A succinct synthesis of valganciclovir hydrochloride, a cytomegalovirus (CMV) retinitis inhibitor #

A concise and efficient synthesis of valganciclovir hydrochloride 1 , a CMV retinitis inhibitor, without involving protection-deprotection sequences, is described. The synthetic utility of (2 S )- azido-3-methylbutyric acid, which acts as a masked L -valine equivalent, is demonstrated in the synthesis of 1


Introduction
Cytomegalovirus (CMV) is a common and opportunistic infection in adult population. 1 When the human immune defences are weak, CMV can attack several parts of the body and cause severe damage. 2 Patients with immunosuppression caused by various diseases including acquired immunodeficiency syndrome (AIDS) are at higher risk of being affected by CMV. 3 The most common illness caused by CMV is retinitis i.e. the death of cells in retinas, which can lead to blindness unless treated.
Valganciclovir hydrochloride (Valcyte ® ) 1 has been used for the treatment of CMV retinitis in patients with weakened immune systems. 4Valganciclovir hydrochloride is the hydrochloride salt of an ester of L-valine with ganciclovir 2 that exists as a mixture of two diastereomers. 5After oral administration, intestinal and hepatic esterases hydrolyze both diastereomers to ganciclovir 2, which inhibits replication of the human cytomegalovirus in vitro and in vivo.It is evident from the literature that valganciclovir hydrochloride 1 is a recognized substrate of the intestinal peptide transporter PEPT1, which instigates the tenfold higher bioavailability of ganciclovir subsequent to valganciclovir compared to oral ganciclovir administration. 6In Dr Reddy's Laboratories, we have been dynamically involved in the synthesis and process development of various active pharmaceutical ingredients (APIs).While working on the synthesis of antiretroviral drugs, we became fascinated by valganciclovir hydrochloride.In this communication, we divulge our efforts toward the synthesis of valganciclovir hydrochloride 1.

Results and Discussion
Since valganciclovir is a L-valyl ester of ganciclovir 2, most of the hitherto known literature on the synthesis of 1 involves the coupling of the protected form of ganciclovir 3 with N-protected L-valine derivative 4 followed by deprotection (Scheme 1).

Scheme 1. Known synthesis of 1.
Scientists from Syntex Inc. have published a number of patent applications, in which 1 was obtained through coupling of N and O protected ganciclovir with Boc or Cbz protected L-valine followed by deprotection. 7Nestor and co-workers have reported the synthesis of 1 from guanine. 8Rao's team has described an alternative method for the preparation of 1 by coupling mono O-acetyl protected ganciclovir with Cbz-L-valine followed by hydrogenolysis. 9Recently, Sharma and co-workers have filed a patent application, in which the synthesis of 1 was achieved by partial hydrolysis of bis Cbz-L-valine ester of ganciclovir followed by deprotection. 10he precedented literature methods for the synthesis of 1 clearly show that the protection of amino group in L-valine is inescapable to coerce the coupling reaction.It was anticipated that bulky protecting group on L-valine might cause some sort of steric influence while coupling with ganciclovir 2 or its protected derivative, thereby reducing the reaction rate or yield of the coupled product.Moreover, the protection-deprotection strategy may further increase the risk of racemization at the amino acid chiral center.These problems could be avoided by the use of small size versatile precursor of amino group such as azide functionality.The azido group is relatively stable under acidic and basic conditions, and can be readily converted to amine by simple reduction.Since there are few potential advantages of replacing the protected L-valine by the azido equivalent; we decided to utilize this protocol for the synthesis of valganciclovir hydrochloride 1.
The required α-azido acid, (2S)-azido-3-methylbutyric acid 5, was prepared from L-valine through azido-transfer reaction with imidazole sulfonyl azide. 11Initially, we were interested to learn the coupling reaction of 5 with N,O-ditrityl protected ganciclovir 6. 7a The DCC mediated coupling reaction of 6 with 5 in DMF gave the coupled product 7 in 74% yield (Scheme 2).Deprotection of trityl groups in 7 was carried out using trifluoroacetic acid in DCM.Unfortunately, the yield of the deprotected compound 8 was disappointing (60%).Attempts to increase the yield of 8 by using different conditions and reagents failed to give satisfactory results.Then, the azido group in 8 was reduced with Zn and acetic acid to give valganciclovir acetate, which in turn was converted in to the hydrochloride salt 1 through anion exchange with aqueous HCl.Whilst the yield of the product in the final stage was moderate (82%), it was observed that the product was always contaminated with zinc acetate.Since valganciclovir hydrochloride is highly soluble in water, the removal of zinc salts from 1 became cumbersome, and we were not able to isolate 1 in pure form. 12Scheme 2. Synthesis of valganciclovir hydrochloride 1.
In view of the fact that the yields of the intermediates in most of the stages in Scheme 2 were unacceptable, we have decided to amend the synthetic strategy.It is evident from the above scheme that the overall yield of the product 1 was significantly diminished by the integration of protection-deprotection sequence.So, it is obvious that the protection-deprotection strategy should be evaded in order to increase the overall yield of 1.This triggered us to investigate the selective acylation of ganciclovir 2 with (2S)-azido-3-methylbutyric acid 5. Thus, the coupling reaction of 2 with 5 was carried out using DCC conditions to provide the desired monoester derivative 8 (36%) along with bisester derivative 9 (25%) and unreacted ganciclovir 2 (20%) under the optimized conditions (Scheme 3).Although, the yield of the monoester derivative 8 was unimpressive, the overall yield of 8 from ganciclovir was better when compared to Scheme 2, i.e. the protection-deprotection strategy.Since the reduction of azido group of 8 with Zn and acetic acid was found to be problematic in Scheme 2, catalytic hydrogenation method was opted (Pd-C/H2).In this case, the reduction of 8 proceeded smoothly and 1 was isolated in 85% yield with good purity (Scheme 3). 12Scheme 3. Direct coupling route to 1.
After completing the synthesis of 1 through direct coupling route (Scheme 3), we turned our attention towards optimization of the coupling stage to improve the yield of 8 under different conditions.Unfortunately, none of the reaction conditions facilitated in improving the yield of 8.This observation provoked us to investigate the synthetic scheme 3 further in detail.In the direct coupling route (Scheme 3), the bisester derivative 9, which was considered as a waste byproduct, was always associated with a considerable quantity of monoester derivative 8. To improve the efficiency of the reaction, it was decided to convert the by-product 9 to the useful monoester intermediate 8 by partial hydrolysis.A number of bases and reagents were tried for this transformation, and finally N, N-diisopropyl ethylamine (Hünig's base) was found to be the better option.The partial hydrolysis of 9 was optimized with equimolar quantity of Hünig's base to give the required product 8 in 70% (in HPLC) 13 along with over-hydrolyzed product 2 (20%) and unreacted 9 (7%) [Scheme 4].

Scheme 4. Partial hydrolysis of diester derivative 9.
The yield and the quality of 8 were satisfactory and the over-hydrolyzed product 2, which is nothing but the starting ganciclovir, was recycled by crystallization.Since the yield of 8 through partial hydrolysis route (Scheme 4) was better than that of selective esterification method (Scheme 3), the strategy of synthesis was slightly modified.In the modified route, ganciclovir 2 was converted to the bisester derivative 9 using excess of 5 under DCC mediated coupling conditions (Scheme 5).The bisester derivative 9 was subsequently converted to the monoester derivative 8 (by partial hydrolysis method) followed by reduction with Pd-C/H2 to provide 1 in good yield and purity.

Conclusions
An efficient synthesis of valganciclovir hydrochloride 1, which is an inhibitor of CMV retinitis, has been developed without involving protectiondeprotection strategy.The synthetic utility of (2S)-azido-3-methylbutyric acid 5, which acts as a masked L-valine equivalent, has been demonstrated in the synthesis of 1.We have also established a proficient method for the partial hydrolysis of 9 to 8.

Experimental Section
General. 1 H NMR spectra were recorded at 400 MHz Varian FT-NMR Spectrometer.The chemical shifts are reported in δ ppm relative to TMS.The IR spectra were obtained using Perkin Elmer, Spectrum One FTIR spectrophotometer, with substances being pressed in a KBr pellet.The mass analyses have been performed on AB-4000 Q-trap LC-MS/MS mass spectrometer (MDS SCIEX, Applied Biosystems, California, USA).All the solvents and reagents were used without further purification.

Preparation of valganciclovir hydrochloride (1)
2-(2-amino-1,6-dihydro-6-oxo-purin-9-yl)methoxy-3-hydroxy-1′-propyl-2′-(S)-azido-3′-methyl butanoate 8 (10.0 g, 0.026 mol) and zinc dust (30.0 g, 0.459 mol) are charged into a round bottom flask and cooled to 5-10 °C.A solution of acetic acid (100 mL) in isopropanol (400 mL) was slowly added to the reaction mixture and the resulting suspension was stirred at 25-30 °C for about 45 minutes.It was filtered through celite and washed with a mixture of acetic acid (10 mL) and isopropanol (40 mL).The solvent was evaporated to dryness at 35 °C and the residue was suspended in acetone (200 mL) and stirred for 30 min.The solid obtained was filtered, washed with acetone (50 mL) and dried under reduced pressure.The solid is suspended in water (50 mL) and stirred for 5 minutes at about 5 °C.Acetone (330 mL) is added to the suspension and stirred for about 5 minutes.The obtained white solid was filtered, washed with acetone (40 mL), and dried under reduced pressure to afford the valganciclovir acetate.It was suspended in isopropanol (140 mL) and acetic acid (7 mL) was added to it followed by trifluoroacetic acid (7 mL) and stirred for 5 minutes.The resulting mixture was filtered through celite and washed with isopropanol (10 mL).The filtrate was treated with 4% HCl in isopropanol (14 mL) for 30 minutes and filtered.The obtained solid was again washed with isopropanol (20 mL) and dried under reduced pressure to afford 1 as a white solid.Yield

Synthesis of monoester derivative (8) by partial hydrolysis
To a stirred solution of bisvaline ester of ganciclovir 9 (50 g, 0.099 mol) in methanol (750 mL) was added N,N-diisopropyl ethylamine (12.8 g, 0.099 mol) at 25-30 o C and the resulting reaction mass was maintained for 14 h at RT.It was quenched with acetic acid (5.94 g, 0.099 mol) and the solvent was removed up to 90%.To this solution, was added hexane (200 mL) and the resulting suspension was stirred for 30 min at RT.The solid thus formed was filtered and washed with hexane (50 mL).The wet material was dried under vaccum to provide a residue, which contains 70% of 8, 20% of 2 and 7% of 9 (by HPLC).The crude material was then purified repeatedly using a mixture of n-butanol and water, followed by methanol to give pure mono ester of ganciclovir 8 as a white solid.Yield 13.0 g (34.5 %).

Synthesis of valganciclovir hydrochloride (1)
To a stirred solution of mono ester derivative 8 (5g, 0.013 mol) in methanol (100 mL) was added 10% Pd/C (1 g) followed by 33% aqueous hydrochloric acid (1.6 mL, 0.014 mol).The resulting suspension was hydrogenated in the autoclave at 4 kg/cm 2 pressure for about 2 h.The reaction mixture was filtered through celite® and washed with methanol (25 mL).The solvent was removed under vacuum and the residue was recrystallized from water: isopropanol (1:10) mixture to provide 1 as a white solid.Yield 3.6 g (70.6 %).