Acridizinium salts. Preparation from 1-(benzylic)-2-formyl and 1-(benzylic)-2-acetyl pyridinium bromides and ring-openings reactions with nucleophilic reagents

Treatment of pyridine-2-carboxaldehyde or 2-acetyl pyridine with several benzylic bromides led to the corresponding 1-(benzylic)-2-formyl pyridinium bromides or 1-(benzylic)-2-acetyl pyridinium bromides, respectively. The use of MeOH as a solvent in the quaternizations of pyridine-2-carboxaldehyde with benzylic bromides ( or in one case benzyl chloride) led to the corresponding pyridinium hemiacetals. These salts were effectively cyclodehydrated to afford the corresponding acridizinium salts. The ring opening reactions of several of the acridizinium salts with hydroxide, oximes, primary amines and a hydrazine have been investigated.


Introduction
A recent review 1 summarizes the status of the synthesis of benzo[b]quinolinium salts (acridizinium salts, Chart 1) and several previous reviews discuss the synthesis and reactions of these salts. 2,3,4It was reported in 1955 5 that treatment of pyridine 2-carboxaldehyde with benzyl bromide or 4-methyl benzyl bromide for 2-3 weeks at room temperature yielded dark red glassylike materials containing the corresponding pyridinium bromide salts.These crude glasses on refluxing with 48% aqueous HBr underwent cyclodehydrations to afford 60% and 39% (overall) yields of the corresponding acridizinium bromides (Chart 1).A similar glassy product was reported on treatment of α-bromomethyl naphthalene with pyridine 2-carboxaldehyde which on refluxing with 48% aqueous HBr led to the substituted acridizinium bromide (52% overall). 6lthough it was initially reported that pyridine 2-carboxaldehyde with 3,4-dimethoxybenzyl bromide in DMF led to a crude glassy red product 7 , a more recent paper describes the isolation of an uncharacterized solid which on cyclization with 48% aqueous HBr yielded the substituted acridizinium salt. 8Br The use of methanol as a solvent for the quaternization of pyridine 2-carboxaldehyde with several methoxy substituted benzylic bromides led to crude uncharacterized products which could be cyclized to their respective acridizinium salts with 48% aqueous HBr. 9 A prior study has reported that treatment of 2-acetylpyridine with benzyl bromide (DMF, 10 o C, 23 days) led to a small amount of an oil which on cyclization with liquid HF led to 11-methylacridizinium (isolated as the perchlorate salt). 10he quaternizations of 2-(1,3-dioxolan-2-yl)pyridine [11][12][13][14][15][16] were then reported to be much superior in reactions with benzylic bromides and the respective pyridinium salts could be converted into the acridizinium salts on acidic cyclodehydrations.
Our prior report documented that pyridine-2-carboxaldehyde with α,α'-dibromo-o-or pxylenes led to the corresponding bis-pyridinium aldehydes which could be cyclized into the corresponding bis-acridizinium salts. 17he goal of the present research was to clearly establish the structures of the products formed from the quaternizations of pyridine 2-carboxaldehydes and 2-acetyl pyridine with benzylic halides and the evaluation of the use of methanol as a solvent in these reactions.These substituted pyridinium salts would be converted into the respective acridizinium salts on acidic cyclodehydrations This methodology would obviate the necessity for the use of pyridine acetals or oximes for the preparation of acridizinium salts.In addition, we wish to report on the ringopening reactions of several of the acridizinium salts with aqueous NaOH solutions, primary amines, oximes and hydrazines.

Results and Discussion
Formation of the pyridinium salts Treatment of pyridine-2-carboxaldehyde (1) with benzyl bromide (2a), 4-methyl benzyl bromide (2b) or 2,5-dimethoxy benzyl bromide (2c) (1:1 molar ratios) in DMF as solvent at 40-45 o C for 20-24 h led to the corresponding pyridinium salts 3a, 3b and 3c, respectively .These salts crystallized from the reaction mixture and were readily isolated by filtration in 50-57% yields.Although sensitive to moisture, they could be stored in the freezer for long periods without any perceptible decomposition.The isolation of the free aldehyde (or hydrate) is also dependent on the humidity during the filtration of the solids as under high humidity conditions the hydrates are the predominant product.In Vermont the winter months, when the humidity is low, were clearly the best times to run these reactions.
In the 1 H NMR spectra (DMSO-d 6 as solvent), the CHO proton of the aldehydes are found at δ 10.3 and in the 13 C nmr spectra the carbon appears at δ 183.The facile hydrations of the aldehydes are easily detected by a comparison of the 1 H and 13 C nmr spectra in the absence and presence of D 2 O, respectively, which in the former indicates the disappearance of the signal at δ 10.3 and in the latter the disappearance of the signal at δ 183.
The expected reactivity of the aldehyde group is demonstrated by treatment of a suspension of crystalline 3a in ethyl acetate followed by addition of MeOH to effect solution.On allowing this solution to stand in the freezer for several days, the resultant crystals were collected by filtration to afford the hemiacetal 4a.
It was then observed that the quaternizations of 1 with 2a or 2b in refluxing MeOH as solvent directly led to the hemiacetals salts 4a (67%) and 4b (30%), respectively.Treatment of 1 with benzyl chloride in refluxing methanol led to hemiacetal chloride salt 4c (30%).
Attempts to prepare 5 by treatment of pyridine 2-carboxaldehyde (1) with αbromoethylbenzene in DMF were unsuccessful.Styrene is rapidly formed by β-elimination from the α-bromoethylbenzene.The preparation of 6 was accomplished by treatment of the corresponding acetal of 2-acetyl pyridine with α-bromoethylbenzene following a modified literature procedure. 12e next turned our attention to the reactions of 2-acetyl pyridine (7) with benzylic bromides.Treatment of 2-acetyl pyridine (7) with an equimolar amount of benzyl bromide (2a) at 60 o C in DMF for 48 h led to 1-(benzyl)-2-acetyl pyridinium bromide (8a) in a poor yield (13%).An attempt was made to perform this quaternization in 2-propanol as solvent.After 65 h at room temperature and 45 o C for 48 h, the only identifiable material ( 1 H NMR) was the hydrobromide salt of 2-acetyl pyridine.Clearly the rate of reaction of benzyl bromide with 2propanol (formation of HBr) is competitive with the anticipated quaternization process.

Acridizinium salts
The cyclodehydrations of the pyridinium salts 3a-c, 6 and 8a,b to the corresponding acridizinium salts 9 were accomplished by heating in 48 % aqueous HBr and precipitation of the salts by the addition of THF.The yields and reactions conditions are tabulated in Table 1.Only 8a cyclized in an extremely poor yield (10%) while 3c readily underwent cyclization.Hemiacetal 4a also readily cyclized to afford 9a.The corresponding tetrafluoroborate salt 9a (where Br = BF 4 ) could be isolated on heating 3a in aqueous tetrafluoroboric acid.Cyclodehydration of 4c in concentrated HCl led to acridizinium chloride 9a (where Br = Cl).
The 1 H NMR absorptions for the singlets (DMSO-d 6 ) at H-6 (except 6) and H-11 (except 9e and f) are also tabulated in Table 1.

Nucleophilic ring openings of the acridizinium salts
It has been previously reported that treatment of acridizinium bromide (9a ) in water with a 10% aqueous NaOH solution led to an orange precipitate which was tentatively identified as a mixture of the pseudo base 10 and ring-opened aldehyde 11a. 18,19H OH 10 We initially evaluated the stability of acridizinium bromide (9a) in aqueous solutions of varying pH.The acridizinium bromide was dissolved in 4 aqueous solutions at different pH values: 8 (NaHCO 3 ), 9-10 (NaHCO 3 /Na 2 CO 3 ), 11 (NaHCO 3 /Na 2 CO 3 ) and 12 (Na 2 CO 3 ).The salt was stable in NaHCO 3 solutions at pH 8, the solution at pH 9-10 formed a orange precipitate after a few minutes while the solutions at pH 11 and 12 formed orange precipitates immediately on addition of the salt.
Attempts to establish the structure of the orange solid via 1 H NMR spectroscopy, as prepared via the procedure described by Bradsher, 18 were only partially successful.It was impossible to obtain a pure sample of the pseudo base previously formulated as 10.The stability of the original solid isolated from the reactions was dependent on the solvent used in the 1 H NMR analysis such as CDCl 3 or C 6 D 6 .If the initial crude solid obtained from treatment of acridiziniium bromide (9a) was collected and dried for a short period and the 1 H NMR recorded in CDCl 3 , a complex spectra indicated the presence of the ring-opened aldehyde [absorptions at δ 10.29 (CHO) and 4.60 (CH 2 )] along with numerous additional absorptions which were difficult to assign.If the same procedure was repeated and the 1 H NMR recorded in C 6 D 6 [ absorptions at δ 10.2 (CHO) and 4.4 (CH 2 )] the proportion of ring-opened product was reduced ARKAT USA, Inc.
(approximately10%) but the integrals for the remaining protons amounted to approximately 25-27 protons, indicative of the presence of several species.If the solutions of the CDCl 3 and C 6 D 6 were allowed to stand for 20 hours, the predominant product was the ring-opened aldehyde 11a.On allowing the crude orange product to stand in the air for 2 days, the 1 H NMR spectrum (CDCl 3 or C 6 D 6 as solvents) did not show much change in composition.
Treatment of an aqueous solution of acridizinim bromide (9a) with 0.1% ( pH = 12.5), 2% (pH = 14) or 10% (pH = 14) aqueous NaOH led to an orange solid which was then extracted into CDCl 3 in each case.The 1 H NMR spectra in each case indicated the presence of the ringopened aldehyde 11a with only relative minor absorptions for other species.These solutions on removal of the CDCl 3 led to black residues which on dissolving in CDCl 3 and 1 H NMR still indicated the presence of the aldehyde 11a.Treatment of the 9-methyl acridizinium bromide (9b) with a 0.1% aqueous solution of NaOH followed by extraction into CDCl 3 and 1 H NMR analysis indicated the presence of the ring-opened aldehyde 11b.Similarly the 6-methyl acridizinium bromide (9d) with 2% aqueous NaOH followed by CDCl 3 extraction and 1H NMR analysis indicated the major product as the ring-open ketone 11c.
The cyclization of the crude orange product, obtained from 9a on treatment with aqueous NaOH, using 48% aqueous HBr readily led to an excellent yield of 9a.
Since the pH of the aqueous hydroxylamine solution remains at 8, this suggests that the oximes arise from an initial attack of the nitrogen electron pair at C-6 of the corresponding acridizinium bromide to yield intermediates which rapidly undergo ring-openings.
In order to more fully evaluate the ring-opening reactions, the acridizinium bromides were treated with several primary amines and N,N-dimethylhydrazine.Treatment of 9-methyl acridizinium bromide (9b) with 1-aminopropane led to 11h (78%) while acridizinium bromide (9a) on addition of N,N-dimethylethylenediamine led to the imine 11i (quantitative).Both imines were readily identified by 1 N NMR and 13 C NMR analysis.These imines were unstable on exposure to air for short periods.It might be noted that imine 11h on heating in 48% aqueous HBr led to 9-methyl acridizinium bromide (9b) which was contaminated with the hydrobromide salt of 1-aminopropane ( 1 H NMR analysis).On the other hand treatment of 9a with the secondary amine pyrrolidine led to an immediate red coloration but which on workup, no identifiable products could be isolated or characterized by 1 H NMR. A similar result has previously been reported on treatment of 9a with piperidine. 20cridizinium bromide (9a) on treatment with N,N-dimethylhydrazine afforded hydrazone 11j (78%).
In conlusion, we have shown that the use of pyridine 2-carboxaldehyde and 2-acetyl pyridine in reactions with benzylic bromides leads to the direct synthesis of pyridinium salts.The use of methanol as solvent in the reactions of pyridine 2-carboxaldehyde with benzylic halides led to the corresponding pyridinium hemiacetals.These salts undergo cyclizations in 48% aqueous HBr (also HBF 4 and HCl) ) to afford the corresponding acridizinium salts.The acridizinium salts undergo facile ring opening reactions with aqueous NaOH, oximes, primary amines and 1,1,-dimethylhydrazine to afford the corresponding ring-opened products.The structure of the intermediate obtained from treatment of acridizinium bromide with aqueous NaOH, could not be firmly established because of facile ring-openings.However, the ring opened products and acidic cyclizations of the crude orange intermediate to the acridizinium salt are quite consistent with the proposed structure 10.

Experimental Section
General Procedures: The pyridine 2-carboxaldehyde (1), 2-acetyl pyridine and benzylic bromides 2a and 2b were purchased from Acros.Melting points were taken on a Fisher-Johns or a Mel Temp II apparatus and are uncorrected.The 1 H NMR and 13 C NMR data were acquired on a Bruker ARX-500 pulsed spectrometer with TMS or the solvent as an internal standard.Pyridinium salts 1-Benzyl-2-formyl pyridinium bromide (3a).Pyridine-2-carboxaldehyde (1) (1.22 g, 11.1 mmol) and benzyl bromide (2a) (1.73 g, 10.0 mmol) were dissolved in DMF (1 mL) and the solution was heated in an oil bath held at 45 o C for 24 h during which time a crystalline, yellow solid separated.After cooling, the crystals were collected by filtration and washed thoroughly with ether to afford 3a as a orange crystalline solid (1.70 g, 61%) which was hygroscopic, mp 179-184 (dec., black liquid).Attempts to recrystallize this product were unsuccessful but the product could be stored in the freezer for long periods, 1 H NMR (DMSO-d 6 ) δ 10.39 (s, 1H), 9.30 (d, J = 5.9 Hz, 1H), 8.91 (t, J = 7. 7

1-(2,5-Dimethoxybenzyl)-2-formylpyridinium bromide (3c).
A mixture of pyridine-2carboxaldehyde (1) (0.31 g, 2.5 mmol) and 2,5-dimethoxy benzyl bromide (2c) (0.65 g, 2.8 mmol) on DMF ( 0.05 mL) was placed in an oil bath at 40 o C.After 6 hours crystals commenced to separate and the mixture was held at this temperature for a total of 24 hours.The solid was collected by filtration and washed with ether to afford beautiful orange crystals of 3c (0.48 g, 50%) which on 1 H NMR analysis indicated a residual trace of DMF and a small amount of the hydrate.This material was crystallized from acetonitrile to afford yellow-orange crystals, mp 138-140 o C. 1

Procedure 2. From pyridine-2-carboxaldehyde and benzyl bromide in MeOH (4a).
The pyridine-2 carboxaldehyde (1) (1.07 g, 10 mmol) and benzyl bromide (2a, 1.71 g, 10 mmol) were dissolved in methanol (4 mL) and the mixture was refluxed for 2.3 h and allowed to stand overnight.The methanol was removed by rotary evaporation to afford a viscous yellow oil.Tetrahydrofuran (50 mL) was added to yield a yellow solid which was collected by filtration, washed with THF (30 mL) and dried to afford 2.08 g (67%) of crude product.A 1 g portion of this material was treated with ethyl acetate (10 ml) and MeOH (1-2 mL) was added to give a white solid which was collected by filtration and dried (4a, 0.70 g, 70% recovery), mp 122-124 o with an identical 1 H NMR as the compound prepared in Procedure 1. Attempts to further purify by heating in ethyl acetate followed by addition of methanol led to nice crystals which on 1 H NMR analysis indicated the presence of a small amount of pyridinium aldehyde 3a.

6-Methylacridizinium bromide (9d).
The pyridinium salt 6 (0.58 g, 1.73 mmol) was treated with 48% aqueous HBr (1.2 mL) and the mixture placed in an oil bath at 70 o C. The mixture was held at 90 o C temperature for 4 h , cooled and allowed to stand overnight.The orange solution was then quenched into THF (40 mL) on which a bright yellow solid formed.

-(2-formylbenzyl)pyridine (11a ).
A solution of acridizinium bromide (9a) (29 mg, 0.09 mmol) in water (0.6 mL) was treated rapidly with 10% aqueous NaOH ( 0.2 mL).An orange solid immediately formed and the mixture was stirred for 1 h during which period the color of the solid changed to a lighter orange color.The solid was collected by filtration, washed with water and dried to yield 15 mg (85%) of product as a light pink material.The 1 H NMR data were obtained in CDCl 3 and C 6 H 6 .Both indicated the presence of the ring-opened aldehyde 11a ( absorptions at δ 10.A solution of acridizinium bromide (9a) (10 mg) in water (0.3 mL) was treated with 10% aqueous NaOH ( 8 drops, pH = 14) to yield an orange precipitate.The mixture was stirred for 20 minutes and then CDCl 3 (0.6 mL) was added.The CDCl 3 layer was separated, dried over Na 2 SO 4 and the 1 H NMR recorded.The presence of mainly the ring-opened aldehyde 11a was indicated.Similar results were found using 0.1% (pH = 12.5) or 2% (pH = 14) aqueous NaOH.

Table 1 .
Reaction conditions, yields and 1 H NMR data for the acridizinium salts 9a-f