Syntheses of Hindered-Polymethylacridinium Esters with Potential for Biological Probe Nanoarchitectonics

and the rate of the chemi-luminescent reactions. We have successfully synthesized the two novel methylated AEs 2 and 3 and compared their chemiluminescent properties with those of 1a （ Fig. 1 ） , and also prepared the substituted acridine esters 4 and 5 , but were unable to convert those acridines into the corresponding acridinium salts.


Introduction
Chemiluminescent immunoassay is an important analytical tool for in-vitro diagnostics 1−3) .Chemiluminescent compounds can replace radioactive isotopes that are hazardous and expensive to produce 4) .Acridinium esters (AEs) show high chemiluminescent sensitivity and quantum yields and low background.They are safe, stable, can be produced in large quantities, and are easily attached to biological molecules by means of an appropriate linker group.Furthermore, their chemiluminescence can be stimulated by alkaline peroxide without a catalyst.Therefore, the synthesis and use of AEs have gained much attention over the years 5−20) .
The AE 1a (Fig. 1) , which contains a succinimidyl (NHS) ester as the active linker group, was developed many years ago 21) .Various studies have been carried out to produce analogues of 1a that have better chemiluminescent properties.A variety of AEs having different leaving groups has been synthesized 22,23) .In particular, substitution at the ortho-positions of the phenoxy leaving group leads to better stability of immunoglobulin (IgG) conjugates in comparison with those of 1a by hindering hydrolysis of the AEs 24) .Such substitution also affects chemiluminescent properties such as quantum yield and rate of development of emission.For example, the dimethoxy AE 1b (Fig. 1) shows slightly better quantum yield than 1a 24) , while the dimethyl and dibromo AEs (1c and 1d, respectively) are slightly less efficient compared with 1a 24) .We have shown that the intro-duction of the linker chain at the ortho-or meta-position of the phenoxy ring also affects some of the chemiluminescent properties 25) .
Modification of the leaving group, however, does not affect the emission wavelength or the efficiency of energy transfer to a given acceptor, which depend on the nature of the excited acridone emitter.Therefore, substitution on the acridinium ring is required in order to influence those properties.The recent publication of the syntheses of AEs with methyl groups at the 2, 3, 6 and/or 7 positions of the acridinium ring 6) prompts us to report our own studies of polymethyl AEs, which were studied as part of our longterm interest in the synthesis of AEs 24−28) .In particular, we were interested to see whether introduction of methyl groups in positions 1 and/or 8 would significantly influence the stability towards hydrolysis and the rate of the chemiluminescent reactions.We have successfully synthesized the two novel methylated AEs 2 and 3 and compared their chemiluminescent properties with those of 1a (Fig. 1) , and also prepared the substituted acridine esters 4 and 5, but were unable to convert those acridines into the corresponding acridinium salts.point apparatus.A Perkin Elmer Fourier Transform Infra-Red (FTIR) Spectrometer was used to record IR spectra (KBr disk) .A Bruker AV 400 Spectrometer was used to record 1 H (400 MHz) and 13 C (100 MHz) NMR spectra; tetramethylsilane (TMS) was used as internal standard.Chemical shifts (δ) are in ppm, and coupling constants (J) are in Hz.A VG 12-250 mass spectrometer was used to record the low and high-resolution mass spectra.Chromatographic purifications were performed using Fisons Matrix silica 60.The purity of products was checked using Whatman silica gel plates with the aid of a UV lamp (254 nm) .A Ciba-Corning Magic Lite analyzer luminometer was used to measure the chemiluminescence in relative light units (RLU) .Reagent A (1 L) contained HNO 3 (70％; 6.    25) .

Synthesis of N-phenyl-4,6-dimethylisatin (16)
The procedure used to produce 10 was also used to produce 16.It involved the use of 15 (7.504 g, 38.1 mmol) , (COCl) 2 (3.75 mL, 43.0 mmol) , and AlCl 3 (16.784g, 125.9 mmol) and gave pure 16 (5.704g, 60％) .Mp 152-153℃.FTIR (ν; cm −1 ) : 1723 (C＝O) , 1748 (C＝O) . 1  to give a concentration of 250 µg/mL.A solution of the appropriate AE in dimethyl sulfoxide (5 µL of concentration 1 mg/mL) was added to a portion (200 µL) of the IgG solution, and the mixture was stirred and then left to stand in the dark for 15 min.A quench buffer (0.1M NaH 2 PO 4 adjusted to pH 8 by addition of 5M NaOH, plus lysine monohydrochloride ［10 mg/mL of buffer］ ) ( 100 µL) was added and the mixture was left for a further 5 min.The mixture was subjected to gel column chromatography (Sephadex G50, eluted with a pH 6.3 buffer ［0.1M Na 3 PO 4 , 0.15M NaCl, adjusted to pH 6.3 by addition of 5M NaOH, plus 0.1％ NaN 3 and 0.1％ bovine serum albumin］ ) and the fractions containing the first (major) chemiluminescent material were combined.Aliquots (10 µL) of this solution were diluted with the appropriate pH buffer (pH 5, 6, 7, or 8) until the concentration was sufficiently low to allow detection within the limit of the luminometer (ca. 5 million RLU) .These solutions were stored at the appropriate temperature and aliquots were extracted and their chemiluminescence was measured at appropriate intervals.

Chemistry
The synthetic route for the 1,3,6,8-tetramethyl AE 2 is shown in Scheme 1. Treatment of 3,5-dimetylaniline (6) with acetic anhydride (Ac 2 O) gave the corresponding acetanilide 7 in 85％ yield.Reaction of this with 1-bromo-3,5dimethylbenzene, copper iodide (CuI) , and potassium carbonate (K 2 CO 3 ) gave a 58％ yield of 8, which on alkaline (KOH) hydrolysis gave 9 in 73％ yield.The NMR spectra of 7 and 8 showed the presence of two categories of methyl groups (aryl and acetyl) but those of 9 showed only a single methyl group signal.
Treatment of 10 with aqueous KOH for a long period under reflux gave the tetramethylacridinecarboxylic acid 11 in 87％ yield.The IR spectrum showed absorption bands at 1733 (C＝O) and 3100 (OH) cm −1 .The FAB mass spectrum of 11 showed pseudo molecular ion peaks at m/z＝ 302 ( ［M＋Na］ ＋ ) and 280 (MH ＋ ) .High-resolution mass spectrometry (HRMS) confirmed the formula of the 280 peak as C 18 H 18 NO 2 .
Attempted conversion of 11 into its corresponding ester 12 in a single step by refluxing it for 17 h with thionyl chloride (SOCl 2 ) and benzyl 3- The presence of two different carbonyl groups was confirmed by the IR spectrum (1732 and 1754 cm −1 ) and the acridine and phenolic moieties.Debenzylation of 12 using hydrogen bromide (HBr) in acetic acid (AcOH) for 3 h at 100℃ gave a 77％ yield of the acid 13.Its CI mass spectrum showed a pseudo molecular ion (MH ＋ ) peak at m/z＝428.A singlet at 12.20 ppm in its 1 H NMR spectrum confirmed it was a carboxylic acid.In addition, its 13 C NMR spectrum showed two signals at low field (168.2 and 173.5 ppm) corresponding to two different carbonyl carbons.
The reaction of 13 and N-hydroxysuccinimide (NHS) in the presence of dicyclohexylcarbodiimide (DCC) in dimethylformamide (DMF) at 20℃ gave 14 in 50％ yield.The IR and 13 C NMR spectra of 14 confirmed the presence of two different carbonyl groups.A singlet integrating for 4 protons at 2.86 ppm in the 1 H NMR spectrum verified that the NHS unit had been incorporated.The purity of 14 was confirmed by elemental analysis.
Finally, N-methylation of 14 using methyl trifluoromethanesulfonate in DCM at 20℃ gave the target AE 2 in 51％ peak at m/z＝539 and its formula was confirmed by the HRMS as C 32 H 31 N 2 O 6 .The synthesis of AE 3, as outlined in Scheme 2, followed the same approach as for the synthesis of 2. The reaction of 7 and bromobenzene under reflux under basic conditions gave 15 in 77％ yield without isolation of the intermediate diarylacetamide (analog of 8) .The structure of 15 was confirmed by a singlet at 5.58 ppm (NH proton) in its 1 H NMR spectrum and a molecular ion peak at m/z＝ 197 in its mass spectrum.
Treatment of 15 with (COCl) 2 and AlCl 3 gave a 60％ yield of 16.Its IR spectrum showed two carbonyl groups (1723 and 1748 cm −1 ) and its molecular ion in the mass spectrum was at m/z＝251.
Hydrolysis of 16 with boiling aqueous KOH gave carboxylic acid 17 in an 83％ yield.The carboxylic acid group was confirmed by its IR spectrum (bands at 1732 and 3100 cm −1 ) and the molecular formula was indicated by HRMS of the pseudo molecular ion (MH ＋ ) peak in the CI mass spectrum at m/z＝252, corresponding to C 16 H 14 NO 2 .
A 32％ yield of 18 was obtained from the reaction of 17 with benzyl 3-(4-hydroxyphenyl) propionate under reflux for 18 hours in benzene with 4-toluenesulfonyl chloride and triethylamine.The purity of 18 was confirmed by elemental analysis; there were also two carbonyl carbon signals in its 13 C NMR spectrum and its CI mass spectrum showed a pseudo molecular ion (MH ＋ ) peak at m/z＝490, confirming its structure.
Debenzylation of 18 gave the corresponding acid 19 in The syntheses of compounds 4 and 5 (Scheme 3) followed similar routes to those of compounds 14 and 20.However, since compound 24 has been reported previously 31) , the approach to that compound followed the literature route.Also, for the esterification steps it was possible to use a standard method for formation of the appropriate polymethylacridinecarboxylic acid chloride followed by addition of benzyl 3-(4-hydroxyphenyl) propionate, which had not been successful with compounds 12 and 18. Otherwise the procedures were very similar to those applicable to Schemes 1 and 2. The experimental details, and the spectroscopic and spectrometric data used to characterize the various compounds are given in the Experimental Section.It had been intended to convert compounds 4 and 5 into their corresponding AEs, but unfortunately stirring compound 4 with methyl trifluoromethanesulfonate in DCM at room temperature produced no precipitate and the starting material (4) was recovered unchanged.Use of more forcing conditions (overnight reflux) still did not produce the desired methylated derivative, but some decomposition of the starting material occurred under such conditions.It has previously been noted that both 4,5-dimethylacridine and 1,4,5,8-tetramethylacridine fail to react with electrophilic boron reagents (BH 3 and BF 3 ) because of hindrance by the methyl groups at positions 4 and 5 31) , so perhaps the failure to methylate 4 is not surprising.Compound 5 was also not methylated with methyl trifluoromethanesulfonate over a period of 2.5 h at room temperature.Under more forcing conditions, some methylation did occur, as evidenced by the emergence of a peak at δ＝4.9 ppm for the N-Me group in the 1 H NMR spectrum and characteristic changes in the aromatic region of the spectrum.However, there was also evidence of breakdown of the ester group under such conditions and it was not possible to obtain pure N-methylated product suitable for either complete characterization or meaningful study of its chemiluminescent properties.
Following the synthetic studies, therefore, just the two AEs, 2 and 3, were available for studies of chemiluminescence and hydrolytic stability.

Chemiluminescence and stability
AEs 2 and 3 were found to give lower light output (31 and 35％ of that for AE 1a, respectively) under typical alkaline conditions, as a result of the presence of the additional methyl groups on the acridine ring.In particular, the methyl groups at the 1-and 8-positions of the acridinium ring of AE 2 probably cause the carboxylic ester group to twist out of planarity and become orthogonal to the acridinium ring, which leads to its loss of conjugation and increased electrophilic character.In addition, attack at position 9 must be highly encumbered, so that the first step of the chemiluminescent reaction, involving attack by peroxide anion at position 9, would be significantly slowed in comparison to attack on 1a.This would allow greater competition from other, non-radiative, pathways, such as attack at the ester carbonyl group, which is not only more electrophilic than that in 1a because of the loss of conjugation, but also presents a relatively open face because of its orthogonal orientation.Consequently, attack by a nucleophile such as hydroperoxide anion or hydroxide anion directly at the carbonyl group, leading to non-radiative pathways, becomes more significant.AE 3, having only one methyl group in such an encumbering situation, will be subject to these same considerations, but to a lesser extent.
If AEs are to be used as biological probes, it is important that their conjugates with biological targets are sufficiently stable to be able to withstand the conditions to which they will be subjected during the monitoring process.In order to assess how AEs 2 and 3 compared to 1a in this respect, all three AEs were reacted with immunoglobulin G (IgG) and the products were subjected to gel chromatography to give solutions of the corresponding pure AE-IgG conjugates.10 µL aliquots of these solutions were separately diluted in buffers of different pH values (5, 6, 7, and 8) , samples of which were incubated at three different temperatures (0, 20, and 37℃) for different periods of time (1, 2, 4, 8, 16, and 32 days) .At the end of the incubation period, the chemiluminescence of each sample was measured and compared with the figure for an identical sample at time 0. The results are shown graphically in Fig. 2.
A few of the readings are clearly outliers (e.g., the 4 day result for IgG-3 at 0℃ and pH 5, the 32 day reading for IgG-1a at 20℃ and pH 5, and the readings for IgG-2 after 8 days at 20℃ and pH 7 and after 4 days at 0℃ and pH 8) , but from the rest of the readings several trends can be deduced.Unsurprisingly, the rate of reduction in the level of emission over time for otherwise similar samples was least for incubation at 0℃.It was also typically the case that incubation at 37℃ led to the greatest rate of reduction, although for some samples the differences between the 20 and 37℃ incubations were quite small.
Under most conditions, IgG-3 lost performance more rapidly than the other two conjugates, which can probably be understood in terms of the acridine-attached carbonyl group twisting out of conjugation somewhat, leading to greater electrophilicity, while at the same time presenting an open face for attack by a nucleophile.Also, degradation of IgG-3 was relatively insensitive to the conditions of both temperature and pH.However, IgG-1a, which was relatively slowly degraded at pH 5-7 under most conditions, was much more sensitive to degradation at the highest pH studied (pH 8) , and this led to its becoming the most rapidly degraded at the higher temperatures at pH 8.Under most conditions, IgG-2 was at least competitive with the best other conjugate in terms of the slowness of degradation in light output, and at higher temperatures, particularly at higher pH, it was the least rapidly degraded of all.The one case where it was not competitive with the best was for pH 7 at 0℃, where IgG-1a was clearly the least rapidly degraded.There were no conditions under which IgG-3 would have advantages over IgG-1a in terms of stability, but IgG-2 was more stable in many circumstances, particularly at higher temperatures and/or higher pH values.Therefore, AE 2 has the potential to be used as an acridinium label in such conditions.

Conclusions
Two acridinium esters (2 and 3) , containing methyl substituents on the acridinium ring, were successfully synthesized and their chemiluminescent properties were measured.Although the synthesized AEs showed substantial chemiluminescence, they displayed lower quantum yields of light output than the corresponding AE that contains no methyl substituents (1a) .The new AEs also possess an Nhydroxysuccinimide ester unit, which enables their easy attachment to many biologically important species, so in principle they could be useful as chemiluminescent labels.For such purposes, the conjugates with biological molecules would need to be relatively stable in buffer solutions for significant amounts of time.However, the 1,3,10-trimethylacridinium ester conjugate with immunoglobulin G (IgG-3) showed a more rapid drop in chemiluminescent output than IgG-1a under most conditions tried (pH 5-7; 0-37℃) , although it was somewhat more stable than IgG-1a at pH 8, particularly at the higher temperatures.The 1,3,6,8,10-pentamethylacridinium ester conjugate IgG-2 showed greater resistance to loss of chemiluminescent output over time, and was significantly more stable than either IgG-1a or IgG-3 at 37℃ at all pH values from pH 5 to pH 8.It was also more stable at 20℃ in pH 7 or pH 8 buffers.Therefore, acridinium ester 2 has the potential to be useful as a chemiluminescent label under such conditions.Two acridine esters (4 and 5) , having methyl groups at positions 4 and/or 5 on the acridine ring, were also synthesized, but it proved impossible to convert them into the corresponding acridinium esters by treatment with methyl trifluoromethanesulfonate, presumably because of the increased steric hindrance around the acridine nitrogen as a result of the proximity of those methyl groups.
3 mL) and H 2 O 2 (30％; 16.5 mL) in distilled H 2 O. Reagent B (1 L) contained NaOH (10 g) and cetyltrimethylammonium chloride (25％) in distilled H 2 O.Both reagents were provided by Molecular Light Technology, Cardiff.The buffer solution (pH＝8) used for the labelling contained sodium dihydrogen orthophosphate (0.1 M) and NaOH (5 M) .The quenching buffer contained lysine (10 mg/mL) in addition to the labelling buffer.Dilute solutions of 1-3 (1×10 −4 mg/mL) were made using MeCN for labelling.Standard procedures were used for the labelling of AEs
(4-hydroxyphenyl) propionate in anhydrous pyridine containing 4-dimethylaminopyridine (DMAP) gave only a 2％ yield, possibly due to slow and incomplete formation of the acid chloride intermediate as a result of hindrance by the methyl groups at positions 1 and 8 and competitive reaction of the phenol group of benzyl 3-(4-hydroxyphenyl) propionate with the thionyl chloride.Even when the acid chloride of 11 was prepared in a separate step and then refluxed for 17 h with benzyl 3-(4-hydroxyphenyl) propionate and DMAP in dry pyridine, product 12 was still not produced in significant yield.Use of toluenesulfonyl chloride (TsCl) instead of thionyl chloride was also unsuccessful, producing the toluenesulfonyl ester instead of the tetramethylacridinecarboxylic ester.However, replacing pyridine (which acted as both solvent and base) with benzene as solvent and triethylamine (Et 3 N) as base in the reaction of 11 with benzyl 3-(4-hydroxyphenyl) propionate in the presence of 4-toluenesulfonyl chloride for 18 h under reflux gave 12 in 12％ yield, and when the reflux period was extended to 2 days the yield was 28％.The chemical ionization (CI) mass spectrum of 12 revealed a pseudo molecular ion (MH ＋ ) peak at m/z＝518.