fragment was concentrated and stored at 4~C.
The pellet was s~ hili7~A in 8M g~ n~ ne-Hcl. The solution was centrifuged for 20
...i..~l~esat20,000rpmandthe:~ul)e...~t~ntwasaddeddropwisetoafoldingbuffer
comprising 50 mM Tris-HCl (pH 7.6), 10 mM CaC12 0.5 mM ZnC12 and the inhibitor
coc~ctail of AEBSF, Leupeptin~, A~lo~ , and Pepstatin(~ (all at concentrations of 1
,ug/mL). The volume of folding buffer was ten times ehat of the sup~rn~t~nt The mixture of
supernatant and folding buffer was centrifuged at 20,000 rpm for 30 min~ltes
The supem~t~nt from this centrifugation was stored at 4~C and the pellet was
subjected twice to the steps described above of sol~lhili7~tion in g~ niflin~-HCl, refolding in
buffer, and centrifugation. The final :~Up~Lllaklllt~ from each of the three centrifugations were
combined and solid ammonium sulfate was added to the point of 20% sZ~tnr~tion The
resulting solution thus derived from the insoluble fraction was subjected to pnrific~tinn on
phenyl Sepharose and Q-Sepharose as described above for the soluble fraction.
The purffled soluble and insoluble fractions were combined to produce about 1.8 mg
of purified isotopically-labeled stromelysin 81-256 fragment per gram of original cell paste.
B. Human papillomavirus ~HPV~ E2 Inhibitors
The papillomaviruses are a family of small DNA viruses that cause genital warts and
cervical carcinomas. The E2 protein of HPV regulates viral transcription and is required for
viral replication. Thus, molecules that block the binding of E2 to DNA may be useful
therapeutic agents against HPV. The protein rather than the DNA was chosen as a target,
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/183i2
- 18-
bccau~.e it is expected that agents with greater selectivity would be found that bind to the
protein rather than the DNA.
The DNA-binding domain of human papillomav~rus E2 was cloned from the full
length DNA that codes for E2 using PCR and ovc~ lcssed in bacteria using the T7
5 c;~ ion system. Uniforrnly 15N-labeled protein was isolated from b?~et~.ri~ grown on a
minim~l n,eJiu~-l c~ i.lg lSN-labeled protein was isolated from b~f~t~ori~7 grown on a
minim~l m-qAinm co~ ;rlg lSN-labeled arnmonium chlc ririe The protein was purified from
the bacterial cell lysate using an S-sepharose FastFlow column ~l~-e~l..;lihr~te~i with buffer
(50 mM Tris, 100 mM NaCl, 1 mM EDTA, pH = 8.3).
The protein was eluted with a linear gradient of 100-500 mM NaCl in buffer, pooled,
and applied to a Mono-S column at a pH = 7Ø The protein was eluted with a salt gradient
(100-500 mM), conc~ LIaLed to 0.3 mM, and exchanged into a TRIS (50 mM, pH = 7.0buffered H20/D20 (9/1) solution conL~...ir-g sodium azide (0.5%).
C. RAF
Uniformly lSN-labeled Ras-binding domain of the RAF protein was ~.c~,d as
described in Emerson et al., Biochemistry. 34 (21): 6911-6918 (1995).
D. FKBP
Uniformly lSN-labeled recombinant human 3:~K binding protein (FKBP) was
prepared as described in Logan, et al., J. Mol. Biol., 236: 637-648 (1994).
Example 2
Screening Compounds Using Two-Dirnensional 15N/IH
NMR Correlation Spectral ~n~lysis
Tne catalytic domain of stromelysin was ~ al~i in accordance with the proceduresof Example 1. The protein solutions used in the screening assay contained the unilol~-ly
lSN-labeled catalytic domain of stromelysin (0.3 mM), aceLohy~ xarnic acid (S00 mM),
CaC12 (20 mM), and sodium azide (0.5%3 in a H20/D20 (9/1~ TRIS buffered solution (50
mM, pH=7.0).
Two--limen~i-)nal 15N/IH NMR spectra were generated at 29~C on a Bruker
AMX500 NMR ~.~e~Ll~nl;;L~l equipped with a triple resonance probe and Bruker sample
changer. The 15N/lH HSQC spectra were acquired as 80 x 1024 complex points usingsweep widths of 2000 Hz (15N, tl) and 8333 E~z (lH, t2). A delay of 1 second l)e~wt;en
scans and 8 scans per free induction decay(fid) were employed in the data collection. All
NMR spectra were processed and analyzed on Silicon Graphics COIllpU~ . using in-house-
written software.
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
- 19-
A first two-~ cion~l l5N/lH NMR correlation S~c il~unl was ac~luucd for the
l5N-labeled stromelysin target molecule as described above. The stromelysin target was then
exposed to a d~t~h~e of test compounds. Stock solutions of the compounds were made at
lO0 mM and l M. In addition, a combin~t;on library was ~lG~d that contained 8-lO5 compounds per sample at a concentration of lO0 mM for each compound.
The pH of the l M stock solution was adjusted with acetic acid and eth~nol~mine so
that no pH change was observed upon a l/lO dilution with a lO0 mM phosphate burr~ ,d
solution (pH = 7.0). It is important to adjust the pH, because small ch~n~s in pH can alter
the ch~mi~l shifts of the biomolecules and complicate the i~ ylckllion of the NMR data.
The compounds in the ~l~t~bz-~e were selected on the basis of size (mnl~cnlslr weight =
100-300) and molecular diversity. The molecules in the collection had (li~r,.c~-t shapes (e.g.,
flat a~ulll~Lic rings(s), ~u~;h~,.cd aliphatic rings(s), straight and branched chain ~iph~ti~s with
single, double, or triple bonds) and diverse functional groups (e.g., carboxylic acids, esters,
ethers, amines, aldehydes, ketones, and various heterocyclic rings) for m~ximi7.ing the
possibility of discovering compound that interact with widely diverse binding sites.
The NMR samples were ~l'c~al~d by adding 4 ~l of the DMSO stock solution of the
compound n~i~Lulcs that contained each compound at a cul~clllla~ion of lO0 mM to 0.4 rnl
H20/D20 (9/1) burr~l~ solution of the uniformly 15N-labeled protein. The final
concentration of each of the compounds ;n the NMR sample was about l mM.
In an initial screen, two compounds were found that bind to the catalytic domain of
stromelysin. Both of these compounds contain a biaryl moiety. Based on these initial hits,
structurally simila-r compounds were tested against stromelysin. The s~ucture of those biaryl
compounds is represented by the structure I, below. (See Table I for definitions of Rl-R3
and A l-A3).
~ A~A3- R3
R2
In the second round of screening, binding was assayed both in the absence and in the
presence of sa~ulaLi~lg amounts of acetohydroxamic acid (500 mM).
Many of the biaryl compounds were found to bind the catalytic domain of
stromelysin. FIG. 4 shows a r~l~s~ntative two--iimton~ional l5N/lH NMR correlation
spectrum before and after exposure of stromelysin to a biaryl test compound. It can be seen
30 from FIG. 4 that the compound caused chemical shifts of lSN-sites such as those ~lesi~n;-t.od
Wl24, Tl87, Al99 and G204.
These sites correspond to a tryptophan (Trp) residue at position 124, a threonine (Thr)
at position l 87, an alanine (Ala) at position l99, and a glycine (Gly) at position 204 of SEQ
ID NO. l. FIG. 9 shows the correlation between the NMR binding data and a view of the
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/1831Z
-20-
NMR-derived three~lim~n~ional ~u~ uc of the catalytic domain of stromelysin. The ability
to locate the specific binding site of a particular ligand is an advantage of the present
invention.
Some compounds only bound to stromelysin in the presence of hyd~ alllic acid.
Thus, the bin~ling affinity of some compounds was enh~nccd in the presence of the
hydroxarnic acid (i. e. cooperative). These results ~Y~mrlify another important c~r~bility of
the present S~;l~.lillg assay: the ability to identify compounds that bind to the protein in the
presence of other molecules.
Various biaryl compounds of ~ u~ c I were tested for binding to stromelysin at
10 differing concentrations. The 15N/lH spectra ~ d at each concer~ tinn were evaluated
to quantify ~liLr~nces in the spectra as a function of compound conf~ntr~ti~n A binding or
dissociation constant (KD)was calcnl~tçA, using standard procedulcs well known in the art,
from those dirrc.~nces. The rçsults of this study are shown in Table 1. The values for Rl-
R3 and Al-A3 in Table 1 refer to the corresponding positions in the structure I, above.
Table 1
Compound No. Rl R2 R3 Al A2 A3KD(mM)
H OH H C C C 1.1
2 CH2OH H H C C C 3.2
3 Br H OH C C C 1.3
4 H H H N N C 1.6
CHO H H C C C 1.7
6 OCH3 NH2 H C C C 0.4
7 H H H N C C 0.2
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
-21 -
Table 1 (Continued)
Compound No. Rl R2 R3 Al A2 A3 KD(mM)
8 OCOCH3 H H C C C~ 0.3
9 OH H OH C C C0.01
H H H N C N0.4
11 OH H H C C C0.3
12 OH H CN C C C0.01
The data in Table 1 show the utility of a process of the present invention in
.lCt~ ~g dissociation or binding constants between a ligand and a target molecule.
Another advantage of an NMR sl;lce~ g assay of the present invention is the
ability to correlate observed chP.mic~,~l shifts from the two-~ n~ n~l 15N/lH NMR
correlation spectra with other spectra or pro~ections of target molecule configuration. The
results of a representative such correlation are shown in FIG. 9, which depicts regions within
10 the polypeptide at which binding ~,vith the substrate molecule is most likely occurring. In ~his
Figure, the ~pd~ Wnding regions in stromelysin are shown for Compound 1 (from Table
1).
Compounds from the ~l~t~b~e were screened in a similar manner for binding to theDNA-binding ~lon~z~in of the E2 protein. Those compounds had the structure II below, where
Rl-R4 and A are defined in Table 2.
R~
~A~ R3
R2 R4
II
NMR exp~nmt~nf~ were performed at 29~C on a Bruker AMX500 NMR specllollle~
equipped with a triple resonance probe and Bruker sample changer. The 15N-/lH HSQC
- spectra were acquired as 80 x 1024 complex points using sweep widths of 2000 Hz (15N,t~ )
and 8333 Hz (lH, t2). A delay of 1 second between scans and 4 scans per free induction
20 decay were employed in the data collection. All NMR spectra were processed and analyzed
on Silicon Graphics computers.
CA 02237336 1998-0~-
wo 97/18469 PCT/US96/1831
-22 -
PIGs. 2 and 3 show represent~tive two-Aimen~io~ SN/IH NMR correlation spectra
before and after exposure of the DNA-binding domain of E2 to a f~st and second test
compound, respectively.
It can be seen from FIG. 2 that the first test compound caused chemic~1 shifts of l~N-
sites such as those desi~n~t~d I15, Y21, R22 and L23. Those sites col.~s~und to an
isoleucine (~e) residue at position 15, a tyrosine residue (Tyr) at position 21, an arginine (Arg
residue at position 22 and a leucine (Leu) residue at position 23 of SEQ ~ NO. 6.
It can be seen from FIG. 3 that the second test compound caused ch.o.mi~ ~l shifts in
the particular 15N-sites de~ign~t.-d I6, G1 l, H38, and T52. Those sites correspond to an
1 Q isoleucine (Ile) residue at position 6, a glycine (Gly) residue at position 11, a hictiAine (EI;s3
residue at position 38 and a threonine (Thr) at position 52 of SEQ ID NO. 6.
FIGs. 7 and 8 show the correlation b~,L~ ,n those NMR binding data and a view ofthe NMR-derived three-Aim~,n~ional ~LLUC~U1C of the DNA-binding domain of E2.
Several structurally similar compounds caused ch~mic~l shift changes of the protein
signals when screened at a c~-n~ . "*-ln of 1 mM. Two distinct sets of amide resonances
were found to change upon the addi*ion of the compounds: one set of signals corresponding
to amides located in the ~-barrel formed between the two monomers and a second set
cc,-l~,syollding to amides located near the DNA-binding site.
For ex~mrlP,, compounds cOI .li. h~ two phenyl rings with a carboxylic acid ~tt~Chf!A
to the carbon linking the two rings only caused chemi~l shift changes to the amides in the
DNA-binding site. In contrast, benzophenones and phenoxyphenyl-cont~ining compounds
only bound to the ~3-barrel. Other compounds caused chemic~l shift changes of both sets of
signals but shifted the signals in each set by different amounts, suggesting the presence of
two distinct binding sites.
By monitoring the chemical shift changes as a function of ligand conce~ ion,
binding constants for the two binding sites were also measured. The results of those studies
are summarized below in Table 2.
CA 02237336 1998-05-ll
WO 97/18469 PCT/US96/183i2
-23 -
Table 2
Comp. A ~1 R2R3 R4 DNA13-baIrelFilter
No. KD(m~KD(tT~ binding
a~.say
- 13 CO H H H OH >50 û.6
14 O H H H CH2OH >50 2.0
a H H COO H 2.0 >50 +
16 a Cl ClCOO H 0.1 >50 +
17 a H HCH2COO H 4.2 4.9 +
18 a H HCH=CHCOO H 1.2 6.2 +
19 0 H HCH2cH2cH(cH3) H 0.5 0.2 +
-CH2COO
O H HCOCH2CH2COO H 2.7 4.8 +
a a dash (-) for A in~lic~t~s no atom (i.e., byphenyl linkage)
Uniforrnly 1 5N-labeled Ras-binding domain of the RAF protein was prepared as
5 rlesçrikerl in Example 1 and screened using two--~im~n~i-nal lSN/lH NMR correlation
spectral analysis in accordance with the NMR procedures described above. The results of a
represçm~tive study are shown in FIG. 5, which depicts two-~limen~ional 15N/lH NMR
correlation spectra both before and after exposure to a test compound.
Uniformly 15N-labeled FKBP was plc~alcd as described in Example 1 and screerEe dlo using two--1imen~ional lSN/lH NMR correlation spectral analysis in accordance with the
NMR procedures described above. The results of a repres~nt~tive study are shown in FIG.
6, which depicts two-tlim-oncional 15N/lH NMR correlation spectra both before and after
exposure to a test compound.
Exarnple 3
Comparison of NMR~ Enzymatic, Filter
Bindin~ and Gel Shift Screening Assays
Studies were performed to compare binding constants of ligands to various
biomolecules, determined by the NMR method of the present invention, to similar results
20 obtained from prior art methods.
In a first study, binding constants were determined, both by the NMR method of the
present invention, and by a prior art enzymatic assay. The target molecule was the catalytic
domain of stromelysin prepared in accordance with the procedures of Example 1. The NMR
binding constants, KD, were derived using two-dimensional 15N/lH NMR correlation
-
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
-24 -
spectroscopy as described in Example 2. The KD values so obtained were co~ d to an
inhibition COI~ KI as det~ ....i..eA in an ~,n~y~latic assay.
The en~ylnatic assay measured the rate of cleavage of a fluorogenic ~,ulJ~ by
following the ffuorescence increase upon peptide cleavage which causes a s~ ;nn be~n
5 the fluorophore and quencher. Enzymatic activity was measured using a matrix of different
concentrations of acetohyd,~;~l,ic acid and biaryl compounds. The assay is a motlifi~tion
of the method described by H. Weingarten, et al. in Anal. Biochem.. 147: 437-440 (1985)
employing the fluorogenic substrate properties ~l~sçriberl by E. Matayoshi, et al. in Science:
247: 954-958 (1990).
Eight acetohy~ amic acid concentrations were used ranging from 0.0 to 1.0 M, andsix compound concentrations were used, rçsnlting in a total of 48 points. Individual
compound concenl.a~ion varied due to solubility and potency.
All NMR mea,ul~n~en~s were ~ Çull"ed in the presence of 500 mM acetohy~u~a",ic
acid, except for the titration of acetohydroxamic acid itself. Dissociation con.~t~nt~ were
5 obtained from the dependence of the observed t~h~mi~l shift changes upon added ligand.
Inhibition const~nt.~ were then obtained from the inhibition data using standard procedures.
The results of these studies are ~7. ~ ed below in Table 3, which shows the
comparison of NMR-derived dissociation cl-n~ s (E~D) with inhibiti~n con~t~nt~ measured
in the enzyme assay (KI)~ using a fluorogenic substrate.
2~
Table 3
Compound No. NMR KD (rnM3 Assay KI (mM3
4 1.6 7.4
7 0. 17 0.32
9 0.16 0-70
0.40 1.8
12 0.02 0. 1 1
Acetohycllu~nic acid 17.0 21.1
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
-25 -
The data in Table 3 show that a NMR process of the present invention provides a
rapid, efficient and accurate way of fl~ t~ l;ll;ng dissociation or binding con~Lanls of ligands
to target biomolecules. Comparison of the binding c~ n~t~ntS clet~ ;..~1 by the two methods
result in the same ranking of potencies of the compounds tested. That is, while the values for
a given substrate as clc t~ ed by the two methods are not equal, they are proportional to one
another.
In a second study, the results for binding of the DNA-b;n-ling domain of E2 to its
target DNA were obtained by prior art methods and con~altd with results obtained by the
methQ~1 of the present invention. The target was the DNA-binding domain of E2, ~c~,d in
accordance with the procedures of Fxample 1. NMR S~ illg assays and NMR processes
for ~1G~ ;ng ligand dissociation con~t~nt~ were ~ rol.ned as set forth above in F.Y~mple
2.
The binding constant from the NMR process was com~d to the results of a
physical, filter binding assay that measured binding of DNA to the target. The high-
throughput filter binding assay was performed using E2, ~l~e;l according to Example 2
above. The 33P-labeled DNA construct comprised a 10,329 base pair plasmid formed by
inserting the HPV-11 genome, con~i~ g three high affinity and one low aff~ity E2 binding
sites, into the PSP-65 plasmid (Promega, Madison, WI).
The binding ~ffinhi~s at the dirr~ sites as ~let~-rminecl by NMR were conll,~Gd for
a subset of the compounds to the inhibition of E2 binding to DNA as measured in the filter
binding assay. As shown in Table 2 above, the activities deL~ ed in the filter binding
assay correlated closely wi~ the binding ~ffinit;çs calculated from the amides of the DNA-
binding site but noe to the ~ffinities measured for the 13-barrel site. This is con~ictent with the
relative locations of each site.
In an alternative study, a comparison of the NMR-~let~rmine~l binding results was
made with similar results obtained by a prior art gel-shift assay using techniques well known
in the art. The gel-shift assay was pGlr~llned using a GST fusion protein which contained
full length E2 and a 33P-labeled 62 base pair DNA fragment cont~ining two E2 binding sites.
The method identified numerous compounds which gave positive results in the gel-shift assay. Some of these positive results, however, were believed to be due to binding to
the DNA, since in these cases, no binding to the E2 protein was observed using the NMR
method of this invention. These compounds were shown to indeed bind to DNA rather than
to E2, as evi~çnçe-l by changes in the cl~n~cal shifts of the DNA rather than the protein upon
the addition of the compounds. These data show that yet another advantage of the present
3~ invention is the ability to minimi7e the occurrence of false positives.
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
- 26 -
FY~mr1e 4
DP~;~n of a potent. non-peptide inhibitor of stromelysin
Studies were pG~Çu l~-ed to design new ligands that bound to the catalytic domain of
stromelysin. Because stromelysin undergoes autolysis, an inhibitor was sought to block the
5 ~legr~ tion of stromelysin. That inhibitor would f~ it~te the screening of other potential
ligands that bind to other sites on the enzyme.
The criteria used in selecting con~L,ou,lds in the scr~nil~g for other binding sites was
based pnmarily on the size of the ligand. The sm~llPst ligand was sought that had enough
solubility to Sd~uldlG (>98% occupancy of enzyme) and inhibit the enzyme.
The cloning, G~.Gssion, and pnrif;~ticln of the catalytic domain of stromelysin was
accompli~hP~I using the procedures set forth in FY~mple 1. An initial step in the design of the
new ligand was the i-lentific~ti~n of a first ligand that bound to the stromelysin target. Such
identification was carried out ul accordance with a two-~iim~-n~ n~ 15N/lH NMR correlation
screening process as disclosed above.
A variety of llydlu~al~c acids of the general formula R-~CO)NHOH were s-, ~,~d
for binding to stromelysin using the procedures set forth in Fx~mple 2. Of the compounds
tested, acetohydroxamic acid [CH3(CO)NHOHl best s~ticfipA the selection criteria: it had a
binding affinity for strûmelysin of 17 mM and had good water solubility. At a concentration
of 500 mM, acetûhydroxamic acid inhihited the degradation of the enzyme, allowing the
20 screening of other potential ~ n~l~
The second step in the design process was the iclentific~tlnn of a second ligand that
bound to the target stromelysin at a site dirr~ from the binding site of acetohyd~ ~ic
acid. This was accompli~hed by screening compounds for their ability to bind stromelysin in
the presence of saturating amounts of acetohydroxamic acid. Details of procedures and
25 results of this second i(lPntific~ti~n step are set forth above in Example 2.The compound idçntifi~ as a second ligand from these studies and used in
subsequent design steps was the compound de~ip;n~ted as Compound #4 in Table 1 (See
Example 2).
The next step in the design process was to construct a ternary complex of the target
30 stromelysin, the first ligand and the second ligand. This was accomplished by exposing the
stromelysin target to the two ligands under conditions that resulted in complex formation.
The three-rlim~n~innal structure of the ternary complex was then ~let- rrninPd using NMR
spectroscopy as described below.
The 1H, 13c~, and 15N backbone resonances of stromelysin in the ternary complex
35 were assigned from an analysis of several 3D double- and tIiple-resonance NMR spectra (A.
Bax, et al.. Acc. Chem. Res.. 26: 131-138 (1993)). The Ca resonances of adjacent spin
systems were illentifi~d from an analysis of three-dimensional (3~) HNCA (L. Kay, et al., L
Ma~n. Reson.. 89: 496-514 (1990)) and HN(CO)CA (A. Bax, et al., J. Bio. NMR. 1: 99
CA 02237336 1998-05-11
WO g7/18469 - 27 - PCT/US96/18312
(1991)) spectra recorded with i(~en*~l spectral widths of 1773 Hz (35.0 ppm), 3788 Hz
(30.1 ppm), and 8333 Hz (16.67 ppm) in the Fl(l5N), F2(13C) and F3( H) dimen.cion.c,
lc~G..~ rely.
The data matrix was 38(tl) x 48(t2) x 1024(t3) complex points for the HNCA
7~1G1LIUIII, and 32(tl) x 40(t2) x 1024(t3) complex points for the HN(CO)CA spectrum. Both
spectra were acquired with 16 scans per i.._rc~--ent~ A 3D QCA(CO)NH ~L)G~;LIUIII (S.
.r7~Si~, et al., J. Am. Chem. Soc.. 114: 6261-6293 (1992)) was collected with 32(tl,
15N) x 48(t2, 13C) x 1024(t3, lH) complex points and 32 scans per in~ cn~ Spectral
widths were 1773 Hz (35.0 ppm), 7575.8 Hz (60.2 ppm), and 8333 Hz (16.67 ppm) in the
0 N, C and H dirnensions, respectively.
For all three spectra, the lH carrier frequency was set on the water iGsona"ce and the
15N carrier frequency was at 119.1 ppm. The 13C carrier ~ u~ cy was set to 55.0 ppm in
HNCA and HN(CO)CA c"~ .nl.~, and 46.0 ppm in the QCA(CO)NH G~)~ t
The backbone ~ were colLL"...ed from an analysis of the crosspeaks
observed in an lsN-sG~ Gd 3D NOESY-HSQC spectrum and a 3D HNHA-J ~e~ u.~.
The 15N-separated 3D NOESY-HSQC spectrum (S. Fesik, et al., J. Magn. Reson.~ 87:588-593 (1988)); D. Marion, etal., J. Am. Chem. Soc.. 111: 1515-1517 (1989)) wascollected with a mixing time of 80 ms. A total of 68(tl, 15N) x 96(t2, lH) x 1024(t3, lH)
complex points with 16 scans per increment were collected, and the spectral widths were
1773 Hz (35.0 ppm) for tne 1 N dimension, 6666.6 Hz (t2, H, 13.3 ppm), and 8333 Hz
(16.7 ppm) for the lH dimension.
The 3D HNHA-J spectrum (G. Vuister, et al., J. Am. Chem. Soc.~ 115: 7772-7777
(1993)), which was also used to obtain 3JHNHa coupling con~t~nt~ was acquired with
35(tl, 15N) x 64(t2, lH) x 1024(t3, lH) complex points and 32 scans per increment.
Spectral widths and carrier frequencies were identical to those of the lSN-sr~ 1 NOESY-
HSQC spectrum. Several of the H~ signals were assigned using the HNHB wc~ t.
The sweep widths were the same as in the 15N-separated NOESY-HSQC spectrum that was
acquired with 32(tl, N) x 96(t2, H) x 1024(t3, H) complexpoints.
The lH and 13C chemical shifts were assigned for nearly all sidechain resonances. A
3D HCCH-TOCSY spectrum (L. Kay, et al., J. Ma~n. Reson.. 101b: 333-337 (1993)) was
acquired with a mixing time of 13 ms using the DIPSI-2 sequence (S. Rucker, et al., Mol.
Phvs., 68: 509 (1989)) for 13C isotropic mixing. A total of 96 (tl, 13C) x 96(t2, lH) x
1024(t3, lH) complex data points were collected with 16 scans per increment using a spectral
width of 10638 Hz (70.8 ppm, wl), 4000 Hz (6.67 ppm, w2), and 4844 (8.07 ppm, w3).
3~ Carrier positions were 40 ppm, 2.5 ppm, and at the water frequency for the 13C, indirectly
detected lH, and observed lH (iimen~ions7 respectively.
Another 3D HCCH-TOCSY study was performed with the 13C carrier at 122.5 ppm
to assign the aromatic r~sid~les The spectra were collected with 36(tl, C) x 48(t2,1H) x
= = = == = . ===.= = = = = =
CA 02237336 l998-05-ll
WO 97/18469 PCT/US96/183i2
- 28 -
1024 ~t3, H) complex points with spectral widths of 5263 Hz (35.0 ppm, wl), 3180 Hz
(5.30 ppm, w2), and 10,000 (16.7 ppm, w3). Carrier positions were 122.5 ppm, 7.5 ppm,
and at the water frequency for the 13C, indirectly d~tectefl lH, and observed lH ~limen~ ns,
respectively.
A 13C-s~alat~d 3D NOESY-HMQC ~e:CllUln (S. Fesik, etal., J. Magn. Reson
87: 588-593 (1988)); D. Marion, etal., J. Am. Chem. Soc.. 111: 1515-1517 (1989)) was
recorded using a rnixing time of 75 ms. A total of 80 (tl, 13C) x 72 (t2, lH) x 1024 (t3, lH)
complex data points with 16 scans per ill~ en~ were coll~cteA over ,epec~l widtns of 10638
Hz (70.49 ppm, wl), 6666.6 Hz (13.3 ppm, w2), and 8333.3 Hz (16.67 ppm, w3). The10 lH carrier frequencies were set to the water resonance, and the 13C carrier frequency was
placed at 40.0 ppm.
Stereospecific ~ ignm~nt~ of methyl groups of the valine and leucine residues were
obtained by using a biosynthe~ic approach (Neri etal., Biochem.. 28: 7510-7516 (1989)) on
the basis of the 13C-13C one-bond coupling pattern observed in a high-resolution IH, 13C-
HSQC spectrum (G. Bo-lçnh~lsen, etal., J. Chem. Phys. Lett.~ 69: 185-189 (1980)) of a
fractionally C-labeled protein sample. The ~e~ u~ll was acquired with 200( 3C, tl) x
2048( H, t2) complex points over spectral widths of 5000 Hz (39.8 ppm, 13C) and 8333 Hz
(16.7 ppm, lH). Carrier positions were 20.0 ppm for the 13C dimrn~ion~ and at the water
frequency for the lH ~iim~n~ion
To detect NOEs between the two ligands and the protein, a 3D 12C-filtered, 13C-
edited NOESY spectrum was collected. The pulse scheme consisted of a double 13C-~llter
sequence (A. Gçmmt ~r, et al., J. Ma~n. Reson.. 96: 199-204 (1992)) concatenated with a
NOESY-E~MQC sequence (S. Fesik, et al., J. Ma~n. Reson.. 87: 588-593 (1988)); D.Marion, et al., J. Am. Chem. Soc.. 111: 1515-1517 (1989)) . The spectrum was recorded
with a mixing time of 80 ms, and a total of 80 (tl, 13C) x 80 (t2, lH~ x 1024 (t3, IH)
complex points with 16 scans per increment. Spectral widths were 8865 Hz (17.73 ppm,
wl), 6667 Hz (13.33 ppm, w2), and 8333 Hz (16.67 ppm, w3), and the carrier positions
were 40.0 ppm for the carbon ~lim~ncion and at the water frequency for both proton
dimensions.
To identify amide groups that exchanged slowly with the solvent, a series of lH,15N-HSQC spectra (G. Bodenhausen, et al., J. Chem. Phys. Lett.. 69: 185-189 (1980))
were recorded at 25~C at 2 hr intervals after the protein was exchanged into D2O. The
acquisition of the first HSQC spectrum was started 2 hrs. after the ~ n of D20.
All NMR spectra were recorded at 25~C on a Bruker AMX500 or AMX600 NMR
35 spectrometer. The NMR data were processed and analyzed on Silicon Graphics computers.
~n all NMR experiments, pulsed field gradients were applied where a~p.ol,~iate as described
(A. Bax, et al., J. Ma~n. Reson.. 99: 638 (1992)) to afford the suppression of the solvent
signal and spec~al artifacts. Quadrature detection in indirectly detected t~imencions was
CA 02237336 1998-05-ll
WO 97/18469 PCT/US96/18312
-29 -
accomplished by using the States-TPPI method (D. Marion, et al., J, Am. Chem. Soc.. 111:
1515- 1517 (1989)). Linear prediction was employed as described (E. Olejnic7~k, et al., J.
Magn. Reson.. 87: 628-632 (1990)).
The derived three-~1im~n~ional structure of the ternary complex was then used to5 define the spatial ori~nt~tinn of the first and second ligands to each other as well as to the
target stromelysin molecule.
Distance ~ derived from the NOE data were cl~ifiçd into six categories based
on the NOE cross peak intensity and given a lower bound of 1.8 A and upper bounds of 2.5
A, 3.0 A, 3.5 A, 4.0 A, 4.5 A, and 5.0 A, respectively. Restraints for q, torsional angles
10 were derived from JHNHa coupling constants measured from the 3D HNHA-J spectrum
(G. Vuister, et al., J. Am. Chem. Soc.. 115: 7772-7777 (1993)). The q3 angle wased to 120%+40% for 3JHNHa > 8.5 Hz, and 60%+40% for 3JHNHa < 5 Hz.
Hydrogen bonds, iflentifi~l for slowly ex~h~nginp amides based on initial structures,
were defined by two restr~int~: 1.8-2.5 A for the H-O distance and 1.8-3.3 A for the N-O
~ t~n~e- Structures were calculated with the X-PLOR 3.1 program (A. Brunger, "XPLOR
3.1 Manual," Yale University Press, New Haven, 1992) on Silicon Graphics co~ u~
using a hybrid distance geometry-~im~ teA ~nne~ling approach (M. Nilges, et al., FEBS
L~., 229: 317-324 (1988)).
A total of 1032 ~l~lo~"aL~ in~l~lu~o~ t~nce restraints were derived from the
NOE data. In addition, 21 lln~mhiguous int~rmnl~:cular distance le~L-ahlts were derived from
a 3D 12C-filtered, 13C-edited NOESY spectrurn. Of the 1032 NOE l~ involving the
protein, 341 were intra-residue, 410 were sequential or short-range between residues
separated in the primary sequence by less than five amino acids, and 281 were long-range
involving residues s~ted by at least five re~i-1nes.
In addition to the NOE distance restraints, 14 ~ dihedral angle r~ ints were included
in the structure c~ tic)ns that were derived from three-bond coupling con~nt~ (3JHNH0c)
det~l",ined from an HNHA-J spectrum (G. Viioster, et al., J. Am. Chem. Soc.. llS: 7772-
7777 (1993)). The c;~ e~ ldill~ also included 120 distance restraints
corresponding to 60 hydrogen bonds. The amides involved in hydrogen bonds were
identified based on their characteristically slow exchange rate, and the hydrogen bond
partners from initial NMR structures calculated without the hydrogen bond restraints. The
total number of non-reAIln-l~nt exp~rirrl~nt~lly-derived restraints was 1166.
- The ~L~u;~ ;S were in excellent agreement with the NMR experiment~l restraints.
There were no distance violations greater than 0.4 A, and no dihedral angle violations greater
- 35 than 5 degrees. In addition, the cimll~ ed energy for the van der Waals repulsion term was
small, in~lic~ting that the structures were devoid of bad inter-atomic contacts.The NMR structures also exhibited good covalent bond geometry, as indicated by
small bond-length and bond-angle deviations from the corresponding i-le~li7eA parameters.
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
-30-
The average atomic root mean square deviation of the 8 ~LlucLul~S for residues 93-247 from
the mean coordinates was 0.93 A for backbone atoms (Ca, N, and C'), and 1.43 A for all
non-hydrogen atoms.
A ribbon plot of the ternary complex involving stromelysin, acetohydlu,~amic acid
(the first ligand), and the second ligand is shown in Fig 10. The structure is very sirnilar to
the global fold of other matrix metallo~ Leillases and consists of a five-s~n~led ~-sheet and
three a-helices.
The catalytic zinc was located in the binding pocket. It was cooldil~a~d to three
hi~fi~in~s and the two oxygen atom of acetohy~ a,l ic acid. A biaryl group of the second
10 ligand was located in the S 1 ' pocket ~Lwcen the second helix and the loop formed from
residues 218-223. This deep and narrow pocket is lined with hydrophobic residues which
make favorable contacts with the ligand.
Based on the three~ n~ LIUl;Lu~c of the ternary complex as ~let~ yl above
and the structure/activity relationships observed for ~e binding to stromelysin of ~LIuCluLal
15 analogs of the second ligand (i.e., other biaryl compounds), new molecules were cle~igned
that linked together the acetohydroxamic acid to biaryls.
As shown in Table 4 below, the initial biaryls chosen contained an oxygen linker and
the ~bs~n~e or presence of CN para to the biaryl link~g~ Initial lilLkers collL~.ed varying
lengths of methylene units. Means for linking compounds with linkers having varying
20 lengths of methylene units are well known in the ar~
Table 4
H
HO' N~X~
R
CA 02237336 l998-05-ll
WO 97/18469 PCT/US96/18312
- 31 -
X R Stromelysin
Compound Tnhibiti~n
21 (cH2)2 H 0.31 ~M
22 ~CH2)3 H 110 ,uM
23 (CH2)4 H 38%@ 100 ~I
24 (CH2)5 H43%@ 100 ,uM
(cH2)2 CN 0.025 ~LM
26 (CH2)3 CN 3.4 ~LM
27 (cH2)4 CN 3.5,uM
28 (CH2)5 CN 1.7 IlM
As ~;~e ;~d based on the better binding of the CN substi~uted biaryls to stromelysin,
the CN derivatives exhibited better stromelysin inhibition. The compound that exhibited the
5 best inhibition of stromelysin con~ l a linker with two methylene units.
The present invention has been described with .c~r~,nce to ~l~r~ ,d embof1imf~nt.c
Those emboflimPntc are not limiting of the claims and specification in any way. One of
ur~ ~y skill in the art can readily envision changes, modifications and alterations to those
embo-1imt-ntc that do not depart from the scope and spirit of the present invention.
F.x~mple S
Desi~n of potent. novel inhibitors of FKBP
Studies were ~3t;l~llned to design novel ligands that bound to FK-binding protein
(FKBP) .
The cloning, expression and plmfication of FKBP was accomplished as set forth inExample 1. An initial step in the design of the new ligand was the i~entific:~tion of a first
ligand that bound to the FKBP target. Such ir~ntific~tion was carried out in accordance with a
two-dirnensional 15N/lH NMR correlation s~ g process as disclosed above.
A variety of low -molecular weight fragments and analogs of several known potent20 immnnnsuppressants (i.e. ascomycin, rapamycin) were screened for binding to FKBP using
the procedures as set forth in example 2. Of the compounds tested~ compound 29, below
CA 02237336 1998-05-11
WO 97118469 PCT/US96/t8312
-32-
o~\ /
ro
/--~ ,p OCI~3
~_~N~ OCH3
OCH3
best s:~ti~fi~l the selection criteria: it had a binding affinity for FKBP of 2 ~I (rrlca~u~ by
fluoresence by the methods known in the art) and s~tllr~t~ the protein (~ 98% occupancy of
the binding site) at ligand conc~nfrations of l mM.
The second step in the design process was the i~ nhfil~atif~n of a second ligand that
5 bound to the target FKBP at a site di~r~ from the binding site of compound 29. This was
accomplished by s~ g compounds for their ability to bind to FKBP in the ~,~,sellce of
.c~hlr~ting amounts of the ascomy(;ill fr~gm~nt analog (compound 29). Details of procedures
for this second i~l~ntific~tion step are as set forth in ex~mr1e 2.
In an initial screen, a compound was found that co.lt~ a ben7~nili~1e moiety. Based
10 on this initial hit, structurally similar compounds were obtained and tested against FKBP. The
structure of these ben7~nili~e. compounds is r~l~,sell~ed by the structure m, below (see Table
5 for definitions of Rl-R4).
Rl~ H Fi2 R4
m
In the second round of screening, binding was assayed both in the presence and in the
absence of saturating amount of compound 29 (l mM).
A structure-activity relationship was developed for these diphenyl amide compounds
as set forth in Table A. Fig. 6 shows a reprçsçnt~tive two--lim--n~ional 15N/lH correlation
spectrum before and after exposure of FE~BP to a diphenyl amide test compound. It can be
seen from Fig. 6 that the compound caused ch~omic~l shifts of 15N sites such as those
~lesi~n~te~l I50, Q53, E54, and V55. These sites correspond to an isoleucine (Ile) at position
50, a gl~ min~ (Gln) at position 53, a glutamate (Glu) at position 54, and a valine (Val) at
position 55 of SEQ ID NO # 7. Figure l l shows the correlation between the NMR binding
data and a view of the NMR-derived three-dimensional structure of FKBP. The ability to
locate the spec~fic binding site of a particular ligand is an advantage of the present invention.
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/183i2
-33 -
Some compounds only bound to FKBP in the presence of compound 29. Thus ~e
binding affinity of some compounds was ~nh~ncel in the presence of compound 29. These
results exemplify yet another i.npo~ t c~p~hility of the present screening assay which is the
ability to identify compounds that bind to the protein in the ~lcsw-ce of other molecules.
Various bon7~nili~1e compounds were tested for binding to FKBP at mul*ple ligandconcentra*ons. The 15N/IH correla*ion spectra ge~ cl at each c~n~entr~*on were evaluated
to quan*fy dirr~,rences in the spectra as a function of compound concentration. A binding or
dissociation constant (Kd) was calculated, using standard procedures well known in the art,
from those dirr r~,nces. The results of this study are shown in Table 5. The values for Rl-R4
in Table 5 refer to the corresponding positions in the structure m, above.
Table 5
Compound RI R2 R3 R4 Kd (mM)
No.
OH OH H H 0.8
31 H H OH H 1.4
32 H H H OH 0.5
33 OH H H OH 0.1
34 OH H H H 0.6
OH H CH3 OH 0.5
36 H H H H >5.0
37 H OH H H >5.0
The data in Table S show the utility of a process of the present invention in
determining dissociation or binding con~ between a ligand and a target molecule.The next step in the design process was to construct a ternary complex of the target
FKBP, the first ligand and the second ligand. This was accompliched by exposing the FKBP
target to the two ligands under conditions that resulted in complex formation. The location
and orientation of the ligands were then determined as described below.
The lH, 13C and 15N resonances of FKBP in the ternary complex were ~igne~l
from an analysis of several 3D double- and triple-resonance NMR spectra. The ~ nm~ont
process was aided by known ~si~nments of FKBP when complexed to ascomycin (R. Xu,
et al., Biopolymers. 33: 535-550, 1993). lH sidechain and lSN/lH backbone resonances
were i(içn~ifie-l from an analysis of three-dimensional (3D) HC(CO)NH spectra recorded with
spectral widths of 200û Hz (39.5 ppm), 6250 Hz (12.5 ppm) and 8333 Hz (16.7 ppm) in the
Fl(l5N~, F2(1H) and F3(1H) ~imencions, respectively, and with a data matrix of 46(tl) x
CA 02237336 l998-0~-ll
WO 97/18469 PCT/US96/183i2
-34 -
80(t2) x 1024(t3) complex points and 16 scans per increment. lH andl3C ~ ,ch~in and Ca
resonances were i~lçntifie-1 from an analysis of 3D HCCH-TOCSY spectra (L. Kay, et al. J.
M~n. Reson., 101b:333-337, 1993) recorded with spectral widths of 7541.5 Hz (60.0
ppm), 6250 Hz (12.5 ppm) and 8333 Hz (16.7 ppm) in the Fl(l3C), F2(1H) and F3(1H)
~im~,n~ion~, respectively, and with a data matrix of 48(tl) x 64(t2) x 1024(t3) complex points
and 16 scans per in~ ~nl. Intermolecular NOEs be~n the ligand and FKBP were
obtained from an analysis of a 3D 12C-filtered, 13C edited NOESY ~e.iL,unl. The pulse
scheme con.~i~tçfl of a double 13C filter sequence (A. Ge,mmç~ .r, et al., J. Ma~n. Reson.,
96:199-204, 1992) conc~tto.n~ted with a NOESY-HMQC sequence (S. Fesik, et al-, I. A~M.
Chem. Soc., 111:1515-1517, 1989). The ~L)e~L.um was recorded with a mixing tirne of 350
ms and a total of 46(tl, 13C) x 64(t2, llI) x 1024(t3, lH) complex points and 16 scans per
ine,clllent. Spectral widths of 7541.5 Hz (60.0 ppm), 6250 Hz (12.5 ppm) and 8333 Hz
(16.7 ppm) were used in the ~1(13C), F2(1H) and F3(~ im~n~ions, ~cspc.,Lively.
In all spectra, the 15N carrier frequency was set at 117.4 ppm, the 13C carrier
frequency was set at 40.0 ppm, and the lH carrier frequency was set on the water resonance.
All spectra were recorded at 303K on a Bruker AMX500 NMR s~e~ mc,~l. The NMR data
were processed and analyzed on Silicon Graphics COIII~)U~ ;. In all NMR exp~
pulsed field gradients were applied where applopiiate as described (A. Bax, et al., J. M~n
~eson, 99:638, 1992) to afford the ~u~plcssion of the solvent signal and spectral artifacts.
Q~ lr?.~lre detection in the indirecdy detected ~lim~,n~ion was accompli~h~1 by using the
States-TPPI method (D. Marion, etal. ~. Am. Chem. Soc. 87: 1515-1517, 1989). ~inear
prediction was employed as described (E. Olejniczak, et al., J. Magn. Reson., 87: 628-632,
1 990).
Distance restraints derived from the NOE data were cl~if-iPd into ~ree categories
based on the NOE crosspeak intensity and were given a lower bound of 1.8 A and upper
bounds of 3.0, 4.0 and 5.0 A. A total of 17 intermolecular distance restraints between the
protein and cc,~ o~ d 33 and 10 intermol~cular distance ~ beL~ the protein and
compound 29 were used to define the location and orientation of the compounds when bound
to FKBP using the known three-dimensional coordinates for the FKBP protein stlucture. A
ribbon plot of the ternary complex involving FKBP, a fragment analog of ascomycin
~compound 29), and a bçn7~nilicle compound (compound 33) is shown in Figure 12.
Based on the three--lim~n~ional structure of the ternary complex as determined above
and the structure activity relationships observed for the binding to FKBP of structural analogs
of the second compound, a new molecule was ~le~ignt A that linked the ascomycin ~r~gn~ent
analog to the ben7~nilic~e compound. This compound, shown below,
CA 02237336 1998-05-11
WO 97/18469 PCT/US96/18312
Ho~N9~ '~
C~OCH3
OCH3
IV
has a 19 nM affinity for FKBP as dclf . ~ d by fluolcsc~ ce titrations. This is a 100-fold
5 increase in potency over the as~ ycil~ fragment analog (compound 29) alone (Kd = 2 ~).
As shown by the above non-limitin~ examples, the present invention relates to a process for
clesi,~nin~ a high-affinity ligand to a given target molecule, cnmI ri~ing
a) iclentirying by the sclce~ g processes dçsrribef1 herein at least two ligands10 which bind to disting binding sites on the target molecu}e using mnlti-Tim~,n~innal NMR
spectroscopy;
b) forming at least a ternary complex by exposing the at least two ligands to the
target molecule;
c) clel~,. ~--i--;--~ the three flimen~iQn~l structure of the complex and thus the spatial
15 orientation of the at least two ligands on the target mnlec~ ; and
d) using the spatial orien~tinn ~let~,rminerl in step c) to design the high affinity
ligand which ~LI u~;Luldlly resembles a combination of the at least two ligands which bind to
distinct sites on the target molecule. Preferrably, the high-affinity ligand designe~l in the
above process serves as or is the basis for a drug which binds to a given target molecule and
20 pclrcJlllls, in vitro and in vivo, a targeted theld~t;u~ic effect in m~mm~l~ incln~1ing hnm~n,~ in
need of LlcaLnlellt thereof.
The process also relates to de~igning a high-affinity ligand to a given target molecule
comprising:
a) identifying a first ligand to the target molecule using mnltirlimt-,n~innal NMR
25 spectroscopy;
b) idellLiry..lg a second ligand to the target molecule using mlll*~lim~-,n~innal
NMR spectroscopy wherein the second ligand may be the same or different than the first
ligand and wherein the second ligand binds to a different site on the target molecule than the
first ligand;
CA 02237336 1998-05-11
WO 97/18469 PCT/IJS96/18312
c) forming a ternary complex by binding the first and second ligands to ~~e target
molecule;
d) cle~ . . . i ..; . .g the three ~lim~n~iona'~ structure of the complex and thus the spatial
ori~.nt~tion of the first ligand and the second ligand on the target molecule; and
e) ~le~igning the ligand wh~ ,;n the spatial oritont~tinn of step d) is "~ u~l
In the process ~lescrihe l above, the first and second ligands may have the identicle mnlecnl~r
structure or formula wherein the moiety binds to at least two hinding sites on the target
molecule. The ligand that is based upon the structural comhin~tion of the first and second
ligands then serves as a drug lead or drug upon actual synthesis of that comhin~A compound
and evaluation in the ay~~uL,liate biological assays. The synthesis of the comhinç~ ligand,
high-affinity ligand, drug lead or drug is achieved through synthetic or biological means.
CollcG~Lually, as in~lie~tf~ ~~u~ lloUt the spe~ifin~tion, the first and second ligands are linked
(joined) together by carbon atoms, he~loa~ll~s, or a comhin~tion thereof to form the ligand
or drug lead. The processes described herein, of course, include ~yll~heses of the high-
affinity ligand by linear or non-linear (convergent) means which nltim~t~ly produce the linked
(combined) first, second or more ligands.
The first and second ligands may also have dirr~,lGnt molecular SLIU~;~UIGS and either of
the ligands may or may not bind to the other (&stinct~ bin&ng site on the target molecule.
In more detail, the process of the invention also relates to ~le~igning a high affinity
ligand to given target mnlecllle, compri~ing
a) plel,~ing an isotopically-labeled target mnleclllF wherein said molecule is
enriched with an NMR cletect~ble isotope;
b) generating a mlllti~1im~n~inn~ NMR spectra of the isotopically-~ l~1 target
molecule;
c) screening ~e isotopically-labeled target molecule by exposing the target
molecule to a plurality of compounds to identify by mnlti~1im~n~iQnal NMR spectroscopy at
least a first and second ligand which bind to &stinct sites on the target molecule;
d) forming at least a ternary complex by exposing at least the first and second
ligand to the isotopically-labeled target molecule;
e) ~lelf ~ g the spatial nri~n~hnn of the at least first and second ligand on the
isotpically-labeled target molecule;
f) using t'ne spatial ori~nt~tion clet.-rmine~l in step e) to design the high affinity
ligand based upon the comkin~tion of the at least first and second ligands. Of course, a
plurality of ligands (1 + n) can be combined to form a high affinity ligand which has the
spatial orientation of the I + n (n = 1 -~) combined lig~n(1.~. After the high-affinity ]igand has
been llesigned~ the process may further include the step f) of making the high affinity ligand
by synthetic or biological means. The at least two ligands (first and second ligand) may be
linlced by carbon atoms (e.g. by methylene or alkylene units) or by heleluatollls (e.g. by
CA 02237336 1998-05-ll
WO 97/18469 PCT/US96/18312
- 37 -
nitrogen, oxygen, sulphur) or by other atoms which ,.,~ ;.. or a~plu~ e the spatial
orientation of the 1 + n ligands to the target molecule. Depending upon the lig~n~e, the
molecules may also be colllbi lled or joined (linked) d-- e~;lly to each other without intervening
alkylene or het~,.ualulll linker units. The high affinity ligand produced from the 1 + n
5 combined ligands pç~ dbly shows an increase in binding potency to the target molecule in
relation to any one of the 1 + n lig~n~l~, The present invention, therefore, incl~l~es hign-
affinity ligands designed by the processes shown herein wherein said high-affinity ligand has
an in.;l~,ase in binding potency (Kd) to the given target molecule over the at least two ligands
which bind to distinct sites on the given target molecule.
The present invention also relates to a method for discovering high affinity ligands
using ~, U~;IUl~ ,-activity relationships obtained from nuclear m~gntotic resonance wherein said
method cc..~-p. ;~es constructing a high-affinity ligand from ligands which bind to a subsite of
a target molecule by;
i) S~ g low molecular weight ( < 450 MW) compounds which bind to a~5 subsite 1 of the target molecule;
iu) screening analogs prepared from the initial results obtained in step i) to
optimize binding to subsite 1;
iii) s~ ,nillg for low molecular weight (< 450 mw) compounds and
cull~,sponding analogs which bind to a nearby binding site, subsite 2, of the target molecule
20 using mnltiAim.oneil~nal NMR spectroscopy to measure binding affinity; wh~,leill, after steps
(i) - (iii), lead fragmentc are generated;
iv) combining lead fr~gment~ generated from steps i) - iii) to design a high affinit~y
ligand. ~ombining can be achieved by synthetic or biological means. Synthetic means
includes organic synthesis of the combined ligand. Biological means inchlcies ~r.l I l lr.11~ n or
25 generation of the combined ligand through a cellular vehicle or system. P~cr~ldbly, the target
molecule is a polypeptide. The present invention also relates to the method as recited above
wherein the combination of fr~gment~ produces a ligand with a higher binding potency (Kd)
than the individual fragments to the target molecule.
CA 02237336 1998-05-11
W 0 97/18469 PCT~US96/18312
-38-
u~:~ LISTING
~l) GENERAL INFORMATION:
(i) APPLICANT: Fesik, Stephen W.
Hajduk, Philip J.
Olejniczak, Edward T.
(ii) TITLE OF INVENTION: Use o~ Nuclear Magnetic
Resonance to Design Ligands to Target Biomolecules
(iii) NUMBER OF SEQUENCES: 7
(iv) CORRESPONDENCE ADDRESS:
(A) ADDRESSEE: Steven F. Weinstock,
Dept. 377 AP6D, Abbott Laboratories
(B) STREET: l00 Abbott Park Road
(C) CITY: Abbott Park
(D) STATE: Illinois
(E) COUNTRY: USA
(F) ZIP: 60064
(vi) COMPUTER READABLE FORM:
(A) MEDIUM TYPE: Floppy disk
(B) COMPUTER: IBM PC compatible
(C) OPERATING SYSTEM: PC-DOS/MS-DOS
(D) SOFTWARE: PatentIn Release #l.0, Version #l.30
(vi) CURRENT APPLICATION DATA:
(A) APPLICATION NUMBER:
(B) FILING DATE:
(C) CLASSIFICATION:
(viii) ATTORNEY/AGENT INFORMATION:
(A) NAME: Anand, Mona
(B) REGISTRATION NUMBER: 34537
(ix) TELECOMMUNICATION INFORMATION:
(A) TELEPHONE: (847) 937-4559
(B) TELEFAX: (847) 938-2623
CA 02237336 l998-0~-ll
W O 97/18469 PCT~US96/18312
-39-
(2) INFORMATION FOR SEQ ID NO:l:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 174 amino acids
(B) TYPE: amino acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:l:
Phe Arg Thr Phe Pro Gly Ile Pro Lys Trp Arg Lys Thr His Leu Thr
1 5 10 15
Tyr Arg Ile Val Asn Tyr Thr Pro Asp Leu Pro Lys Asp Ala Val Asp
Ser Ala Val Glu Lys Ala Leu Lys Val Trp Glu Glu Val Thr Pro Leu
Thr Phe Ser Arg Leu ~yr Glu Gly Glu Ala Asp Ile Met Ile Ser Phe
Ala Val Arg Glu His Gly Asp Phe Tyr Pro Phe Asp Gly Pro Gly Asn
Val Leu Ala His Ala Tyr Ala Pro Gly Pro Gly Ile Asn Gly Asp Ala
His Phe Asp Asp Asp Glu Gln Trp Thr Lys Asp Thr Thr Gly Thr Asn
100 105 110
Leu Phe Leu Val A$a Ala His Glu Ile Gly His Ser Leu Gly Leu Phe
115 120 125
Hi~ Ser Ala Asn Thr Glu Ala Leu Met Tyr Pro Leu Tyr ~is Ser Leu
130 135 140
Thr Asp Leu Thr Arg Phe Arg Leu Ser Gln Asp Asp Ile Asn Gly Ile
145 150 155 160
Gln Ser Leu Tyr Gly Pro Pro Pro Asp Ser Pro Glu Thr Pro
165 170
CA 02237336 1998-05-11
WO 97/18469 PCTnUS96/18312
-40-
(2) INFORMATION FOR SEQ ID NO:2:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 83 amino acids
(B) TYPE: amino acid
(C) STRAWDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:2:
Met Ala Thr Thr Pro Ile Ile His Leu Lys Gly Asp Ala Asn Ile Leu
l 5 l0 15
Leu Cy5 Leu Arg Tyr Arg Leu Ser Lys Tyr Lys Gln Leu Tyr Glu Gln
Val Ser Ser Thr Trp His Trp Thr Cys Thr Asp Gly Lys His Lys Asn
Ala Ile Val Thr Leu Thr Tyr Ile Ser Thr Ser Gln Arg Asp Asp Phe
Leu Asn Thr Val Lys Ile Pro Asn Thr Val Ser Val Ser Thr Gly Tyr
65 70 7S 80
Met Thr Ile
(2) INFORMATION FOR SEQ ID NO:3:
(i) SEQUEN OE CHARACTERISTICS:
(A) LENGTH: 18 base pairs
(B~ TYPE: nu~leic acid
(C) STRAWDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: DNA (genomic)
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:
GAAATGAAGA ~ CAA l8
CA 02237336 1998-0~-11
WO 97/18469 PCTAUS96/18312
-41-
(2) INFORMATION FOR SEQ ID No:4:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 18 base pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: DNA (genomic)
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:4:
GCGTCCCAGG ~ l~GAG 18
(2) INFORMATION FOR SEQ ID NO:5:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 28 ba~e pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: DNA (genomic)
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:5:
ATACCATGGC CTATCCATTG GATGGAGC 28
(2) INFORMATION FOR SEQ ID NO:6:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 30 base pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: DNA (genomic)
(xi) SEQUENCE DESCRIPTION: SEQ ID No:6:
ATAGGATCCT TAGGTCTCAG GGGAGTCAGG 30
CA 02237336 l998-05-ll
WO 97/18469 PCT~US96/183i2
-42-
(2) INFORMATION FOR SEQ ID NO:7:
(i) SEQUENCE CHARACTERISTICS:
~A) LENGTH: 107 amino acids
(B) TYPE: amino acid
(C) STRA~n~n~cs: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:7:
Gly Val Gln Val Glu Thr Ile Ser Pro Gly Asp Gly Arg Thr Phe Pro
1 5 10 15
Lys Arg Gly Gln Thr Cys Val Val His Tyr Thr Gly Met Leu Glu Aqp
Gly Lys Lys Phe Asp Ser Ser Arg Asp Arg Asn Lyq Pro Phe Lys Phe
Met Leu Gly Lys Gln Glu Val Ile Arg Gly Trp Glu Glu Gly Val Ala
Gln Met Ser Val Gly Gln Arg Ala Lys Leu Thr Ile Ser Pro Asp Tyr
Ala Tyr Gly Ala Thr Gly His Pro Gly Ile Ile Pro Pro His Ala Thr
Leu Val Phe Asp Val Glu Leu Leu Lys Leu Glu
100 105
