Organometallic scaffolds as protein tyrosine phosphatase 1b inhibitor

Summary This thesis describes the synthesis and evaluation of two novel classes of organoruthenium and organogold complexes as inhibitors of protein tyrosine phosphatases PTPs.. The orga

ORGANOMETALLIC SCAFFOLDS AS PROTEIN TYROSINE

PHOSPHATASE 1B INHIBITOR

ONG JUN XIANG

(B.Sc.(Hons.), NUS

A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF SCIENCE

DEPARTMENT OF CHEMISTRY NATIONAL UNIVERSITY OF SINGAPORE

2012

Thesis Declaration

The work in this thesis is the original work of Ong Jun Xiang, performed independently under the supervision of Dr Ang Wee Han, (in the laboratory S15-05-01), Chemistry Department, National University of Singapore, between 1/8/2010 and 24/8/2012

Ong Jun Xiang 23/8/2012

My sincere thanks to the technical staff at X-ray diffraction (Prof Koh, Ms Tan and Ms Hong), Nuclear Magnetic Resonance (Mdm Han), Mass spectrometry and Elemental Analysis laboratories for their technical support

I would like to thank my group members Chee Fei, Daniel, Mun juinn, Diego and Tian Quan for their assistance and helpful discussions

I am grateful to NUS for my research scholarship

Lastly, I am very thankful to my family members and friends for their love and support

Table of Contents

Summary V

List of Tables VII

List of Figures VIII

List of Schemes IX

List of Abbreviations X

1 Introduction 1

1.1 Protein Tyrosine Phosphatase 1B as drug target 1

1.2 Organic inhibitors of Protein Tyrosine Phosphatases 4

1.3 Metal complexes as inhibitors of Protein Tyrosine Phosphatases 6

2 Design concept of metallo-inhibitors 9

2.1 Design of Ruthenium-based inhibitor 9

2.2 Design of Gold-based inhibitor 10

3 Synthesis of organoruthenium and organogold PTP inhibitors 12

3.1 Synthesis of ligands 12

3.2 Synthesis of organoruthenium inhibitors 15

3.3 Synthesis of organogold inhibitors 21

4 Biological evaluation of organoruthenium PTP inhibitors 24

4.1 Aqueous stability of organoruthenium PTP inhibitors 24

4.2 Inhibition of PTP-1B and TC-PTP by organoruthenium PTP inhibitors 25

4.3 Determination of inhibition constants of 4a towards PTP-1B and TC-PTP 30

5 Conclusion 32

6 Experimental Section 33

6.1 Preparation of organoruthenium complexes 35

6.2 Preparation of organogold complexes 44

Appendix 50

References 67

Summary

This thesis describes the synthesis and evaluation of two novel classes of organoruthenium and organogold complexes as inhibitors of protein tyrosine phosphatases (PTPs) The findings of the research are presented in four chapters

Chapter 1 gives an introduction to protein tyrosine phosphatase 1B (PTP-1B) and its potential as a therapeutic drug target The major challenge in the attempt to design inhibitors of PTPs arose from their highly conserved enzymatic active-site across the PTP family Selectivity against closely related PTPs could be low due to the structural homologies This is particularly problematic for PTP-1B since it is 80% homologous with T-cell protein tyrosine phosphatase (TC-PTP) in their catalytic domains So far, the organic and metallo-inhibitors synthesized to date have yet to achieve high selectivity towards PTP-1B against TC-PTP

Chapter 2 describes the concepts behind the design of the organoruthenium and organogold complexes respectively Since PTPs possess an active-site which is highly conserved across the PTP family, targeting the active-site to achieve selectivity among the various PTPs will not be a feasible approach We reasoned that the proximal space to the PTP active-site must

be important for the molecular recognition of their protein substrate By capitalizing on the potential difference in the structural environment proximal to the active-site, we hope to develop metallo-inhibitors that are selective towards specific PTPs To do so, we would rationally build molecules which can simultaneously bind to the active-site as well as being able to interact with the chemical space proximal to the active-site of the PTPs

Chapter 3 describes the synthesis of the ligands and their respective organoruthenium and organogold complexes The reaction of imidazole/benzimidazole difluoromethylphosphonate

esters, 1 and 3, and imidazole/benzimidazole difluoromethylphosphonic acids, 2 and 4, with

[(η 6

-arene)RuCl2]2 where arene = cymene or 1,3,5-triisopropyl-benzene, yielded mononuclear

Ru(II) complexes 1a-b/3a-b and 2a-b/4a-b respectively The molecular structures of complexes

1b , 3b, 4a and 4b were also determined by X-ray crystallography Based on a different approach, ethynyl difluoromethylphosphonate ester 9 was reacted with [Au(PR3)Cl] where R = phenyl in

the presence of base to yield mononuclear alkynyltriphenylphosphinegold(I) complex 9a

Chapter 4 describes the biological studies of the organoruthenium complexes as inhibitors

of PTP-1B and TC-PTP The organoruthenium complexes containing phosphonic acid groups were found to inhibit PTP-1B at low micromolar level with 7-10 fold selectivity towards PTP-1B over TC-PTP whereas the inhibitors containing phosphonate ester groups were found to be inefficacious Steady-state kinetics experiments have shown that the complexes competitively bind to the enzymes at their active-sites In addition, molecular docking studies have also shown that the organoruthenium complexes bind better than their parent ligands in PTP-1B due to additional hydrophobic interactions with Phe182 Overall, these results suggest that this novel class of organoruthenium complexes may be promising therapeutic agents to target PTP-1B The manuscript entitled “Rational Design of Selective Organoruthenium Inhibitors of Protein Tyrosine Phosphatase 1B” have been submitted for publication

List of Tables

Table 4.1 Initial screening of inhibitors against PTP-1B and TC-PTP and IC50 of

List of Figures

Figure 1.4 Selected PTP-1B inhibitors of difluoromethylphosphonic acid class of

Figure 1.5 Mimicking protein kinase inhibitor staurosporine with octahedral metal

Figure 4.2 Inhibition of PTP-1B and TC-PTP at inhibitor concentration of 100 µM 27

Figure 4.3 Dose-response curves of inhibition of compounds towards PTP-1B and

Figure 4.5 Steady-state kinetic studies of complex 4a towards PTP-1B and TC-PTP 31

List of Schemes

Scheme 3.3 Synthetic route for the synthesis of ruthenium(II)-arene complexes 17

Chapter 1 Introduction

Sedentary lifestyle and lack of physical exercises, which are common in many developed and developing countries, have been the major contributors to obesity in both the adult and child populations.[1-3] Obesity is strongly linked to Type 2 diabetes mellitus (TD2M) which is a chronic disease where the body loses its sensitivity to the blood glucose-regulating hormone insulin People with diabetes have elevated blood glucose level and over time, their eyes, kidney, nerves and blood vessels will be damaged Consequently, they incur long-term health problems leading to kidney failure, heart disease and stroke World Health Organization reported in 2004 that an estimated of 3.4 million people died from consequences of high blood sugar, and that diabetes-linked death will double between 2005 and 2030 To address these problems, there has been an intensified search for new therapeutic treatments for T2DM and obesity

1.1 Protein Tyrosine Phosphatase 1B as drug target

Protein tyrosine phosphatases (PTPs) belong to a large family of 107 enzymes that play a vital role in the regulation of various signaling transduction pathways in mammalian systems.[4, 5]PTP enzymes catalyze the dephosphorylation from phosphorylated tyrosine residues and in conjunction with protein tyrosine kinases (PTKs), they are responsible for managing the levels of phosphorylation within the cells.[6, 7] Studies have shown that the dysregulation of PTP can lead

to several pathological conditions including diabetes, obesity, cancer and autoimmune disorders.[8-10] Amongst the members of the PTP family, PTP-1B is a key negative regulator of the insulin and leptin signaling pathways associated with obesity and diabetes PTP-1B is responsible for dephosphorylation of activated insulin receptor (IR) or insulin receptor substrates (IRS) in insulin signaling[11, 12] and dephosphorylates JAK2, which is downstream of the ObR

receptor in the leptin signaling pathway[13, 14] Cell cultures and gene coding studies have shown that aberrant expression of PTP-1B can contribute to obesity and diabetes.[15-17] In addition, PTP-1B knockout mice experiments showed that PTP-1B deficiency lead to increased sensitivity towards insulin and resistance to diet-induced obesity, suggesting that inhibition of PTP-1B could address obesity and insulin resistance.[18, 19] These pioneering studies have validated the notion that inhibition of PTP-1B could serve to address issues of obesity and diabetes, and there have been growing interest in the development of PTP-1B inhibitors as potential therapeutic agents.[20-22]

Figure 1.1 Role of PTP-1B in insulin and leptin signaling pathways

One major challenge in the attempt to design inhibitors of PTPs arose from their highly conserved enzymatic active site across the PTP family.[4] Within the conserved PTP catalytic domain, a unique signature sequence motif, CX5R, which is invariant among all PTPs, can be found This motif contains residues 214-221 which bind to the phosphate group of pTyr and the cysteine 215 (Cys215) residue which is responsible for catalyzing the dephosphorylation of

transforms into a closed conformation and stabilizes the substrate within the active site This brings the pTyr substrates into close proximity to Cys215 which is in a position to undergo a nucleophilic attack on the substrate phosphorous atom The activity of PTPs is regulated by Cys215 through a deactivation/activation pathway mediated by signaling molecules, hydrogen peroxide and glutathione, in the body.[23-26] The oxidation (deactivation) of the thiol of Cys215 leads to the formation of a sulfenyl-amide intermediate rendering it ineffective in catalyzing the dephosphorylation of pTyr The subsequent reduction (deactivation) of the sulfenyl-amide intermediate restores the active thiol

Figure 1.2 Overlap of protein crystal structures of TC-PTP and PTP1B

Selectivity against closely related PTPs could be low due to the structural homologies This is particularly problematic for PTP-1B since it is 80% homologous with TC-PTP in their catalytic domains.[27] TC-PTP is widely distributed throughout the body and is responsible for modulating the immune functions of the body.[28] The consequences of using a non-selective PTP-1B inhibitor that also inhibits TC-PTP can lead to severe side-effects and recent studies have shown that TC-PTP deficient mice die within 3-5 weeks of age.[29, 30] Developing PTP-1B inhibitors with high selectivity towards PTP-1B, as opposed to TC-PTP, remain a daunting task

1.2 Organic inhibitors of Protein Tyrosine Phosphatases

The organic inhibitors that target the active site of PTPs can be classified into several classes based on the type of functional groups that binds to the catalytic site One of the most potent classes of organic inhibitors is the difluoromethylphosphonic acids and they have been the core of many inhibitor designs The main strategy behind the design of this class of active-site inhibitors is to build molecules that contain units that mimic pTyr The fact that it is almost structurally identical to pTyr is the key reason why inhibitors incorporating difluoromethylphosphonic acids moieties are able to exhibit high inhibitory activities These units are non-hydrolysable and compete for the active-site In this way, it prevents the over-dephosphorylation of pTyr on protein substrates

Figure 1.3 Mimetics of pTyr A) pTyr; B) Methylenephosphonic acid; C) Difluoromethylphosphonic acid

The structure of difluoromethylphosphonic acid, a non-hydrolyzable mimetic of pTyr, is shown in Figure 1.3 The difluoromethylphosphonic acid has been shown to be a 1000-fold more superior than its non-fluorinated derivative.[31] The increased binding affinity of

been attributed to interactions between the fluorine atoms and residues in the active site.[32] In addition, the geometry brought about by the CF2 group makes the Ph−CF2−PO3H2 angle resemble that of Ph−O−PO3H2 observed in pTyr This allowed for a better fit to the PTP active site which would otherwise not be observed for methylenephosphonic acid or even the mono-fluorinated derivative.[33] Some selected inhibitors of the difluoromethylphosphonic acid class of

PTP-1B inhibitors are shown in Figure 1.4 Compound I, which is the most potent and selective

PTP-1B inhibitor identified to date (Ki = 2.4 nM), exhibits a 1000- to 10000-fold selectivity against a panel of other PTPs, but only 10-fold against the structurally similar TC-PTP.[34]

Figure 1.4 Selected PTP-1B inhibitors of difluoromethylphosphonic acid class of inhibitors

Some other classes of active-site inhibitors include the 2-carbomethoxybenzoic acids[35]

and the 2-oxalylaminobenzoic acids[36] The strategy adopted in designing these inhibitors was the same, which was to design pTyr mimetics to bind to the catalytic pocket However, as these compounds were often negatively charged owing to the high polar nature of the active-site, they exhibited low cellular penetration levels This was often the drawback associated with such

active-site inhibitors An interesting class of inhibitors was the allosteric inhibitors which bind to

a novel site located ~20Å away from the catalytic pocket.[37] This site is amenable to binding small molecules, considerably less polar and not well-conserved among PTPs, thus affording an opportunity to circumvent the problems associated with active-site inhibitors The binding of an allosteric inhibitor to the allosteric site prevents the closure of catalytic WPD loop (Trp179, Pro180 and Asp181) at the active-site This in turn rendered the catalytic site inactive as the WPD loop needed to close over the active-site in order to facilitate the cleavage of pTyr substrates.[38] Interestingly, it was found that the allosteric site also differed from the corresponding region in the closely related TC-PTP at the central position 280 (cysteine instead

of phenylalanine) Hence, this difference at this central position between PTP-1B and TC-PTP can potentially be exploited to develop selective inhibitors

1.3 Metal complexes as inhibitors of Protein Tyrosine Phosphatases

Although the field of synthesizing organic inhibitors of PTPs has been extensively studied, the use of metal complexes to target PTPs remains largely unexplored So far, several metallo-complexes have been investigated as potential PTP inhibitors A series of vanadium and copper complexes containing Schiff base ligands were found to be very potent PTP inhibitors but low selectivity was observed between PTP-1B and TC-PTP.[39-44] These Schiff-based vanadium and copper complexes exhibited at most a 2-fold and 3-4 fold selectivity towards PTP-1B over TC-PTP respectively, although several higher magnitudes of selectivity were observed against other PTPs such as SHP-1, SHP-2 and PTP-MEG2 Recently, a library of gold-phosphine and gold-carbene complexes was screened and several found to exhibit good PTP-1B inhibitory activity with modest levels of selectivity.[45, 46] However, these gold complexes were only selective

towards lymphoid tyrosine phosphatase (LYP) and protein tyrosine phosphatase PEST PEST) Their mechanism were not known but given the affinity of these metal centres, especially

(PTP-Au, towards S-containing cysteine residues within the enzyme active-site, it was possible that the metal centre was directly binding to the S-atom which could account for the poor selectivity between the two enzyme homologues Indeed, some reports have shown that inhibition of PTPs was attributed to binding of the metal centres in these organo-metallic complexes to the sulphur atom of the Cys215 at the active-site.[47, 48] The coordination of the metal center to Cys215 led to the retardation of the redox pathway which was critical in modulating the activity of PTPs The search for a selective metal-based PTP inhibitor remained elusive

Earlier on, Meggers et al had shown that highly selective active-site inhibitors of PTKs can

be prepared using a known kinase inhibitor, staurosporine, as a template By replacing the glycoside motif with an organoruthenium fragment, improved inhibitory profiles against specific PTKs were achieved, representing some of the most efficacious PTK inhibitors reported In this manner, the octahedral ruthenium framework provided the scaffold upon which ligands could be structurally organized and as a basis for structure-activity studies.[49-59]

Figure 1.5 Mimicking protein kinase inhibitor staurosporine with octahedral metal complexes

We were interested in applying these principles in the design of PTP inhibitors using metal-ligand interactions as a structural element PTPs presented a different challenge since the active-site was small, designed to bind a pTyr motif, as compared to PTKs which had a large cleft capable of accommodating a bulky ATP substrate We reasoned that the proximal space to the PTP active site must be important for the molecular recognition of their protein substrate By capitalizing on the potential difference in the structural environment proximal to the active site,

we hope to develop metallo-inhibitors that are selective towards specific PTPs

Chapter 2 Design concept of metallo

2.1 Design of Ruthenium

Figure 2.1 Approach to designing ruthenium

As the active-site is highly conserved across the PTP fa

achieve selectivity among the various PTPs would not be a feasible approach As such, our approach is to target the structural environment proximal to

achieve selectivity We hypothesize t

environment proximal to the substrate binding site of various PTPs which can be exploited to obtain selectivity Our strategy was

active-site as well as being able to interact

the PTPs Figure 2.1 depicted

design of the organoruthenium

linker One portion of the molecule consisted

moiety which was a well-studied

PTPs.[33] The other portion of the molecule will consist of a

Design concept of metallo-inhibitors

Design of Ruthenium-based inhibitor

Approach to designing ruthenium-based PTP inhibitors

site is highly conserved across the PTP family, targeting the activeachieve selectivity among the various PTPs would not be a feasible approach As such, our

is to target the structural environment proximal to the active-site of the various PTPs to achieve selectivity We hypothesize that there is potential difference in the st

the substrate binding site of various PTPs which can be exploited to ain selectivity Our strategy was to build molecules which can simultaneously bind to the

l as being able to interact with the chemical space proximal to

the approach to designing ruthenium-based inhibitors ofruthenium inhibitors consist of two fragments joined toge

portion of the molecule consisted of the phenyl difluoromethylphosphonic acid

studied non-hydrolyzable mimetic of pTyr that bindtion of the molecule will consist of a 3-D globular fragment which had

targeting the active-site to achieve selectivity among the various PTPs would not be a feasible approach As such, our

site of the various PTPs to hat there is potential difference in the structural the substrate binding site of various PTPs which can be exploited to

multaneously bind to the with the chemical space proximal to the active-site of

based inhibitors of PTPs The inhibitors consist of two fragments joined together by a bi-dentate

yl difluoromethylphosphonic acid that bind to the active-site of

D globular fragment which had the

potential to interact with amino-acid residues surrounding the substrate binding site The ruthenium metal center in the design solely serves a structural role to organize the ligands in three-dimensional space thereby creating molecules with unique and well-defined 3-D globular shapes, owing to its ability to form octahedral coordination geometry With the ligands organize

in a three dimensional fashion, interactions with amino-acid residues can be maximized leading

to better binding to the PTPs This was otherwise not easily achievable in purely organic molecules as the carbon center is only limited to a coordination number of four By capitalizing

on the potential difference in structural environment proximal to the active site of PTP-1B and TC-PTP, we hope to achieve selectivity between the two PTPs arising from contrasting interactions with the globular ruthenium fragment In addition, with the established coordination chemistry of ruthenium metal, structurally diverse complexes can be assembled with relative ease as compared to organic molecules whose synthesis often involved multiple synthetic steps which is an arduous process

2.2 Design of Gold-based inhibitor

active site

PTP

peripheral structural space

Figure 2.2. Approach to designing gold-based inhibitors

The strategy adopted to designing gold-based inhibitors towards PTPs is somewhat similar

to the approach previously described for the design of the ruthenium-based inhibitors This novel class of gold(I) complexes will consist of the phenyl difluoromethylphosphonic acid moiety and

a 3-D globular fragment which was provided by a bulky phosphine ligand The acetylene group

in conjunction with the gold metal center served to link these two fragments together as depicted

in Figure 2.2 The gold metal center in the design solely served a structural role and was not expected to interact with the enzymes Similarly, this class of gold(I) inhibitors was expected to bind at the active-site of PTPs with the globular phosphine ligand at the exterior, poised to interact with amino-acid residues surrounding the substrate binding site By capitalizing on the potential difference in structural environment proximal to the active-site of the various PTPs, we hope to achieve selectivity between the various PTPs arising from contrasting interactions with the globular phosphine ligand This approach was attractive as it allowed an easy access to a library of structurally diverse compounds due to the commercial availability of a wide range of phosphine ligands In addition, the typically strong gold-phosphorous and gold-carbon bonds will give rise to the formation of stable gold(I) complexes which was critical in biological evaluation

Chapter 3 Synthesis of organoruthenium and organogold PTP inhibitors 3.1 Synthesis of ligands

The ligands 1, 2, 3 and 4 were synthesized as shown in Scheme 3.1 from p-tolualdehyde

p-Tolualdehyde was first treated with diethylphosphite in the presence of a catalytic amount of

base to give the hydroxyl-phosphonate ester P1 Subsequent oxidation of P1 with Dess-Martin periodinane (DMP) afforded the keto-phosphonate ester P2 Treatment of compound P2 with

diethylaminosulfur-trifluoride (DAST) resulted in the replacement of the ketone functional group

by geminal fluorine atoms to give the difluoromethylphosphonate ester P3 It has been reported that P3 can be synthesized via Shibuya coupling using 4-iodotoulene and

diethyl(bromodifluoromethyl)phosphonate with Zn dust and CuBr as the coupling agents,[60, 61]

but after repeated unsuccessful attempts, the described three-step approach was adopted

Bromination was first carried out on P3 with N-bromosuccinimide (NBS) in the presence of

azobisisobutyronitrile (AIBN) as the catalyst The bromination reaction resulted in the formation

of both mono-brominated P4 as the major product as well as di-brominated side-product which could not be effectively separated using flash-column chromatography However, only P4

reacted with the 2-(2-pyridyl)imidazole and 2-(2-pyridyl)benzimidazole to afford the desired

ligands 1 and 3 respectively in good yields, while the di-brominated side-product remained unreacted and can be removed using flash-column chromatography Treatment of 1 and 3 with

N ,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and iodotrimethylsilane (TMSI) followed by

MeOH resulted in the hydrolysis of the phosphonate ester groups to give ligands 2 and 4

respectively as the phosphonic acids

Scheme 3.1 Synthetic route for the synthesis of ligands 1, 2, 3 and 4

The ligand 9 was synthesized as shown in Scheme 3.2 from p-iodo benzaldehyde Initial

attempts to obtain compound 9 through the Stille coupling reaction of ethynyltributylstannane

with compounds P7a or P7b, which were synthesized from p-triflate benzaldehyde and p-bromo

benzaldehyde respectively through the hydroxyl- and keto- phosphonate ester intermediates

P5a-b and P6a-b, did not work well Although several previous reports have shown that coupling reaction of ethynyltributylstannane with P7a or P7b was possible,[62-64] it was observed that there

was no reaction with the triflate diethylphosphonate ester P7a after extended hours of reaction at

elevated temperatures with a variety of polar solvents such as acetonitrile, 1,4-dioxane, THF,

DMF, etc On the other hand, coupling reaction with the bromo diethylphosphonate ester P7b

resulted in a mixture of starting material and product which could not be readily separated using flash-column chromatography Since iodo-substrates are known to be more activated towards palladium-catalyzed coupling reactions compared to the corresponding triflate- and bromo-

substrates, iodo-diethylphosphonate ester P7c was chosen as the coupling reagent for the Stille

coupling reaction p-iodo benzaldehyde was first treated with diethylphosphite in the presence of

a catalytic amount of base to give the hydroxyl-phosphonate ester P5c Subsequent oxidation of

P5c with DMP afforded the keto-phosphonate ester P6c Treatment of compound P6c with

DAST resulted in the replacement of the ketone functional group by geminal fluorine atoms to

give the difluoromethylphosphonate ester P7c Again, attempts were carried out to synthesize the difluoromethylphosphonate ester compounds P7a-c by the Shibuya coupling but they were unsuccessful after repeated attempts Finally, compound P7c was subjected to the Stille coupling

reaction with ethynyltributylstannane in DMF with PdCl2(PPh3)2 (5 mol%) as the catalyst to

afford ligand 9 in moderate yield Alternatively, ligand 9 can also be synthesized through the direct coupling reaction of P7c with trimethylsilylacetylene (TMSA) under palladium-catalyzed conditions to yield the trimethylsilyl (TMS) protected alkyne P8 De-protection of the compound

P8 under basic conditions afforded compound 9 in relatively higher yield

Scheme 3.2 Synthetic route for the synthesis of ligand 9

3.2 Synthesis of organoruthenium inhibitors

Reaction of two equivalents of imidazole-diethylphosphonate ester 1 with

[(η 6-arene)RuCl2]2 where arene = cymene or 1,3,5-triisopropyl-benzene (TIPB), yielded

mononuclear Ru(II) complexes 1a and 1b whereas treatment of imidazole-phosphonic acid 2 yielded complexes 2a and 2b as shown in Scheme 3.3 The pyridyl-imidazole ligand cleaved the

dinuclear ruthenium precursor as well as displacing a chlorido ligand by forming a stable

5-membered chelate with the metal centre Increased steric encumbrance afforded by the larger

TIPB ligand did not adversely affected the formation of the desired products Notably, 1a-b were

obtained as monocationic complexes with hexafluorophosphate as the counter-anion after anion exchange with NH4PF6 Anion exchange was necessary to improve the yields and purity of the

product On the other hand, complexes 2a-b were isolated directly from the reaction mixture

without anion exchange as the chloride salts Under similar reaction conditions,

benzimidazole-diethylphosphonate ester 3 and benzimidazole-phosphonic acid 4 gave similar organoruthenium complexes 3a-b and 4a-b in good yields All the complexes were isolated as yellow or orange- yellow solids Complexes with the diethylphosphonate ester group 1a-b and 3a-b were soluble in

common organic solvents namely dichloromethane, chloroform, MeOH, DMSO and acetone In

contrast, 2a-b and 4a-b are only soluble in polar solvents MeOH and DMSO presumably due to

the highly polar phosphonic acid groups All of the synthesized ruthenium complexes were soluble in water to concentrations exceeding 1 mM which is important for subsequent biological investigations

N N N

F F

P OOEt EtO

N N N

F F

P OOH HO

F F

P O OEt EtO

F F

P O OH HO

R R'

2a :R = imidazole, R' = cymene

2b : R = imidazole, R' = 1,3,5-iPr-benzene 4a : R = benzimidazole, R' = cymene 4b :R = benzimidazole, R' = 1,3,5-iPr-benzene

Cl

Cl Ru Ru

Cl

Cl

NH4PF6R'

R'

Cl

Cl Ru Ru Cl

Cl R'

R'

R

Cl

Scheme 3.3 Synthetic route for the synthesis of ruthenium(II)-arene complexes

The compounds were analyzed by 1H, 31P{1H}, 19F{1H}-NMR, ESI-MS and RP-HPLC A distinct feature of the ruthenium-arene complexes is the presence of resonances at 5-6 ppm due the aryl-CH protons of the facially-bound arene ligand The resonances were shifted upfield from the aromatic region, indicating a more shielded environment in the presence for the metal centre

In addition, the arene protons of the cymene ligand of complexes 1a-4a could be observed as

four sets of doublets, compared to two sets of doublets of the precursor, indicating the

desymmetrization of the arene ligand upon coordination of the imidazole or benzimidazole ligands Likewise, the six methyl protons on the isopropyl group were observed as two sets of doublets, as opposed to a set of doublet in the precursor This indicated that the ligands were bound at more than one coordination site around the metal centre Similar observations were

made for complexes 1b-4b with the TIPB ligand A set of triplet and a set of doublet were

observed in the 31P{1H}- and 19F{1H}-NMR respectively for all the complexes due to 2JPF

coupling All the complexes were observed as M+ parent molecular ions in the ESI-MS and confirmed with MS/MS fragmentation analysis However, organic CHN elemental analyses did not yield results consistent with the molecular formula of the desired compounds despite using crystalline and highly-pure samples In comparison, Ru content analysis on the samples by ICP-OES was within error limits We hypothesized that the discrepancy was due to the presence of the –CF2– group which could yield interfering HF on combustion, giving rise to inaccurate results In order to ascertain purity, RP-HPLC were carried out and the newly synthesized compounds were found to be >95% pure

We investigated whether it was possible to hydrolyze the phosphonate ester groups in these

organoruthenium complexes directly as facile entry to 2a-b and 4a-b The direct hydrolysis of

phosphonate ester groups has been reported for several octahedral polyaromatic ruthenium complexes[65] but not ruthenium-arene compounds which are more reactive and susceptible to

ligand displacement Treatment of 1a directly with TMSI followed by MeOH yielded

quantitatively an unknown species with a similar 1H NMR profile to 2a The 1H NMR spectrum

of the unknown complex revealed the disappearance of the ethyl peaks of the phosphonate ester

group indicating hydrolysis, while other peaks remained unchanged suggesting that organometallic scaffold remained intact Closer inspection by ESI-MS analysis however

suggested that the chlorido ligand coordinated to the Ru(II) centre had been displaced by an

iodido ligand, with a single peak observed at m/z 728 corresponding to [(ŋ6-cymene)RuI(2)]+ In

contrast, 2a is observed by ESI-MS as the parental molecular ion at m/z 636 The source of

iodido ligand was presumably excess TMSI reagent which underwent halide exchange reaction

at the Ru(II) centre after hydrolysis Although direct hydrolysis of phosphonate groups was technically feasible, further steps would be required to convert the iodido ligand back to the chlorido ligand and the method was thus abandoned

Single crystals of 1b, 2a and 3b suitable for X-ray diffraction studies were grown by layer

diffusion of diethyl ether into methanolic solutions of the complexes whereas single crystals of

4a and 4b was grown by slow evaporation of methanolic and aqueous solutions, respectively To

the best of our knowledge, these complexes are the first metallo-complexes reported which contain the difluoromethylphosphonate ester and difluoromethylphosphonic acid functional

groups respectively The structures of complexes 1b, 2a, 3b, 4a and 4b with atomic numbering

are depicted in Figure 3.1 and Figure 3.2 respectively and selected X-ray crystallographic data

are shown in Table A1 in the appendix Selected bond lengths and angles of complexes 1b and

4a are also shown in Table 3.1 The high R1 value observed for complex 3b was due to poor

quality of the crystal Some problems were also encountered when solving the crystal structures

of complexes 2a and 4b As the solvent molecules and chlorido ligand in 2a and 4b respectively

were highly disordered, their identities cannot be determined unambiguously Given the poor

X-ray diffraction data, analysis of the bond parameters of 2a, 3b and 4b was not carried out

Figure 3.2 Molecular representations of 2a, 4a and 4b; atoms are represented as spheres of arbitrary radii

The bond lengths and angles observed in these complexes are typical values of piano-stool

ruthenium-arene complexes The pyridyl-imidazole and pyridyl-benzimidazole rings in 1b and

4a were essentially in the planar conformation The solid state structure of 4a showed that the

complex crystallised as a zwitterionic structure with deprotonated monobasic phosphonate group The negative charge on the deprotonated phosphonate group was delocalized between O2-P1-O3 with similar O2-P1 and O3-P1 bond lengths of 1.481(3) Å and 1.512(3) Å as

observed in 4a The distances were consistent with a bond order of 1 to 2 In comparison, the

protonated O1-P1 bond was significantly longer at 1.558(3)

Table 3.1 Comparison of bond distances [Å] and angles [°] of 1b and 4a

3.3 Synthesis of organogold inhibitors

The reaction of one equivalent of ligand 9 with [Au(PR3)Cl] (where R = phenyl) precursor

in the presence of an equimolar of base yielded the mononuclear complex 9a as shown in Scheme 3.4 Compound 9 was first treated with base to deprotonate the acidic acetylene proton

to form the negatively charged ligand, followed by reaction with Au(I) precursor to afford the

neutral mononuclear complex 9a by displacement of the chlorido ligand from the labile Au-Cl

bond The complex was isolated as pale-yellow solid and was soluble in common organic

solvents namely dichloromethane, chloroform, MeOH and DMSO However, complex 9a was

only moderately soluble in water presumably due to the complex being in the neutral charge

Compound 9a was analyzed by 1H, 31P{1H}, 19F{1H}-NMR and ESI-MS The acetylene proton

of ligand 9 which was at a chemical shift of 3.16 ppm was no longer observed in complex 9a after complexation indicated the coordination of the alkyne to the gold metal center Ligand 9 is

most likely to coordinate to the gold metal center in a σ fashion In addition, a set of singlet and a set of triplet were observed in the 31P{1H}-NMR due to the presence of two phosphorous atoms

in complex 9a each coming from the triphenylphosphine ligand and ligand 9 respectively The

singlet resonance of triphenylphosphine at around 40 ppm is typical of that found for other alkynylphosphinegold(I) complexes.[66, 67] The set of triplet observed in the 31P{1H}-NMR together with another set of doublet observed in the 19F{1H}-NMR were due to 2

JP-F coupling

The formation of complex 9a was also supported by ESI-MS in which 9a showed as its M+

parent molecular ion in the ESI-MS mass spectrum, although the spectrum was dominated by signal corresponding to [Au(PR3)2]+ (100%) fragment, as has previously been reported for other alkynyl(phosphine)gold(I) complexes.[68, 69]

With the synthesis of the alkynyltriphenylphosphinegold(I) complex 9a established, a library of compounds can be generated by reaction of ligand 9 with various Au(I) precursors of

the form [Au(PR3)Cl] (where R = methyl, ethyl, cyclohexyl, etc.) following the general reaction

procedure Work is also on-going to synthesize the phosphonic acid derivative of ligand 9 and its

corresponding gold(I) phosphine complexes This novel class of gold(I)-based inhibitors will subsequently be evaluated for their inhibition towards PTPs, in particular PTP-1B and TC-PTP

Scheme 3.4 Synthetic route for the synthesis of gold(I) complexes

Chapter 4 Biological evaluation of organoruthenium PTP inhibitors

4.1 Aqueous stability of organoruthenium PTP inhibitors

The aqueous stabilities of 3a and 4a were investigated using UV-Vis spectroscopy over a

24 h period There were no significant shifts in their UV-Vis spectra in water over the 24 h period indicating good aqueous stability In the presence of 1 mM glutathione (GSH), an endogenous intracellular thiol-containing tripeptide, however, there was a significant blue shift

in both spectra of both compounds suggesting that organoruthenium complexes could potentially react with these nucleophiles Such reactions have been reported via a myriad of different reaction pathways resulting in both mononuclear and dinuclear species In the presence of 200

mM NaCl however, this reactivity was suppressed suggesting that the aquation of the Ru-Cl bond was important for their reactivity and not via the degradation of the imidazole linker or phosphonate group Therefore under physiological conditions at high chloride concentrations, the organoruthenium PTP inhibitors can be expected to maintain their high stability even in the presence of nucleophiles Upon cell entry, where the chloride levels are lower, reaction with intracellular nucleophiles may occur

0.5

0 0.7

0 0.7

220 270 320 370 420

0h 24h

Figure 4.1 UV-Vis aqueous stability studies of complex 4a (Left) UV-Vis spectrum of complex 4a in H2O (Right)

UV-Vis spectrum of complex 4a in 1mM GSH (Inset) UV-Vis spectrum of complex 4a in 1mM GSH + 200mM

NaCl

4.2 Inhibition of PTP-1B and TC-PTP by organoruthenium PTP inhibitors

The organoruthenium inhibitors are expected to bind to the PTP active-sites through their phenyl difluoromethylphosphonic acid moiety, which is a good non-hydrolyzable mimetic of pTyr.[33] The bulky organoruthenium scaffold can then be utilized to achieve selectivity amongst the PTPs by exploiting on the potential difference in structural environment peripheral to the substrate binding site Several metal complexes have been previously been reported to be strong inhibitors of PTP enzymes, presumably via direct covalent binding to the Cys215 at the active-site However, such an approach would render the metallo-inhibitor to be poorly selective across broad spectrum of PTPs since the enzymatic active-site would be highly conserved and homologous amongst with PTP family

The compounds were subjected to an initial screen at a concentration of 100 µM against

PTP-1B and TC-PTP using the para-nitrophenylphosphate (pNPP) assay Phosphatases catalyze the hydrolysis of the phosphate group on the pNPP substrate to yield para-nitrophenol which can

be quantitated by UV-Vis spectroscopy at A405 Briefly, the enzymes were treated with the

inhibitors for 30 min and pNPP substrate was added, incubated for a further 30 min before levels

of para-nitrophenol were determined ICP-OES was carried out to determine the amount of

ruthenium metal content in the stock solutions so as to determine the actual concentrations of the prepared solutions and to account for weighing error and purity of the compounds In addition, the same stock of organoruthenium solution was applied against two different enzymes to minimize errors arising to sample preparation Based on the initial screen as shown in Figure 4.2,

compounds containing the diethylphosphonate ester ligands, namely 1a-b and 3a-b, were

ineffective regardless of the nature of their Ru-arene fragment or linker groups Indeed only

compounds containing the phosphonic acid moiety, namely 2a-b and 4a-b, were efficacious

This was expected since the ester groups would increase the steric encumbrance around site targeting moiety and preventing effective binding In addition, it rendered the phosphonate group strongly hydrophobic and lowers the affinity towards the hydrophilic enzyme pocket Because the inhibitor contains a second-row transition metal with known affinity towards soft nucleophiles such cysteine, we investigated whether the Ru-arene fragment can itself

active-directly inhibit enzymatic activity Therefore, [(ŋ6-cymene)Ru(pyridylimidazole)Cl]PF6 5 and

[(ŋ6-cymene)Ru(pyridylbenzimidazole)Cl]PF6 6 which model the organoruthenium imidazole and organoruthenium benzimidazole fragments respectively and themselves do not contain phenyl difluoromethylphosphonic acid moiety were prepared These model compounds were themselves inactive indicating that the phenyl difluoromethylphosphonic acid groups were important and that the organoruthenium imidazole/benzimidazole groups were not solely

responsible for inhibitory activity observed in 2a-b and 4a-b

0 20 40 60 80 100 120

Figure 4.2. Inhibition of PTP-1B and TC-PTP at inhibitor concentration of 100 µM

Comparing between the organoruthenium inhibitors and their parent ligands, it was immediately obvious that at high concentrations, the parent ligand was equally efficacious against both PTP-1B and TC-PTP When the organoruthenium moiety was incorporated, selectivity against PTP-1B was drastically improved with the organoruthenium inhibitors being 7-11 fold more efficacious against PTP-1B This observation prompted a more detailed investigation examining a dose-response of the organoruthenium inhibitors against both PTP enzymes The IC50 values, which depicted the concentration at which the enzymatic activity is reduced to 50% level vs untreated controls, was shown in Table 4.1 As with the initial screen, the inhibitors were some 7-10 fold effective against PTP-1B than TC-PTP In general, with the incorporation of the ruthenium scaffold, the inhibition of PTP-1B is enhanced with respect to the parent ligands This enhanced effect could be due to favorable interactions of the ruthenium

scaffold in the inhibitors with amino-acid residues surrounding the active site in PTP-1B leading

to better binding of the complexes in PTP-1B

Figure 4.3 Dose-response curves of inhibition of compounds towards PTP-1B (bold) and TC-PTP (dotted)

Molecular docking studies of PTP-1B against 4a with ligand 4 as comparison were done

by our collaborator, Dr Yap Chun Wei from the Department of Pharmacy Results from the

docking study, as shown in Figure 4.4, indicated that the cymene ligand in 4a was able to interact

with Phe182 This additional interaction increased its binding energy to -10.84 kcal/mol, which

was stronger than the -9.90 kcal/mol seen in the parent ligand 4 On the other hand, the observed

diminishing effect in the potency of the inhibitors compared to their parent ligands towards

TC-PTP is postulated to be attributed to steric effects

moiety impeded the effective

Figure 4.4. Conformation of parent ligand

of 4a in PTP-1B substrate binding site (

substrate binding site are highlighted in yellow.

From these results, we can see that with the presence of the ruthenium scaffold in the structures, the complexes are able to inhibit PTP

towards TC-PTP was diminished

otherwise not prominent in the parent ligands Interestingly, the imidazole analogues are more potent in the inhibition of TC

postulated that the steric effect which

enables the compounds to bind relatively better in TC

However, this decrease in steric bulk ha

be attributed to steric effects whereby the bulkiness of the rutheniumeffective binding of the complexes at the active-site of the enzyme

Conformation of parent ligand 4 in PTP-1B substrate binding site (-9.90 kcal/mol) (left); conformation

1B substrate binding site (-10.84 kcal/mol) (right) The hydrophobic regions around of the phosphatase substrate binding site are highlighted in yellow

From these results, we can see that with the presence of the ruthenium scaffold in the

are able to inhibit PTP-1B with increased potency while the inhibiPTP was diminished As such, selectivity towards PTP-1B is achie

otherwise not prominent in the parent ligands Interestingly, the imidazole analogues are more potent in the inhibition of TC-PTP as compared to the benzimidazole counterparts We

he steric effect which the imidazole analogues exhibit is less pronounced and this enables the compounds to bind relatively better in TC-PTP thereby giving lower IC

decrease in steric bulk has no effect on PTP-1B as both the imidazole and

by the bulkiness of the ruthenium-arene

f the enzyme

9.90 kcal/mol) (left); conformation The hydrophobic regions around of the phosphatase

From these results, we can see that with the presence of the ruthenium scaffold in the

potency while the inhibition 1B is achieved which is otherwise not prominent in the parent ligands Interestingly, the imidazole analogues are more

PTP as compared to the benzimidazole counterparts We

exhibit is less pronounced and this PTP thereby giving lower IC50 values

1B as both the imidazole and