Key words
SQ109 - tuberculosis - synthesis - reduction - trimethylsilyl chloride - geranylamine - lithium aluminum hydride
Tuberculosis (TB) remains the most lethal disease worldwide, caused by the bacillus of Mycobacterium tuberculosis (Mtb). An estimated 10 million people fall ill with TB every year and it is a leading infectious agent for carriers of HIV.[2] Several compounds have been developed as antitubercular agents and N-geranyl-N′-(2-adamantyl)ethane-1,2-diamine (SQ109, 10)[3]
[4] is a promising second-generation ethylenediamine agent after the first-line-drug ethambutol (Figure [1]).[5] SQ109 (10) blocks an essential step of the synthesis of the mycobacterial outer membrane, via direct or indirect inhibition of MmpL3 transporter of trehalose monomycolate (TMM),[6]
[7]
[8]
[9]
[10] which is a basic component of the mycobacterial cell envelope.[9]
[11]
[12] SQ109 (10) is in phase IIb/III of clinical trials showing high potency against resistant Mtb strains and other pathogens.[13]
[14]
[15] The importance of SQ109 (10) has triggered interest into the synthesis of analogues[5]
[6]
,
[16]
[17]
[18]
[19] aiming at the improvement of potency and pharmacokinetic properties.
Figure 1 Chemical structures of ethambutol and SQ109 (10)
The reported methods for the synthesis of SQ109 (10) couple geranylamine 4 with 2-adamantanamine (6) through an ethylenediamine linker and include condensation of one of these amines[5]
,
[16]
[17]
[18] with chloroacetyl chloride[5] or bromoacetyl chloride,[16–18] following a condensation reaction with the second amine, and then reduction of the amide group of the two carbon linker with LiAlH4 in refluxing tetrahydrofuran (THF)[16]
[17]
[18] or with Red-Al in THF under reflux[16] or room temperature.[5] Alternatively, condensation of 2-adamantanone with N-geranyl ethylenediamine (prepared from geranyl bromide 2 and ethylenediamine) is followed by reductive amination with NaBH4 in methanol.[17]
[18] This procedure is much more convenient and can be applied to the synthesis of SQ109 analogues having a second substituent, such as an alkyl group, at the 2-adamantyl position.
The preparation of geranylamine 4 has been reported using the low-cost precursor geraniol (1) and phthalimide, which were subjected to a Mitsunobu reaction to afford N-geranyl-phthalimide.[16]
[20] Geraniol (1) can also be converted quantitatively into geranyl bromide (2)[17]
[18] that can be reacted with potassium phthalimide using the Gabriel reaction[17]
[18]
[20]
[21] or with phthalimide under microwave conditions,[22] affording N-geranyl phthalimide (Schemes 1 and 2). The N-geranyl phthalimide is then treated with an aqueous solution of hydrazine 60% w/v to afford 4 (without chromatographical purification according to the literature) in 37–49%,[16] 69 %,[20] or 62%[17]
[18] yield from geraniol (1) using the Mitsunobu or Gabriel reaction, respectively. In the original literature procedure we found geranylamine 4 was synthesized by Gabriel reaction in 20% yield from geranyl chloride[23] and 77% from geranyl bromide.[17]
[18]
[20]
[21] Reduction of 1-geranylazide (5) with Lindlar catalyst has been reported as effective for the selective reduction of the azido group to an allylic azide.[24] Additionally, the Staudinger reaction of geraniol acetate,[25] using PPh3 and NH3 has been applied for the preparation of 4 in 59% yield.
Scheme 1
Reagents and conditions: (a) PBr3, dry Et2O, –5 °C, 3 h, (quant.); (b) Ph3P, DIAD, phthalimide, anhydrous THF, rt, 24 h (81%); (c) phthalimide, K2CO3, anhydrous THF, reflux, 24 h, (88%); (d) N2H4·H2O, EtOH, reflux, 6 h (81%).
Scheme 2
Reagents and conditions: (a) ΝaΝ3, EtOH, reflux, 5 h (92%); (b) PPh3, THF/H2O, rt, 12 h (34%); (c) LiAlH4, anhydrous Et2O, rt, 24 h (24%).
We managed to synthesize and test the anti-Mtb activity of SQ109 (10) and analogues according to the first of the above-mentioned procedures for SQ109 (10) preparation,[5]
,
[16]
[17]
[18] which also enables the preparation of analogues substituted at the 2-adamantyl position. We report herein our observations of inconsistencies and improvements regarding previously reported results.
We firstly describe our observations during preparation of the commercially available but expensive geranylamine 4. We reacted geraniol (1) with phosphorus tribromide in dry diethyl ether at –5 °C and obtained geranyl bromide 2 in 99.5% yield (Scheme [1]), which was converted into N-geranyl phthalimide 3 using both the Mitsunobu reaction at room temperature[16]
[20] or Gabriel reaction under reflux.[17,18,20,21] Hydrazinolysis of N-geranyl phthalimide[16]
[17]
[18]
,
[20] under refluxing conditions afforded geranylamine 4 in 81% yield. During workup use of aqueous hydrochloric acid should be avoided since the double bond at 6-position of geranylamine 4 reacts, as evidenced by the observation of a singlet at δ = 1.19 ppm in the 1H NMR spectrum, which is attributed to a t-Bu group, and the disappearance of the unsaturated CH proton resonance at ca. δ = 5 ppm leading to a secondary amine contamination of the product.
We also tested the reaction of geranyl bromide 2 with sodium azide in refluxing ethanol to afford geranylazide 5 in 92% yield. We then examined reduction of 1-geranylazide (5) by Staudinger reaction,[26]
[27] using PPh3 in THF/H2O (10:1) and isolated geranylamine 4 in 34% yield as a solid hydrochloride salt (Scheme [2]). We were not able to improve the yield and did not isolate any other amine side product. However, we observed minor peaks in the 1H NMR spectrum of geranylazide. It has been shown[25] that geranylazide is in equilibrium with linalyl azide in a ratio of ca. 80:20, as is shown in the spectrum of geranylazide in the Supporting Information. Perhaps, the equilibrium between geranylazide and linalyl azide of different stability (the last having a tertiary carbon azide bond) is the reason for the low yield of geranylamine 4 which is formed through the corresponding triphenyl(N-geranylimino)phosphorane. This may the reason why the authors[25] improved the Staudinger reaction conditions using geranyl acetate and Ph3P/NH3 to obtain geranylamine 4 in 59% yield. However, we did not apply these reaction conditions since we obtained geranylamine 4 in a better yield using a procedure using Mitsunobu or Gabriel conditions as described above.
Next, we examined the reduction of 1-geranylazide (5) with LiAlH4 at room temperature and obtained geranylamine 4 as the free amine in 24% yield. It should be borne in mind that the Staudinger reaction has the disadvantage of the need to remove the phosphine oxide side product from the resulting reaction mixture.
We then proceeded to the connection of geranylamine 4 and commercially available 2-adamantanamine (6, alternatively prepared by reduction of the oxime of 2-adamantanone[28] with LiAlH4, see the Supporting Information) through the two-methylene linker (Scheme [3]). We treated either 2-adamantanamine (6) or geranylamine 4 with a mixture of 1-bromoacetylcloride and potassium carbonate at room temperature and produced the corresponding bromoacetamides 7 or 11 in 83% or 91% yield, respectively (Scheme [3]). Subsequently, adding solution of 7 or 11 in dry THF to a solution of geranylamine 4 or 2-adamantanamine (6) and triethylamine in dry THF at room temperature furnished 8 or 12, which are aminoamide precursors of SQ109 (10), in 86% or 72% yield, respectively.
Scheme 3 Synthesis of SQ109. Reagents and conditions: (a) ClCOCH2Br, K2CO3 (aqueous), DCM, rt, 24 h (7: 83%, 11: 91%); (b) 4 or 6, Et3N, dry THF, rt, 48 h, (8: 86%, 12: 72 %); (c) (i) Me3SiCl, LiAlH4, dry DCM, 0–5 °C, Ar, 2.5 h; (ii) NaOH 10%, 0 °C (31–38%).
We applied the literature conditions for the reduction of aminoamide 8, which involves the reaction with LiAlH4 in refluxing dry THF under inert atmosphere for 16 h.[6]
[17]
[19] Although it was reported that this reduction yielded SQ109 (10) in 50% yield after column chromatography, our two repeats afforded SQ109 (10) in 8% and 14% yield after column chromatography. We also observed the formation of geranylamine 4 and a byproduct formed due to the partial reduction of the geranyl chain since the double bond close to the reaction center becomes saturated. This is evidenced immediately by the disappearance of the corresponding unsaturated CH proton resonance at δ = 5.25 ppm (m) in the 1H NMR spectrum. Additionally, an aqueous hydrochloride solution should not be used during reaction workup since, as mentioned previously, we observed that the distant double bond from the reaction center becomes saturated. We also tested the reduction with LiAlH4 in dry tetrahydrofuran or diethyl ether at room temperature, but we obtained again the same mixture of amines.
We then explored more selective reduction conditions of the amido group and examined the use of Me3SiCl/LiAlH4
[29] (see experimental section in ref. 17) This reagent can activate the amide carbonyl functionality through the formation of the trimethylsilyl enol intermediates 9 or 13 (Scheme [3]); the in situ generated imine 9 or 13 being reduced efficiently with LiAlH4 to the amino group under mild conditions. Thus, when we treated aminoamides 8 or 12 with a mixture of LiAlH4 and distilled Me3SiCl in dry dichloromethane (DCM) at 0–5 °C for 2.5 h under an inert atmosphere, we isolated SQ109 (10) in 31–38% yield after column chromatography (Scheme [3]). Increasing the reaction time or temperature to gentle reflux did not improve the yield of SQ109 (10).
The yield for converting geraniol (1) into geranylamine 4 was 20% in ref. 17 and 37% in ref. 16 compared to 66% in our work. The overall yield for SQ109 (10) starting from geraniol (1) in ref. 17 is 5% and in ref. 16 is 16% compared to our ca. 20% yield. The yield for SQ109 (10) synthesis in ref. 5 was reported to be 24% from geranylamine 4, or the overall yield from geraniol (1) was ca. 5–8% (based on the yields for converting geraniol (1) into geranylamine (4) using the Gabriel[17] or Mitsunobu reaction[16] reported in ref. 17 and 16, respectively).
The yield of a second reported procedure including the condensation of 2-adamantanone with N-geranyl ethylenediamine (prepared from geranyl bromide 2 and ethylenediamine), followed by reductive amination with NaBH4 in methanol,[17]
[18] is 18% from geranyl bromide 2, and the overall yield from geraniol (1) is 17%. However, as mentioned, this procedure is not general for SQ109 (10) analogues with a second substituent at 2-adamantyl group position.
In conclusion, SQ109 (10) is currently in phase Ib/III of clinical trials against Mycobacterium tuberculosis. Herein, we report our investigations into the literature procedures for the synthesis of SQ109 (10) that enables the development of structure–activity relationships through the preparation of SQ109 (10) analogues at 2-adamantyl position.[30] We report inconsistencies and improvements for the more efficient preparation of important intermediates such as geranylamine 4 and the reduction of aminoamides 8 or 12 leading to SQ109 (10). Thus, while we used both 2-adamantanamine (6) and geranylamine 4 as the first or second amine shown in Scheme [3], we found that the reduction of the amide bond in the intermediate aminoamide 8 or 12, respectively, is problematic, yielding a mixture of amines, due to decomposition of the substrate and partial reduction of the geranyl chain, significantly lowering the yield of SQ109 (10). We identified that the application of Me3SiCl/ LiAlH4 under mild conditions can be used for the efficient reduction of the aminoamide precursors 8 or 12 to SQ109 (10). The findings will hopefully assist in the investigation of the efficacy of SQ109 (10) and analogues with potentially improved pharmacokinetic properties.
