WO1999043409A1 - Hplc fractionation of complex chemical mixtures - Google Patents

Abstract

The rapid deconvolution of complex chemical mixtures, such as combinatorial libraries prepared via solution-phase simultaneous addition of functionalities is demonstrated using HPLC. Libraries containing 25-216 members were screened for biological activity and active libraries are fractionated into discreet pools using preferred reversed-phase HPLC. The entire process can be completed from a single synthesis of 15-50 mg, without the need for fixed positions, serial synthesis, or tagging.

Description

HPLC FRACTIONATION OF COMPLEX CHEMICAL MIXTURES

FIELD OF THE INVENTION

The present invention relates to the identification of active pharmacophores and other useful chemicals from mixtures, especially complex mixtures such as diverse combinatorial libraries. High pressure liquid chromatography (HPLC) processes are used to fractionate mixtures such as combinatorial libraries, yielding pools for screening. Active pools are further fractionated, preferably by HPLC, to yield active pharmacophores.

BACKGROUND OF THE INVENTION

The screening of complex chemical mixtures such as combinatorial mixtures or libraries of compounds, prepared individually or simultaneously has fundamentally changed the way drugs are discovered. Such strategies have been shown to be useful for identifying agricultural, veterinary and other useful chemicals as well. The scope of combinatorial chemistry is broad and the methods are suited for identification of unique new pharmacophores as well as for veterinary drugs and agricultural chemicals optimization of lead compounds against biological targets. A variety of approaches for the synthesis and deconvolution of combinatorial libraries have been described in recent years. See Szostak, J. ., Chem. Rev. , 1997, 97, 347-348; Gordon, E.M., et al., J. Med. Chem . 1994, 31, 1385-1401; Fruchtel,

J.S., Jung, G., Angew. Chem . Intl . Ed. Engl . , 1996, 35, 17-

42; Wilson, S.R., In Combinatorial Chemistry: synthesis and application, S.R. Wilson and A.W. Czarnik, Eds., John Wiley, New York, 1997, pp. 1-24; Combinatorial librari es -Synthesis , Screening, and Applica ti on Potential , Ed.: R. Cortese, Walter de Gruyter, Berlin, 1996; Janda, and K.D., Proc . Na tl . Acad. Sci . USA, 1994, 91 , 10779. The synthetic routes can be categorized based on whether the preparation, is solid-phase or solution-phase. Solid-phase synthesis methods generate compounds from scaffolds (core structures) that can be elaborated off solid supports, such as peptides, peptide mimetics, and small heterocycles . See Pinilla, C, et al . , Bi oTechniques 1992, 13 , 901; Geysen, H. ., et al . ,

Proc . Na tl . Acad. Sci . USA 1984, 81 , 3998; Houghten, R.A., Proc . Na tl . Acad. Sci . USA 1985, 82 , 5131 and Thompson, L.A., Ellman, J.A. , Chem. Rev. 1996, 96, 555. Solution- phase synthesis methods have been used for the preparation of mixtures of compounds from scaffolds containing several reactive sites. See An, H., et al . , J. Am . Chem. Sσ . 1997,

119, 3696-3708 and An, H., et al . , J. Org. Chem . 1997, 62 ,

5156-5164.

Much of the significance of combinatorial chemistry revolves around the ability to prepare and screen large mixtures of compounds in solution against biological targets, both simultaneously and rapidly. A number of deconvolution strategies have been proposed for the identification of the active species following identification of an active mixture. Ecker, D.J., et al . , Nucl . Acids Res . 1993, 21 , 1853; Pirrung, M.C., et al . ,

Indexed combinatorial libraries: nonoligomeric chemical diversity for the discovery of novel enzyme inhibitors, In Combinatorial Chemistry: synthesis and application; S.R. Wilson and A.W. Czarnik, Eds., John Wiley; New York, 1997, pp. 191-206; Xiao, X., Nova, M.P., "Combinatorial Chemistry: synthesis and application", S.R. Wilson and A.W. Czarnik, eds., pg. 135-152, John Wiley and Sons, New York 1997; - 3 -

Balkenhohl, F., et al . , Angew. Chem . Int . Ed. Engl . 1996,

35, 2288. The screening of large mixtures has not been previously pursued by most research groups for a number of reasons. Artificial suppression of biological activity can be generated from "quasi-species" interactions among compounds (i.e. false negatives). Moreover, the screening of mixtures can produce misleading, high biological activity in a sample when the low individual activities of many compounds sum and are incorrectly attributed to a single, very active compound (false positives) . Accordingly, researchers conventionally now screen single compounds or diverse groups of only about 5-10 compounds which have been created non-synthetically, usually via mixing. The great promise of combinatorial chemistry, hence, is not being fully realized.

The iterative synthetic deconvolution of "winners" from large combinatorial libraries can require complicated protection/deprotection schemes. Several rounds of synthesis and screening may be required to fix the functional group at each position on the scaffold. The time and effort required to identify a "winner" can be significant. Other strategies have been evaluated, including positional scanning, indexing, and tagging with labels or chemical codes. These techniques place limits on the complexity of a library, such as the minimal biological activity which can be detected, or the types of reaction chemistry which do not degrade the label or tag.

There is, accordingly, a great need for methods which can manipulate complex mixtures of chemical species in a fashion which can lead to the identification of useful compounds for pharmaceutical, diagnostic, veterinary, agricultural, industrial, nutritional and other uses. Such methods, together with their related apparatus, reagents, control systems and technologies, are desired to amenable to automation, to be capable of economically employment and to be adaptable to widely diverse synthetic chemistries . These objects can be attained through the present invention.

SUMMARY OF THE INVENTION

Important problems associated with the deconvolution of complex chemical mixtures including combinatorial libraries have now been solved in accordance with this invention. In accordance with preferred embodiments, alternative deconvolution methods have been developed which take advantage of the increased complexity which can be built into a combinatorial library using solution phase simultaneous addition methods. These methods can eliminate the need for serial synthesis and can greatly improve economy and effectiveness of identifying drug candidate compounds . One effective strategy utilizes an initial HPLC separation of a selected complex library into new fractions using a reversed-phase solid support and an organic gradient with an aqueous buffer. This approach puts less stringency on the yield and purity required from the initial synthesis, since impurities can be removed during the chromatographic step. Several examples of HPLC fractionation of libraries are illustrated in the examples below. In some cases, the biological activity is distributed among many fractions, while in other cases a single fraction contains the majority of activity present in the library.

A second, orthogonal separation of active fractions can often yield single compounds, whose identity can be established using mass spectrometry, NMR spectroscopy and other conventional methodologies. Hence, in accordance with this invention, a library can be deconvoluted without fixing functionalities serially and without the need for tagging or coding. The HPLC fractionation methods of this invention are believed to be broadly applicable to many types of combinatorial libraries prepared via solution phase synthesis and otherwise. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is an HPLC profile observed for Library 2. Figure 2 is an HPLC profile observed for Library 1. Figure 3 is an HPLC profile observed for fraction 14 of Library 1.

Figure 4 is an HPLC profile observed for Library 3. The present invention provides methods for the deconvolution, biological testing and subsequent identification of active compounds from complex combinatorial and other chemical mixtures . The present invention is applicable to both combinatorial libraries and mixtures of compounds prepared from non-combinatorial methods. A combinatorial library (known to the art per se) can synthesized using, for example, solution phase simultaneous addition of functional groups to a scaffold or mixture of scaffolds. U.S. Patent 5,571,902, incorporated herein by reference, illustrates certain approaches to solution phase synthesis of such complex chemical mixtures. A crude mixture of products can be separated by a first pass through HPLC. The fractions are then screened to determine biological activity. Active fractions are separated into individual compounds via a second pass through HPLC . The individual compounds are then each individually tested in biological assays. The active compounds are identified using e.g. mass spectroscopy and NMR spectroscopy or other conventional methods .

In accordance with one embodiment of the invention, a chemical mixture is deconvoluted so as to identify active components of the mixture so that their chemical identity can be learned. A mixture of chemical compounds is subjected to high pressure liquid chromatography . This chro atography yields an array which comprises a plurality of subgroups of the chemical mixture. The chemical subgroups thus formed are then assayed in a biological test probative of activity in a desired biological system. For - 6 -

exa ple, the assay may determine the ability of chemical species to inhibit growth or development of an infectious agent or disease state. Such assay may also determine activity in a veterinary, agricultural or other chemical context as well .

Following the assay of some or all of the chemical subgroups comprising the array resulting from the HPLC step, those subgroups having activity in the assay are selected for further actions. It is preferred that HPLC again be performed upon the subgroups, preferably one at a time, to give rise to second level arrays of chemical subgroups. At least some of these second array chemical subgroups are then assayed and active chemical species in the subgroups identified. Further assay and HPLC iterations may also be performed.

It will be appreciated that employing different HPLC media, solvents and conditions in the two (or more) HPLC steps may give rise to a preferred, improved separation of subgroups in the arrays. Accordingly, use of differing media, differing solvents or differing conditions of temperature, pressure and the like may be employed and may, in some cases be preferred.

It has also been found preferred for some uses to employ reverse phase gels as the stationary phase for HPLC in accordance with the invention.

It may also be useful to include a further chromatography step, usually one other than HPLC, to improve separations. Thus, liquid, gas, gel, plate, electrophoretic or other chromatography may be used, often before the initial HPLC step, to "clean up" or otherwise improve the mixtures for separation and evaluation/assay. Such additional chromatography may also be performed before a second or further HPLC step. Such additional chromatography may optionally be accompanied by a biological assay step. When leading compounds have been identified in accordance with the invention through iterative HPLC and - 7 -

assay, it is preferred that such compounds be identified. It will be appreciated that the chemical compounds making up the chemical mixtures or libraries are not known ab ini ti o . Further, the moiety or moieties which have shown activity in the assays are not known either. Indeed, the fact that compounds having biological activity can be identified from complex mixtures of compounds , even though the members of the mixture are numerous and unknown, is a highly important aspect of the present invention. Thus, elucidation of the chemical structure or identity of "winning" compounds is preferably performed. Such identification may employ any of the panoply of techniques known to chemistry such as mass spectrometry, NMR, IR, UV and visible spectrophotometry, derivatization and otherwise. As will be appreciated, if necessary, additional quantities of winning compounds may be isolated by using the HPLC in a preparative mode, knowing which arrays give rise to the winners .

In accordance with some preferred embodiments, chemical mixtures or libraries are formulated in such a fashion that the members thereof share certain features. Thus, common nuclear structures (backbones) , common substituents or other common features may be employed. This result is a normal feature of combinatorial chemistry efforts, giving rise to combinatorial libraries. This technique facilitates elucidation of chemical structures of particular utility in the desired biological assays, however such commonality renders more difficult the separation and elucidation of structure. Notwithstanding this, the present invention is capable of handling even complex mixtures of relatively similar compounds.

For example, a combinatorial library formed through the solution phase addition of a plurality of functional groups to a scaffold molecule having one or a plurality of reactive sites reactive with the functional groups can be accommodated by the present invention. Indeed, the present invention can operate on chemical mixtures having 20 to 50 compounds therein. In preferred used, mixtures with 50 to 100, 100 to 500 or even from 500 up to 1000 chemical compounds can be employed. Mixtures where a majority of the compounds forming them have a common functional group may also be employed.

While the present invention may identify specific active compounds, its use is even more expansive. Through identification of common factors present in compounds which show activity, the present invention can identify pharmacophores (or corresponding veterinary or agricultural chemical moieties.) Such knowledge is valuable per se and can be used to construct active compounds.

In one application of the present invention, complex combinatorial libraries are synthesized using solution-phase simultaneous addition of functionalities as referenced previously. For a new library, two or three initial analytical HPLC separations are generally performed with different solid phases and solvent systems to optimize resolution. The preparative HPLC separation can be performed with any commercial or special purpose apparatus such as a Gilson (Chicago, IL) model 215 liquid handler and pump system. In a typical preparation, a 10 mm x 250 mm reversed phase C18 or C8 column is used, and 30 mg of material is loaded from a library dissolved in the starting buffer. The elution profile is monitored e.g. at 200-280 nm. Fractions containing 1-6 mL are collected into 7 mL tubes . Solvent is removed under reduced pressure to give a volume of ~0.2 mL, and the liquid is transferred to a pretared 1.4 mL tube (Micronics) in a 96-well rack. The fractions are taken to dryness under reduced pressure, and weights of material in each tube are determined automatically using a robotic weigh station (Bohdan Automation, Mundelien, IL) . Typical first-round fractions contain 1-15 compounds, depending on the starting complexity of the library and the nature of the functional groups attached to the scaffold. Biological screening is typically performed by growing E. coli or S. pyogenes to log phase in minimal media containing HPLC fractions at e.g. 2-30 μv concentrations (see Procedures 1-8) . Screening for other biological activity can be accomplished in accordance with established procedures for such assays.

The HPLC fractionation of combinatorial libraries prepared via solution-phase simultaneous addition of functionalities offers many advantages as a deconvolution strategy. Enough material can be fractionated and purified from one HPLC run for evaluation in 10-20 biological assays. If biological activity is observed from a fraction, individual compounds can be isolated in a second round of HPLC purification wi thout the need for resynthesis . The fractionation process can be used to grade compounds on hydrophobicity, charge, or other physical properties. This information can be of value when the biological activity tracks with a characteristic of the separation process . The fractionation process provides insight into the quality of potential "winners" in the library. For example, when all of the fractions from a library demonstrate low levels of biological activity, the library may be dropped from further consideration for some or all uses since the probability of finding a single "winner" having the root characteristics of the particular is low. In cases where the biological activity is observed in a single first round fraction, additional fractionation is warranted. In the second round fractions, the biological activity may again be distributed among several individual compounds. In favorable cases, a single active compound can be isolated in a single secondary fraction, whose identity can be established using methods described below. As with other combinatorial strategies, the biological activity of a fraction and the compound identities can be used to perform QSAR analyses. The conditions developed for HPLC separation of libraries or sublibraries can be transferred to larger - 10 -

scales as required for toxicology or animal pharmacology.

Determination of the structures of "winning" compounds may be achieved through employment of any of the techniques known in the chemical arts. Electrospray and atmospheric pressure chemical ionization mass spectrometry can be used to determine the molecular masses of compounds present in each fraction. Given the known masses of the starting scaffold and the functionalities, the composition of individual compounds can be determined. In some complex cases, mass spectrometry of mass spectrometrically- separated species (MS/MS) can be employed to assign the regiochemistry of a functional group. If required, sufficient material can be isolated to perform a multidimensional NMR analysis of structure. Differences can be observed in the HPLC retention times for regioisomers having the same letters and molecular masses, which simplifies the determination of structure .

In one aspect of the present invention, combinatorial libraries are synthesized yielding from 20 to about 100 mg of crude material containing essentially equimolar amounts of from 25 to about 1000 compounds. It is preferred that mixtures having from 20-50 compounds be prepared. Pools having from 50-100 on even 100-500 and more are still more preferred. Mixtures of from 500-1000 chemical compounds can also be handled in this way. A preferred method of synthesis of combinatorial libraries is via solution-phase simultaneous addition of functionalities. See, e.g. U.S. Patent 5,571,902 issued to the assignee of this invention and incorporated herein by reference . These combinatorial libraries are fractionated into arrays or pools of e.g. from

1 to 10 compounds by HPLC fractionation techniques. Each pool or array subgroups of the combinatorial library is then evaluated such as by using high-throughput biological assays

(see Procedures 1-8 below) to determine biological activity. Array subgroups that exhibit biological activity are separated preferably using orthogonal HPLC methods to give - 11 -

single isolated compounds. These compounds are the individually screened to confirm the active members of the pool. The structure of the active compound is then assigned using mass spectrometry, NMR spectroscopy or other combinatorial techniques. The overall processes provide rapid routes to identification of active pharmacophores from combinatorial libraries and other mixtures.

The present fractionation procedures offer several benefits relative to conventional deconvolution of combinatorial libraries. When 25-50 mg of a library is prepared using the solution-phase simultaneous addition of functionalities method, sufficient material is isolated for screening and a second round of HPLC purification. The second round of HPLC purification yields adequate material for both mass spectrometric analysis and high-resolution NMR determination of structure. Thus, the identity of the active species are established without additional, systematic synthesis of compounds with fixed positions as required in serial deconvolution/rescreening schemes. The number of samples which must be screened in biological assays to identify the active pool is reduced compared to serial deconvolution schemes. This saves time and money.

In one aspect of the present invention, 30 mg of a library containing 25-1000 compounds is dissolved in 0.1-0.2 L of a convenient starting buffer. The preparative HPLC separation is performed with a Gilson (Chicago, IL) model 215 liquid handler and pump system. A 10 mm x 250 mm reversed phase C18 or C8 column (Altima) is used. The starting aqueous buffer (A) contains 50 mM ammonium acetate, pH 4.6 and the organic buffer (B) contains 50 mM ammonium acetate in 90% methanol. A gradient is run from 30% B to 100% B from 4 to 20 minutes. The elution profile is monitored at 200-280 nm. Fractions of 1-6 mL are collected into racks of 24 7 mL tubes. Following collection of 20-40 - 12 -

fractions, the solvent is removed from each tube under reduced pressure to give a volume of -200 μL, and the remaining liquid is transferred to a pretared 1.4 μL tube (Micronics) in a 96-well rack. The fractions are taken to dryness under reduced pressure, and weights of material in each tube are determined automatically using a robotic weigh station (Bohdan Automation, Mundelien, IL) .

Typical first-round fractions contain 1-15 compounds, depending on the starting complexity of the library and the chemical diversity of the functional groups attached to the scaffold or scaffolds. Biological screening is performed by growing E. coli or S. pyogenes to log phase in minimal media containing the fractions at 2-30 μm concentrations (as illustrated in Procedures 1-8). Fractions which demonstrate biological activity are separated as single compounds in a second HPLC purification step. Typically, 2-3 columns and varied gradient conditions are evaluated to optimize the resolution of the peaks in the active fraction. In a preferred separation, a 4.6 x 150 mm column packed with 5 μM particles of perfluorinated C8 supports, Zorbax C8, or Hypersil phenyl or C8 supports is used. Many aqueous buffer systems are amenable to the present process. A preferred aqueous buffer system ranges from 50 mM triethylammonium acetate at pH 6.8 to 25 mM ammonium formate, pH 3.5. The organic phases may include methanol or acetonitrile . Following optimization of conditions, 200-1000 μg of sample is loaded, and eluted with a gradient of 30-60% to 95% organic buffer. Fractions are collected into pretared 1.4 mL tubes and evaporated to dryness under reduced pressure. These individual compounds undergo a second round of screening to determine the active species .

Following isolation and identification of active species, the structures are established using mass -13-

spectrometry and NMR spectroscopy. For example, the identities of the functional groups attached to the scaffold are established from measurement of the molecular mass. The regiochemistry of the functional groups are established using conventional one-dimensional or two-dimensional NMR methods such as NOESY and TOCSY experiments.

EXAMPLE 1

3- [N- (N,N- (Bis-t-Boc) -acridinyl] -a inobenzyl alcohol

A solution of 3 -aminobenzyl alcohol (Aldrich) (2.0 g, 16.2 mmol) and N, N' -bis ( tert-butoxycarbonyl) -1H- carboxamidine (prepared as per the procedure of Bernatowicz, M.S., Tetrahedron Lett., 1993, 34, 3389-3392) (4.53 g, 14.62 mmol) in 30 mL of dry THF was stirred at room temperature for 2 days. The solvent was evaporated, and the residue was dissolved in CHC1₃. The solution was washed with water and brine, dried (Na₂S0₄) and concentrated. The residue was purified by flash chromatography on a silica gel column. Elution with 200:1 CH₂Cl₂-MeOH and then 4:1 hexanes-EtOAc gave the title compound as a white foam, yield 4.95 g (91.0%) . ^αH NMR δ 1.51 (s, 9H) , 1.54 (s, 9H) , 4.68 (s, 2H) , 7.12 (d, 1H, J = 7.8 Hz) , 7.34 (t, 1H, J = 7.8 Hz) , 7.58 (d, 2H, J = 7.8 Hz) , 10.36 (s, 1H) , 11.61 (s, 1H) . ¹³C NMR δ 25.5, 27.9, 28.2, 28.6, 64.0, 79.6, 83.6, 120.6, 120.9, 121.0, 123.2, 123.5, 128.6, 128.9, 136.4, 142.3, 153.2, 153.7,

163.3. HRMS (FAB) ffl/z 366.201 (M + H) ⁺ (C₁₈H₂₈N₄0₅ requires 366.202) .

EXAMPLE 2

3- [N- (N,N- (Bis-t-Boc) -acridinyl] -aminobenzylbro ide (L_X1) To a solution of 3- (N-acridinyl) -aminobenzyl alcohol (1.45 g, 3.97 mmol) in 15 mL of CH₂C1₂ were added N- bromosuccinimide (ΝBS) (0.78 g, 4.4 mmol) and PPh₃ (1.25 g, -14-

4.76 mmol) . The resulting solution was stirred at room temperature for 2 h and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column. Elution with 50:1 hexanes-EtOAc gave the title compound as a white foam: yield 1.21 g (71.6%) .

^XH NMR δ 1.51 (s, 9H) , 1.54 (s, 9H) , 4.47 (s, 2H) , 7.14 (d, 1H, J = 7.6 Hz) , 7.32 (t, 1H, J = 7.9 Hz) , 7.58-7.64 (m, 2H) , 10.37 (s, 1H) , 11.63 (s, 1H) . HRMS (FAB) /z 428.117 (M + H)⁺ (C₁₈H₂₇BrN₃0₄ requires 428.118) . Anal. Calcd. for C₁₈H₂₆BrN₃0₄ : C, 50.57; H, 6.14; N, 9.84. Found: C, 50.55; H, 6.12; N, 9.99.

EXAMPLE 3

N- (N,N-Bis- t-Boc) -guanidinyl piperazine

A mixture of piperazine (34.46 g, 0.4 mol), 1,3- bis ( tert-butoxy carbonyl) -2-methyl-2-thiopseudourea

(Aldrich) (29.0 g, 0.1 mol) in 260 mL of DMF was stirred at 50-60 °C for 2 h. The solvent was evaporated to dryness and the residue was dissolved in water-chloroform. The organic phase was separated and the aqueous phase was extracted with chloroform. The combined organic phase was washed with brine, dried (Na₂S0₄) and concentrated. The residue was purified by flash chromatography on silica gel using 1:2 hexanes-EtOAc, and then 1:1 EtOAc-MeOH as eluents to afford 27.2 g (83%) of the title compound as a white solid; silica gel TLC R_f 0.34 (100% MeOH) . ^αH NMR δ 1.42 (s, 18H) , 2.78-2.91 (m, 4H) , 3.40-3.66 (m, 4H) . HRMS (FAB) /z 329.218 (M + H) ⁺ (C₁₅H₂₉N₄0₄ requires 329.218). Anal. Calcd. for C₁₅H₂₈N₄0₄ : C, 54.86; H, 8.58; N, 17.06. Found: C, 54.87; H, 8.48; N, 17.20.

EXAMPLE 4

N-Bromoacetyl-N- [ (N,N-Bis- -Boc) -guanidinyl] piperazine (L₁₂) - 15 -

A solution of bromoacetyl bromide (2.06 g, 10.2 mmol) in 20 mL of THF was added dropwise to a stirred solution of N- (N,N-Bis- -Boc) -guanidinyl piperazine (3.28 g, 10 mmol) and diisopropylethylamine (2.1 mL, 1.56 g, 12 mmol) in 50 mL of THF at -30 °C. The cooling bath was removed and the reaction mixture was stirred for 1.5 h. After the solvent was evaporated, the residue was dissolved in chloroform. This solution was washed with water, brine, dried (Na₂S0₄) and concentrated. The residue was purified by flash chromatography on silica gel column. Elution with 5:1, 2:1 and then 1:1 hexanes-EtOAc afforded 3.5 g (78%) of the title compound as a white solid.

Silica gel TLC R_f 0.45 (1:2 hexanes-EtOAc) . ²H NMR δ 1.46 (s, 18H) , 3.48-3.75 (m, 8H) , 3.85 (s, 2H) ; HRMS (FAB) m/z 449.141 (M + H) ⁺ (C₁₇H₃₀BrN₄O₅ requires 449.140). Anal. Calcd. for C₁₇H₂₉BrN₄0₅ : C, 45.44; H, 6.49; N, 12.46. Found: C,45.54; H, 6.25; N, 12.66.

EXAMPLE 5

N' -ethylenediamine- (bis-N-tert-butoxycarbonyl) -1-carbox- amidine

1, 3 -Bis (tert-butoxycarbonyl) -2 -methyl -2- thiopseudourea (11.62 g, 40.0 mmol) was dissolved in THF and added dropwise over 30 min to a stirred solution of ethylenediamine (26.74 mL, 400 mmol) in 2% water (5 mL) and THF (255 mL) which was preequilibrated and maintained at 50 °C. The reaction mixture was stirred at 50 °C for an additional 10 min until the absence of thiopseudourea by TLC. Prolonged reaction times resulted in degradation of the product . The solvent was evaporated in vacuo to give an oil which was dissolved in CHC1₃ and washed with saturated NaHC0₃ (x3) . The organic layer was separated, dried with MgS0₄, and the solvent was evaporated in vacuo to give an oil . The oil was purified by flash chromatography using hexane :CH₂C1₂ :NEt₃ (20:80:1, v/v/v) to -16-

afford the title compound as a white solid (9.30 g, 77%) . ^XH NMR (CDC1₃) : δ 11.5 (br, 1H) , 8.66 (br, 1H) , 3.62 (m, 2H) , 3.48 (m, 2H) , 2.90 (m, 2H) , 1.50 (s, 18H) . ¹³C NMR (CDCI₃) : δ 163.4, 156.2, 153.0, 82.9, 79.0, 43.2, 40.8, 28.3, 28.1. HRMS (ES+) : M + H, calcd 303.2032, found 303.2039. Anal. Calcd. for C₁₃H₂₆N₄0₄ : C, 51.64, H, 8.67, N, 18.53. Found: C, 51.44, H, 8.63, N, 18.13.

EXAMPLE 6

2-Bromo-N' - [2' - (bis-N- -Boc) et ylguanidino] -acetamide (L₁₃)

To a solution of N' -ethylenediamine- (bis-N-tert- butoxycarbonyl) -1-carboxamidine (8.55 g, 28.3 mmol) dissolved in CH₂C1₂ (140 mL) and THF (140 mL) was added pulverized NaHC0₃ (23.8 g, 283 mmol) and the mixture was stirred for 5 min at ambient temperature then cooled to - 50°. Bromoacetylbromide (2.71 mL, 31.1 mmol) was added dropwise and the reaction mixture was allowed to slowly warm to ambient temperature over several hours . The solvent was evaporated in vacuo to give a solid which was suspended in CH₂C1₂. The suspension was washed with water (x3) . The organic layer was separated, dried with MgS0₄, and the solvent evaporated in vacuo to give an oil . The oil was purified by flash chromatography using hexane:EtOAc (30:70, v/v) to give the title compound as a white foam (7.84 g, 65%) . ^lH NMR (CDCI₃) : δ 11.45 (br, 1H) , 8.65 (br, 1H) , 8.20 (br, 1H) , 3.85 (s, 2H) , 3.60 (m, 2H) , 3.45 (m, 2H) , 1.50 (S, 18H) . ¹³C NMR (CDCI₃) : δ 166.1, 162.6, 157.2, 152.7, 83.2, 79.2, 77.6, 77.0, 76.3, 41.7, 39.4, 28.5, 28.2, 27.8, 27.6. HRMS (ES+) : M + Cs, calcd 555.0219/557, found 555.0231/557. M + H, calcd 423/425, found 423/425. Anal. Calcd for C₁₅H₂₇BrN₄0₅: C, 42.56; H, 6.43; N, 13.24. Found: C, 43.12, H, 6.24, N, 12.88. - 17 -

EXAMPLE 7

4-Bromopyridine-2, 6-dimethanol

The title compound was synthesized via modification of Acta Chem. Scand. B42, 373, 1988. A mixture of NaBH₄ (6.8 g, 180 mmol), diethyl 4-bromopyridine-2 , 6- dicarboxylate (12.1 g, 40 mmol) in absolute EtOH (500 mL) was heated under reflux for 18 h. The solvent was removed in vacuo and the residue was treated with hot saturated aqueous NaHC0₃. The mixture was extracted with EtOAc, the combined organic phase was dried (Na₂S0₄) and evaporated. The dried material was recrystallized from EtOAc to give 7.3 g (83%) of the title compound as white crystals.

R_f 0.62 (EtOAc/MeOH 20:1). ^XH NMR (DMSO-d₆, 200 MHz) δ 4.54 (d, 4 H, J = 5.4 Hz), 5.55 (5, 2 H, J = 5.8 Hz), 7.53 (s, 2 H) . ¹³C NMR (DMSO-d δ 63.69, 118.08, 121.10, 163.19.

EXAMPLE 8

4-Piperazinylpyridine-2, 6-diethyldicarboxylate

A mixture of diethyl 4-bromopyridine-2 , 6- dicarboxylate (14.0 g, 46 mmol) and piperazine (12.2 g, 142 mmol) in dioxane (700 mL) was refluxed for 2 days. The solid was filtered and the collected solvent was evaporated in vacuo to a residue. The residue was diluted with CH₂C1₂ and washed with NaCl . The layers were separated and the aquous phase was extracted twice with CH₂C1₂. The combined organic phase was washed with brine, dried (Na₂S0₄) and concentrated under reduced pressure. The crude white solid was purified by silica gel flash column chromatography using CH₂Cl₂:MeOH (9:1, 6:4) as the eluent. Further purification by recrytallization using hexanes/EtOAc gave 12.3 g (87%) of the title compound as white crystals.

R_f 0.20 (MeOH:30% NH₄0H, 100:1). ^ NMR (CD₃OD) δ -18-

1.41-1.48 (m, 6 H) , 1.63-1.80 (br, 1 H) , 2.99-3.04 (m, 4 H) , 3.42-3.47 (m, 4 H) , 4.39-4.50 (m, 4 H) , 7.63 (s, 2 H) . ¹³C NMR (CD₃OD) δ 14.19, 45.53, 46.96, 62.16, 111.34, 149.25, 156.38, 165.62. HRMS (FAB) m/z 308.162 (M + H) ⁺ (C₁₅H₂₂N₃0₄ requires 308.161) .

EXAMPLE 9

4-Piperazinylpyridine-2, 6-dimethanol

A mixture of 4-Bromopyridine-2 , 6-dimethanol (3.2 g, 15 mmol) and piperazine (13.0 g, 150 mmol) in dioxane (200 mL) was refluxed for 4 days. The solvent was evaporated, and the residue was thouroughly dried at elevated temperature under high vacuum to remove excess piperazine. The crude solid was purified by silica gel flash column chromatography using EtOAc :MeOH (10/1 - l/l, v/v) and MeOH:NH₄OH (1/0 -50/1, v/v) as the eluent to afford 2.9 g (86%) of the title compound as a white solid. —

Mp 178.0 - 179.5 °C. R_f 0.35 (MeOH/NH₄0H 50:1). ^XH NMR (CD₃OD) δ 2.88-2.98 (m, 4 H) , 3.33-3.42 (m, 4 H) , 4.57 (S, 4 H) , 6.86 (s, 2 H) . ¹³C NMR (CD₃0D) δ 46.15, 47.72, 65.55, 104.35, 158.58, 162.11. HRMS (FAB) m/z 224.139 (M + H)⁺ (CnH₁₉N₃0₂ requires 224.139).

EXAMPLE 10

4- [N4 ' - (t-Boc) -piperazin-Nl" -yl]pyridine-2, 6-dimethanol A solution of di- tert-butyl dicarbonate (15.0 g, 69 mmol) in THF (100 mL) was added to a stirred solution of 4-Piperazinylpyridine-2 , 6-dimethanol (10.8 g, 48 mmol) and Et₃N (14 mL, 100 mmol) in THF:MeOH (1/2, v/v) (300 mL) at 0 °C. The reaction mixture was allowed to warm to rt and stirred for 24 h. The solvent was removed in vacuo, the residue diluted with CH₂C1₂ and washed with H₂0. The layers were separated and the aquous phase was extracted twice with CH₂C1₂. The combined organic phase was washed with brine, dried (Na₂S0₄) , and concentrated under reduced - 19 -

pressure. The crude product was purified by column chromatography using EtOAc :MeOH (7/3 - 1/9, v/v) as the eluent to give 14.3 g (92%) of the title compound as a white solid. R_f 0.70 (MeOH/NH₄OH 100:1). ^XH NMR (CD₃0D) δ 1.49

(s, 9 H) , 3.44 - 3.47 (m, 4 H) , 3.55 - 3.58 (m, 4 H) , 4.57 (S, 4 H) , 6.88 (S, 2 H) . ¹³C NMR (CD₃OD) δ 28.7, 44.0, 46.9, 65.5, 81.6, 104.5, 156.4, 158.1, 162.1. HRMS (FAB) m/z 346.173 (M + Na)⁺ (C₁₆H₂₅N₃0₄Na requires 2346.174. Anal. Calcd for C₁₆H₂₅N₃0₄ : C, 59.42; H, 7.79; N, 12.99. Found: C, 59.37; H, 7.86; N, 12.71.

EXAMPLE 11

4- [N4 ' - (t-Boc) -piperazin-Nl' -yl] -pyridine-2, 6-dimethanol- ditosylate 4- [N4 ' - ( t-Boc) -piperazin-Nl ' -yl] pyridine-2 , 6- dimethanol

(3.9 g, 12 mmol) was dissolved in a mixture of THF (80 mL) and NaOH (14 mL, 4.3 M in H₂0) . The reaction mixture was cooled to 0 °C and p-toluenesulfonyl chloride (5.5 g, 29 mmol) was added slowly. The reaction mixture was allowed to warm to rt and stirred for 24 h. The solvent was evaporated, and the residue was diluted with CH₂C1₂ and washed with H₂0. The layers were separated and the aquous phase was extracted twice with CH₂C1₂. The combined organic phase was washed with brine, dried (Na₂S0₄) and concentrated under reduced pressure. The crude product was purified by silica gel flash column chromatography using hexanes : EtOAc (8:2 - 6:4, v/v) as the eluent to give 6.0 g (79%) of the title compound as a white solid. R_f 0.52 (hexanes/EtOAc 3:7). ^XH NMR (CDC1₃) δ 1.49

(S, 9 H) , 2.44 (s, 6 H) , 3.30 - 3.32 (m, 4 H) , 3.53 - 3.56 (m, 4 H) , 4.93 (s, 4 H) , 6.63 (s, 2 H) , 7.31 - 7.33 (m, 4 H) , 7.78 - 7.80 (m, 4 H) . ¹³C NMR (CDC1₃) δ 21.6, 28.4, 45.8, 71.7, 80.4, 105.2, 128.0, 130.0, 132.8. HRMS (FAB) -20-

m/z 764.108 (M + Cs)⁺ (C₃₀H₃₇N₃O₈S₂Cs requires 764.107) . Anal. Calcd for C₃₀H₃₇N₃O₈S₂ : C, 57.04; H, 5.90; N, 6.65. Found: C, 56.77; H, 6.06; N, 6.41.

EXAMPLE 12 N1,N6,N10-Tris (2-nitrobenzenesulfonyl) spermidine

A solution of 2-nitrobenzenesulfonyl chloride (50.4 g, 227 mmol) in CH₂C1₂ (150 mL) was added to a stirred solution of spermidine (10.0 g, 69 mmol) and Et₃N (34 mL, 244 mmol) in CH₂C1₂ (500 mL) at 0 °C. The reaction mixture was allowed to warm to rt and stirred for 24 h. The solvent was removed in vacuo, and the residue was diluted with CH₂C1₂ and washed with H₂0. The layers were separated and the aquous phase was extracted with CH₂C1₂. The combined organic phase was washed with brine, dried (Na₂S0 ) and concentrated under reduced pressure. The crude product was purified by silica gel flash column chromatography using hexanes : EtOAc (1:1 - 9:1, v/v) as the eluent to afford 42.2 g (88%) of the title compound as a pale green foam R_f 0.33 (hexanes/EtOAc 2:8). ^XH NMR (CDC1₃) δ 1.56 - 1.61 ( , 4 H) , 1.84 - 1.91 (m, 2 H) , 3.11 - 3.23 (m, 4 H) , 3.28 - 3.43 (m, 4 H) , 5.34 - 5.38 (m, 1 H) , 5.63 - 5.69 (m, 1 H) , 7.64 - 7.92 (m, 9 H) , 8.02 - 8.06 (m, 1 H) , 8.13 - 8.18 (m, 2 H) . ¹³C NMR (CDC1₃) δ 25.2, 26.6, 28.9, 40.8, 43.1, 45.1, 47.5, 124.3, 125.4, 127.6, 130.6, 131.0,

132.1, 132.8, 133.0, 133.4, 133.8, 134.1, 148.0. HRMS (FAB) m/z 832.995 (M + Cs)⁺ (C₂₅H₂₈N₆0₁₂S₃Cs requires 832.998) . - 21 -

EXAMPLE 13

N1,N3,N6-Tris (2-nitrobenzenesulfonyl) -1, 6-diamino-3- azahexane

The title compound was synthesized following the procedure illustrated in Example 51. Using N-(2- aminoethyl) -1, 3-propanediamine (8.0 g, 8.6 mL, 68 mmol), 2-nitrobenzenesulfonyl chloride (53.0 g, 239 mmol), and Et₃N (33 L, 237 mmol) in CH₂C1₂ (600 mL) afforded 43.5 g (95%) of the title compound as a pale green foam. R_f 0.33 (hexanes/EtOAc 2:8). ^αH NMR (CD₃OD) δ 1.86 - 1.92 (m, 2 H) , 3.12 - 3.16 (m, 8 H) , 5.61 - 5.75 (m, 2 H) , 7.61 - 8.30 (m, 12 H) . ¹³C NMR (CD₃OD) δ 28.7, 40.7, 42.2, 46.0, 47.8, 124.4, 124.8, 125.5, 131.0, 132.2, 133.0, 133.1, 133.8, 133.9, 134.2, 135.9, 148.1. HRMS (FAB) m/z 804.963 (M + Cs)⁺ (C₂₃H₂₄N₆0₁₂S₃Cs requires 804.966).

EXAMPLE 14

2,9- (bis-o-nitrobenzenesulfonyl) -5- (o- nitrobenzenesulfonyl) triazadecane [2.6] - [4- (N4-t-Boc- piperazine-1-yl) ] pyridinophane A mixture of 4- [N4 '-( t-Boc) -piperazin-Nl ' -yl] - pyridine-2 , 6-dimethanolditosylate (14.5 g, 23 mmol) ,Nl,N3,N6-tris (2-nitrobenzenesulfonyl) -1, 6-diamino-3- azahexane (16.9 g, 25 mmol) and Cs₂C0₃ (29.8 g, 92 mmol) in DMF (750 mL) was stirred at rt for 24 h. The solvent was removed in vacuo and the residue was diluted with

CH₂C1₂ and washed with H₂0. The layers were separated, and the aqueous phase was extracted with CH₂C1₂. The combined organic phase was washed with brine, dried (Na₂S0₄) , and concentrated in vacuo . The residue was purified by silica gel flash column chromatography using hexanes :EtOAc (6/4 - 1/9, v/v) to give 16.3 g (72%) of the title compound as a white foam.

R_f 0.35 (EtOAc 100%). H NMR (CDC1₃) δ 1.49 (s, 9 -22-

H) , 1.99 - 2.01 (m, 2 H) , 2.92 - 2.94 (m, 2 H) , 3.21 - 3.23 (m, 2 H) , 3.25 - 3.39 (m, 8 H) , 3.52 - 3.59 (m, 6 H) 4.42 - 4.44 (m, 4 H) , 6.85 (s, 1 H) , 6.96 (s, 1 H) , 7.59 7.77 (m, 9 H) , 7.92 - 8.09 (m, 3 H) . ¹³C NMR (CDC1₃) δ 27.5, 28.4, 45.7, 47.1, 47.5, 49.0, 55.9, 56.0, 80.3, 108.0, 108.4, 124.2, 124.3, 130.7, 131.1, 131.8, 131.9,

132.0, 132.1, 132.3, 133.6, 133.8, 148.0, 148.3, 154.5,

156.1, 156.6, 157.2. HRMS (FAB) m/z 1092.126 (M + H) ⁺ (C₃₉H₄₅N₉0₁₄S₃Cs requires 1092.130) .

EXAMPLE 15

2,5, 9-Triazadecane[2.6] - [4- (N4- t-Boc-piperazine-1- yl) ] pyridinophane

Thiophenol (2.8 mL, 27 mmol) was added to a stirred mixture of 2 , 9- (bis-o-nitrobenzenesulfonyl) -5- ( o- nitrobenzenesulfonyl) triazadecane [2.6] - [4- (N4- -Boc- piperazine-1-yl) ] pyridinophane (7.4 g, 7.5 mmole) and anhydrous K₂C0₃ (10.4 g, 75 mmol, 10 equiv) in DMF (80 mL) . The reaction mixture was stirred at rt for 4 h, and the solvent was removed under reduced presure . The residue was diluted with CH₂C1₂ and washed with H₂0. The layers were separated, and the aqueous phase was extracted with CH₂C1₂ (3x) . The combined organic phase was washed with brine, dried (Na₂S0₄) , and concentrated in vacuo. The crude product was purified by silica gel flash column chromatography using MeOH (100%), and MeOH:NH₄OH (100:1 - 9/1, v/v) as the eluent to give 4.65 g (65%) of the title compound as a pale yellow foam.

R_f 0.10 (MeOH/NH₄OH 4:1). ^αH NMR (CDC1_.J δ 1.47 (s, 9 H) , 1.75-1.85 (m, 2 H) , 2.78-2.86 (m, 2 H) , 2.89-2.90 (m, 4 H) , 3.19-3.24 (m, 2 H) , 3.39-3.42 (m, 4 H) , 3.52-

3.53 (m, 4 H) , 3.87-3.89 (m, 4 H) , 6.65 (s, 1 H) , 6.68 (s, 1 H) . ¹³C NMR (CDC1₃) δ 26.78, 27.42, 28.62, 46.88, 47.39, 52.23, 53.55, 55.02, 81.60, 106.09, 106.79, 156.34, 157.78, 159.91, 164.41. HRMS (FAB) m/z 405.297 (M + H) ^{+ ■} 23

( C₂₁H₃₆N₆0₂H requires 405 . 297 ) .

EXAMPLE 16

Preparation of 2, 9- (bis-L-_L-._!,*,) -5- (3-fluorobenzyl) -2, 5, 9* triazadecane [10] (2, 6) -pyridinophane (Library 1)

Formula I

A mixture of benzyl bromide (Br-L-_t) (147 μL, 207 mg, 1.2 mmol), α-bromo-m-xylene (Br-L₂) (169 μL, 222 mg, 1.2 mmol), 3-fluorobenzyl bromide (Br-L₃) (149 μL, 227 mg, 1.2 mmol), 3-cyanobenzyl bromide (Br-L₄) (237 mg, 1.2 mmol), cinna yl bromide (Br-L₅) (237 mg, 1.2 mmol), 3- chlorobenzyl bromide (Br-L₆) (163 μL, 248 mg, 1.2 mmol), 3-nitrobenzyl bromide (Br-L₇) (259 mg, 1.2 mmol), methyl 3- (bromomethyl) benzoate (Br-L₈) (276 mg, 1.2 mmol), oJ - bromo-α, , α-trifluoro-m-xylene (Br-L₉) (185 μL, 288 mg, 1.2 mmol), and 3-bromobenzyl bromide (Br-L₁₀) (300 mg, 1.2 mmol) (total 12 mmol, 2.4 equiv) in anhydrous CH₃CN (50 mL) was added to a stirred mixture of 2 , 9-diaza-5- ( - Boc) azadecane [10] (2,6) pyridinophane (1.60 g, 5.0 mmol) and anhydrous K₂C0₃ (10.0 g, 72 mmol) in 130 mL of CH₃CN. The resulting reaction mixture was stirred at rt overnight and concentrated. The residue was dissolved in a mixture of chloroform-water. The organic phase was separated, and the aqueous phase was extracted with CHC1₃. The combined organic phase was washed with brine, dried - 24 -

(Na₂S0₄) , and concentrated. The residue was purified by flash chromatography on a silica gel column using hexanes : ethyl acetate (15/1, v/v), ethyl acetate (100%) and then ethyl acetate :methanol (20/1, v/v) as eluents. The library fractions, monitored by TLC, were collected and concentrated to afford 2.80 g (97%) of an intermediate library that still has a t-Boc blocking group on the N5 position as a pale yellow oil; MS (ES) m/z 501-659 (M + H⁺) . Trifluoroacetic acid (TFA) (35 mL) was added to a stirred solution of the intermediate library (2.80 g, 4.87 mmol) in 7 L of CHC1₃ at 0 °C. The resulting reaction mixture was stirred at rt for 3 h and concentrated to remove excess amount of TFA. The residue was dissolved in CHC1₃, washed with an aqueous solution of Na₂C0₃, washed with brine, dried (Na₂S0₄) , and then concentrated. The residue was purified by flash chromatography on a silica gel column using Me0H:NH₄0H (100/0 to 40:1, v/v) followed by MeOH:30% NH₄0H (5:1, v/v) to afford 2.30 g (88%) of the deblocked library as a light yellow oil; silica gel TLC R_f 0.35 (5:1 MeOH-30% NH₄0H) ; ^λE NMR δ 1.65-1.82 (m, 2H) , 2.28-2.35 (m, 0.3H), 2.45-2.65 (m, 4H) , 2.64-2.96 (1, 4H) , 3.40-4.00 ( , 10.3H), 6.20-8.40 ( , 11.3H); MS (ES) m/z 401-559 (M + H) ⁺ . A mixture of the deblocked library (105 mg, 0.22 mmol), 3-fluorobenzyl bromide (L₃) (0.266 mmol), and K₂C0₃ (0.5 g, 3.6 mmol) in CH₃CN (5 mL) was stirred at rt overnight. After the solvent was evaporated, the residue was dissolved in a mixture of water-chloroform. The layers were separated, and the aqueous phase was extracted with CHC1₃. The combined organic phase was washed with brine, dried (Na₂S0₄) , and concentrated. The residue was purified by preparative TLC using MeOH: 30% NH₄0H (90:1, v/v) as a developing agent to afford the title library ^■ 25 *

(Formula I) as a colorless oil in 96% yield.

EXAMPLE 17

Preparation of 2, 9- (bis-Ln.x_s) -2, 5, 9- triazadecane

[10] (2, 6) -pyridinophane (Library 2)

N

Ll l-15 — N N—Ll l-15

Formula II

A solution containing equal molar amounts of 3- [N- (N,N- (Bis-t-Boc) -acridinyl] -aminobenzylbromide (Br-L₁₂) , N-bromoacetyl-N- [ (N, N-Bis- -Boc) -guanidinyl] piperazine (Br-L₁₂) , 2-bromo-N' - [2 ' - (bis-N- t-Boc) ethylguanidino] - acetamide (Br-L₁₃) , benzoic anhydride (Aldrich) (C₆H₅C(=0)0- L₁₄) , and oJ -bromo-α, α, α-trifluoro-m-xylene (Aldrich) (Br- L₉) (total 1.1 equiv per reactive site on each scaffold) in CH₃CN was added to a mixture of scaffold 2,9-diaza-5- ( -Boc) azadecane [10] (2,6) pyridinophane (1.0 equiv) and K₂C0₃ (5.0 g, 36 mmol) in CH₃CN (20 mL) . The resulting reaction mixture was stirred at rt for 2-4 hours

(monitored by TLC) . The reaction mixture was worked up, purified and deblocked as illustrated above for Library 1. Following the deblocking step excess TFA was removed by evaporation and the Library was dissolved in hydrochloride saturated MeOH. Evaporation of the solvent gave the hydrochloride salts of the title library as a foam. -26 *

EXAMPLE 18

Preparation of 2, 9- (bis-L^_o) -5- (3-fluorobenzyl) -2, 5, 9 - triazadecane [10] (2, 6) -pyridinophane (Library 3)

111 Ll M5

Foπnula III

Library 3 was prepared as above for Library 2 using 2,5, 9 -triazadecane [2.6] - [4- (N4- -Boc-piperazine-1- yl) ] pyridinophane (Example 15), the corresponding bromides and benzoic anhydride. The intermediate library obtained was deprotected with TFA and treated with HCl to afford the hydrochloride salt as a pale yellow foam.

EXAMPLE 19

HPLC Fractionation of Library 1

A quantity of Library 1, containing about 100 members, was prepared via solution phase simultaneous addition of functionalities. This library was chosen as a test example where the properties of the functional groups, meta-substituted benzyl functions, are similar.

The HPLC trace is presented in Figure 2. An Altima C18 reversed-phase column (10x250 mm; 5μM particles) was - 27 -

employed. The starting aqueous buffer (A) contained 50 mM ammonium acetate, ph 4.6 and the organic buffer (B) contained 50 μM ammonium acetate in 90% methanol. A gradient was run from 60% B to 100% B from 4 to 30 minutes. The separation was adequate to resolve 19 peaks, and 25 fractions were collected. Activity was observed for fractions 12-15 in the screening procedure for E. coli (imp^") (see Procedure 1 below) . In addition, fractions 12-15 inhibited the growth of S. pyogenes at 5 μM. Mass spectrometric analysis showed that fraction 14 contained 11 compounds. Three HPLC conditions were compared in attempts to resolve the compounds. Separation with a Fluofix C18 column and an Intertsil C8 column failed to resolve all of the compounds. However, a second HPLC separation could be performed on fraction 14 using a Zorbax C8 solid support, with a flow rate of 1 mL/min.

Eleven individual compounds were resolved (Figure 3) in the second chromatogram. A Zorbax C8 reversed-phased column (4.6x150 mm; 5 μM particles) was employed. A 200 L portion of a 2 μM stock solution was loaded onto the column. The starting aqueous buffer (A) contained 50 mM triethylammonium acetate, pH 6.8, and the organic buffer (B) contained 50 mM triethylammonium acetate, pH 6.8 in 95% acetonitrile . Some of these fractions inhibited growth of S. pyogenes at 5 μM concentrations. The activities of the individual fractions summed to give 75- 88% of the biological activity found in the initial screening of the library. These results suggest that the inhibition of microbial growth does not result from synergistic activity among the compounds in the starting fraction. - 28 -

EXAMPLE 20

HPLC Fractionation of Library 2

A quantity of Library 2 (Example 17), containing ca. 25 charged and aromatic members was prepared via solution phase simultaneous addition of functionalities and the crude product was fractionated.

The HPLC chromatographic fractionation resulted in 19 fractions (Figure 1) . An Altima C18 reversed-phase column (10x250 mm; 5 μM particles) was employed. The starting aqueous buffer (A) contained 50 μM ammonium acetate, pH 4.6 and the organic buffer (B) contained 50 μM ammonium acetate in 90% methanol. A gradient was run from 30% B to 100% B from 4 to 20 minutes. The chromatogram contained eight major peaks and twelve lesser peaks, and fractions were pooled to generate 9 wells for screening. The contents of the individual fractions were redissolved at 5 μM based on the mass of material in each fraction. The fractions were screened for their ability to inhibit the growth of E. coli (imp-) and S. pyogenes (see Procedure 1 below) . Fractions 5, 7, and 8 inhibited the growth of S. pyogenes (>90%) at 20 μM, while only fraction 8 inhibited growth by >90% at 5 μM. Fractions 7 and 8 inhibited growth of E. coli by >90% at a 20 μM concentration.

Mass spectrometric analysis of the composition showed less than 10% compound overlap between adjacent fractions. The early, polar fraction 2 containing compounds with guanidino letters also inhibited the binding of Tar RNA with tat peptide at 100 μM in a scintillation proximity assay system (Procedure 1, Tier IV) . Mass spectrometric analysis demonstrated that fraction 8 contained a single compound with a molecular mass of 537.3 Da. The structure was established as 2 , 9- (bis-benzoyl) -2 , 5 , 9 -triazadecane [10] (2 , 6) -pyridinophane . - 29 -

EXAMPLE 21

HPLC Fractionation of Library 3

A quantity of Library 3 (Example 18) , containing 216 members was prepared via solution phase simultaneous addition of functionalities and 22 mg of the crude product was fractionated using a gradient of ammonium acetate/acetonitrile and a C8 reversed phase silica column. The HPLC fractionation resulted in over sixty peaks being resolved, and 34 fractions were collected for screening (Figure 4) . An Alti a C18 reversed phase column (10x250 mm; 5 μM particles) was employed. The starting aqueous buffer (A) contained 25 mM ammonium formate, pH 3.5 and the organic buffer (B) contained 25 mM ammonium formate in 85% acetonitrile . A gradient was run form 0% B to 100% B over 40 minutes. Fractions 14-16 and 21-22 inhibited the growth of E. coli and S. pyogenes at 20 and 10 μM, respectively. Several of the more polar early fractions also inhibited transcription/translation in a rabbit reticulocyte lysate system.

PROCEDURE 1

Antimicrobial Assays TIER I

A. Streptococcus pyogenes (Gram Positive Specie)

S . pyogenes [American Type Culture Collection (ATCC) # 14289] is used in this bacterial growth assay. To initiate the exponential phase of bacterial growth prior to the assay, a sample of bacteria is grown for 6 hours in Todd Hewitt Broth (Difco 0492-17-6) at 37 °C then re- inoculated into fresh media and grown overnight at 37 °C. The bacterial cells are collected by centrifugation for 10 minutes at 3200 rp , diluted and absorbance read at 595 nm. Bacteria diluted in 2x Todd-Hewitt Broth (75 μL) are added to the compound mixtures (75μL) for a total volume 30

of 150 μL. The assays are performed in 96-well microplates with approximately 1 x 10⁴ colony forming units (CFU) per well .

The plates are incubated at 37 °C and growth monitored over a 24 hour period by measuring the optical density at 595 nm using a BioRad model 3550 UV microplate reader. The percentage of growth relative to a well containing no test compound or library is determined. Ampicillin and tetracycline antibiotic controls are concurrently tested in each screening assay.

Compounds are assayed in duplicate at a single dose. Compounds which show inhibitory activity are re-tested in duplicate at multiple doses to determine minimum inhibitory concentration (MIC) .

B. E. coli imp- (Gram Negative Specie)

The strain E. coli imp- obtained from Spencer Benson (Sampson, B.A., Misra, R. & Benson, S.A... Genet i cs , XL.89 ,

122, 491-501, Identification and characterization of a new gene of Escheri chia coli K-12 involved in outer membrane permeability) is used in this bacterial growth assay. To initiate the exponential phase of bacterial growth prior to the assay, a sample of bacteria is grown for 6 hours in Mueller Hinton II Broth (BBL 12322) at 37^rC then re- inoculated into fresh media and grown overnight at 37 °C. The bacterial cells are collected by centrifugation for 10 minutes at 3200 rpm, diluted and absorbance read at 595 nm.

Bacteria diluted in 2x Mueller Hinton II Broth (75 μL) are added to the compound mixtures (75μL) for a total volume of 150 μL. The assays are performed in 96-well microplates with approximately 1 x 10⁴ colony forming units (CFU) per well. The plates are incubated at 37 °C and growth monitored over a 24 hour period by measuring the optical density at 595 nm using a BioRad model 3550 UV - 31-

microplate reader. The percentage of growth relative to a well containing no compound is determined. Ampicillin and tetracycline antibiotic controls are concurrently tested in each screening assay. Compounds are assayed in duplicate at a single dose. Compounds which show inhibitory activity are re-tested in duplicate at multiple doses to determine minimum inhibitory concentration (MIC) . Such compounds may be further teste dwith one or more gram positive bacteria such as but not limited to the Tier II et seq. organisms described in the following sections.

TIER II

A. Gram Positive

The following gram positive strains are used to test compounds which showed activity in at least one of the Tier I organisms: Staphylococcus aureus (ATCC #13709), Entercoccus hirae (ATCC #10541) , Streptococcus pyogenes (ATCC #49399) . To initiate the exponential phase of bacterial growth prior to the assay, a sample of bacteria is grown for 6 hours in Todd Hewitt Broth (Difco 0492-17- 6) at 37 °C then re-inoculated into fresh media and grown overnight at 37 °C. The bacterial cells are collected by centrifugation for 10 minutes at 3200 rpm, diluted and absorbance read at 595 nm. Bacteria diluted in 2x Todd Hewitt Broth (75 μL) are added to the compound mixtures (75μL) for a total volume of 150 μL.

The assays are performed in 96-well microplates with approximately 1 x 10⁴ colony forming units (CFU) per well. The plates are incubated at 37 °C and growth monitored over a 24 hour period by measuring the optical density at 595 nm using a BioRad model 3550 UV microplate reader. The percentage of growth relative to a well containing no compound is determined. Ampicillin and tetracycline antibiotic controls are concurrently tested in each - 32 -

screening assay. Compounds are assayed in duplicate at multiple doses to determine minimum inhibitory concentration (MIC) .

B. Gram Negative The following gram negative strains are used to test compounds which showed activity in at least one of the Tier I organisms: Escherichia coli (ATCC #25922), Klebsiella pneumoniae (ATCC #10031) , Proteus vulgaris (ATCC #13315) , and Pseudomonas aeruginosa (ATCC #9027) . To initiate the exponential phase of bacterial growth prior to the assay, a sample of bacteria is grown for 6 hours in Mueller Hinton II Broth (BBL 12322) at 37 °C then re-inoculated into fresh media and grown overnight at 37 °C. The bacterial cells are collected by centrifugation for 10 minutes at 3200 rpm, diluted and absorbance read at 595 nm. Bacteria diluted in 2x Mueller Hinton II Broth (75 μL) are added to the compound mixtures (75μL) for a total volume of 150 μL.

The assays are performed in 96-well microplates with approximately 1 x 10⁴ colony forming units (CFU) per well. The plates are incubated at 37 °C and growth monitored over a 24 hour period by measuring the optical density at 595 nm using a BioRad model 3550 UV microplate reader. The percentage of growth relative to a well containing no compound is determined. Ampicillin, tetracycline and ciprofloxacin antibiotic controls are concurrently tested in each screening assay. Compounds are assayed in duplicate at multiple doses to determine minimum inhibitory concentration (MIC) . - 33 -

TIER III

A. ANTIFUNGAL ASSAY (CANDIDA ALBICANS)

The strain Candida albi cans (ATCC #10231) is used. To initiate the exponential phase of yeast growth prior to the assay, a sample of yeast is grown overnight at 25 °C in YM Broth (Difco 0711-17-1) . The yeast cells are collected by centrifugation n for 10 minutes at 3200 rpm, diluted and absorbance read at 595 nm. Yeast diluted in 2x YM Broth (75 μL) are added to the compound mixtures (75μL) for a total volume of 150 μL . The assays are performed in 96-well microplates with approximately 1 x 10⁴ cells per well. The plates are incubated at 25 °C and growth monitored at 48 hours by visual inspection of yeast growth. Amphotericin B anti- fungal control is concurrently tested in each screening assay.

Compounds are assayed in duplicate at multiple doses to determine minimum inhibitory concentration (MIC) .

B. Red Blood Cell Lysis Assay

Compounds are tested for hemolysis of mammalian red blood cells. Horse red blood cells (Colorado Serum Co. #CS0004) are diluted 1:5 in IX phosphate buffered saline (PBS) . 50 μL diluted RBC ' s are added to 50 μL of test compound in IX PBS (total volume = 100 μL) in a round bottom 96-well microplate, mixed gently, and incubated 1 hour at 37 °C. The microplate is then centrifuged for 5 minutes at 1000 rpm. The supernatant is diluted 1:5 (20 μL supernatant + 80 μL IX PBS) into a clean flat bottom 96-well microplate. Absorbance at 540 nm is read using a BioRad model 3550 UV microplate reader. Compounds are tested in duplicate at multiple doses to determine the minimum hemolytic concentration (MHC) . - 34 -

TIER IV

RNA BINDING ASSAY (IN VITRO)

The effect of libraries on tat/TAR interactions

SPA method (scintillation proximity assay) A fast assay targeting tat/TAR interactions was developed for high through-put screening. The assay is used to rapidly identify compounds which are capable of disrupting the interaction of HIV-1 tat protein with the TAR RΝA stem/loop structure. 1. Materials

The C terminal basic binding domain of the tat protein (a 39 residue tat peptide, aa 48-86 of HIV-1 tat protein) was synthesized by a contract lab and further labeled with ¹²⁵I (specific activity 100 μCi/mL) at Amersham Life Sciences.

A 30 base RΝA oligonucleotide (TAR oligonucleotide) consisting of the bulge and stem/loop structure of HIV TAR was synthesized at ISIS Pharmaceuticals and furthere labeled via conjugation with Biotin at the 3' end. A PRB buffer was prepared consisting of: 50 mM Tris- HC1 (pH 8.0), 0.01% ΝP-40, 10% glycerol, 1.5 mM MgCl₂, and 50 mM KC1.

Streptavidin coated SPA beads were purchased from Amersham Life Sciences. Opaque 96 well plates were used purchased. 2. Methods

Streptavidin coated SPA beads are incubated for 20 minutes at room temperature in a PRB buffer with 0.1 μCi of the labeled peptide and 100 nM of the biotin conjugated RNA oligonucleotide. Incubations are performed in the presence or absence of test samples in a volume of 50 μl in an opaque 96 well plate. Following the incubation the plates are spun at 1000 rpm for 5 minutes to settle the - 35 -

SPA beads. The biotintylated TAR oligonucleotide binds the steptavidin coated SPA bead.

The labeled tat peptide associated with the biotintylated TAR oligonucleotide excites the scintillant in the SPA bead, resulting in a quantifiable signal which are read in the TopCount 96 well scintillation counter. Compounds that interfere with the tat/TAR interaction result in ¹²⁵I tat floating free in buffer where excited electrons are quenched before transferring energy to scintillant in the SPA bead. This is observed as a decrease in signal.

PROCEDURE 2

ANTIMICROBIAL MECHANISTIC ASSAY

Bacterial DNA Gyrase DNA gyrase is a bacterial enzyme which can introduce negative supercoils into DNA utilizing the energy derived from ATP hydrolysis. This activity is critical during DNA replication and is a well characterized target for antibiotic inhibition of bacterial growth. In this assay, libraries of compounds are screened for inhibition of DNA gyrase. The assay measures the supercoiling of a relaxed plasmid by DNA gyrase as an electrophoretic shift on an agarose gel.

Initially all library pools are screened for inhibitory activity at 30 μM and then a dose response analysis is effected with active pools. Novobiocin, an antibiotic that binds to the β subunit of DNA gyrase is used as a positive control in the assay. The sensitivity of the DNA gyrase assay was determined by titrating the concentration of the know DNA gyrase inhibitor,

Novobiocin, in the supercoiling assay. The IC₅₀ was determined to be 8 nM, sufficient to identify the activity of a single active species of comparable activity in a - 36 -

library having 30 μM concentration.

PROCEDURE 3

USING LIBRARIES FOR IDENTIFYING METAL CHELATORS AND

IMAGING AGENTS This procedure is used to identify compounds of the invention from libraries of compounds constructed to include a ring that contains an ultraviolet chromophore. Further the chemical functional groups attached to the compounds of the invention are selected from metal binders, coordinating groups such as amine, hydroxyl and carbonyl groups, and other groups having lone pairs of electrons, such that the compounds of the invention can form coordination complexes with heavy metals and imaging agents. The procedure is used to identify compounds of the invention useful for chelating and removing heavy metals from industrial broths, waste stream eluents, heavy metal poisoning of farm animals and other sources of contaminating heavy metals, and for use in identifying imaging agent carriers, such as carriers for technetium 99.

An aliquot of a test solution having the desired ion or imaging agent at a known concentration is added to an aliquot of standard solution of the pool under assay. The UV spectrum of this aliquot is measured and is compared to the UV spectrum of a further aliquot of the same solution lacking the test ion or imaging agent. A shift in the extinction coefficient is indicative of binding of the metal ion or imaging ion to a compound in the library pool being assayed. - 37 -

PROCEDURE 4

ASSAY OF COMBINATORIAL LIBRARY FOR PLA_a INHIBITORS

A preferred target for assay of combinatorially generated pools of compounds is the phospholipase A₂ family. Phospholipases A₂ (PLA₂) are a family of enzymes that hydrolyze the sn-2 ester linkage of membrane phospholipids resulting in release of a free fatty acid and a lysophospholipid (Dennis, E.A., The Enzymes, Vol. 16, pp. 307-353, Boyer, P.D., ed., Academic Press, New York, 1983) . Elevated levels of type II PLA₂ are correlated with a number of human inflammatory diseases. The PLA₂-catalyzed reaction is the rate-limiting step in the release of a number of pro-inflammatory mediators. Arachidonic acid, a fatty acid commonly linked at the sn-2 position, serves as a precursor to leukotrienes, prosta- glandins, lipoxins and thromboxanes . The lysophospholipid are a precursor to platelet-activating factor. PLA₂ is regulated by pro-inflammatory cytokines and, thus, occupies a central position in the inflammatory cascade (Dennis, ibid. ; Glaser et al . , TiPs Revi ews 1992, 14, 92; and Pruzanski et al . , Inflammati on 1992, 1 6, 451). All mammalian tissues evaluated thus far have exhibited PLA₂ activity. At least three different types of PLA₂ are found in humans: pancreatic (type I), synovial fluid (type II) and cytosolic. Studies suggest that additional isoenzymes exist. Type I and type II, the secreted forms of PLA₂, share strong similarity with phospholipases isolated from the venom of snakes. The PLA₂ enzymes are important for normal functions including digestion, cellular membrane remodeling and repair, and in mediation of the inflammatory response. Both cytosolic and type II enzymes are of interest as therapeutic targets. Increased levels of the type II PLA₂ are correlated with a variety of inflammatory disorders - 38 -

including rheumatoid arthritis, osteoarthritis, inflammatory bowel disease and septic shock, suggesting that inhibitors of this enzyme would have therapeutic utility. Additional support for a role of PLA₂ in promoting the pathophysiology observed in certain chronic inflammatory disorders was the observation that injection of type II PLA₂ into the footpad of rats (Vishwanath et al . , Inflamma tion 1988, 12, 549) or into the articular space of rabbits (Bomalaski et al . , J. Immunol . 1991, 1 46, 3904) produced an inflammatory response. When the protein was denatured before injection, no inflammatory response was produced.

The type II PLA₂ enzyme from synovial fluid is a relatively small molecule (about 14 kD) and are distinguished from type I enzymes (e.g. pancreatic) by the sequence and pattern of its disulfide bonds. Both types of enzymes require calcium for activity. The crystal structures of secreted PLA₂ enzymes from venom and pancreatic PLA₂, with and without inhibitors, have been reported (Scott et al . , Sci ence 1990, 250, 1541). Recently, the crystal structure of PLA₂ from human synovial fluid has been determined (Wery et al . , Nature 1991, 352, 79) . The structure clarifies the role of calcium and amino acid residues in catalysis. Calcium acts as a Lewis acid to activate the scissile ester carbonyl bond of 1, 2-diacylglycerophospholipids and binds to the lipid, and a His-Asp side chain diad acts as a general base catalyst to activate a water molecule nucleophile. This is consistent with the absence of any acyl enzyme intermediates, and is also comparable to the catalytic mechanism of serine proteases. The catalytic residues and the calcium ion are at the end of a deep cleft (ca. 14 A) in the enzyme. The walls of this cleft contact the hydrocarbon portion of the phospholipid and - 39 -

are composed of hydrophobic and aromatic residues. The positively-charged amino-terminal helix is situated above the opening of the hydrophobic cleft. Several lines of evidence suggest that the N-terminal portion is the interfacial binding site (Αchari et al . , Cold Spring Harbor Symp . Quant . Bi ol . 1987, 52, 441; Cho et al . , J. Bi ol . Chem . 1988, 263, 11237; Yang et al . , Bi ochem . J. 1989, 262, 855; and Noel et al . , J. Am. Chem . Soc . 1990, 112, 3704) . Much work has been reported in recent years on the study of the mechanism and properties of PLA₂-catalyzed hydrolysis of phospholipids . In in vi tro assays, PLA_: displays a lag phase during which the enzyme adsorbs to the substrate bilayer and a process called interfacial activation occurs. This activation may involve desolvation of the enzyme/lipid interface or a change in the physical state of the lipid around the cleft opening. Evidence favoring this hypothesis comes from studies revealing that rapid changes in PLA₂ activity occur concurrently with changes in the fluorescence of a membrane probe (Burack et al . , Biochemistry 1993, 32, 583). This suggests that lipid rearrangement is occurring during the interfacial activation process. PLA₂ activity is maximal around the melting temperature of the lipid, where regions of gel and liquid-crystalline lipid coexist. This is also consistent with the sensitivity of PLA₂ activity to temperature and to the composition of the substrate, both of which can lead to structurally distinct lipid arrangements separated by a boundary region. Fluorescence microscopy was used to simultaneously identify the physical state of the lipid and the position of the enzyme during catalysis (Grainger et al . , FEBS Lett . 1989, 252, 73) . These studies clearly show that - 40 -

PLA₂ binds exclusively at the boundary region between liquid and solid phase lipid. While the hydrolysis of the secondary ester bond of 1, 2-diacylglycerophospholipids catalyzed by the enzyme is relatively simple, the mechanistic and kinetic picture is clouded by the complexity of the enzyme-substrate interaction. A remarkable characteristic of PLA₂ is that maximal catalytic activity is observed on substrate that is aggregated (i.e. phospholipid above its critical micelle concentration) , while low levels of activity are observed on monomeric substrate. As a result, competitive inhibitors of PLA₂ either have a high affinity for the active site of the enzyme before it binds to the substrate bilayer or partition into the membrane and compete for the active site with the phospholipid substrate. Although a number of inhibitors appear to show promising inhibition of PLA₂ in biochemical assays (Yuan et al . , J. Am . Chem . Soc . 1987, 205, 8071; Lombardo et al . , J. Bi ol . Chem . 1985, 260, 7234; Washburn et al . , J. Biol . Chem . 1991, 266, 5042; Campbell et al . , J. Chem . Soc , Chem . Commun . 1988, 1560; and Davidson et al . , Bi ochem . Bi ophys . Res . Commun . 1986, 131, 587), reports describing in vi vo activity are limited (Miyake et al . , J. Pharmacol . Exp . Ther . 1992, 263, 1302) . In one preferred embodiment, compounds of the invention are selected for their potential to interact with, and preferably inhibit, the enzyme PLA₂. Thus, compounds of the invention are used for topical and/or systemic treatment of inflammatory diseases including atopic dermatitis and inflammatory bowel disease. In selecting the functional groups, advantage can be taken of PLA₂ ' s preference for anionic vesicles over zwitterionic vesicles. Preferred compounds of the invention for assay - 41 -

for PLA₂ include those having aromatic diversity groups to facilitate binding to the cleft of the PLA₂ enzyme (Oinuma et al . , J. Med. Chem. 1991, 34, 2260; Marki et al . , Agents Acti ons 1993, 38, 202; and Tanaka et al . , J. Antibi oti cs 1992, 45, 1071) . Benzyl and 4-hexylbenzyl groups are preferred aromatic diversity groups. PLA₂-directed compounds of the invention can further include hydrophobic functional groups such as tetraethylene glycol groups . Since the PLA₂ enzyme has a hydrophobic channel, hydrophobicity is believed to be an important property of inhibitors of the enzyme.

After each round of synthesis as described in the above examples, the resulting libraries or pools of compounds are screened for inhibition of human type II PLA₂ enzymatic activity. The assay is effected at the conclusion of each round of synthesis to identify the wining pool from that round of synthesis. Concurrently, the libraries additionally can be screened in other in vi tro assays to determine further mechanisms of inhibition.

The pools of the libraries are screened for inhibition of PLA₂ in the assay using E. coli labeled with ³H-oleic acid (Franson et al . , J. Lipid Res . 1974, 15, 380; and Davidson et al . , J. Biol . Chem . 1987, 262, 1698) as the substrate. Type II PLA₂ (originally isolated from synovial fluid) , expressed in a baculovirus system and partially purified, serves as a source of the enzyme. A series of dilutions of each of the library pools is done in water: 10 μl of each pool is incubated for 5 minutes at room temperature with a mixture of 10 μl PLA₂, 20 μl 5X PLA₂ Buffer (500 mM Tris 7.0-7.5, 5 mM CaCl₂) , and 50 μl water. Samples of each pool are run in duplicate. At this point, 10 μl of ³H E. coli cells is added. This - 42 -

mixture is incubated at 37 °C for 15 minutes. The enzymatic reaction is stopped with the addition of 50 μL 2M HCl and 50 μL fatty-acid-free BSA (20 mg/mL PBS), vortexed for 5 seconds, and centrifuged at high speed for 5 minutes. 165 μL of each supernate is then put into a scintillation vial containing 6 ml of scintillant (ScintiVerse) and cp s are measured in a Beckman Liquid Scintillation Counter. As a control, a reaction without the combinatorial pool is run alongside the other reactions as well as a baseline reaction containing no compounds of the invention as well as no PLA₂ enzyme. CPMs are corrected for by subtracting the baseline from each reaction data point.

Confirmation of the "winners" is made to confirm that a compound of the invention binds to enzyme rather than substrate and that the inhibition by a compound of the invention that is selected is specific for type II PLA₂. An assay using ¹⁴C-phosphatidyl ethanolamine (¹⁴C-PE) as substrate, rather than E. coli membrane, is used to insure enzyme rather than substrate specificity. Micelles of ¹⁴C-PE and deoxycholate are incubated with the enzyme and a compound of the invention. ¹⁴C-labeled arachidonic acid released as a result of PLA₂-catalyzed hydrolysis is separated from substrate by thin layer chromatography and the radioactive product is quantitated. The "winner" is compared to phosphatidyl ethanolamine, the preferred substrate of human type II PLA₂, to confirm its activity. PLA₂ from other sources (snake venom, pancreatic, bee venom) and phospholipase C, phospholipase D and lysophospholipase can be used to further confirm that the inhibition is specific for human type II PLA₂. - 43 -

PROCEDURE 5

PROBES FOR THE DETECTION OF SPECIFIC PROTEINS AND MRNA IN

BIOLOGICAL SAMPLES

For the reliable, rapid, simultaneous quantification of multiple varieties of proteins or mRNA in a biological sample without the need to purify the protein or mRNA from other cellular components, a protein or mRNA of interest from a suitable biological sample, i.e., a blood borne virus, a bacterial pathogen product in stool, urine and other like biological samples, is identified using standard microbiological techniques. A probe comprising a compound of a combinatorial library of the invention is identified by a combinatorial search as noted in the above examples. Preferred for the protein probe are compounds synthesized to include chemical functional groups that act as hydrogen bond donors and acceptors, sulfhydryl groups, hydrophobic lipophilic moieties capable of hydrophobic interactions groups and groups capable of ionic interactions . The probe is immobilized on insoluble CPG solid support utilizing the procedure of Pon, R.T., Protocols for Oligonucleotides and Analogs, Agrawal, S., Ed., Humana Press, Totowa, NJ, 1993, p 465-496. A known aliquot of the biological sample under investigation is incubated with the insoluble CPG support having the probe thereon for a time sufficient to hybridize the protein or mRNA to the probe and thus form a linkage via the probe to the solid support. This immobilizes the protein or mRNA present in the sample to the CPG support. Other non- immobilized materials and components are then washed off the CPG with a wash media suitable for use with the biological sample. The mRNA on the support is labeled with ethidium bromide, biotin or a commercial radionucleotide and the amount of label immobilized on the CPG support is - 44 -

measured to indicate the amount of mRNA present in the biological sample. In a similar assay a protein is also labeled and quantified.

PROCEDURE 6 LEUKOTRIENE B₄ ASSAY

Leukotriene B₄ (LTB₄) has been implicated in a variety of human inflammatory diseases, and its pharmacological effects are mediated via its interaction with specific surface cell receptors. Library pools are screened for competitive inhibition of radiolabeled LTB₄ binding to a receptor preparation.

A Nenquest™ Drug Discovery System Kit (NEN Research Products, Boston, MA) is used to select an inhibitor of the interaction of Leukotriene B₄ (LTB₄) with receptors on a preparation of guinea pig spleen membrane. [³H] Leukotriene B₄ reagent is prepared by adding 5 mL of ligand diluent (phosphate buffer containing NaCl, MgCl₂, EDTA and Bacitracin, pH 7.2) to 0.25 mL of the radioligand. The receptor preparation is made by thawing the concentrate, adding 35 mL of ligand diluent and swirling gently in order to re-suspend the receptor homogeneously. Reagents are kept on ice during the course of the experiment, and the remaining portions are stored at -20°C. Library pools prepared as per general procedure of examples above are diluted to 5 μM, 50 μM and 500 μM in phosphate buffer (lx PBS, 0.1% azide and 0.1% BSA, pH 7.2), yielding final test concentrations of 0.5 μM, 5 μM and 50 μM, respectively. Samples are assayed in duplicate. [³H] LTB₄ (25 μL) is added to 25 μL of either appropriately diluted standard (unlabeled LTB₄) or library pool. The receptor suspension (0.2 L) is added to each tube. Samples are incubated at 4°C for 2 hours. Controls - 45 -

include [³H] LTB₄ without receptor suspension (total count vials), and sample of ligand and receptor without library molecules (standard) .

After the incubation period, the samples are filtered through GF/B paper that had been previously rinsed with cold saline. The contents of each tube are aspirated onto the filter paper to remove unbound ligand from the membrane preparation, and the tubes washed (2 x 4 mL) with cold saline. The filter paper is removed from the filtration unit and the filter disks are placed in appropriate vials for scintillation counting. Fluor is added, and the vials shaken and allowed to stand at room temperature for 2 to 3 hours prior to counting. The counts/minute (cpm) obtained for each sample are subtracted from those obtained from the total counts to determine the net cpm for each sample. The degree of inhibition of binding for each library pool is determined relative to the standard (sample of ligand and receptor without library molecules) .

PROCEDURE 7

COUPLED TRANSCRIPTION/TRANSLATION ASSAY

E. coli S30 coupled transcription/translation extract was prepared as described by Burgess (Lesley, S. A., Brow, M. D., and Burgess, R. R., (1991), Journal of Biologi cal Chemi stry 266, 2632-2680) . A 10X complete amino acid mix was prepared at 1 mM each amino acid. Plasmid DNA template (pBestLuc) containing the reporter gene for luciferase was purified on a Qiagen Ultrapure 100 column. S30 Premix without amino acids was special ordered from Promega (non-catalog item) . The assay is performed in a total volume of 35 μL by combining 13 μL premix, 4 μL 10X amino acids, 5 μL S30 extract, 5 μL test compound and 8 μL (1 μg) pBestLuc template DNA in black 96-well microplates. - 46 -

The plates are mixed well and incubated at 37 °C for 35 minutes. The luciferase reporter enzyme is detected with Packard LucLite substrate (Catalog # 6016911) and luminescence is measured with a Packard TopCount . The percentage of transcription/translation inhibition is determined relative to control wells which do not contain test compound. Compounds are assayed in duplicate at a single dose. Compounds which show inhibitory activity are re-tested in duplicate at multiple doses to determine the IC₅₀.

PROCEDURE 8

TRANSCRIPTION TERMINATION FACTOR RHO

The transcription termination factor Rho is a excellent target for antibacterial chemotherapy for three reasons. First, the protein performs a function essential for bacterial growth (Das, A., et al . , Proc . Na tl . Acad. Sci . USA, 1976, 13, 1959-1963) . Second, Rho is present in divergent bacteria (Opperman, T., Richardson, J.P., J. Bacteri ol . 1994, 1 16, 5033-5043; Richardson, J.P., J. Biol . Chem . , 1996, 211 , 1251-1254; Ingham, C.J., et al . , J. Bi ol . Chem . , 1996, 211 , 21803-21807) and thus a compound that inhibits Rho may be useful as a broad- spectrum antibacterial. Finally, no mammalian homologue of Rho has been found. Thus, compounds that specifically inhibit Rho-induced transcription termination most likely would not be toxic to mammalian cells. In support of these assumptions, Rho has been identified as the site of action of a natural product, bicyclomycin, that is used in veterinary medicine (Miyoshi, T., et al . , Nippon Nogei Kagaku Kaishi , 1988, 62, 1319-1324) .

In bacteria, transcription termination is important in regulation of gene expression and is likely necessary - 47 -

for formation of the 3' ends of most messenger RNA (Richardson, J.P., Cri t . Rev. Bi ochem . Mol . Bi ol . 1993, 28, 1-30) . Termination is a result of either intrinsic termination or factor-dependent dissociation of the polymerase from the DNA template. Intrinsic termination results from spontaneous dissociation of the polymerase at specific RNA sequence elements. Factor-dependent termination provides a mechanism for dissociation of polymerase from transcripts lacking sequence elements of an intrinsic terminator. Rho factor is essential for termination at approximately half of the messages in the E. coli genome (Platt, T., Richardson, J.P., Transcripti onal Regula tion, 1992, S.L. McKnight & K.R. Yamamoto, eds, pp. 365-388, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY) .

A poly r (C) -dependent ATP hydrolysis assay (Ingham et al . , 1996) was selected to search combinatorial libraries for inhibitors of Rho. The assay allows detection of inhibitors of binding of RNA to either the tight or weak binding sites, or inhibitors of the conformational change required for ATP hydrolysis, or the hydrolysis reaction itself. The colorimetric assay detects pyrophosphate released upon hydrolysis of ATP via complex formation with malachite green dye. Protein for assay validation was provided by

Professor Terry Platt (University of Rochester) . The assay is performed as follows with final concentrations given in parentheses: ATP (200 μM) and poly r(C) (20 μg/mL) are added to a microtiter well containing compound at 10-fold the screening concentration; Rho (4 μg/mL) and buffer are added to start the reaction. The buffer for the assay is 10 mM MgCl₂, 40 mM Tris (pH 7.5), 50 mM KC1, 0.1 mM DTT and 10% glycerol. The final volume is 40 μL . The reaction is stopped after 30 min. at room temperature -48 -

by addition of a malachite green/ammonium molybdate solution. Absorbance is read on a microtiter plate reader (BioRad) at a wavelength of 655 nm to quantitate pyrophosphate release.

Claims

- 49 -WHAT IS CLAIMED IS:

1. A method for deconvoluting a chemical mixture comprising: a. subjecting said mixture to high pressure liquid chromatography to yield a first array of chemical subgroups; b. assaying at least some of the subgroups of the first array for biological activity in a preselected assay; c. performing high pressure liquid chromatography on each of said subgroups of the first array which have exhibited biological activity to yield, for each of said first array subgroups, a second array of chemical subgroups; d. assaying at least some of the second array subgroups for biological activity in said preselected assay; and e. identifying the chemical species present in the second array subgroups having at least a preselected level of biological activity in said assay.

2. The method of claim 1 wherein the liquid chromatography of steps "a" and "c" are performed on different stationary phases.

3. The method of claim 1 wherein the liquid chromatography of steps "a" and "c" are performed using different eluting solvents.

4. The method of claim 1 wherein at least one liquid chromatography step employs a reverse phase gel stationary phase.

5. The method of claim 1 further comprising at least one additional chromatographic separation of at least some of the chemical subgroups of an array prior to the identification step. - 50 -

6. The method of claim 5 wherein said additional chromatographic separation comprises column, gel, electrophoretic or gas chromatography.

7. The method of claim 5 wherein said additional chromatography is performed prior to step "a".

8. The method of claim 5 wherein said additional chromatography is performed prior to step "c".

9. The method of claim 5 wherein said additional chromatography is accompanied by an assay of the array of chemical compounds deriving therefrom.

10. The method of claim 1 wherein said assay comprises a plurality of assays.

11. The method of claim 1 wherein said assay comprises a pharmaceutical, agricultural chemical or veterinary screen.

12. The method of claim 1 wherein the identification of the chemical species includes mass spectrometry.

13. The method of claim 1 wherein the chemical mixture is formed through combinatorial chemistry.

14. The method of claim 1 wherein the members of the chemical mixture share at least one common chemical functional group.

15. The method of claim 1 wherein the chemical mixture is formed through the solution phase addition of reactive functional groups to a scaffold molecule having at least one reaction site reactive with said functional groups .

16. The method of claim 1 wherein the chemical mixture comprises from 20 to 50 chemical compounds.

17. The method of claim 1 wherein the chemical mixture comprises from 50 to 100 chemical compounds.

18. The method of claim 1 wherein the chemical mixture comprises from 100 to 500 chemical compounds.

19. The method of claim 1 wherein the chemical mixture - 51 -

comprises from 500 to 1000 chemical compounds.

20. The method of claim 1 wherein a majority of the chemical compounds forming the chemical mixture share a common functional group.

21. A method for identifying a chemical compound contained within a mixture of compounds and having activity in a preselected biological assay comprising: a. subjecting said mixture to at least two stages of chromatography, at least one of the stages being high pressure liquid chromatography to yield a first array of chemical subgroups; b. assaying at least some of the subgroups of the first array for biological activity in a preselected assay; c. performing high pressure liquid chromatography on each of said subgroups of the first array which have exhibited biological activity to yield, for each of said first array subgroups, a second array of chemical subgroups; d. assaying at least some of the second array subgroups for biological activity in said preselected assay; and e. identifying the chemical species present in the second array subgroups having at least a preselected level of biological activity in said assay.

22. A method for identifying a pharmacophore active in a preselected biological assay from a mixture of compounds of indeterminate structure comprising performing at least two stages of high pressure liquid chromatography upon said mixture, each stage yielding an array of chemical compounds; assaying each of said array in said biological assay; determining compounds having activity in said assay; -52-

and determining the structure of the active compounds .