TNP Analogues Inhibit the Virulence Promoting IP3-4 Kinase Arg1 in the Fungal Pathogen Cryptococcus neoformans

New antifungals with unique modes of action are urgently needed to treat the increasing global burden of invasive fungal infections. The fungal inositol polyphosphate kinase (IPK) pathway, comprised of IPKs that convert IP3 to IP8, provides a promising new target due to its impact on multiple, critical cellular functions and, unlike in mammalian cells, its lack of redundancy. Nearly all IPKs in the fungal pathway are essential for virulence, with IP3-4 kinase (IP3-4K) the most critical. The dibenzylaminopurine compound, N2-(m-trifluorobenzylamino)-N6-(p-nitrobenzylamino)purine (TNP), is a commercially available inhibitor of mammalian IPKs. The ability of TNP to be adapted as an inhibitor of fungal IP3-4K has not been investigated. We purified IP3-4K from the human pathogens, Cryptococcus neoformans and Candida albicans, and optimised enzyme and surface plasmon resonance (SPR) assays to determine the half inhibitory concentration (IC50) and binding affinity (KD), respectively, of TNP and 38 analogues. A novel chemical route was developed to efficiently prepare TNP analogues. TNP and its analogues demonstrated inhibition of recombinant IP3-4K from C. neoformans (CnArg1) at low µM IC50s, but not IP3-4K from C. albicans (CaIpk2) and many analogues exhibited selectivity for CnArg1 over the human equivalent, HsIPMK. Our results provide a foundation for improving potency and selectivity of the TNP series for fungal IP3-4K.


Introduction
Invasive fungal diseases affect over 300 million people and cause greater than 1.5 million deaths annually around the world, matching deaths from tuberculosis and exceeding those from malaria [1,2]. Cryptococcus neoformans and drug-resistant Candida species are high on a list of a soon-to-be-released fungal priority pathogens established by the WHO. C. neoformans is an environment yeast that initially infects the lungs but has a predilection for the central nervous system where it causes meningoencephalitis. C. albicans is a component of the human mycobiota and causes candidemia.
Despite the high rates of morbidity and mortality due to invasive fungal diseases, our antifungal drug armamentarium has limitations and is confined to four classes that

RNA Extraction and cDNA Synthesis
The mRNA sequences of Arg1 in Cryptococcus neoformans (NCBI Reference Sequence: XM_012198137.1) and Ipk2 in Candida albicans (NCBI Reference Sequence: XM_709458.2) were retrieved from the NCBI database. RNA was extracted from each strain using TRIzol™ (Invitrogen) as per the protocol described in [14,39]. Briefly, YPD overnight-grown fungal cells were pelleted by centrifugation and snap-frozen in liquid nitrogen. TRIzol™ and 425-600 µm glass beads (Sigma, St. Louis, MO, USA) were added to the cell pellets, which were homogenised by bead-beating. RNA was extracted following the manufacturer's instructions, residual DNA was removed by RQ DNaseI treatment (Promega) and cDNA was synthesized using Moloney Murine Leukemia Virus Reverse Transcriptase (Promega).

Cloning of IP 3-4 K into the pGEX-6P Expression Vector
pGEX-6P_CnArg1 and pGEX-6P_CaIpk2: IP 3-4 K cDNA created above was used as a template to PCR-amplify CnArg1 using primers ARG1-BglII-s and ARG-EcoRI-a and CaIpk2 with primers CaIPK2-BamHI-s and CaIPK2-XhoI-a (see Table 1), using Invitrogen™ Platinum™ Taq High Fidelity DNA Polymerase. The PCR products were cloned into the pGEX-6P expression vector, and transformed into TOP10 competent cells. Colonies were screened by colony PCR using primers pGEX-Seq-s and pGEX-Seq-a to identify those with the correct insert. The inserts were sequenced to confirm the absence of PCR-induced mutations by comparison to the NCBI database, and that the cDNA was in-frame with the GST to ensure proper translation.
pGEX-6P_HsIPMK: human IPMK (ORF NM_152230.5) cloned into pGEX-6P. This plasmid was ordered from GenScript using their Express Cloning service including sequencing by GenScript to ensure the correct sequence was provided.

Expression and Purification of IP 3-4 K Proteins
pGEX-6P_CnArg1, pGEX-6P_CaIpk2 and pGEX-6P_HsIPMK plasmids were used to transform chemically competent BL21 (DE3) cells. Transformed BL21 (DE3) cells were used to inoculate LB-Ampicillin broth grown overnight at 37 • C with shaking. This starter culture was used to seed fresh LB-Ampicillin medium (1:200 dilution), which was then incubated at 37 • C with shaking until the OD 600 reached 0.6. IPTG (1 mM) was then added to induce protein expression overnight. Cells were harvested and the cell pellet was resuspended in GST lysis buffer (20 mM HEPES, pH 7.3, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.5% Triton X-100, 2 mM DTT, 1 mM PMSF, Roche cOmplete™ Protease Inhibitor Mini Tablets, EDTA-free), probe-sonicated and centrifuged to remove debris. The cleared lysate (supernatant) containing soluble GST-tagged fusion protein was then collected and subjected to two rounds of purification.
The first purification step involved incubating the clear lysate with Glutathione sepharose ® 4B (GE Life Sciences, Chicago, IL, USA) beads at 4 • C with gentle end-to-end rotation. The suspension was then added to an empty column for gravity flow chromatography, and unbound proteins (flow-through) were separated from bead-bound proteins. The beads were then washed 6 times with 3 column volumes of the following ice-cold buffers each time: twice with lysis buffer, twice with wash buffer 1 (50 mM Tris-HCl, pH 7.5, 500 mM NaCl, 2 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2 mM DTT) and twice with wash buffer 2 (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 2 mM DTT). For protein elution, the washed beads were incubated in elution buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 2 mM DTT, 150 µg/mL GST-HRV 3C protease) at 4 • C overnight with gentle end-to-end rotation. The eluted protein solution was then collected, and the beads were further washed with wash buffer 2 to maximise protein recovery. All the eluted protein was pooled and concentrated using an Amicon ® Ultra-15 mL Centrifugal Filter Unit (Merck Millipore, Burlington, MA, USA) with a 10 kDa molecular weight cut-off (MWCO).
The second purification step involved size exclusion chromatography (SEC) using an AKTA Fast Protein Liquid Chromatography (FPLC) system (GE Life Sciences, Chicago, IL, USA) with a HiLoad 26/600 Superdex 75 pg column (GE Life Sciences, Chicago, IL, USA). Fractions were collected, pooled and concentrated using an Amicon ® Ultra-15 mL Centrifugal Filter Unit (Merck Millipore, Burlington, MA, USA) with a 10 kDa MWCO. Enzyme purity and size were assessed by SDS-PAGE (4-12% Bis-Tris protein gel). The protein multimerization state and a more accurate molecular weight were determined using Biomolecules 2022, 12, 1526 5 of 20 SEC-MALLS (size exclusion chromatography coupled to multi angle laser light scattering (Wyatt Technology, Santa Barbara, CA, USA)).

Determination of K m and V max for ATP
The kinetic properties of the purified recombinant enzymes were determined and compared using a Kinase-Glo ® Max Luminescent Kinase Assay kit (Promega, Madison, WI, USA) in a reaction buffer consisting of 20 mM HEPES pH 6.8, 100 mM NaCl, 6 mM MgCl 2 , 20 µg/mL BSA, 1 mM DTT [14]. The IP 3 concentration was fixed at 200 µM. The ATP starting concentrations ranged between 25 µM to 400 µM and the ATP remaining after up to 10-min reaction time was assessed by adding Kinase-Glo ® reagent. An ATP standard curve (0-500 µM) was also generated using the same kit. The ATP concentration remaining was measured as luminescence using a SpectraMax iD5 plate reader. The relative luminescence unit (RLU) was converted to ATP concentration using the ATP standard curve. The reaction velocity was calculated for each starting ATP concentration and the data was fitted to the Michaelis-Menten equation using GraphPad Prism 9 to obtain V max and K m values.

Enzyme Activity and Inhibition Assay to Screen for ATP-Competitive Inhibitors
The 2,6-disubstituted purine compounds are ATP-competitive IPK inhibitors. The screening assay was therefore set up to favour screening of ATP-competitive inhibitors following the manufacturer's suggestion. This involved using 10 µM ATP and a Kinase-Glo kit that measures up to 10 µM ATP (Kinase-Glo ® Luminescent Kinase Assays, Promega, Madison, WI, USA). The amount of IP 3 and kinase used for each recombinant enzyme was determined following the manufacturer's instructions, and this concentration was used to assess the inhibitory properties of TNP analogues against each of the recombinant enzymes.
The assay was carried out in reaction buffer (20 mM HEPES pH 6.8, 100 mM NaCl, 6 mM MgCl 2 , 20 µg/mL BSA, 1 mM DTT) containing 10 µM ATP and the optimal amount of IP 3 and enzyme determined via the optimisation process in a final volume of 50 µL. The analogues were dissolved in DMSO. For the assay, the inhibitors were added to the reaction mixture to achieve a final concentration of 50 µM in 5% DMSO (1:20 dilution). To achieve inhibitor concentrations lower than 50 µM in the assay, the inhibitor stock was first serially diluted two-fold in DMSO and added to the reaction mixture, to achieve concentrations ranging between 0.8 µM and 50 µM in 5% DMSO. Once all analogues and their dilution series had been added, the reaction was started by adding an optimal amount of recombinant IP 3-4 K enzyme. The reaction was stopped after 10-min incubation at room temperature by adding a 50 µL of Kinase-Glo ® reagent. The mixture was incubated at room temperature in the dark for 10-15 min. Luminescence (RLU) was measured using a SpectraMax iD5 and an integration time of 0.5 s. Percent enzyme activity was defined as (RLU negative − RLU sample )/(RLU negative − RLU positive ) × 100. 'Positive' refers to 100% enzyme activity (no inhibitor used as the positive control), while 'negative' refers to 0% enzyme activity (no enzyme used as the negative control). The IC 50 was calculated using GraphPad Prism 9 by plotting the RLU, which indicates the ATP concentration remaining, against the concentration of TNP analogue.

Surface Plasmon Resonance (SPR) to Assess Binding Affinity of TNP Analogues
SPR was carried out on a Biacore T200 (Cytiva, Marlborough, MA, USA). A streptavidin surface was prepared by amine coupling streptavidin to a CM5 chip using standard procedures in 20 mM HEPES, 150 mM NaCl pH 7.5 at 37 • C. Briefly, the surface was activated by injection of 1:1 NHS:EDC (N-ethyl-N'-(3-(dimethylamino)propyl)carbodiimide/Nhydroxysuccinimide) followed by a 7-min injection of 100 µg/mL streptavidin in 10 mM sodium acetate (pH 4.5) at a flow rate of 2 µL/min. Unreacted groups on the surface were blocked by injection of 1 M ethanolamine (pH 8.0). Purified recombinant CnArg1 was N-terminally biotinylated and immobilised to a level of~4000 RU onto the streptavidin surface at 25 • C in 20 mM HEPES, 200 mM NaCl, 5 mM MgCl 2 (pH 7.5) at a flow rate of 2 µL/min. For analysis, compounds were prepared to 50 mM in 100% DMSO and diluted to the desired concentration in SPR running buffer; 20 mM HEPES (pH 7.5), 200 mM NaCl, 5 mM MgCl 2 , 5% DMSO. Analysis was carried out using multi-cycle kinetics over a compound concentration range of 6.25-200 µM in running buffer at 10 • C, with an association time of 60 s and a dissociation time of 120 s. Data were reference subtracted, and a solvent correction applied. Analysis was carried out using Biacore T200 Evaluation Software and all data fit to a 1:1 Langmuir binding isotherm.
2.8. Synthesis of 2,6-Disubstituted Purine Analogues 2.8.1. General Experimental Procedure All reagents were purchased and used without further purification. Silica gel was used for column chromatography purification. 1 H NMR and 13 C NMR spectra were collected on a Bruker Advance III Nanobay 400 MHz spectrometer ( 1 H at 400 MHz and 13 C at 100 MHz). All spectra were processed using MestReNova 11.0 software. The chemical shifts of 1 H and 13 C are reported in parts per million (ppm) and were measured relative to the expected chemical shifts of the NMR solvents; CDCl 3 , 7.26 (77.16 for 13 C NMR) CD 3 OD, 3.31 (49.00 for 13 C NMR) and DMSO-d 6 , 2.50 (39.52 for 13 C NMR). The format used to report the spectra was as follows: chemical shift (multiplicity, coupling constant (if applicable), integration). Multiplicity was defined as: s = singlet, d = doublet, t = triplet, q = quartet, sd = singlet of doublets, dd = doublet of doublets, dt = doublet of triplets, tt = triplet of triplets and m = multiplet. Apparent splitting was abbreviated as app. and a broad resonance was abbreviated as br. Coupling constants were reported as J in Hertz (Hz).
All analytical HPLC analyses were done on an Agilent 1260 Infinity Analytical HPLC coupled with a 1260 Degasser: G1322A, 1260 Binary Pump: G1312B, 1260 HiP ALS autosampler: G1367E, 1260 TCC: G1316A and 1260 DAD detector: G4212B. The column used was a Zorbax Eclipse Plus C18 Rapid Resolution 4.6 × 100 mm 3.5-micron. The sample injection volume was 2 µL which was run in 0.1% TFA in acetonitrile at a gradient of 5-100% over 10 min with a flow rate of 1 mL/min. Detection methods were with 214 nm and 254 nm.
All HRMS analyses were done on an Agilent 6224 TOF LC/MS Mass Spectrometer coupled to an Agilent 1290 Infinity (Agilent, Palo Alto, CA, USA). All data were acquired, and reference mass corrected via a dual-spray electrospray ionization (ESI) source. Each scan or data point on the Total Ion Chromatogram (TIC) is an average of 13,700 transients, producing a spectrum every second. Mass spectra were created by averaging the scans across each peak and background subtracted against the first 10 s of the TIC. Acquisition was performed using the Agilent Mass Hunter Fata Acquisition software version B.05.00 Build 5.0.5042.2 and analysis was performed using Mass Hunter Qualitative Analysis version B.05.00 Build 5.0.519.13.

General Procedure I for the First Nucleophilic Aromatic Substitution in the 6-Position (2a and 2b)
To a solution of 6-chloro-2-fluoropurine (1 mmol, 1.00 equiv.) in n-butanol (4 mL), was added N,N-diisopropylethylamine (1.41 equiv.). The mixture was stirred at room temperature for 5 min. To the mixture, was added an amine (1.02 equiv.). The mixture was warmed to 65 • C and stirred for 4 h. Then, the reaction mixture was concentrated in vacuo and to the residue, was added cold water. The precipitate was filtered, washed with cold water and purified. To a solution of a N6-substituted-2-fluoro-9H-purin-6-amine (1 mmol, 1.00 equiv.) in n-butanol (4 mL), was added N,N-diisopropylethylamine (2.20 equiv.). The mixture was stirred at room temperature for 5 min. To the mixture, was added an amine (2.00 equiv.). The mixture was heated to reflux and stirred overnight (16 h). Then, the reaction mixture was concentrated in vacuo and purified. To a solution of 2-amino-6-chloropurine (1 mmol, 1.00 equiv.) in ethyl acetate (4 mL), was added a benzaldehyde (1.20 equiv.). The mixture was cooled to 0 • C and was added trifluoroacetic acid (2.24 equiv.) and sodium triacetoxyborohydride (1.50 equiv.). The reaction was warmed to room temperature and stirred overnight (16 h). The mixture was quenched with 10% aq. NaOH solution (20 mL) to pH~8-9 and extracted with ethyl acetate (3 × 20 mL). The organic extracts were combined, washed with brine (50 mL) and dried over MgSO 4 . Then, it was concentrated in vacuo and purified.
Additional synthetic methods and spectroscopic data for individual compounds is provided in the Supplementary Materials S1.

Results
3.1. Purification of IP 3-4 K from C. neoformans, C. albicans and Human IP 3-4 K from C. neoformans (CnArg1) and C. albicans (CaIpk2) and HsIPMK were expressed from pGEX-6P in E. coli as GST fusion proteins and purified in a two-step process involving glutathione-affinity chromatography, followed by GST cleavage by HRV 3C protease and size exclusion chromatography. SDS-PAGE indicated CnArg1 and CaIpk2 have molecular weights close to the predicted values of 49 kDa and 40 kDa, respectively, and a purity of >95% ( Figure 1A). SEC-MALLS was used to assess the oligomeric state and was consistent with both IP 3-4 K enzymes being monomeric. The solution molecular weights of CnArg1 and CaIpk2 were calculated as 52 kDa and 43 kDa, respectively (i.e., within 10% of the predicted monomer molecular weight) ( Figure 1B). From SDS-PAGE, the molecular weight of HsIPMK was~47 kDa ( Figure 1A), in agreement with the predicted molecular weight of 47.2 kDa [40]. HsIPMK purity was~60%. Reduced HsIPMK purity could be due to limited expression levels and contamination with bacterial heat shock proteins. HsIPMK was previously shown to be monomeric [21,22,40,41]. The purification yielded~3.6 mg per litre of CnArg1,~8.8 mg per litre of CaIpk2 and~0.3 mg per litre of HsIPMK proteins.

Establishing the Kinetic Properties of IP 3-4 K
A luminescence assay was optimised from [14] to determine the kinetic properties of CnArg1 and CaIpk2 using IP 3 as the substrate. For CnArg1 and CaIpk2, the K m was 300 ± 67 µM and 213 ± 49 µM, and the V max was 12 ± 1.5 and 21 ± 3 µmol ATP/mg protein/min, respectively ( Figure 2). Using the same assay conditions, HsIPMK had a lower K m (128 ± 35 µM) compared to the fungal IP 3-4 K enzymes and a lower V max (0.6 ± 0.07 µmol ATP/mg protein/min) (Figure 2), consistent with a higher affinity for ATP. This compares to a previously determined K m for ATP of 61 ± 6 µM when PIP 2 was used as the substrate [22] and 10 µM when IP 4 was used as the substrate [31]. The K m of rat IPMK for ATP was previously determined to be 64 µM when IP 3 was used as the substrate [36].

Establishing the Kinetic Properties of IP3-4K
A luminescence assay was optimised from [14] to determine the kinetic properties of CnArg1 and CaIpk2 using IP3 as the substrate. For CnArg1 and CaIpk2, the Km was 300 ± 67 µM and 213 ± 49 µM, and the Vmax was 12 ± 1.5 and 21 ± 3 µmol ATP/mg protein/min, respectively ( Figure 2). Using the same assay conditions, HsIPMK had a lower Km (128 ± 35 µM) compared to the fungal IP3-4K enzymes and a lower Vmax (0.6 ± 0.07 µmol ATP/mg protein/min) (Figure 2), consistent with a higher affinity for ATP. This compares to a previously determined Km for ATP of 61 ± 6 µM when PIP2 was used as the substrate [22] and 10 µM when IP4 was used as the substrate [31]. The Km of rat IPMK for ATP was previously determined to be 64 µM when IP3 was used as the substrate [36].

Optimizing the Enzyme Assay and Determining the IC50 of TNP
Next, we determined the inhibitory properties of TNP. TNP is an ATP-competitive inhibitor [34]. To bias the assay toward ATP-competitive inhibitors, the ATP concentra-

Optimizing the Enzyme Assay and Determining the IC 50 of TNP
Next, we determined the inhibitory properties of TNP. TNP is an ATP-competitive inhibitor [34]. To bias the assay toward ATP-competitive inhibitors, the ATP concentration was reduced from 500 µM to 10 µM, which is a value lower than the K m of all IP 3-4 K enzymes ( Figure 2). IP 3 and enzyme concentrations were also optimised following the manufacturer's instructions. Briefly, this involved measuring ATP consumption using a fixed saturated amount of kinase and varying the concentration of IP 3 from 0 to 200 µM. The IP 3 concentration that resulted in the largest change in luminescence was chosen as the optimal IP 3 concentration. This optimal amount of IP 3 was then used to test varying concentrations of the kinase ranging from 0 to 20 ng/µL. The amount of kinase that produced luminescence values in the linear range of the kinase titration curve was deemed to be the optimal kinase amount. The optimised concentrations of enzyme assay components are summarised in Table 2. Using the optimized assay conditions (Table 2), TNP was found to inhibit CnArg1 and HsIPMK with an IC 50 of 21 ± 6 µM and 7 ± 0.5 µM, respectively ( Figure 3). To our knowledge, this is the first report that TNP inhibits HsIPMK. Interestingly, TNP did not inhibit CaIpk2.  Reactions were carried out at room temperature for 10 min using 10 µM ATP and the optimal amount of enzyme and IP3 as summarized in Table 2. ATP consumption was calculated from the luminescence reading and expressed as percentage enzyme activity according to the methods. The IC50 curves were generated by plotting % enzyme activity against the concentration of TNP. For CnArg1, the IC50 of 21 ± 6 µM (SEM) was calculated from 2 biological replicates, each performed in technical triplicate, error bar = SD. For HsIPMK, the IC50 of 7 ± 0.5 µM (SEM) was calculated from 3 technical replicates, error bar = SD. For CaIpk2, no inhibitory activity was found up to 50 µM (n = 1).

Synthesis of 2,6-Disubstituted Purine Analogues
TNP analogues were synthesized using multiple strategies that expedited the inclusion of substituents at the N2-and N6-positions of the purine scaffold (Tables 3-5). In addition, modification at N9 was included in examples 45 and 46 ( Table 6). Note that several of these compounds were previously reported in studies of human IPKs as well as other enzyme targets [34,35,37,42,43]. Table 3. Summary of inhibitory properties resulting from chemical exploration at the N2 position of the TNP purine core following removal of the NO2 group at the N6 position (21 analogues). Only potent and soluble analogues with an IC50 < 40 µM for CnArg1 were tested for inhibitory properties against HsIPMK. ND = not determined. Reactions were carried out at room temperature for 10 min using 10 µM ATP and the optimal amount of enzyme and IP 3 as summarized in Table 2. ATP consumption was calculated from the luminescence reading and expressed as percentage enzyme activity according to the methods. The IC 50 curves were generated by plotting % enzyme activity against the concentration of TNP. For CnArg1, the IC 50 of 21 ± 6 µM (SEM) was calculated from 2 biological replicates, each performed in technical triplicate, error bar = SD. For HsIPMK, the IC 50 of 7 ± 0.5 µM (SEM) was calculated from 3 technical replicates, error bar = SD. For CaIpk2, no inhibitory activity was found up to 50 µM (n = 1).

Synthesis of 2,6-Disubstituted Purine Analogues
TNP analogues were synthesized using multiple strategies that expedited the inclusion of substituents at the N2and N6-positions of the purine scaffold (Tables 3-5). In addition, modification at N9 was included in examples 45 and 46 ( Table 6). Note that several of these compounds were previously reported in studies of human IPKs as well as other enzyme targets [34,35,37,42,43]. Table 3. Summary of inhibitory properties resulting from chemical exploration at the N2 position of the TNP purine core following removal of the NO 2 group at the N6 position (21 analogues). Only potent and soluble analogues with an IC 50 < 40 µM for CnArg1 were tested for inhibitory properties against HsIPMK. ND = not determined. enzyme activity against the concentration of TNP. For CnArg1, the IC50 of 21 ± 6 µM (SEM) was calculated from 2 biological replicates, each performed in technical triplicate, error bar = SD. For HsIPMK, the IC50 of 7 ± 0.5 µM (SEM) was calculated from 3 technical replicates, error bar = SD. For CaIpk2, no inhibitory activity was found up to 50 µM (n = 1).

Synthesis of 2,6-Disubstituted Purine Analogues
TNP analogues were synthesized using multiple strategies that expedited the inclusion of substituents at the N2-and N6-positions of the purine scaffold (Tables 3-5). In addition, modification at N9 was included in examples 45 and 46 ( Table 6). Note that several of these compounds were previously reported in studies of human IPKs as well as other enzyme targets [34,35,37,42,43]. Table 3. Summary of inhibitory properties resulting from chemical exploration at the N2 position of the TNP purine core following removal of the NO2 group at the N6 position (21 analogues). Only potent and soluble analogues with an IC50 < 40 µM for CnArg1 were tested for inhibitory properties against HsIPMK. ND = not determined. CaIpk2, no inhibitory activity was found up to 50 µM (n = 1).

Synthesis of 2,6-Disubstituted Purine Analogues
TNP analogues were synthesized using multiple strategies that expedited the inclusion of substituents at the N2-and N6-positions of the purine scaffold (Tables 3-5). In addition, modification at N9 was included in examples 45 and 46 ( Table 6). Note that several of these compounds were previously reported in studies of human IPKs as well as other enzyme targets [34,35,37,42,43]. Table 3. Summary of inhibitory properties resulting from chemical exploration at the N2 position of the TNP purine core following removal of the NO2 group at the N6 position (21 analogues). Only potent and soluble analogues with an IC50 < 40 µM for CnArg1 were tested for inhibitory properties against HsIPMK. ND = not determined.

Synthesis of 2,6-Disubstituted Purine Analogues
TNP analogues were synthesized using multiple strategies that expedited the inclusion of substituents at the N2-and N6-positions of the purine scaffold (Tables 3-5). In addition, modification at N9 was included in examples 45 and 46 ( Table 6). Note that several of these compounds were previously reported in studies of human IPKs as well as other enzyme targets [34,35,37,42,43]. Table 3. Summary of inhibitory properties resulting from chemical exploration at the N2 position of the TNP purine core following removal of the NO2 group at the N6 position (21 analogues). Only potent and soluble analogues with an IC50 < 40 µM for CnArg1 were tested for inhibitory properties against HsIPMK. ND = not determined.

Synthesis of 2,6-Disubstituted Purine Analogues
TNP analogues were synthesized using multiple strategies that expedited the inclusion of substituents at the N2-and N6-positions of the purine scaffold (Tables 3-5). In addition, modification at N9 was included in examples 45 and 46 ( Table 6). Note that several of these compounds were previously reported in studies of human IPKs as well as other enzyme targets [34,35,37,42,43]. Table 3. Summary of inhibitory properties resulting from chemical exploration at the N2 position of the TNP purine core following removal of the NO2 group at the N6 position (21 analogues). Only potent and soluble analogues with an IC50 < 40 µM for CnArg1 were tested for inhibitory properties against HsIPMK. ND = not determined.

CH3 Not inhibitory ND ND
For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed. 24 ± 1 82 ± 9 3.4

42
Biomolecules 2022, 12, x FOR PEER REVIEW 13 of 21  For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed. 10 ± 0.5 37 ± 2 3.7

43
Biomolecules 2022, 12, x FOR PEER REVIEW 13 of 21  For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed. 27    For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed. 82 ± 13 ND ND Table 6. Exploration at the N9 position on the TNP purine core abolished inhibition against CnArg1. ND = not determined.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.

46
Biomolecules 2022, 12, x FOR PEER REVIEW 13 of 21  For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.   For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed.

CH 3 Not inhibitory ND ND
For fixed N6-amine substitutions, 2-fluoro-6-chloro-9H-purine, 1 was subjected to consecutive nucleophilic aromatic substitutions (Scheme 1) [37,44,45]. While substitution of 1 occurs in the 6-position preferentially, giving 2 [46], bis-substitution was a common observation initially. Monitoring the temperature and length of the first nucleophilic aromatic substitution reaction gave improved outcomes. The second substitution at the 2-position yielded the target compounds, 3 but also proved to be challenging, with reactions requiring high reagent concentrations to proceed. In order to diversify the 6-position conveniently, an alternate synthetic route was adopted, where the group on the 2-position was introduced first. This has been achieved previously through the use of protecting groups or modification of the 6-chloro group to bypass the more reactive groups [28,37]. However, we pursued a more direct route by a reductive alkylation of 2-amino-6-chloro-9H-purine (Scheme 2). This was then followed by the nucleophilic aromatic substitution by various amines at the 6-chloro position to give 5. Reductive alkylation in the 2-position has previously been demonstrated [47,48], but to our knowledge, reductive alkylation in the 6-position has not been reported before. Substitution of 5 with various amines gave target compounds 6. Scheme 2. General synthesis of 6 from 4 via 5. i, Aldehyde-R1, NaBH(OAc)3, TFA, EtOAc, room temp., 16 h, 50-64%; ii, NHR2R3, triethylamine, n-BuOH, reflux, 5 h, 46-97%.
Lastly, the impact of an alkyl group in the 9-position was explored. This was accomplished through alkylation of 7 using an alkyl halide to give 8 (Scheme 3).

Assessment of the Inhibitory Properties of 2,6-Disubstituted Purine Analogues
Using the assay conditions established for TNP, all analogues were assessed for inhibition of CnArg1, HsIPMK and CaIpk2. We found that most analogues inhibited CnArg1 with IC50 values ranging from 10-100 μM (summarised in Tables 3-6), but none inhibited CaIpk2 (see supplementary Figure S1). Compound structures and their IC50 values against CnArg1 and HsIPMK are summarised in Tables 3-6. Only a selected number of analogues were tested for inhibitory activity against HsIPMK, with a focus on those with similar or better potencies than TNP (i.e., those with an IC50 less than 40 µM). Note that the reduced potencies of some of these compounds could be due to poor solubility in water although this was not investigated further. Enzyme inhibition assays could not be performed on 11, 18 and 19 as they were poorly soluble in 5% DMSO.
In order to diversify the 6-position conveniently, an alternate synthetic route was adopted, where the group on the 2-position was introduced first. This has been achieved previously through the use of protecting groups or modification of the 6-chloro group to bypass the more reactive groups [28,37]. However, we pursued a more direct route by a reductive alkylation of 2-amino-6-chloro-9H-purine (Scheme 2). This was then followed by the nucleophilic aromatic substitution by various amines at the 6-chloro position to give 5. Reductive alkylation in the 2-position has previously been demonstrated [47,48], but to our knowledge, reductive alkylation in the 6-position has not been reported before. Substitution of 5 with various amines gave target compounds 6. In order to diversify the 6-position conveniently, an alternate synthetic route was adopted, where the group on the 2-position was introduced first. This has been achieved previously through the use of protecting groups or modification of the 6-chloro group to bypass the more reactive groups [28,37]. However, we pursued a more direct route by a reductive alkylation of 2-amino-6-chloro-9H-purine (Scheme 2). This was then followed by the nucleophilic aromatic substitution by various amines at the 6-chloro position to give 5. Reductive alkylation in the 2-position has previously been demonstrated [47,48], but to our knowledge, reductive alkylation in the 6-position has not been reported before. Substitution of 5 with various amines gave target compounds 6.
Lastly, the impact of an alkyl group in the 9-position was explored. This was accomplished through alkylation of 7 using an alkyl halide to give 8 (Scheme 3).

Assessment of the Inhibitory Properties of 2,6-Disubstituted Purine Analogues
Using the assay conditions established for TNP, all analogues were assessed for inhibition of CnArg1, HsIPMK and CaIpk2. We found that most analogues inhibited CnArg1 with IC50 values ranging from 10-100 μM (summarised in Tables 3-6), but none inhibited CaIpk2 (see supplementary Figure S1). Compound structures and their IC50 values against CnArg1 and HsIPMK are summarised in Tables 3-6. Only a selected number of analogues were tested for inhibitory activity against HsIPMK, with a focus on those with similar or better potencies than TNP (i.e., those with an IC50 less than 40 µM). Note that the reduced potencies of some of these compounds could be due to poor solubility in water although this was not investigated further. Enzyme inhibition assays could not be performed on 11, 18 and 19 as they were poorly soluble in 5% DMSO.
Lastly, the impact of an alkyl group in the 9-position was explored. This was accomplished through alkylation of 7 using an alkyl halide to give 8 (Scheme 3). In order to diversify the 6-position conveniently, an alternate synthetic route was adopted, where the group on the 2-position was introduced first. This has been achieved previously through the use of protecting groups or modification of the 6-chloro group to bypass the more reactive groups [28,37]. However, we pursued a more direct route by a reductive alkylation of 2-amino-6-chloro-9H-purine (Scheme 2). This was then followed by the nucleophilic aromatic substitution by various amines at the 6-chloro position to give 5. Reductive alkylation in the 2-position has previously been demonstrated [47,48], but to our knowledge, reductive alkylation in the 6-position has not been reported before. Substitution of 5 with various amines gave target compounds 6. Scheme 2. General synthesis of 6 from 4 via 5. i, Aldehyde-R1, NaBH(OAc)3, TFA, EtOAc, room temp., 16 h, 50-64%; ii, NHR2R3, triethylamine, n-BuOH, reflux, 5 h, 46-97%.
Lastly, the impact of an alkyl group in the 9-position was explored. This was accomplished through alkylation of 7 using an alkyl halide to give 8 (Scheme 3).

Assessment of the Inhibitory Properties of 2,6-Disubstituted Purine Analogues
Using the assay conditions established for TNP, all analogues were assessed for inhibition of CnArg1, HsIPMK and CaIpk2. We found that most analogues inhibited CnArg1 with IC50 values ranging from 10-100 μM (summarised in Tables 3-6), but none inhibited CaIpk2 (see supplementary Figure S1). Compound structures and their IC50 values against CnArg1 and HsIPMK are summarised in Tables 3-6. Only a selected number of analogues were tested for inhibitory activity against HsIPMK, with a focus on those with similar or better potencies than TNP (i.e., those with an IC50 less than 40 µM). Note that the reduced potencies of some of these compounds could be due to poor solubility in water although this was not investigated further. Enzyme inhibition assays could not be performed on 11, 18 and 19 as they were poorly soluble in 5% DMSO.
We first identified that 9, which lacked the nitro group of TNP, had a similar potency to TNP against CnArg1, but slightly reduced potency against HsIPMK. 9 had previously been described by [34,37] during their studies around IP3K and IP6K, respectively. They Scheme 3. Synthesis of 8 via 7. i, R 3 X, 68-85%.

Assessment of the Inhibitory Properties of 2,6-Disubstituted Purine Analogues
Using the assay conditions established for TNP, all analogues were assessed for inhibition of CnArg1, HsIPMK and CaIpk2. We found that most analogues inhibited CnArg1 with IC 50 values ranging from 10-100 µM (summarised in Tables 3-6), but none inhibited CaIpk2 (see supplementary Figure S1). Compound structures and their IC 50 values against CnArg1 and HsIPMK are summarised in Tables 3-6. Only a selected number of analogues were tested for inhibitory activity against HsIPMK, with a focus on those with similar or better potencies than TNP (i.e., those with an IC 50 less than 40 µM). Note that the reduced potencies of some of these compounds could be due to poor solubility in water although this was not investigated further. Enzyme inhibition assays could not be performed on 11, 18 and 19 as they were poorly soluble in 5% DMSO.
Midway through this study, a research team at Johns Hopkins University published their strategy of first improving TNP solubility by replacing the N6-nitrobenzyl group with a methoxyethyl chain, before making further chemical modifications [37]. We employed the same strategy to explore the N2-position of the purine core, synthesizing a total of 11 analogues with the methoxyethyl chain at the N6-position (Table 4). It was reported that 30 had improved IP6K inhibition compared to TNP [37] but this was not observed with either CnArg1 or HsIPMK. Swapping benzylamine substitution at N2 gave a range of outcomes. Arylamine substitution as 37 and 40 was tolerated (IC 50 of 17 ± 2 µM and 14 ± 0.5 µM, respectively).
Further variations at the N6-position were also considered ( Table 5). One of these, 42, which had a 4-methyltetrahydropyran group, produced the most potent inhibition against CnArg1 with an IC 50 of 10 ± 0.5 µM. Substitution in the 9-position was not tolerated in the two analogues, 45 and 46, which did not have any inhibitory activity for CnArg1 (Table 6).

Assessing Binding Affinities of TNP and Its Analogues
To complement the enzyme inhibition data, the binding affinities (K D ) of TNP analogues were determined for CnArg1 using surface plasmon resonance (SPR) and the results are summarised in Table 7 (see Figure S2 for representative SPR sensorgrams and doseresponse curves). A K D could not be determined for TNP and its closest analogue 9, due to poor behaviour on the SPR chip, which is most likely attributable to poor aqueous solubility. Chemical modification of TNP and 9 improved compound properties, allowing the binding affinities of 16 analogues to be determined. Apart from 37, these analogues bound to CnArg1, confirming a correlation between binding and inhibition. With the exception of 39 and 40, the analogues generally had comparable IC 50 and K D values. Table 7. Summary to the binding affinities (K D ) of TNP and its analogues to CnArg1 as determined by surface plasmon resonance (SPR). Analogue binding to N-terminally biotinylated CnArg1 immobilised onto a streptavidin chip was assessed between 6.25-200 µM at 10 • C in 20 mM HEPES, 150 mM NaCl, pH 7.5. The IC 50 and K D values are presented as mean ± SEM (IC 50 -n ≥ 3 technical replicates; K D -n ≥ 1). ND = not determined.

Comparing the Selectivity of TNP and Its Analogues for CnArg1
Using the IC50 data for TNP and the seventeen TNP analogues with the lowest IC50 against CnArg1, the HsIPMK/CnArg1 IC50 ratios were calculated (see last column in Tables  3-6) and plotted (Figure 4). Fourteen TNP analogues were found to be more selective for CnArg1 (ratio higher than 1). While unable to identify significantly more potent analogues of CnArg1 than TNP, the collected assessment of enzyme inhibitory activities did show that the relative selectivity over the human orthologue could be influenced by altered substitutions, with the selectivity ratios changing 20-fold in going from TNP (3-CF3 substitution) to 13 (3-OCH3 substitution) ( Table 3).

Discussion
IP3-4K is critical for the virulence of C. neoformans and C. albicans. An IP3-4K (CnArg1) deletion mutant failed to either grow at 37 °C or establish an infection in a mouse model [14]. The mutant was also defective in producing a plethora of virulence-related phenotypes including capsule, melanin and phospholipase B, was more readily phagocytosed by blood-derived monocytes, had a cell wall defect and reduced metabolic flexibility and could not upregulate phosphate acquisition machinery in response to phosphate deprivation [13][14][15][16]49]. Attempts to delete both alleles of IP3-4 kinase (CaIpk2) in the diploid yeast, C. albicans, were unsuccessful, suggesting that IP3-4K is an essential protein. Using a knockdown approach, CaIpk2 was shown to be essential for hyphal development, secretion of degradative enzymes and survival inside macrophages [17]. Targeting fungal IP3-4K is therefore a promising strategy for developing a novel class of antifungal drug. Fungal IP3-4K is also an attractive drug target due to the redundancy that exists in the analogous IP3 to IP5 conversion step in the human pathway.
Our SAR studies revealed that most TNP analogues inhibited CnArg1, but not CaIpk2, and that many TNP analogues were more selective for CnArg1 over HsIPMK (summarised in Figure 4). This suggests that potential differences exist in the active site of each kinase. Like many IPK enzymes, the protein structures of CnArg1 and CaIpk2 have not been experimentally determined. To investigate possible structural differences among the three IP3-4K proteins that could explain our results, we utilized the AlphaFold-predicted structures for CnArg1 (UniProt: J9W3G0) and CaIpk2 (UniProt: Q59YE9) [50,51]

Discussion
IP 3-4 K is critical for the virulence of C. neoformans and C. albicans. An IP 3-4 K (CnArg1) deletion mutant failed to either grow at 37 • C or establish an infection in a mouse model [14]. The mutant was also defective in producing a plethora of virulence-related phenotypes including capsule, melanin and phospholipase B, was more readily phagocytosed by blood-derived monocytes, had a cell wall defect and reduced metabolic flexibility and could not upregulate phosphate acquisition machinery in response to phosphate deprivation [13][14][15][16]49]. Attempts to delete both alleles of IP 3-4 kinase (CaIpk2) in the diploid yeast, C. albicans, were unsuccessful, suggesting that IP 3-4 K is an essential protein. Using a knockdown approach, CaIpk2 was shown to be essential for hyphal development, secretion of degradative enzymes and survival inside macrophages [17]. Targeting fungal IP 3-4 K is therefore a promising strategy for developing a novel class of antifungal drug. Fungal IP 3-4 K is also an attractive drug target due to the redundancy that exists in the analogous IP 3 to IP 5 conversion step in the human pathway.
Our SAR studies revealed that most TNP analogues inhibited CnArg1, but not CaIpk2, and that many TNP analogues were more selective for CnArg1 over HsIPMK (summarised in Figure 4). This suggests that potential differences exist in the active site of each kinase. Like many IPK enzymes, the protein structures of CnArg1 and CaIpk2 have not been experimentally determined. To investigate possible structural differences among the three IP 3-4 K proteins that could explain our results, we utilized the AlphaFold-predicted structures for CnArg1 (UniProt: J9W3G0) and CaIpk2 (UniProt: Q59YE9) [50,51] and compared them to the published HsIPMK crystal structure consisting of amino acid residues 50 to 416 in a complex with ADP (PDB: 5W2H) [21,22]. The HsIPMK protein used to obtain the crystal structure had residues 263-377 deleted, which included a nuclear localisation signal.
Both the fungal IP 3-4 kinases share only a~20% sequence homology with HsIPMK. However, an overlay of the AlphaFold models of CnArg1 and CaIpk2 with HsIPMK (PDB: 5W2H) revealed a conserved catalytic core consisting of the three characteristic domains: an N-terminal α + β domain (N-lobe), a C-terminal α + β domain (C-lobe) and an inositol binding domain [21,22]. The N-and C-terminal domains each contain one anti-parallel beta sheet where ADP is sandwiched in between, making up the ATP binding site ( Figure 5A,B). Seven out of twelve of the amino acids in the HsIPMK active site that interact with ATP are conserved in CnArg1 and CaIpk2 ( Figure 5C,D) and represent Pro 111 , Leu 254 , Asp 144 , Ile 384 , Asp 385 and Lys 75 ( Figure 5E). In HsIPMK, Leu 254 and Ile 384 form van der Waals interactions with each other and Asp 144 forms a hydrogen bond with the ribose group of ADP. Asp 385 interacts with the two magnesium ions, which in turn contact the two phosphate groups of ADP [21]. Ile 384 and Asp 385 are part of an Ile-Asp-Phe tripeptide, which is conserved in the IPK family and known to interact with a metal cofactor [11,21,52]. Lys 75 forms a salt bridge with the alpha-phosphate of ADP [21].
Both the fungal IP3-4 kinases share only a ~20% sequence homology with HsIPMK. However, an overlay of the AlphaFold models of CnArg1 and CaIpk2 with HsIPMK (PDB: 5W2H) revealed a conserved catalytic core consisting of the three characteristic domains: an N-terminal α + β domain (N-lobe), a C-terminal α + β domain (C-lobe) and an inositol binding domain [21,22]. The N-and C-terminal domains each contain one anti-parallel beta sheet where ADP is sandwiched in between, making up the ATP binding site ( Figure  5A,B). Seven out of twelve of the amino acids in the HsIPMK active site that interact with ATP are conserved in CnArg1 and CaIpk2 ( Figure 5C,D) and represent Pro 111 , Leu 254 , Asp 144 , Ile 384 , Asp 385 and Lys 75 ( Figure 5E). In HsIPMK, Leu 254 and Ile 384 form van der Waals interactions with each other and Asp 144 forms a hydrogen bond with the ribose group of ADP. Asp 385 interacts with the two magnesium ions, which in turn contact the two phosphate groups of ADP [21]. Ile 384 and Asp 385 are part of an Ile-Asp-Phe tripeptide, which is conserved in the IPK family and known to interact with a metal cofactor [11,21,52]. Lys 75 forms a salt bridge with the alpha-phosphate of ADP [21].  [3][4] Ks to determine conservation of active site residues. ADP-bound HsIPMK structure (PDB:5W2H) showing the N-lobe in orange, C-lobe in yellow, hinge region in green and inositol binding domain (IP-domain) in blue is superimposed with the AlphaFold models of CnArg1 (cyan) (A) and CaIpk2 (pink) (B). Surface representation (in grey) of the ATP-binding pocket of CnArg1 (C) and CaIpk2 (D), with residues identical to those of HsIPMK in blue and residues non-identical to those of HsIPMK in yellow. Stick models depicting all residues involved in ATP binding (E) are depicted with CnArg1 in cyan, CaIpk2 in pink and HsIPMK coloured green for carbon, blue for nitrogen and red for oxygen. Residues that differ among the three IP 3-4 K proteins being circled. For all panels, ADP is shown as stick representation with green for carbon, blue for nitrogen, red for oxygen and orange for phosphorus, and magnesium ions are depicted as violet spheres.
Differences between the fungal IP [3][4] Ks and HsIPMK ( Figure 5E, circled residues) are the replacement of the negatively charged Asp 132 in HsIPMK with the polar asparagine in CnArg1 (Asn 100 ) and CaIpk2 (Asn 102 ) and replacement of Glu 131 in HsIPMK with Ala 99 in CnArg1 and Ser 101 in CaIpk2. In HsIPMK, Glu 131 and hydrophobic Val 133 form hydrogen bonds with the N6 and N1 atoms of adenine, respectively. Residues within the ATP binding site that are unique to CaIpk2 are Phe 28 , which is replaced by valine in HsIPMK (Val 73 ) and CnArg1 (Val 32 ). In HsIPMK, Val 73 makes a van der Waals interaction with the adenine group of ADP [21]. The aromatic group on Phe 28 could potentially interfere with the binding of TNP and its analogues. Other differences are Ser 103 in CaIpk2 replacing Val 133 (in HsIPMK) and Leu 101 (in CnArg1), and Cys 21 in CaIpk2 replacing Ile 65 (in HsIPMK) and Val 23 (in CnArg1). Ile 65 in HsIPMK makes van der Waals interaction with the adenine group of ADP.
Human IP6K2 was modelled using the Entamoeba histolytica (Eh) IP6KA crystal structure (PDB ID: 4O4D) and was subsequently used for SAR analysis with flavonoids [41] and TNP [35]. Similarly, differences in the fungal IP 3-4 K active site residues can be further explored through molecular docking and/or via site-directed mutagenesis to identify residues responsible for the lack of inhibition observed for CaIpk2 and to facilitate the design of a more potent inhibitor of CnArg1, which also inhibits CaIpk2. This, TNP derivatives may become relevant to CaIpk2 inhibition once the affinity of the compounds is increased to give IC 50 in the nanomolar range.

Conclusions
In summary, we have demonstrated that TNP and a series of analogues derived from it, inhibit purified, tag-free IP 3-4 K produced by the human fungal pathogen, C. neoformans. We also show that the relative selectivity of TNP and its analogues over the human orthologue can be influenced 20-fold by making substitutions at the N2 position of the purine. Recently released AlphaFold protein models of fungal pathogen IP [3][4] Ks have also revealed amino acid differences in the human and fungal ATP binding site, which are being used in combination with the assays developed to guide ongoing SAR studies aimed at improving the potency and selectivity of the TNP analogues for the fungal enzyme targets, and testing new IPK inhibitor scaffolds.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biom12101526/s1, Data S1: Synthetic methods and spectroscopic data for individual compounds. Figure S1: TNP and its analogues did not inhibit CaIpk2 when tested at concentrations up to 50 µM. Figure S2: Surface plasmon resonance (SPR) sensorgrams and dose-response curves of compounds 39 and 41 are shown as representative data. The K D values in Table 7 were obtained using steady-state affinity.