N-Alkylated Iminosugar Based Ligands: Synthesis and Inhibition of Human Lysosomal β-Glucocerebrosidase

The scope of a series of N-alkylated iminosugar based inhibitors in the d-gluco as well as d-xylo configuration towards their interaction with human lysosomal β-glucocerebrosidase has been evaluated. A versatile synthetic toolbox has been developed for the synthesis of N-alkylated iminosugar scaffolds conjugated to a variety of terminal groups via a benzoic acid ester linker. The terminal groups such as nitrile, azide, alkyne, nonafluoro-tert-butyl and amino substituents enable follow-up chemistry as well as visualisation experiments. All compounds showed promising inhibitory properties as well as selectivities for β-glucosidases, some exhibiting activities in the low nanomolar range for β-glucocerebrosidase.


Introduction
Iminoalditols, also called iminosugars, represent polyhydroxylated alkaloids and are structurally related to common carbohydrates in which the endocyclic oxygen atom is replaced by a basic trivalent nitrogen atom. This exchange endows this class of glycomimetics with basic properties that are responsible for their remarkable biological activity. Iminoalditols are widely distributed in nature and can be found in bacteria, fungi or plants [1]. Common structural features are monocyclic systems such as piperidines (e.g., compounds 1 and 2) as well as pyrrolidines (e.g., 3) and bicyclic scaffolds, for example indolizidines (e.g., 4), pyrrolizidines (e.g., 5), and nortropanes (e.g., 6) (see Figure 1).
An interesting enzyme in this respect is human lysosomal β-glucocerebrosidase (glucosylceramidase, GCase, EC 3.2.1.45, glycosidase family 30 [28]). This enzyme catalyses the degradation of the β-glycosidic linkage of glucosylceramide or glucosylsphingosine into D-glucose units and ceramide or sphingosine, respectively, which is the last step of the lysosomal catabolism of glycosylsphingosines in the autophagy-lysosomal pathway. A malfunction of GCase causes Gaucher disease (GD), a severe lysosomal storage disorder with an incidence of approx. 1 in every 50,000 people worldwide [29]. Additionally, Parkinson's as well as Alzheimer's disease have been linked to GCase deficiency [30,31].
In this context, a large variety of iminosugar based glycomimetics have been investigated [4,[32][33][34][35][36][37]. Paradigmatic examples are shown in Figure 2. Compounds 7 [38] have been introduced by Mellet and coworkers, D-xylo configured 1-C-alkyl iminosugars 8 [39] represent another important family which is also true for compounds 9 [40], all showing significant properties as pharmacological chaperones for mutant GCases. We are particularly interested in iminoalditol structures, where the modification is located at the ring nitrogen in order to install customised features for different research related properties of GCase. In this context, well-known N-alkylated examples are compounds 10 and 11, which inhibit GCase with Ki values of 116 µM and 0.3 µM, respectively [41] (Figure 3).A large variety of N-substituted iminosugar based glycomimetics was contributed by Overkleeft and co-workers, their compound collection of 1-DNJ derivatives contain various hydrophobic groups at the ring nitrogen such as alkyl chains 12 [42], different biphenyl derivatives 13 [43], an adamantly group 14 [44] and carborane- In general, representatives of this substance class are known as ligands for carbohydrate-processing enzymes (CPEs), as they interact with the active site of the respective protein as mimics of the natural substrate [7][8][9][10]. Iminosugar-based structures are versatile probes and have been used in several different applications as inhibitors, [11,12] therapeutics [8][9][10] and probes for activity-based protein profiling [13,14]. In particular, these structures have proved interesting for the therapy of CPEassociated diseases such as cancer [15], bacterial infections [16], HIV [17], HPV [18], influenza [19], hepatitis [20], the dengue virus [21], malaria [22] and fungal infections [23]. In addition, iminosugar based glycomimetics have shown encouraging results as so-called pharmacological chaperones (PCs) in the treatment of lysosomal storage diseases (LSDs) [24][25][26]. Due to their strong interaction with the active site of their respective enzyme, they can stabilize the correctly folded structures of mutant lysosomal enzymes, thereby obviating their cellular clearance [24,27].
An interesting enzyme in this respect is human lysosomal β-glucocerebrosidase (glucosylceramidase, GCase, EC 3.2.1.45, glycosidase family 30 [28]). This enzyme catalyses the degradation of the β-glycosidic linkage of glucosylceramide or glucosylsphingosine into d-glucose units and ceramide or sphingosine, respectively, which is the last step of the lysosomal catabolism of glycosylsphingosines in the autophagy-lysosomal pathway. A malfunction of GCase causes Gaucher disease (GD), a severe lysosomal storage disorder with an incidence of approx. 1 in every 50,000 people worldwide [29]. Additionally, Parkinson's as well as Alzheimer's disease have been linked to GCase deficiency [30,31].
An interesting enzyme in this respect is human lysosomal β-glucocerebrosidase (glucosylceramidase, GCase, EC 3.2.1.45, glycosidase family 30 [28]). This enzyme catalyses the degradation of the β-glycosidic linkage of glucosylceramide or glucosylsphingosine into D-glucose units and ceramide or sphingosine, respectively, which is the last step of the lysosomal catabolism of glycosylsphingosines in the autophagy-lysosomal pathway. A malfunction of GCase causes Gaucher disease (GD), a severe lysosomal storage disorder with an incidence of approx. 1 in every 50,000 people worldwide [29]. Additionally, Parkinson's as well as Alzheimer's disease have been linked to GCase deficiency [30,31].
In this context, a large variety of iminosugar based glycomimetics have been investigated [4,[32][33][34][35][36][37]. Paradigmatic examples are shown in Figure 2. Compounds 7 [38] have been introduced by Mellet and coworkers, D-xylo configured 1-C-alkyl iminosugars 8 [39] represent another important family which is also true for compounds 9 [40], all showing significant properties as pharmacological chaperones for mutant GCases. We are particularly interested in iminoalditol structures, where the modification is located at the ring nitrogen in order to install customised features for different research related properties of GCase. In this context, well-known N-alkylated examples are compounds 10 and 11, which inhibit GCase with Ki values of 116 µM and 0.3 µM, respectively [41] (Figure 3).A large variety of N-substituted iminosugar based glycomimetics was contributed by Overkleeft and co-workers, their compound collection of 1-DNJ derivatives contain various hydrophobic groups at the ring nitrogen such as alkyl chains 12 [42], different biphenyl derivatives 13 [43], an adamantly group 14 [44] and carborane-modified moieties 15 [45], just to mention a few. All of these compounds show Ki-values in the low micro to nanomolar range against GCase (respective Ki values see references [42][43][44][45]). Additionally, N,O-dialkylated derivatives of 1-DNJ (e.g., compound 16) were reported recently as potent ligands of ceramide transport protein [46]. Furthermore, multivalent iminosugar-based structures [47] (e.g., compounds 17 and 18) were reported to exhibit multivalency effects on GCase inhibition with Ki values of 55 nM and 285 nM, respectively [48]. 1,5-Dideoxy-1,5-imino-D-xylitol-based N-alkylated derivatives 19 and 20 were found as selective inhibitors of GCase with Ki values of 57 µM and 4.1 µM, respectively. [36]. Now, we present a robust synthetic method for the simple and concise composition of diversity driven N-alkylated iminosugar based inhibitors, which have the potential to be used as powerful tools such as inhibitors, pharmacological chaperones or probes for investigations of glycoprocessing enzymes in general and GCase in particular.

Synthesis
The developed synthetic concept relies on the coupling of three main building blocks, as shown in Figure 4: The iminosugar scaffold (A) acts as an active site ligand; the interface moiety (B) enables variation in length and consequently properties of the final product; the terminal building block (tag) (C) provides various functional groups that can be customised for further applications. Now, we present a robust synthetic method for the simple and concise composition of diversity driven N-alkylated iminosugar based inhibitors, which have the potential to be used as powerful tools such as inhibitors, pharmacological chaperones or probes for investigations of glycoprocessing enzymes in general and GCase in particular.

Synthesis
The developed synthetic concept relies on the coupling of three main building blocks, as shown in Figure 4: The iminosugar scaffold (A) acts as an active site ligand; the interface moiety (B) enables variation in length and consequently properties of the final product; the terminal building block (tag) (C) provides various functional groups that can be customised for further applications.
Since we are interested in inhibiting β-glucosidases, we employed d-gluco as well as d-xylo configured iminosugar scaffolds. The component constructed from the interface (B) and terminal building block or tag (C), including a six-carbon alkyl spacer for constant conjugation to the iminosugar (A), was intended to react either through an N-alkylation reaction or via reductive amination with the ring nitrogen of the respective iminosugar moiety. Therefore, various esters of modified benzoic acids and different ω-halogen alcohols (6-bromo-, and 6-chlorohexanol) act as suitable building blocks. The alkyl halide employed can either react directly in N-alkylation reactions, or can alternatively be converted into the corresponding aldehyde functionality through a Kornblum oxidation, as required for reductive amination reactions. As terminal tags we have chosen nitrile, azide, alkyne, nonafluoro-tert-butyl and N-dansyl functionalities which will enable follow-up chemistry as well as visualisation experiments in further investigations. Since we are interested in inhibiting β-glucosidases, we employed D-gluco as well as D-xylo configured iminosugar scaffolds. The component constructed from the interface (B) and terminal building block or tag (C), including a six-carbon alkyl spacer for constant conjugation to the iminosugar (A), was intended to react either through an N-alkylation reaction or via reductive amination with the ring nitrogen of the respective iminosugar moiety. Therefore, various esters of modified benzoic acids and different ω-halogen alcohols (6-bromo-, and 6-chlorohexanol) act as suitable building blocks. The alkyl halide employed can either react directly in N-alkylation reactions, or can alternatively be converted into the corresponding aldehyde functionality through a Kornblum oxidation, as required for reductive amination reactions. As terminal tags we have chosen nitrile, azide, alkyne, nonafluoro-tert-butyl and N-dansyl functionalities which will enable follow-up chemistry as well as visualisation experiments in further investigations.
To keep the synthetic approach as flexible as possible, the conceptual synthetic strategy has been designed to start with the central hydroxybenzoic acid motif, to which the two handles are introduced. The installation of the terminal nitrile via an ether bond can be realised by a simple Williamson ether synthesis employing 6-bromohexanenitrile and potassium carbonate with methyl hydroxybenzoate (21) to obtain compound 22 (Scheme 1). Saponification of the methyl ester leads to the previously reported benzoic acid derivative 23 [49]. Subsequent esterification under standard Mitsunobu conditions with 6-chlorohexanol, PPh3 and DIAD in THF provided compound 24. In light of the poor reactivity of alkyl chlorides in N-alkylation reactions, a second parallel series of building blocks containing a more reactive aldehyde functionality was prepared via a Kornblum reaction [50,51]. This would allow coupling to the amine via a reductive amination reaction in the event that the alkyl halide route was unsuccessful. Therefore, compound 24 was treated with sodium hydrogen carbonate and DMSO at 120 °C to give aldehyde building block 25. The nitrile can be used in orthogonal follow up chemistry such as (2+3) cycloaddition [52,53] for further labelling of the enzyme-ligand complex.  To keep the synthetic approach as flexible as possible, the conceptual synthetic strategy has been designed to start with the central hydroxybenzoic acid motif, to which the two handles are introduced. The installation of the terminal nitrile via an ether bond can be realised by a simple Williamson ether synthesis employing 6-bromohexanenitrile and potassium carbonate with methyl hydroxybenzoate (21) to obtain compound 22 (Scheme 1). Saponification of the methyl ester leads to the previously reported benzoic acid derivative 23 [49]. Subsequent esterification under standard Mitsunobu conditions with 6-chlorohexanol, PPh 3 and DIAD in THF provided compound 24. In light of the poor reactivity of alkyl chlorides in N-alkylation reactions, a second parallel series of building blocks containing a more reactive aldehyde functionality was prepared via a Kornblum reaction [50,51]. This would allow coupling to the amine via a reductive amination reaction in the event that the alkyl halide route was unsuccessful. Therefore, compound 24 was treated with sodium hydrogen carbonate and DMSO at 120 • C to give aldehyde building block 25. The nitrile can be used in orthogonal follow up chemistry such as (2+3) cycloaddition [52,53] for further labelling of the enzyme-ligand complex. Since we are interested in inhibiting β-glucosidases, we employed D-gluco as well as D-xylo configured iminosugar scaffolds. The component constructed from the interface (B) and terminal building block or tag (C), including a six-carbon alkyl spacer for constant conjugation to the iminosugar (A), was intended to react either through an N-alkylation reaction or via reductive amination with the ring nitrogen of the respective iminosugar moiety. Therefore, various esters of modified benzoic acids and different ω-halogen alcohols (6-bromo-, and 6-chlorohexanol) act as suitable building blocks. The alkyl halide employed can either react directly in N-alkylation reactions, or can alternatively be converted into the corresponding aldehyde functionality through a Kornblum oxidation, as required for reductive amination reactions. As terminal tags we have chosen nitrile, azide, alkyne, nonafluoro-tert-butyl and N-dansyl functionalities which will enable follow-up chemistry as well as visualisation experiments in further investigations.
To keep the synthetic approach as flexible as possible, the conceptual synthetic strategy has been designed to start with the central hydroxybenzoic acid motif, to which the two handles are introduced. The installation of the terminal nitrile via an ether bond can be realised by a simple Williamson ether synthesis employing 6-bromohexanenitrile and potassium carbonate with methyl hydroxybenzoate (21) to obtain compound 22 (Scheme 1). Saponification of the methyl ester leads to the previously reported benzoic acid derivative 23 [49]. Subsequent esterification under standard Mitsunobu conditions with 6-chlorohexanol, PPh3 and DIAD in THF provided compound 24. In light of the poor reactivity of alkyl chlorides in N-alkylation reactions, a second parallel series of building blocks containing a more reactive aldehyde functionality was prepared via a Kornblum reaction [50,51]. This would allow coupling to the amine via a reductive amination reaction in the event that the alkyl halide route was unsuccessful. Therefore, compound 24 was treated with sodium hydrogen carbonate and DMSO at 120 °C to give aldehyde building block 25. The nitrile can be used in orthogonal follow up chemistry such as (2+3) cycloaddition [52,53] for further labelling of the enzyme-ligand complex. Additionally, the reaction sequence can be rearranged in order to start with the installation of the ester bond in the first step followed by formation of the ether (cf. Schemes 1 and 2). To effect this the hydroxyl group of hydroxybenzoic acid was protected with a THP group under standard reaction conditions, giving compound 26 [54] which was then converted into benzoic ester 27 using 6-bromohexanol, PPh 3 and DIAD in THF. Removal of the THP group under acidic conditions provided known compound 28 [55]. Subsequent etherification under standard Mitsunobu conditions Finally, the bromine functionality was transformed into an aldehyde by a Kornblum oxidation to give compound 30. In order to evaluate different types of spacers, and to demonstrate the diversity of the presented synthetic tool box, a 2-(2-hydroxyethoxy)ethanol chain has been introduced in this interface (compound 30). The azide group allows for subsequent click chemistry tagging of the ligand-enzyme complex after incubation for labelling or quantification experiments [57].
Additionally, the reaction sequence can be rearranged in order to start with the installation of the ester bond in the first step followed by formation of the ether (cf. Scheme 1 and Scheme 2). To effect this the hydroxyl group of hydroxybenzoic acid was protected with a THP group under standard reaction conditions, giving compound 26 [54] which was then converted into benzoic ester 27 using 6-bromohexanol, PPh3 and DIAD in THF. Removal of the THP group under acidic conditions provided known compound 28 [55]. Subsequent etherification under standard Mitsunobu conditions with 2-(2-azidoethoxy)ethanol [56] gave intermediate 29. Finally, the bromine functionality was transformed into an aldehyde by a Kornblum oxidation to give compound 30. In order to evaluate different types of spacers, and to demonstrate the diversity of the presented synthetic tool box, a 2-(2-hydroxyethoxy)ethanol chain has been introduced in this interface (compound 30). The azide group allows for subsequent click chemistry tagging of the ligand-enzyme complex after incubation for labelling or quantification experiments [57]. The same approach has been followed for the introduction of a terminal alkyne group to enable orthogonal follow up click chemistry (Scheme 3). Therefore, 6-chlorohexanol has been employed in a Mitsunobu reaction with THP protected hydroxybenzoic acid 26 to yield ester 31. Removal of the THP group under acidic conditions using ion exchange resin liberated alcohol 32. Subsequent etherification with 3-bromoprop-1-yne and potassium carbonate in acetone led to structure 33. To demonstrate an additional procedure for the preparation of the desired aldehyde, the alkyl chloride was first hydrolysed to alcohol 34 using modified Kornblum reaction conditions. Therefore, compound 33 was treated with NaHCO3 in a mixture of DMSO and water at 100 °C to yield alcohol 34 which, in a second step, was oxidized with Dess Martin's reagent (DMP) to desired component 35. X-ray diffraction (XRD) studies unambiguously confirmed the structure of 35 (CCDC 2021385, see SM Figure S1).
The same approach has been followed for the introduction of a terminal alkyne group to enable orthogonal follow up click chemistry (Scheme 3). Therefore, 6-chlorohexanol has been employed in a Mitsunobu reaction with THP protected hydroxybenzoic acid 26 to yield ester 31. Removal of the THP group under acidic conditions using ion exchange resin liberated alcohol 32. Subsequent etherification with 3-bromoprop-1-yne and potassium carbonate in acetone led to structure 33. To demonstrate an additional procedure for the preparation of the desired aldehyde, the alkyl chloride was first hydrolysed to alcohol 34 using modified Kornblum reaction conditions. Therefore, compound 33 was treated with NaHCO 3 in a mixture of DMSO and water at 100 • C to yield alcohol 34 which, in a second step, was oxidized with Dess Martin's reagent (DMP) to desired component 35. X-ray diffraction (XRD) studies unambiguously confirmed the structure of 35 (CCDC 2021385, see SM Figure S1).
Additionally, the reaction sequence can be rearranged in order to start with the installation of the ester bond in the first step followed by formation of the ether (cf. Scheme 1 and Scheme 2). To effect this the hydroxyl group of hydroxybenzoic acid was protected with a THP group under standard reaction conditions, giving compound 26 [54] which was then converted into benzoic ester 27 using 6-bromohexanol, PPh3 and DIAD in THF. Removal of the THP group under acidic conditions provided known compound 28 [55]. Subsequent etherification under standard Mitsunobu conditions with 2-(2-azidoethoxy)ethanol [56] gave intermediate 29. Finally, the bromine functionality was transformed into an aldehyde by a Kornblum oxidation to give compound 30. In order to evaluate different types of spacers, and to demonstrate the diversity of the presented synthetic tool box, a 2-(2-hydroxyethoxy)ethanol chain has been introduced in this interface (compound 30). The azide group allows for subsequent click chemistry tagging of the ligand-enzyme complex after incubation for labelling or quantification experiments [57]. The same approach has been followed for the introduction of a terminal alkyne group to enable orthogonal follow up click chemistry (Scheme 3). Therefore, 6-chlorohexanol has been employed in a Mitsunobu reaction with THP protected hydroxybenzoic acid 26 to yield ester 31. Removal of the THP group under acidic conditions using ion exchange resin liberated alcohol 32. Subsequent etherification with 3-bromoprop-1-yne and potassium carbonate in acetone led to structure 33. To demonstrate an additional procedure for the preparation of the desired aldehyde, the alkyl chloride was first hydrolysed to alcohol 34 using modified Kornblum reaction conditions. Therefore, compound 33 was treated with NaHCO3 in a mixture of DMSO and water at 100 °C to yield alcohol 34 which, in a second step, was oxidized with Dess Martin's reagent (DMP) to desired component 35. X-ray diffraction (XRD) studies unambiguously confirmed the structure of 35 (CCDC 2021385, see SM Figure S1).  Next, we wanted to introduce a reporter group to allow subsequent mass spectrometric monitoring of reactions by introducing a heavy substituent, in particular a nonafluoro-tert-butoxy group (Scheme 4) [58]. Therefore, diethylenoxy benzoic acid methyl ester 36 was synthesised as described previously from methyl hydroxybenzoate (21) and 2-(2-chloroethoxy)ethanol [59]. Compound 36 underwent a Mitsunobu reaction employing nonafluoro-tert-butanol, PPh 3 and DIAD in THF to give compound 37. Follow-up chemistry for the introduction of the aldehyde at the other terminus of the hydroxybenzoic ester moiety was performed as described before. Saponification of the methyl ester provided benzoic acid derivative 38, followed by esterification under Mitsunobu conditions using 6-bromohexanol to obtain alkyl bromide 39. As seen in the 2 step conversion of the halocarbon into the corresponding aldehyde (compare Scheme 3), the bromine (compound 39) could be successfully hydrolyzed to alcohol 40 then oxidized to aldehyde 41 using Dess Martin´s reagent. Alcohol 40 can be formed starting from either the bromide or chloride of type 39. Scheme 3. Reagents and Conditions: (a) 6-Chlorohexanol, PPh3, DIAD, THF, 51%; (b) Amberlite ® IR-120H + , dioxane/H2O, 40 °C, 96%; (c) 3-bromoprop-1-yne, K2CO3, acetone, 82%; (d) NaHCO3, DMSO/H2O, 100 °C, 77%; (e) DMP, CH2Cl2, 69%.
Next, we wanted to introduce a reporter group to allow subsequent mass spectrometric monitoring of reactions by introducing a heavy substituent, in particular a nonafluoro-tert-butoxy group (Scheme 4) [58]. Therefore, diethylenoxy benzoic acid methyl ester 36 was synthesised as described previously from methyl hydroxybenzoate (21) and 2-(2-chloroethoxy)ethanol [59]. Compound 36 underwent a Mitsunobu reaction employing nonafluoro-tert-butanol, PPh3 and DIAD in THF to give compound 37. Follow-up chemistry for the introduction of the aldehyde at the other terminus of the hydroxybenzoic ester moiety was performed as described before. Saponification of the methyl ester provided benzoic acid derivative 38, followed by esterification under Mitsunobu conditions using 6-bromohexanol to obtain alkyl bromide 39. As seen in the 2 step conversion of the halocarbon into the corresponding aldehyde (compare Scheme 3), the bromine (compound 39) could be successfully hydrolyzed to alcohol 40 then oxidized to aldehyde 41 using Dess Martin´s reagent. Alcohol 40 can be formed starting from either the bromide or chloride of type 39. In order to gain access to an amino substituent at the terminal end ready for further functionalisations with for example fluorescent dyes, benzoic acid moiety 42 [60] carrying a Cbzprotected aminoethyl spacer-arm in its para position was introduced (Scheme 5). Esterification of compound 42 to compound 43 followed by oxidation, as before, yielded aldehyde 44. Assembly of the different building blocks via halides 24, 29, 33, 39 and 43, turned out to be problematic since the halides were rather unreactive, requiring elevated temperatures and extended In order to gain access to an amino substituent at the terminal end ready for further functionalisations with for example fluorescent dyes, benzoic acid moiety 42 [60] carrying a Cbz-protected aminoethyl spacer-arm in its para position was introduced (Scheme 5). Esterification of compound 42 to compound 43 followed by oxidation, as before, yielded aldehyde 44. Next, we wanted to introduce a reporter group to allow subsequent mass spectrometric monitoring of reactions by introducing a heavy substituent, in particular a nonafluoro-tert-butoxy group (Scheme 4) [58]. Therefore, diethylenoxy benzoic acid methyl ester 36 was synthesised as described previously from methyl hydroxybenzoate (21) and 2-(2-chloroethoxy)ethanol [59]. Compound 36 underwent a Mitsunobu reaction employing nonafluoro-tert-butanol, PPh3 and DIAD in THF to give compound 37. Follow-up chemistry for the introduction of the aldehyde at the other terminus of the hydroxybenzoic ester moiety was performed as described before. Saponification of the methyl ester provided benzoic acid derivative 38, followed by esterification under Mitsunobu conditions using 6-bromohexanol to obtain alkyl bromide 39. As seen in the 2 step conversion of the halocarbon into the corresponding aldehyde (compare Scheme 3), the bromine (compound 39) could be successfully hydrolyzed to alcohol 40 then oxidized to aldehyde 41 using Dess Martin´s reagent. Alcohol 40 can be formed starting from either the bromide or chloride of type 39. In order to gain access to an amino substituent at the terminal end ready for further functionalisations with for example fluorescent dyes, benzoic acid moiety 42 [60] carrying a Cbzprotected aminoethyl spacer-arm in its para position was introduced (Scheme 5). Esterification of compound 42 to compound 43 followed by oxidation, as before, yielded aldehyde 44. Assembly of the different building blocks via halides 24, 29, 33, 39 and 43, turned out to be problematic since the halides were rather unreactive, requiring elevated temperatures and extended reaction times, hence preparatively unsatisfying yields below 10% were obtained. Therefore, a reductive amination reaction between the aldehydes of the respective components 25, 30, 35, 41 and 44 and the ring nitrogen of the iminosugars 1-DNJ (1) and 1,5-dideoxy-1,5-imino-d-xylitol (DIX, 2) was employed since this method allowed smoother and faster reactions leading to far better yields compared to the N-alkylation approach (Scheme 6). The two chosen iminosugars, 1 [61] and 2 [62] were synthesised as previously described by our group.
was employed since this method allowed smoother and faster reactions leading to far better yields compared to the N-alkylation approach (Scheme 6). The two chosen iminosugars, 1 [61] and 2 [62] were synthesised as previously described by our group.
Two different reducing methods were employed for the reductive amination reaction between the aldehyde building block and the iminosugar scaffold, depending on the nature of the functional group on the phenyl linker. Reaction of the cyano group-containing building block 25 with iminosugars 1 and 2 was carried out under an atmosphere of H2 with Pd/C (10%) as catalyst (variant a, Scheme 6), yielding the corresponding D-gluco and D-xylo configured structures 45 (63%) and 46 (48%), respectively. Scheme 6. Reagents and Conditions: Overview of various reductive amination reactions and respective conditions: (a) Pd/C (10%), H2-atm., AcOH cat., MeOH; (b) NaBH3CN, AcOH cat., MeOH.
All other reductive aminations were performed with NaBH3CN in methanol (variant b, Scheme 6) to avoid reduction of other substituents present in the respective molecules. In this way compounds 47 and 48 were obtained in yields of 56% and 54%, respectively. Likewise the alkynecontaining building block 35 and NHCbz-carrying component 44 reacted with iminosugars 1 and 2 to give compounds 51 (60%), 52 (35%), 53 (55%) and 54 (27%) in the yields shown using NaBH3CN. Indeed the same method also worked well with nonafluoro-tert-butyl group reagent 41 yielding iminosugar derivatives 49 and 50 in yields of 52% and 62% respectively while the yield of compound 46 was increased up to 71% employing this methodology, compared to 48% with Pd/C and H2.
N-Cbz-Deprotection of compounds 53 and 54 was achieved using Pd/C (10%) under an atmosphere of H2, allowing reaction of the amines with dansyl chloride in the presence of triethylamine in MeOH to give the desired dansylated inhibitors of both D-gluco and D-xylo configuration, compounds 55 and 56, respectively (Scheme 7). All other reductive aminations were performed with NaBH 3 CN in methanol (variant b, Scheme 6) to avoid reduction of other substituents present in the respective molecules. In this way compounds 47 and 48 were obtained in yields of 56% and 54%, respectively. Likewise the alkyne-containing building block 35 and NHCbz-carrying component 44 reacted with iminosugars 1 and 2 to give compounds 51 (60%), 52 (35%), 53 (55%) and 54 (27%) in the yields shown using NaBH 3 CN. Indeed the same method also worked well with nonafluoro-tert-butyl group reagent 41 yielding iminosugar derivatives 49 and 50 in yields of 52% and 62% respectively while the yield of compound 46 was increased up to 71% employing this methodology, compared to 48% with Pd/C and H 2 .
N-Cbz-Deprotection of compounds 53 and 54 was achieved using Pd/C (10%) under an atmosphere of H 2, allowing reaction of the amines with dansyl chloride in the presence of triethylamine in MeOH to give the desired dansylated inhibitors of both d-gluco and d-xylo configuration, compounds 55 and 56, respectively (Scheme 7).

Biological Evaluation
Inhibition constants were determined for the interaction of each of a selected set of glycoside hydrolases (GHs) with the N-alkylated iminosugar based glycomimetics 45-56 (Table 1). With the exception of the terminal alkyne 51, all compounds were better inhibitors of the human lysosomal β-glucosidase GCase (GH30) than any of the other enzymes, including the bacterial β-glucosidase from Agrobacterium species. (Abg, GH1). In all cases, the d-gluco configured iminosugar based compounds showed better activities against both β-glucosidases GCase and Abg compared to the d-xylo analogs. None of the compounds were particularly useful inhibitors of either the human lysosomal α-galactosidase (Fabrazyme, GH27) or the β-galactosidase from Escherichia. coli (E. coli, GH2). This is not surprising since these enzymes are fairly specific galactosidases. However, all compounds turned out to inhibit the GH35 bovine liver β-galactosidase in the micromolar range. The presence of aromatic substituents caused some unexpected inhibition, most notably with α-glucosidase from Saccharomyces cerevisiae (S.cer., GH13) and β-galactosidase from E. coli likely due to adventitious interactions. This additional binding interaction is also seen with the "productive" inhibitor/enzyme combinations, with the 1-DNJ derivatives 53 (NHCbz) and 55 (NHdansyl), increasing affinity for GCase down to K i values of 22 and 18 nM, respectively. This follows a trend we have observed with most of our compounds, that the dansyl moiety contributes significantly to a better interaction with β-glucosidases [37,63]. Interestingly, β-glucosidase from Abg shows a slight preference for a shorter handle between the interface and the terminal building block as in compounds 51-56 with K i values in the nanomolar range (

Biological Evaluation
Inhibition constants were determined for the interaction of each of a selected set of glycoside hydrolases (GHs) with the N-alkylated iminosugar based glycomimetics 45-56 (Table 1). With the exception of the terminal alkyne 51, all compounds were better inhibitors of the human lysosomal βglucosidase GCase (GH30) than any of the other enzymes, including the bacterial β-glucosidase from Agrobacterium species. (Abg, GH1). In all cases, the D-gluco configured iminosugar based compounds showed better activities against both β-glucosidases GCase and Abg compared to the D-xylo analogs. None of the compounds were particularly useful inhibitors of either the human lysosomal αgalactosidase (Fabrazyme, GH27) or the β-galactosidase from Escherichia. coli (E. coli, GH2). This is not surprising since these enzymes are fairly specific galactosidases. However, all compounds turned out to inhibit the GH35 bovine liver β-galactosidase in the micromolar range. The presence of aromatic substituents caused some unexpected inhibition, most notably with α-glucosidase from Saccharomyces cerevisiae (S.cer., GH13) and β-galactosidase from E. coli likely due to adventitious interactions. This additional binding interaction is also seen with the "productive" inhibitor/enzyme combinations, with the 1-DNJ derivatives 53 (NHCbz) and 55 (NHdansyl), increasing affinity for GCase down to Ki values of 22 and 18 nM, respectively. This follows a trend we have observed with most of our compounds, that the dansyl moiety contributes significantly to a better interaction with β-glucosidases [37,63]. Interestingly, β-glucosidase from Abg shows a slight preference for a shorter handle between the interface and the terminal building block as in compounds 51-56 with Ki values in the nanomolar range (Table 1) 56 with Abg = β-glucosidase/β-galactosidase from Agrobacterium sp.; E. coli = lac Z β-galactosidase from E. coli; Bovine liv. = β-galactosidase from bovine liver; Fabrazyme = commercial recombinant human lysosomal α-galactosidase; S. cer. = α-glucosidase from Saccharomyces cerevisiae; GCase = recombinant human lysosomal β-glucocerebrosidase; N.I. = no or weak inhibition, with K i > 1 mM;

General Procedure A: (Mitsunobu Reaction)
A 10% solution of the respective starting material (1.0 equiv.) in THF, Ph 3 P (1.0 equiv.), diisopropyl azodicarboxylate (DIAD, 1.0 equiv.) and the respective alcohol (1.0 equiv.) was stirred until completed conversion of the reactants was detected. Subsequently, the reaction mixture was diluted with CH 2 Cl 2 and washed consecutively with aqueous HCl (2 N) and saturated NaHCO 3 . After drying over Na 2 SO 4 , the filtrate was concentrated under reduced pressure to provide the corresponding crude product.

General Procedure B: (Kornblum Oxidation)
Variant 1: (Conversion of a halocarbon to the corresponding aldehyde) A 10% solution of the respective halocarbon (1.0 equiv.) in DMSO was stirred with NaHCO 3 (4.0-6.0 equiv.) at 120 • C until completed conversion of the starting material was detected. After allowing the system to cool to room temperature, the reaction mixture was diluted with CH 2 Cl 2 and subsequently washed with water. The combined organic layers were dried over Na 2 SO 4 , filtered and concentrated under reduced pressure providing the corresponding crude product.
Variant 2: (Conversion of a halocarbon to the corresponding alcohol) Alternatively, the respective halocarbon was dissolved in a mixture of DMSO and water (10:1 v/v) instead of pure DMSO. The remaining protocol is identical to variant 1.

General Procedure C: (Dess-Martin Oxidation)
To a 10% solution of the respective alcohol (1 equiv.) in CH 2 Cl 2 , Dess-Martin periodinane (1.1 equiv.) was added. After completed conversion of the starting material, the reaction mixture was carefully quenched with saturated NaHCO 3 . After separation and drying over Na 2 SO 4 , the organic layers were filtered off and concentrated under reduced pressure to obtain the corresponding crude aldehyde.

General Procedure D: (Reductive Amination employing NaBH 3 CN)
A 20% solution of the respective aldehyde (1.0 equiv.) and iminosugar (1.0-1.2 equiv.) in MeOH (containing a catalytic amount of AcOH) was stirred for 15 min before NaBH 3 CN (1.5-3.0 equiv.) was added. After completed conversion of the starting materials was detected, the reaction mixture was concentrated under reduced pressure to provide the corresponding crude title compound.

Conclusions
We have developed a robust and flexible conceptual synthetic protocol towards N-alkylated iminosugar based inhibitors for glycoside hydrolases and have probed this concept on D-gluco (45, 47, 49, 51, 53, 55) and d-xylo (46, 48, 50, 52, 54, 56) configured iminosugar scaffolds. The sequence of the composition of the different building blocks allows for flexibility in choosing the spacer length and terminal tag on the non iminosugar hemisphere of the compound. Furthermore, we introduced different terminal tags such as nitrile, azide, alkyne, nonafluoro-tert-butyl and amino substituents, which allows for simple follow-up chemistry customised for different applications such as orthogonal labelling with fluorescent dyes as reporter groups or ligation reactions. The biological evaluation with a set of different glycoside hydrolases showed that all synthesised compounds proved to bind tightly to GCase with K i values in the low micro and nanomolar range. Most of them exhibit also good selectivities, thereby clearly underlining the potential of this compound class to be used as tools and therapeutics in the context of human lysosomal β-glucocerebrosidase. Potential applications are their use as enzyme inhibitors, pharmacological chaperons and active site directed ligands for enzyme labelling.