Anthraquinone derivatives as ADP-competitive inhibitors of liver pyruvate kinase European of Medicinal Chemistry

Liver pyruvate kinase (PKL) is a major regulator of metabolic ﬂ ux and ATP production during liver cell glycolysis and is considered a potential drug target for the treatment of non-alcoholic fatty liver disease (NAFLD). In this study, we report the ﬁ rst ADP-competitive PKL inhibitors and identify several starting points for the further optimization of these inhibitors. Modeling and structural biology guided the optimization of a PKL-speci ﬁ c anthraquinone-based compound. A structure e activity relationship study of 47 novel synthetic derivatives revealed PKL inhibitors with half-maximal inhibitory concentration (IC 50 ) values in the 200 nM range. Despite the dif ﬁ culty involved in studying a binding site as exposed as the ADP site, these derivatives feature expanded structural diversity and chemical spaces that may be used to improve their inhibitory activities against PKL. The obtained results expand the knowledge of the structural requirements for interactions with the ADP-binding site of PKL.


Introduction
Non-alcoholic fatty liver disease (NAFLD) refers to the accumulation of excess fat in liver cells unrelated to alcohol consumption [1,2]. NAFLD is considered the most common cause of chronic liver disease worldwide [3], and people with NAFLD have a higher risk of developing cardiovascular diseases and type 2 diabetes [1]. The global prevalence of NAFLD is currently 25% and is rapidly increasing worldwide [3,4]. No approved pharmacotherapy is currently available to treat NAFLD, indicating an urgent unmet need for the development of therapeutic strategies. NAFLD pathogenesis has been associated with the altered expression of a gene encoding pyruvate kinase (PK) in the liver and red blood cells (PKLR). Systems-level analyses have indicated that PKLR could be targeted for the development of an efficient NAFLD treatment strategy [2,5].
PK is an enzyme involved in the last step of glycolysis, catalyzing the irreversible transfer of phosphate from phosphoenolpyruvate to adenosine diphosphate (ADP). This transfer results in the production of pyruvate and adenosine triphosphate (ATP). PK is a major regulator of metabolic flux and ATP production in glycolysis and is considered a potential drug target [6e8].
Four mammalian PK isoforms have been identified: PKL, PKR, PKM1, and PKM2 [9]. The L and R isozymes are expressed in the liver (L) and red blood cells (R), respectively, whereas PKM2 is expressed in early embryonic cells and other proliferating cells, and PKM1 is expressed in the brain, skeletal muscle, and heart [10,11]. The pyruvate kinase L and R isoforms are both encoded by the PKLR gene, whereas PKM1 and PKM2 are transcribed from the PKM gene via alternative splicing [12,13]. PKL is regulated allosterically by fructose-1,6-bisphosphate, an intermediate in the first stage of glycolysis. PKL also contains an active site where ADP binds to produce ATP [14]. Recently, due to the discovery of an abundance of the vertebrate PKM2 isozyme expressed in cancer cells, this isoform has been the focus of considerable research [10,11,13,15,16].
However, no PKL inhibitors or activators have been reported in the literature. Here, we report the synthesis and initial structureeactivity relationship (SAR) for novel ADP-competing PKL inhibitors, comprising an anthraquinone scaffold originating from 1, which was identified after screening 60,000 compounds for their ability to modulate PKL capacity (data not shown). The iterative design, synthesis, and testing cycle relied on modeling and X-ray structure analyses of PKLeinhibitor complexes.

Results and discussion
2.1. Substituents on the anthraquinone core Compound 1 was identified as an ADP-competitive inhibitor with a sub-micromolar affinity for PKL (Table 1) but no selectivity for the other PK isoforms, PKR, PKM1, and PKM2 (Fig. S1).
Initially, commercially available derivatives of 1, including 2, 3, and 4, were tested to determine the importance of the catechol and sulfonyl moiety for the activity of 1 against the PKL enzyme. Significant decrease in ability to inhibit PKL was observed upon the removal of these groups from 1, indicating that the sulfonyl and catechol groups were essential for binding to the enzyme. The binding of compound 1 to the ADP-binding site of PKL was analyzed by X-ray crystal structure (Fig. S2). We found that the SO 2 group interacted with Lys379, whereas the catechol groups interacted with the backbone Asn87.

Anthraquinones with a secondary sulfonamide linker
The sulfonamide was seen as the most likely position for derivatization of 1, a series of derivative that would go towards the active site were designed, with an aim to mimic the interactions of phosphate groups of ADP, the recipient of phosphate from phosphoenolpyruvate. The docking of compound 17, a simple sulfonamide derivative of 1 with a terminal carboxylic acid group, indicated that 17 should preserve the key interactions observed between PKL and 1 while allowing for additional interactions deeper in the cavity, mimicking the substrate phosphate groups (Fig. 1).
A series of anthraquinones with a secondary sulfonamide linker was synthesized and evaluated (Scheme 1). The direct chlorination of 1 was ineffective, presumably due to the catechol functionality; therefore, dimethylated 5 was chosen as the key intermediate instead, as a good yield could be obtained on a large scale after methylation and chlorination using oxalyl chloride. The coupling of 5 with various amino acid esters (a-, b-, g-, and d-) and a benzylamine resulted in intermediaries 6e13 in good yield following demethylation (typically with BBr 3 ) with very good yields afforded a series of derivatives (14e22). However, the g-glycine adduct 8 exclusively formed a y-lactam under these conditions. Switching to demethylation using HBr/AcOH partially prevented this intramolecular ring-closing reaction, allowing for 16 to be isolated, albeit at a poor yield (19%). The BBr 3 treatment of the serine-adduct 10 resulted in the O-demethylation of the serine oxygen and the formation of 18 with a moderate yield (57%). Switching to AlCl 3, which is a milder demethylating reagent, allowed for the isolation of the O-methylated serine derivative 19, with a very good yield (84%). Finally, although the di-and tri-glycine derivatives were also compatible with BBr 3 -based deprotection conditions, the inherent instability of the amide bonds under low pH conditions made the isolation difficult, resulting in poor yields for 21 and 22.
The effects of derivatives 14e22 on PKL activity were determined using an in vitro biochemical assay using recombinant PKL enzyme. The half-maximal inhibitory concentration (IC 50 ) of 14 was similar to the IC 50 of 1, indicating that these compounds have similar inhibitory effects against PKL activity. However, the addition of hydrophilic groups to 1 generally reduced the inhibitory effects against PKL activity ( Table 2). Increasing the distance between the amine and carboxylic acid functionalities, through the addition of an aliphatic chain, afforded weaker inhibition of PKL activity (cf. 15e17). The addition of smaller a-groups, generating compounds 18e20, also decreased potency of the IC 50 values by 4.5e7-fold.  Notably, the weakened IC 50 observed for the di-glycine derivative 21 was somewhat mitigated by the tri-glycine derivative 22.
Modeling the interaction of 22 with PKL indicated that this group most likely aligned along the highly polar polyphosphate binding pocket of PKL (data not shown). Finally, the aromatic dicarboxylic acid derivative 25, with IC 50 of 0.3 mM, showed only a minor reduction in PKL inhibition compared to 1, indicating that this derivative might bind differently than the aliphatic derivatives. We subsequently evaluated sulfonamides from six-member heterocycles (Scheme 2), as modeling predicted that a sixmembered ring would fit well into the active binding site (Fig. S3). Coupling the key intermediate, 5, with the respective piperidine carboxylate esters resulted in the 3-and 4-substituted derivatives 28 and 29, respectively, after demethylation with BBr 3 . Compounds 28 and 29 were further reacted with methyl-protected amino acids aspartate, glutamate, and histidine, to obtain 35e39 after deprotection with LiOH. A series of piperazine derivatives were also prepared, starting with the coupling of 5 with 1-Boc-piperazine to produce 40. The removal of the Boc-group and demethylation with BBr 3 yielded 42, which was used in a carbonyldiimidazole (CDI)-based coupling with two diacids to produce the urea-linked derivatives 43 and 45 after BBr 3 demethylation. Finally, a one-pot Boc-deprotection and 1-ethyl-3-(3dimethylaminopropyl) carbodiimide (EDC)-mediated coupling to 1,2-dimethyl citrate afforded the amide-linked citrate derivative 47 after demethylation with BBr 3 .
Compound 28 showed a minor improvement in the inhibitory potency on PKL activity, but all three further modifications of this position (35e37) led to a significant increase in the IC 50 (Table 3). Compound 29 showed a 5.5-fold reduction in the IC 50 value, and Scheme 1. Synthesis of anthraquinone derivatives with a secondary sulfonamide linker. Reagents and conditions: (a) i) CH 3 I, Na 2 CO 3 , NMP, 100 C, 3.5 h, yield 71%, ii) (COCl) 2 (Table S2).

Replacement of catechol
To improve our understanding of the binding mode utilized by the anthraquinone derivatives, we used the co-crystal structure of PKL in complex with 47 (PDB: 7QZU) to identify key interactions between the anthraquinone moiety and the ADP-binding site of the enzyme ( Fig. 2A). The tricycle core occupied the hydrophobic pocket of the adenosine binding site, with one oxygen of the sulfone group creating a favorable H-bond (2.11 Å) interaction with Lys379 and the second oxygen intramolecularly interacting with the catechol. The piperazine linker formed an unconventional CHep interaction [17] with His90. The carboxylate group linked to the quaternary carbon of the citrate moiety was stabilized through H-bonds formed with Asn87 (1.91 Å) and His90 (2.01 Å). The carbonyl also interacted with a water molecule (1.88 Å). The crystal structure indicated that R 1 -and R 2 -OH groups did not participate in any key interaction. Therefore, we investigated whether these groups could be removed without affecting the inhibitory effects against PKL activity (Fig. 2B).
We replaced R 2 with a hydrogen atom and an NH 2 group, as shown in 52 and 55 (Scheme 3), respectively. We also prepared derivative 58 (Scheme 4), which features no R 1 -OH, to determine its effects on PKL inhibitory activity. The synthesis of compounds 52, 55, and 58 started from commercially available bromamine acid (48). Compound 52 was obtained in a 72% yield over five steps. Compound 55 was obtained in a 58% yield over three steps, and compound 58 was obtained in a 74% yield over four steps.
Compound 52 showed a significantly reduced inhibitory effect than 47 (Table 4). By contrast, derivatives of the sodium sulfonate salt, 54, 55, and 58, demonstrated moderate to good inhibition, ranging from 65% to 95%. Compounds 55 and 58 inhibited PKL by 77% and 95%, respectively, suggesting that a hydrogen bond donator in the R 2 position was important for inhibitory activity. Interestingly, comparing the IC 50 values for 54 and 58 (7.3 and 3.4 mM, respectively) with 1 (0.2 mM) indicate that the improved inhibitory effect of 58 over 54 is caused by the presence of an OH group in position-R 2 .

Replacing the sulfonyl substituent
The PKL crystals were characterized while bound to 47 (Fig. 2B), which showed that the sulfonyl group interacted with Lys379. A series of derivatives replacing the sulfonyl group was designed to investigate whether the sulfonyl group was necessary for enzymatic inhibition. Using molecular docking (Fig. S7), we explored various functional groups, such as amine, amide, and urea, as linkers for position 2 of the anthraquinone moiety, which occupies a space similar to the anthraquinone. These modeled compounds reached the polar cavity due to the presence of carboxylic acid groups in the molecule. We also explored a series of derivatives in which the anthraquinone core was connected directly to a functionalized carbonyl or aryl group (Fig. S8), which were selected due to similarity with 47 in the anthraquinone binding site, and the aryl-aryl occupation of a different space from the polar binding site in the cavity.
Compound 64 was prepared from commercially available alizarin (59) in a 57% yield over seven steps (Scheme 4). First, regioselective bromination afforded the intermediate 60, and methylation of the catechol group was immediately followed by lithiation and carbonation to generate the carboxylic acid 61. The carboxylic acid was then coupled to Boc-piperazine, followed by subsequent acidic treatment to generate the intermediate 62.
Finally, 62 was treated with 1,2-dimethyl citrate, followed by deprotection of the methyl-ether and ester to yield 64.
Compounds 67, 68, 71, and 72 were prepared from the intermediate 60 in moderate yields. Suzuki-Miyaura cross-coupling between 60 and boronic acids afforded the target compounds 67 and 68 and the intermediates 69 and 70. Treatment with BBr 3 yielded the target compounds 71 and 72.
The urea derivative 76 was obtained from intermediate 60 in a 39% yield over four steps. The protection of the catechol moiety using neat CCl 2 Ph 2 provided the intermediate 73, and 74 was  produced by palladium coupling using benzophenone imine, followed by acid treatment at an elevated temperature. Intermediate 75 was produced by forming the carbamoyl chloride under basic conditions in the presence of triphosgene, followed by the addition of amine 23. Hydrolysis of the ester and cleavage of the catechol protection group yielded 76. Amine derivatives 80e82 were prepared from intermediate 73 at moderate yields. Palladium coupling using selected amines (23, dimethyl glutamate, or methyl 2-(4-aminophenyl)acetate) provided three precursor molecules (77e79). Products 80e82 were obtained using two steps: the first step was the basic hydrolysis of ester groups, and the second step was the use of acidic conditions to deprotect the catechol group. Methanolysis of 80 and 82 yielded 83 and 84, respectively, as methyl ester derivatives.
Compounds 64 and 65 were weak inhibitors (Table 5) compared with their sulfonamide analogs 47 and 42, respectively. The sodium carboxylate salt (66) inhibited PKL by 94% in a cell-free assay at 10 mM. Derivatives of the anthraquinone-aryl group 67 and 68, which are both SO 2 -containing groups, were weak to moderate inhibitors, and the para-substituted 68 provided slightly better (66%) PKL inhibition than the meta-substituted 67 (31%). Compounds 71 and 72, which contained hydroxyl groups, did not show significant inhibitory effects against PKL.
Compounds from the series featuring the N-containing group (76, 80e82) inhibited PKL by 61%e96%. Compounds featuring the urea linker (76) and substituted-benzylamine (80) presented slightly lower IC 50 values (1.7 and 1.3 mM, respectively) than compounds featuring an amino acid (81) or an aromatic amine (82; 2.4 mM and 3.8 mM, respectively). We selected compounds based on their enzymatic potency for further in vitro profiling using cell lysates obtained from the HepG2 PKM2 KO liver cell line as a model to identify selective whole-cell PKL inhibition.
Although compound 66 showed very good inhibition of PKL activity during the biochemical assay, no inhibitory effect was observed after the incubation of this compound with cell lysates. Compounds 76 and 80e82 showed inhibitory correlations between the biochemical (61%e96%) and cell lysate (34%e84%) assays. We observed that 80 (without SO 2 ) showed 5-fold greater inhibitory activity in the cell lysate than its analog 25, which contained a sulfone linker. Therefore, we suggest that the removal of the sulfone group improved the in vitro inhibitory activity. We subsequently evaluated the effects of the most active compounds after incubation with HepG2 PKM2 KO cells. The treatment of cells for 4 and 48 h with compounds at 10 mM concentration showed no inhibitory effects against PKL activity (data not shown). This inactivity might be due to the polar nature of the compounds, which could prevent cell entry.
Two isomers of carboxylesterases, which metabolize esters, are known to be present in liver cells [18]. Therefore, we also detected carboxylesterase activity in HepG2 PKM2 KO cells (data not shown). To enhance the ability of the compounds to enter cells, we consequently prepared corresponding methyl esters (83 and 84). However, these prodrugs did not show any inhibitory effects against PKL activity after incubation with cells (data not shown).  (c) i) t-BuLi, THF, À78 C to À10 C, 10 min, ii) CO 2    were incubated with DMSO. The PK activity of the control cell lysates was set to 100%. The results represent mean ± SD, n ¼ 3, ****p < 0.0001, One-way ANOVA followed by Dunnett's multiple comparison test. n.d ¼ not determined.

Derivatives of most active amino substitutions without an OH group at the R 1 position
We then merged the structural modifications from derivatives 58 and 80e82 (Fig. 3) to generate 91e94, with no R 1 -OH or sulfonyl groups.
The synthesis of 91e94 is shown in Scheme 5. We started with commercially available 2-chloroanthraquinone (85) and performed three reaction steps to obtain the intermediate 86. Treatment with LiOH in 1,4-dioxane water (1:1) at 170 C under microwave irradiation conditions for 2.5 h produced 2-hydroxyanthraquinone in a 55% yield. Regioselective iodination by microwave irradiation afforded 2-hydroxy-3-iodoanthraquinone in a 40% yield. Upon treatment with iodomethane, intermediate 86 was obtained in a 65% yield. Buchwald-Hartwig couplings using various amines generated 87e90 at yields of 29%e70%. Treatment with BBr 3 produced the target compounds 91e94 at yields ranging from 11% to 75%.
The compounds 91e94 (10 mM) inhibited PKL activity by 57e94% when assessed using the in vitro biochemical assay ( Table 6). Consistent with these results, the inhibitory activity in the cell lysate was 41e83%; however, no inhibition of enzyme activity was observed after the treatment of cells with these compounds (data not shown).
2.6. Derivatives of the most active sulfonamide without an OH group at the R 1 position Previous structural modifications (91e94) did not improve biological activity; therefore, we explored new derivatives of 58 based on the most potent inhibitors (14,28,47), which all contained the sulfonamide group.
Compounds 96, 97, and 101 demonstrated >50% inhibitory activity against PKL in the biochemical assay (Table 7). Additionally, compounds showed similar trends in their capacity to inhibit PKL in the cell lysate (23e79% inhibition). However, no whole-cell activity was observed for these compounds, presumably due to poor uptake by cells.
To better understand the binding mode of compounds from this new series, the X-ray co-crystal structure of human PKL in complex with compound 96 (PDB: 5SCF) was obtained (Fig. 4). As the tricyclic structure shifted, the oxygen of sulfone no longer participated in the interaction with Lys379. Instead, the sulfone oxygen participated in an H-bond network with Arg85 (1.97 Å) and Asn87 (1.74 Å). The oxygen from phenol participated in a favorable H-bond interaction with Lys379 (1.67 Å). The oxygen from the carboxylate group formed a network interaction through a water molecule and oxalate, forming an H-bond interaction with Arg85 (1.63 Å). However, the carbonyl group demonstrated H-bond interactions with Asn87 (1.93 Å) and His90 (1.79 Å).

Conclusions
A hit identified by a high-throughput screen was optimized using a SAR study, which resulted in the development of 47 novel first-in-class PKL inhibitors. We used a combination of modeling and structural biology to develop these inhibitors (the SAR is summarized in Fig. S103).
We identified two sets of functional groups that were beneficial for improving the interaction with the ADP-binding site of PKL.
First, the sulfone was removed from compound 25 to provide derivative 80 (amine), which was favorable for achieving PKL inhibition in cell lysates, and the inhibitory activity measured in the biochemical and cell lysate assays were correlated for this compound.
Second, the catechol moiety was exchanged for a phenol substituent, which afforded good inhibition values for both the biochemical and cell lysate assays. This series also was analyzed by X-ray, which showed that compound 96 features a flipped tricyclic core in the ADP-binding site, allowing for interactions with Lys379 or Asn87.
We found that the tricyclic anthraquinone core binds in the same location as the adenosine moiety in the active site of PKL, based on the X-ray crystal structures of PKL in complex with several compounds (15,17,29,47, and 96). We also identified that long aliphatic chains for sulfonamide derivatives, such as those in 16 and 17, are not convenient, possibly due to poor compatibility with the polar cavity of the active site. We also identified the low inhibitory activity of sulfonamide derivatives (14e25 and 28e47) when combined with cell lysates, possibly due to the rapid metabolism of the catechol moiety by liver enzymes.
The results from the SAR studies set the stage for the further development of PKL inhibitors with improved properties.

General information
Unless otherwise specified, all reagents were obtained from commercial supplier Sigma-Aldrich and used without further  Table 6 The inhibitory effects of derivatives 91e94 (10 mM) on PKL activity.

Entry % PKL inhibition
Biochemical assay Cell lysates þ 91 57 ± 0.9 59.7 ± 3.5**** þ Before performing the PKL activity assay, the compounds were incubated for 5 min at rt with cell lysates obtained from HepG2 PKM2 KO cells. The control cell lysates were incubated with DMSO. The PK activity of the control cell lysates was set to 100%. The results represent mean ± SD, n ¼ 3, ****p < 0.0001, One-way ANOVA followed by Dunnett's multiple comparison test. nd ¼ not determined.
Scheme 6. Synthesis of ester anthraquinone derivatives without an OH(R 1 ) substituent. Reagents and conditions: (a) i) CH 3  Reactions run in water were acidified to pH 1e2 by addition of HCl (aq., 1 M), and filtered to afford the desired carboxylic acid. Reactions run in water/THF were concentrated to approx. half their volume on a rotavap, acidified to pH 1e2 by addition of HCl (aq., 1 M), and filtered to afford the desired carboxylic acid.
To a solution of the selected carboxylic acid in DMF (0.1e0.3 M), HOBt (1.2 equiv.), DIPEA (3 equiv.) and the desired amine (1.2 equiv.) were added. The reaction mixture was stirred for 5 min at 0 C before EDC$HCl (1.2 equiv) were added. After 5 min, the reaction was returned to rt, and was left stirring for 5e15 h. The reaction mixture was then concentrated, mounted on silica and purified by flash chromatography to afford the desired amide. The selected methyl aryl ether was placed in a microwave vial and sealed with a cap. The vial was cooled to 0 C, and BBr 3 (1 M in DCM; 3e20 equiv.) was added by syringe. After 5 min, the reaction was returned to, and was left stirring for 1e48 h until the demethylation was complete. The reaction was followed by LCMS. The reaction mixture was then cooled to 0 C and quenched by the addition of water (1e4 mL), and the resulting suspension was concentrated and purified by reverse-phase chromatography to afford the desired compound. Table 7 The inhibitory effect of derivatives without OH(R 1 ) substituent (10 mM) on PKL activity.

101
-H 63 ± 0.6 22.6 ± 2.7**** þ Before performing the PKL activity assay, the compounds were incubated for 5 min at rt with cell lysates obtained from HepG2 PKM2 KO cells. The control cell lysates were incubated with DMSO. The PK activity of the control cell lysates was set to 100%. The results represent mean ± SD, n ¼ 3, ****p < 0.0001, One-way ANOVA followed by Dunnett's multiple comparison test. General procedure E e Demethylation by boron tribromide with subsequent LiOH treatment. The selected methyl aryl ether was placed in a microwave vial and sealed with a cap. The vial was cooled to 0 C, and BBr 3 (1 M in DCM; 10e20 equiv.) was added by syringe. After 5 min, the reaction was returned to rt, and was left stirring for 24e48 h until the aryl demethylation was complete. The reaction was followed by LCMS. The reaction mixture was then cooled to 0 and quenched by the addition of LiOH (aq. 1 M; 4e5 Â BBr 3 equiv.). After 5 min, the resulting dark purple solution was brought to rt and left for 30 min with stirring. Then the reaction mixture was acidified (pH~3) with formic acid, concentrated and purified by reverse-phase chromatography to afford the desired compound.
The selected aryl amine was dissolved in HCl (aq. 1 M; 50 equiv.) and cooled to 0 C in an ice bath. A solution of NaNO 2 (aq. 0.3 M; 3 equiv.) was added dropwise, 5 min after the addition, the ice bath was removed, and the reaction mixture was allowed to slowly return to rt. After 1 h, EtOH (50 mL/mmol aryl amine) was added, followed by Zinc dust (6 equiv.), and the reaction was stirred at rt for 1 h. The reaction mixture was concentrated to ca. 30% of its volume, and the resulting dark red solution was purified by RP column chromatography to afford the desired compound.
4.1.1.7. General reaction G e Fischer esterification. The selected carboxylic acid was suspended in the desired alcohol. Conc. H 2 SO 4 (~10 mL) was added, and the reaction was heated to reflux with stirring. After 30 min, the reaction was cooled to rt, and the mixture was concentrated to~0.5 mL and filtered. The filtered solids were washed with water (2 Â 0.5 mL) and dried to afford the desired ester.
4.1.1.8. General reaction H -Palladium catalysed CeC cross coupling [19]. A suspension of aryl bromide 60 (1.0 equiv.), Pd(OAc) 2 (0.05 equiv.), PPh 3 (0.2 equiv.), KOH (2.5 equiv.), and the corresponding aryl boronic acid (1.5 equiv.) in THF (10 mL) and water (3 mL) under N 2 on a sealed tube were left stir at 80 C for 18 h. The reaction mixture was cooled to 21 C and transferred in a 50 mL beaker with 30 mL of H 2 O and stirred for 30 min. Then a 10% v/v HCl solution was added until pH 2 was reached. The mixture was filtered under vacuum and the crude obtained was purified by chromatography on silica gel. 4.1.1.9. General reaction I -BuchwaldeHartwig coupling [20]. To a mixture of correspond aryl halide 73 or 86 (1.0 equiv.), Cs 2 CO 3 (1.6 equiv.), rac-BINAP (0.15 equiv.), Pd 2 (dba) 3 (0.1 equiv.) and correspond amine or amino acid ester hydrochloride (1.2 equiv) in a MW flask (10e20 mL) under N 2 was added anhydrous toluene (6 mL). The reaction mixture was stirred and refluxed at 120 C for 18 h in a heating plate. The resulting solution was cooled to 25 C and the reaction mixture was filtered on silica gel eluting with pure acetone. To the filtrate was added a small amount of silica gel and the solvent was removed under reduced pressure. The residue was purified by column chromatography to give the pure cross-coupled product.
4.1.1.10. General reaction J -BuchwaldeHartwig coupling [21]. To a mixture of correspond aryl halide 73 or 86 (1.0 equiv.), Cs 2 CO 3 (1.4 equiv.), XPhos (0.4 equiv.), Pd(OAc) 2 (0.2 equiv.) and methyl 2-(4-aminophenyl)acetate (1.2 equiv.) in a MW flask (10e20 mL) under N 2 was added anhydrous toluene (6 mL). The reaction mixture was stirred and refluxed at 120 C for 18 h in a heating plate. The resulting solution was cooled to 25 C and the reaction mixture was filtered on silica gel eluting with pure acetone. To the filtrate was added a small amount of silica gel and the solvent was removed under reduced pressure. The residue was purified by column chromatography to give the pure cross-coupled product.
4.1.1.11. General reaction K -. Hydrolysis of ester and diphenylmethylene ketal groups. To a stirred solution of compound to be deprotected (1.0 equiv.) in 12 mL of 1,4-Dioxane:H 2 O (1:1 v/v) at rt was added LiOH (10 equiv.) and the resulting solution is stirred at reflux for 1 h or until complete hydrolysis of ester. The reaction was followed by LCMS. The mixture is cooled to rt and the reaction mixture is acidified to pH 1 using concentrate HCl, the resulting mixture is stirred at reflux for 1 h or until complete hydrolysis of diphenylmethylene ketal group. After complete deprotection the solvent is evaporated, and the residue is purified by reverse phase chromatography or trituration to give desired compound.
To a suspension of corresponding methyl ether (1.0 equiv.) in AcOH (1 mL), HBr 48% wt (1 mL). The mixture was heating at 130 C in a MW reactor until complete reaction. The reaction was followed by LCMS. The reaction mixture was cooled to 21 C and diluted with water and the solid was separated and washed several times with water.

Sodium 3,4-dimethoxyanthraquinone-2-sulfonate (4)
Alizarin Red S (1) (Fluorochem, 86 wt-% by qNMR; 9.55 g, 24.0 mmol), Na 2 CO 3 (6.36 g, 60.0 mmol, 2.5 equiv.), NMP (480 mL) and iodomethane (9 mL, 144 mmol, 6 equiv.) was combined in a 1L RB flask, which was subsequently sealed with a fresh septum (locked in place with a zip-tie) and heated to 100 C in a heating block. After 3h, the reaction mixture cooled and subsequently separated between Et 2 O (900 mL) and water (80 mL). The organic phase was removed by decantation, washed with water (100 mL) and re-used to extract more NMP from the aqueous phase. Six repeats of this process resulted in~100 mL of a dark aqueous phase, after any residual Et 2 O was removed on the rotavap (at rt). This aqueous phase was diluted to 490 mL and was loaded onto a C18column in several portions for purification by RP chromatography (SNAP C18 60g; MeOH in water: 0e13%). Fractions containing the desired product were combined and concentrated to afford 7.36 g of a dark red solid, which was recrystallized from MeOH to afford 4.70 g of a yellow solid. An additional 1.65 g was recovered from the supernatant following repurification by reverse-phase chromatography (SNAP C18 60g; MeOH in water: 0e13%) and subsequent recrystallization from MeOH. Combined yield: 6
202e204 C (dec). 1   Compound 10 (60 mg, 0.13 mmol) and anhydrous AlCl 3 (104 mg, 0.78 mmol, 6.0 equiv.) was placed in a microwave vial and sealed with a cap. CH 2 Cl 2 (2 mL) was added by syringe and the vial was heated to 70 C for 3 h in a microwave reactor. Then HCl (aq. 37%; 4 mL) was added, and the vial was heated to 100 C for 30 min in a microwave reactor. The pH of the reaction mixture was adjusted tõ 3 by addition of NaOH (aq., 6 M) and extracted with acetone (2 Â 10 mL). The organic phases were combined, concentrated and purified by RP column chromatography (12g C18 column; MeOH in water: 0e70%) to afford 36 mg (84%) of a yellow solid after trituration with DCM. M.p. 125e126 C (dec). 1

(1-((3,4-Dihydroxyanthraquinon-2-yl)sulfonyl)piperidine-3carbonyl)glutamic acid (36)
Following general procedure B, 31 (30 mg, 0.051 mmol) was reacted with LiOH (6 mg, 0.25 mmol, 5 equiv.) in water (0.50 mL), to afford 25 mg (88%) of an orange solid after acidification (pH~1) with 1 M HCl (aq) , filtration and subsequent recrystallization from water. M.p. 80e81 C (dec). 1    To a solution of 5 (0.6 g, 1.07 mmol) in methylene chloride (15 mL) and cooled to 0 C trifluoroacetic acid (2.5 mL) was added over 1 min and the mixture was warmed to rt and stirred for 1 h. The mixture was then evaporated in vacuum. The residue was dissolved in water, the pH adjusted to 12e13 using 6 N sodium hydroxide (5 mL) and extracted with DCM. The organic phase was washed with brine, dried over sodium sulfate and concentrated to give of the desired product as a pale orange solid (  To a solution of glutamic acid dimethyl ester hydrochloride (55 mg, 0.26 mmol) in DMF 0.5 mL was added DIPEA (0.05 mL, 0.26 mmol). The resulting mixture was stirred at rt for 10 min, after which time was added 1-1 0 -carbonyldiimidazole (42 mg, 0.26 mmol). The mixture was stirred at rt for 1 h, after which time was added 42 (100 mg, 0.26 mmol). The mixture was stirred overnight. The reaction mixture is diluted with 10 mL of acetone and solvent is removed by rota evaporation. Then LiOH (43 mg, 1.03 mmol) in 2 mL of water is added to the residue and the resulting mixture is stirred at rt for 1 h the mixture was neutralized with HCl 1 M and the crude material was purified by reverse phase using H 2 Oþ0.1% FA/ACN to afford title compound as a red solid (50 mg, 0.09 mmol, 35%). M.p. 168e169 C. 1   To a solution of methyl 5-(aminomethyl)-2-(2-methoxy-2oxoethyl)benzoate (61 mg, 0.26 mmol) in DMF 0.5 mL was added DIPEA (0.05 mL, 0.26 mmol). The resulting mixture was stirred at rt for 10 min, after which time was added 1-1 0 -carbonyldiimidazole (42 mg, 0.26 mmol). The mixture was stirred at rt for 1 h, after which time was added 42 (100 mg, 0.26 mmol). The mixture was stirred at rt overnight. The reaction mixture was diluted with 10 mL of acetone and solvent was evaporated by rota-vapor. The residue was purified by reverse phase using H 2 Oþ0.1%FA/MeOH, to afford title compound as a red solid (63 mg, 0.097 mmol, 38%). M.p.    13
The resulting purple solid was dissolved in DMF (28 mL) and filtered. Toluene (220 mL) was added to the filtrate, and the resulting purple suspension was filtered, and the solids washed with MeCN (2 Â 15 mL) to afford 1.01 g crude sodium 1-amino-4methoxyanthraquinone-2-sulfonate as a dark purple solid, which was used directly in the next step without further purification.

Sodium 4-bromoanthraquinone-2-sulfonate (53)
Bromaminic acid sodium salt (48) (3.23 g, 8.00 mmol) was dissolved in HCl (aq. 1 M; 400 mL) by gentle heating, and then cooled to 0 C in an ice bath. A solution of sodium nitrite (1.66 g, 24 mmol, 3 equiv.) in 60 mL distilled water was added dropwise. After 5 min, the reaction was brought to rt, and was left for 30 min with stirring. Then EtOH (400 mL) was added to the resulting orange suspension, followed by zinc dust (3.12 g, 48 mmol, 6 equiv.), and the reaction was stirred at rt for 1 h.
The resulting red solution was concentrated to~200 mL, loaded onto a 60g C18 column in three portions for purification by RP column chromatography (MeOH in water: 20e55%), to afford 2.23 g (72%) of the title compound as an orange solid. M.p. 151e154 C (dec). C. 1  Compound 54 (70 mg, 0.215 mmol) was dissolved in dry DMF (1 mL) at rt. Bromine (30 mL, 3 equiv.) was added dropwise, and the mixture was left at rt with stirring for 15 min.
The reaction mixture was diluted with water (10 mL) and was loaded onto a 30g C18 column for purification by RP column chromatography (MeOH in [water þ0.1% formic acid]: 0e40%) to afford 75 mg of 54 (~90% pure by HPLC) as a red solid, which was used directly in the next step without further purification.
A microwave vial was charged with NMP (2 mL) and sealed with a septum. Ammonia (generated from the addition of aq. ammonia (28%) to KOH flakes in a separate flask) was bubbled through this solution for 10 min. Then crude sodium 3-bromo-4hydroxyanthraquinone-2-sulfonate (60 mg,~0.14 mmol) and CuI (5.3 mg, 0.028 mmol, 0.2 equiv.) was added, and the vial was sealed with a cap. The vial was heated to 130 C for 3 h in a microwave reactor. The resulting dark purple reaction mixture was acidified (pH~0) with HCl (aq. 2 M, 2.5 mL) and heated to 130 C for 5 min in a microwave reactor. The resulting dark red reaction mixture was diluted with water (7 mL) and loaded onto a 12g C18 column for purification by RP column chromatography (MeOH in water: 35e50%) to afford 13 mg (22% over 2 steps) of the title compound as a red solid. M.p. >300 C. 1
Then a solution of sodium sulfite (2.45 g, 19.5 mmol, 4 equiv.) in water (15 mL) was added and the vial was left at rt for 15 min with stirring. The reaction mixture was then diluted with water (120 mL) and was loaded in two portions onto a 60g C18 column for purification by RP column chromatography (MeCN in water: 0e25%) to afford 1.15 g (58%) of 56 as a red solid. M.p. 241e243 C (dec). 1

Molecular modelling
Molecular modelling studies were conducted using Molecular Operating Environment (MOE) [25] version MOE 2019.0102 (Chemical Computing Group, Montreal, CA). The X-ray crystallographic complexes were prepared based on the chain C of the protein using MOE quick preparation tool. We used as parameters a forcefield: PFROSST and solvatation: Born. Each compound data was prepared by minimizing energy, adding hydrogen atoms, calculating partial charges and potential energy. Dockings were carried out as a triangle matcher, rigid receptor, score London dG and 30 poses. Validation was achieved by restoring the original orientation upon redocking 47.

Determination of PKL activity in a biochemical assay
The enzymatic reactions were conducted in triplicates at RT for 30 min in a 25 mL mixture containing 50 mM Tris, pH 7.4, 10 mM MgCl 2 , 100 mM KCl, 0.05% Tween, 0.1 mM ADP, 0.125 mM PEP, PKL 3.0 ng and the test compounds at 10 mM (or from 100 mM to 0.003 mM for IC 50 determination). The final DMSO in the reaction was 1%. Thereafter, 25 ml of Kinase-Glo Max reagent (Promega) was added to each well and luminescence was measured using a BioTek Synergy 2 microplate reader. The Kinase-Glo Max luminescence assay kit measures kinase activity by quantitating the amount of ATP produced following a kinase reaction.
The luminescence data were analyzed using the computer software, GraphPad Prism. In the absence of the compound, the intensity (C e ) in each data set was defined as 100% activity. In the absence of enzyme, the intensity (C 0 ) in each data set was defined as 0% activity. The percent activity in the presence of each compound was calculated according to the following equation: % activity ¼ (CeC 0 )/(C e eC 0 ), where C ¼ the luminescence in the presence of the compound.
The values of % activity versus a series of compound concentrations were plotted using non-linear regression analysis of Sigmoidal dose-response curve generated with the equation Y] Bþ(T-B)/1 þ 10 ((LogEC50ÀX)ÂHill Slope) , where Y ¼ percent activity, B ¼ minimum percent activity, T ¼ maximum percent activity, X ¼ logarithm of compound and Hill Slope ¼ slope factor or Hill coefficient. The IC 50 value was determined by the concentration causing a half-maximal percent activity. number in parenthesis after each code): 1abc (1). All figures for crystal structures are found in the Supplementary information ( Fig. S89-S102).

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.