Solution-Phase Parallel Synthesis of a Library of Δ²-Pyrazolines

The solution-phase parallel synthesis of primary, or focused libraries is an established practice in medicinal chemistry.¹ The solution-phase approach avoids the need to re-optimize the chemistry to the solid-phase prior to library generation. Δ²-Pyrazolines² (2-pyrazolines or 4,5-dihydropyrazoles) are an important class of heterocyclic small molecules³ that have shown potential bioactivity in numerous screening tests.⁴ For example, pyrazolines 1 have demonstrated moderate to good MIC₉₀ values against Helicobacter pylori.⁵ The optimized pyrazoline 2 showed nanomolar inhibition (IC₅₀ = 26 nM) against kinesin spindle protein (KSP); inhibitors of this protein constitute a novel approach to cancer treatment.⁶ Pyrazoline 3 displayed 70% inhibitory activity against neuronal nitric oxide synthase (nNOS) and was inactive against kynurenine 3-hydroxylase.⁷ Such selective inhibition is indicative of potential neuroprotective properties. Compounds containing the pyrazoline core have also been examined for antidepressant activity through screening against monoamine oxidases (MAOs),⁸ treatment of obesity as CB₁ antagonists,⁹ antiviral activity against West Nile virus,¹⁰ and multidrug resistance (MDR) modulators in tumor cells.¹¹

Most of these studies were limited to structures lacking substitution at N1 and C4; additionally, only a few modifications at C5 have been reported. Such constraints have been predominantly due to the similarity of existing synthetic strategies to the dihydropyrazole framework, namely, the cyclocondensation of acrylates or chalcones with hydrazines.² Carreira and coworkers have developed a different strategy, the dipolar cycloaddition of TMSCHN₂ to enoates generating 2–pyrazolines lacking substitution at N1 and C3.¹² It must be noted that these approaches are advantageous when specific modification at N1 is desired.

Recently, one of us reported the enantioselective [3+2] dipolar cycloaddition of nitrilimines to α,β-unsaturated enoates 4 under chiral Lewis acid catalysis for the construction of 2–pyrazolines 8 with high yields and high enantioselectivities (Scheme 1).¹³ The E geometry of the dipolarophile is translated to trans stereochemistry at C4 and C5 positions of 2-pyrazolines 8. This study suggested that the method could rapidly supply a diverse set of 2-pyrazolines. Specifically, the development of this methodology to parallel synthesis would clearly provide access to tetrasubstituted pyrazolines and allow for their primary high-throughput biological screening.¹⁴ Additionally, there is only one report of a solution-phase parallel synthesis of 2–pyrazolines and the authors utilized the cyclocondensation approach.¹⁵

Scheme 1.

Enantioselective Nitrilimine Cycloaddition

Hence, it was pertinent to evaluate the accessibility of tetrasubstituted 2-pyrazolines through solution-phase parallel synthesis employing the nitrilimine cycloaddition. At the outset, the preparation of racemic compounds was targeted to avoid influencing the biological screening process toward either enantiomer. The demonstration of this methodology in parallel synthesis is documented here through the synthesis of 80 compounds containing the 2-pyrazoline moiety.

Results and Discussion

The strategy for library production is shown in Scheme 2. The libraries were designed to contain both alkyl and aryl substituents at the β-carbon of the alkenoyl oxazolidinones 12 and varying aryl groups on the hydrazonyl halides 13, providing two diversity elements. In addition, a third diversification opportunity presents itself in the preparation of the hydrazonyl halides using N-chlorosuccinimide or N-bromosuccinimide.¹⁶ Thus, chlorination provided hydrazonyl chloride 13a whereas bromination with NBS provides 13b, containing bromine at the para position of the electron-rich aromatic ring. The libraries with structure 14 could then be reduced to library 17, the hydroxyl group of which could be further diversified.

Scheme 2.

Library Design for 2-Pyrazolines

The initial cycloaddition experiments were carried out in the absence of the chiral ligand to obtain the racemic cycloadducts (Scheme 3). In addition to the expected product (20C-isomer), the regioisomeric cycloadduct (20N-isomer) was also obtained.¹⁷ This unexpected result further diversifies the number of library members in this endeavor.¹⁸ Such N-regioisomers were not observed in the enantioselective reactions, leading to the speculation that regiocontrol is provided by complexation to the chiral Lewis acid. Attempts to control the regioselectivity through activation with main-group, transition-metal, or lanthanide Lewis acids [Mg(ClO₄)₂, Mg(NTf₂)₂, TiCl₄, SnCl₄, or Yb(OTf)₃] or altering the amine base provided products with decreased regioselection. Additionally, the regioisomers were inseparable through either normal or reverse phase HPLC. However, an increase in regioisomeric ratio could be obtained through a decrease in temperature to −78 °C.

Scheme 3.

Synthetic Route to 2-Pyrazoline Libraries

With these preliminary results, we proceeded to the library synthesis as described in Scheme 3. The dipolarophiles and dipole precursors were chosen to demonstrate the scope of the methodology (Figure 1). Both alkyl and aryl (with both electron-donating and electron-withdrawing substituents) groups at the β-carbon of 18 were chosen. Similar considerations were applied in the choice of hydrazonyl halides. The dipolarophiles 18{1-7} were prepared using literature procedures starting from either the commercial alkenoyl acid chlorides or alkenoic acids. The hydrazone precursors to the dipoles were obtained through a simple condensation of the aldehydes and phenyl hydrazine using magnesium sulfate or 10 mol% magnesium perchlorate.¹⁹ These hydrazones were then converted to the hydrazonyl chlorides 19{1-4} by treatment with N-chlorosuccinimide and Me₂S and to hydrazonyl bromides 19{5-8} with N-bromosuccinimide and Me₂S.¹⁶

Figure 1.

Library components 18{1-7} and 19{1-8}.

The dipolar cycloaddition between 18{1-7} and 19{1-8} to provide 56 compounds was carried out in two runs with either an Innovasyn SynthArray-24 reactor or 6×4 Bohdan MiniBlock™ XT. The MiniBlock™ reaction system was cooled to −78 °C prior to the addition of Et₃N. The 56 library members were easily purified by filtration through a SPE cartridge containing silica gel. These crude products were analyzed by LC-MS followed by purification by preparative LC with UV trigger to provide 20{1-28} and 21{1-28} as mixtures of the C and N regioisomers. The regioisomeric ratios were determined from the purified products before cataloging them for biological screening. The isolated yields, purities determined by UV (215 nm), MS (ELSD), and C:N regioisomeric ratios are collected for these 56 samples (Tables 1 and 2). For the dipolarophiles containing alkyl substituents at the β–carbon, the C-regioisomer was predominantly observed. In comparison to 18{3}, electron-withdrawing groups on the aryl ring decreased the regioselectivity. However, 18{7} containing the electron-donating 2-OMe group, provided a higher ratio of C-regioisomer. In general, the reactions with hydrazonyl chlorides provided slightly better regioselectivities compared to the hydrazonyl bromides.

Table 1.

Library data for compounds 20{1-28} from 18{1-7} and 19{1-4}^a

compound	R1	Ar1	yield (%)b	purity (%)c	purity (%)d	C:N ratioe
20{1}	Me	Ph	46	100	100	>30:1
20{2}	Et	Ph	56	100	98	30:1
20{3}	Ph	Ph	49	100	100	6.2:1
20{4}	4-FC6H4	Ph	46	100	100	4.8:1
20{5}	4-ClC6H4	Ph	59	99	100	5.2:1
20{6}	4-NO2C6H4	Ph	47	96	100	2.9:1
20{7}	2-OMeC6H4	Ph	51	100	92	5.6:1
20{8}	Me	3-MeC6H4	73	98	100	27:1
20{9}	Et	3-MeC6H4	44	100	100	32:1
20{10}	Ph	3-MeC6H4	51	99	100	5.5:1
20{11}	4-FC6H4	3-MeC6H4	48	100	100	4.4:1
20{12}	4-ClC6H4	3-MeC6H4	44	100	100	4.6:1
20{13}	4-NO2C6H4	3-MeC6H4	47	99	100	2.6:1
20{14}	2-OMeC6H4	3-MeC6H4	48	100	93	5.7:1
20{15}	Me	3-ClC6H4	85	97	100	23:1
20{16}	Et	3-ClC6H4	37	99	100	34:1
20{17}	Ph	3-ClC6H4	32	99	100	8.4:1
20{18}	4-FC6H4	3-ClC6H4	31	100	100	6.2:1
20{19}	4-ClC6H4	3-ClC6H4	36	100	100	8.0:1
20{20}	4-NO2C6H4	3-ClC6H4	42	100	100	4.3:1
20{21}	2-OMeC6H4	3-ClC6H4	46	100	97	9.1:1
20{22}	Me	4-FC6H4	76	98	100	22:1
20{23}	Et	4-FC6H4	52	100	100	32:1
20{24}	Ph	4-FC6H4	53	99	100	8.5:1
20{25}	4-FC6H4	4-FC6H4	51	100	100	6.5:1
20{26}	4-ClC6H4	4-FC6H4	60	99	100	7.1:1
20{27}	4-NO2C6H4	4-FC6H4	69	100	100	3.7:1
20{28}	2-OMeC6H4	4-FC6H4	61	100	96	5.8:1

^aSee supporting information for experimental details.

^bIsolated yields after purification on LC.

^cUV purity determined at 215nm.

^dELSD purity.

^eEstimated from ¹H NMR.

Table 2.

Library data for compounds 21{1-28} from 18{1-7} and 19{5-8}^a

compound	R1	Ar1	yield (%)b	purity (%)c	purity (%)d	C:N ratioe
21{1}	Me	Ph	58	100	97	>30:1
21{2}	Et	Ph	57	100	100	30:1
21{3}	Ph	Ph	51	92	94	5:1
21{4}	4-FC6H4	Ph	31	86	91	3.6:1
21{5}	4-ClC6H4	Ph	38	96	99	4.5:1
21{6}	4-NO2C6H4	Ph	49	100	83	3.0:1
21{7}	2-OMeC6H4	Ph	44	99	98	5.4:1
21{8}	Me	3-MeC6H4	37	96	100	>30:1
21{9}	Et	3-MeC6H4	21	97	100	30:1
21{10}	Ph	3-MeC6H4	46	96	91	4.8:1
21{11}	4-FC6H4	3-MeC6H4	36	92	85	3.3:1
21{12}	4-ClC6H4	3-MeC6H4	49	97	92	3.9:1
21{13}	4-NO2C6H4	3-MeC6H4	59	100	71	2.5:1
21{14}	2-OMeC6H4	3-MeC6H4	49	100	99	6.6:1
21{15}	Me	3-ClC6H4	94	98	100	>30:1
21{16}	Et	3-ClC6H4	31	100	100	30:1
21{17}	Ph	3-ClC6H4	32	100	100	4.6:1
21{18}	4-FC6H4	3-ClC6H4	53	100	100	4.4:1
21{19}	4-ClC6H4	3-ClC6H4	54	100	100	4.3:1
21{20}	4-NO2C6H4	3-ClC6H4	61	100	100	3.3:1
21{21}	2-OMeC6H4	3-ClC6H4	39	100	100	12.1:1
21{22}	Me	4-FC6H4	72	99	100	>30:1
21{23}	Et	4-FC6H4	57	100	100	30:1
21{24}	Ph	4-FC6H4	53	86	96	6.5:1
21{25}	4-FC6H4	4-FC6H4	52	86	93	4.8:1
21{26}	4-ClC6H4	4-FC6H4	47	94	99	5.2:1
21{27}	4-NO2C6H4	4-FC6H4	53	99	100	4.3:1
21{28}	2-OMeC6H4	4-FC6H4	15	100	100	3.9:1

^aSee supporting information for experimental details.

^bIsolated yields after purification on LC.

^cUV purity determined at 215nm.

^dELSD purity.

^eEstimated from ¹H NMR.

The next library, chemset 22, was obtained from a subset of the library 20 through a reduction of the oxazolidinone group using NaBH₄ as shown in Scheme 3. This parallel synthesis was also performed in a 6×4 Bohdan MiniBlock™ XT followed by parallel purification using aqueous work-up to remove inorganic salts. The aqueous work-up was performed using Alltech^® phase separator columns containing hydrophobic frits followed by drying through polypropylene tubes containing MgSO₄. The crude products were analyzed by LC-MS and subjected to purification by preparative LC with UV triggered separation. The isolated yields, purities (UV at 215 nm and ELSD), and C:N ratios of the compounds over two steps are presented in Table 3. In general, the C:N ratios in this library were slightly higher than the previous libraries due to enrichment obtained through the fraction-trigger during preparative HPLC purification.

Table 3.

Library data for compounds 22{1-24} from 18{1-5,7} and 19{1-4}^a

compound	R1	Ar1	yield (%)b	purity (%)c	purity (%)d	C:N ratioe
22{1}	Me	Ph	35	99	100	30:1
22{2}	Et	Ph	21	99	100	30:1
22{3}	Ph	Ph	55	98	100	6.2:1
22{4}	4-FC6H4	Ph	72	100	100	5.1:1
22{5}	4-ClC6H4	Ph	49	98	100	5.0:1
22{6}	2-OMeC6H4	Ph	77	93	100	6.7:1
22{7}	Me	3-MeC6H4	32	93	99	30:1
22{8}	Et	3-MeC6H4	24	99	100	30:1
22{9}	Ph	3-MeC6H4	72	99	100	5.5:1
22{10}	4-FC6H4	3-MeC6H4	59	99	100	4.1:1
22{11}	4-ClC6H4	3-MeC6H4	71	99	100	4.4:1
22{12}	2-OMeC6H4	3-MeC6H4	54	97	100	5.5:1
22{13}	Me	3-ClC6H4	27	99	100	30:1
22{14}	Et	3-ClC6H4	27	100	100	30:1
22{15}	Ph	3-ClC6H4	48	100	100	8.0:1
22{16}	4-FC6H4	3-ClC6H4	50	98	100	6.0:1
22{17}	4-ClC6H4	3-ClC6H4	59	100	100	6.3:1
22{18}	2-OMeC6H4	3-ClC6H4	54	96	100	10.1:1
22{19}	Me	4-FC6H4	80	99	100	30:1
22{20}	Et	4-FC6H4	48	96	100	30:1
22{21}	Ph	4-FC6H4	40	98	100	8.4:1
22{22}	4-FC6H4	4-FC6H4	30	99	100	7.0:1
22{23}	4-ClC6H4	4-FC6H4	56	100	100	7.5:1
22{24}	2-OMeC6H4	4-FC6H4	44	100	100	8.6:1

^aSee supporting information for experimental details.

^bIsolated yields after purification on LC.

^cUV purity determined at 215nm.

^dELSD purity.

^eEstimated from ¹H NMR.

The compounds prepared in these parallel syntheses were evaluated in silico for their drug-likeness based on the Lipinski's “rule of five”.²⁰ Molecular weight, clogP, number of hydrogen bond donors, and acceptors were calculated using SYBYL²¹ and are presented in Figure 2. Overall, 24% of the library had 2 violations, 66% of the library had 1 violation, and 10% of the library had 0 violations. Many of the compounds in the collection had high clogP values (<5 is normal), which would be addressed in future libraries directed towards orally active agents. In addition, standard absorption-distribution-metabolism-excretion (ADME) properties were calculated using the VolSurf²² program and are presented in the Supporting Information. Chemical diversity analysis relative to the PubChem collection (as performed via DiverseSolutions²³) suggests that the present collection occupies regions of PubChem space that already have substantial populations. However, the library members do not directly duplicate compounds in PubChem and will thus be a unique addition to the database and compound collection.

Figure 2.

Analysis of Lipinski rule parameters.

In summary, the efficient parallel synthesis of a library of eighty tetrasubstituted 2-pyrazolines containing either oxazolidinone or hydroxymethyl groups at C5 have been described. Further diversification of the current library compounds will be explored. The compounds will be evaluated in high-throughput screens and modified accordingly.

Supplementary Material

1si20060821_03. Supporting Information Available.

Experimental procedures for library synthesis, characterization data for new compounds, LC traces, ¹H/¹³C NMR for library members, and ADME property table. See any current masthead page for ordering information and Web access instructions.