Search

Although CD1d and NKT cells have been proposed to have highly conserved functions in mammals, data on functions of CD1d and NKT cells in species other than humans and rodents are lacking. Upon stimulation with the CD1d-presented synthetic antigen α-galactosylceramide, human and rodent type I invariant NKT cells release large amounts of cytokines. The two bovine CD1D (boCD1D) genes have structural features that suggest that they cannot be translated into functional proteins expressed on the cell surface. Here we provide evidence that despite an intron–exon structure and signal peptide that are different from all other known CD1 genes, boCD1D can be translated into a protein that is expressed on the cell surface. However, in vivo treatment of cattle (Bos taurus) with 0.1, 1, or 10 µg kg–1 of the most commonly used α-galactosylceramide, which has a C26 fatty acid, did not lead to an increase in body temperature and serum cytokine levels of the animals. This lack of reactivity is not due to a complete inability of boCD1d to present glycosphingolipids because α-galactosylceramide variants with shorter fatty acids could be presented by boCD1d to human NKT cells in vitro. This suggests that the natural ligands of boCD1d are smaller lipids.

BackgroundTo date, only few compounds targeting the AI-2 based quorum sensing (QS) system are known. In the present study, we screened cinnamaldehyde and substituted cinnamaldehydes for their ability to interfere with AI-2 based QS. The mechanism of QS inhibition was elucidated by measuring the effect on bioluminescence in several Vibrio harveyi mutants. We also studied in vitro the ability of these compounds to interfere with biofilm formation, stress response and virulence of Vibrio spp. The compounds were also evaluated in an in vivo assay measuring the reduction of Vibrio harveyi virulence towards Artemia shrimp.ResultsOur results indicate that cinnamaldehyde and several substituted derivatives interfere with AI-2 based QS without inhibiting bacterial growth. The active compounds neither interfered with the bioluminescence system as such, nor with the production of AI-2. Study of the effect in various mutants suggested that the target protein is LuxR. Mobility shift assays revealed a decreased DNA-binding ability of LuxR. The compounds were further shown to (i) inhibit biofilm formation in several Vibrio spp., (ii) result in a reduced ability to survive starvation and antibiotic treatment, (iii) reduce pigment and protease production in Vibrio anguillarum and (iv) protect gnotobiotic Artemia shrimp against virulent Vibrio harveyi BB120.ConclusionCinnamaldehyde and cinnamaldehyde derivatives interfere with AI-2 based QS in various Vibrio spp. by decreasing the DNA-binding ability of LuxR. The use of these compounds resulted in several marked phenotypic changes, including reduced virulence and increased susceptibility to stress. Since inhibitors of AI-2 based quorum sensing are rare, and considering the role of AI-2 in several processes these compounds may be useful leads towards antipathogenic drugs.

A series of novel 3'-amido-3'-deoxy-N(6)-(1-naphthylmethyl)adenosines was synthesized applying a polymer-assisted solution phase (PASP) protocol and was tested for anti-malarial activity versus the Dd2 strain of Plasmodium falciparum. Further, this series and 62 adenosine derivatives were analyzed regarding 1-deoxy-d-xylulose 5-phosphate (DOXP) reductoisomerase inhibition. Biological evaluations revealed that the investigated 3',N(6)-disubstituted adenosine derivatives displayed moderate but significant activity against the P. falciparum parasite in the low-micromolar range. On the molecular level, DOXP reductoisomerase utilizing an adenosyl-containing substrate was identified as a promising metabolic target for ligands of adenosine binding motifs.

With focus on exploring the structural demands of adenosine triphosphate pockets of different target enzymes, the 3'-amido-3'-deoxyadenosines 1-7 were synthesized. The inhibitory activity of these compounds on a selected cyclin-dependent kinase as model target in anticancer research was evaluated in vitro. A starfish oocyte enzyme based assay revealed a decreased inhibitory activity in comparison to adenosine. Consequently, the introduction of spacefilling lipophilic 3'-amido substituents alters the enzyme inhibition in an unfavorable manner.

P1-receptors, or adenosine receptors, and P2Y receptors, are G-protein-coupled receptors, belonging to the purinergic receptor family. Adenosine is the natural ligand of the P1 receptors, while P2Y receptors are activated by nucleotide ligands (ATP,ADP, UTP and UDP). The adenosine receptors (P1) consist of 4 receptor subtypes: A1, A2A, A2B and A3. The P2Y receptors are subdivided in 8 subtypes: P2Y1, P2Y2,P2Y4, P2Y6, P2Y11, P2Y12, P2Y13, P2Y14. In the first part of this work (chapter 2 and 3), we present the synthesis and biological evalutation of two series of adenosine A3 receptor ligands. In the second part (chapter 4), we discuss the synthesis and biological evaluation of new P2Y2 receptor ligands. Adenosine receptors are not only the major targets of caffeine, the most commonly consumed drug in the world, they could also be promising therapeutic targets in a wide range of conditions, including cardiovascular, inflammatory and neurodegenerative diseases. As a consequence of their numerous therapeutic possibilities and the ubiquity of the ARs, synthetic ligands need high selectivity with respect to receptor subtype and tissue to be of therapeutic value. The main approach for the discovery of selective A3AR ligands, has been the modification of the natural, non-selective ligand adenosine. Previous investigations showed that introduction of a 3’-amino group improved the selectivity for the human A3AR, while enhancing the water solubility. By introducing extra modifications, we evaluated the possibility of overcoming the A3AR affinity drop caused by this 3’-modification. We synthesized a series of hypermodified adenosine derivatives (chapter 2) and demonstrated that introduction of a 2-phenylethynyl substituent in concert with a N6-methyl group restored the A3AR receptor affinity. Based on the good results obtained with the 2-phenylethynyl modification, both on A3AR receptor selectivity and affinity, we decided to further explore the modifications at the 2-position. Therefore, we synthesized a series of 4-substituted 2-(1,2,3)-triazol- 1-yl N6-methyl adenosine analogues via Huisgen [3+2]cycloaddition (chapter 3). Combined with an unmodified ribose moiety, such 2-(1,2,3-triazol-1-yl) substitution resulted in very potent and selective A3AR antagonists, for example 2-(4- cyclopentylmethyl-1,2,3-triazol-1-yl)-N6-methyl-9-(-D-ribofuranosyl)adenine (3.10, Ki (A3AR) = 1.3 nM). The reduced efficacy resulting from such modifications at the 2-position were restored by the 4’-ethylcarbamoyl ribose modification, resulting in a series of potent an selective A3AR agonists, with 9-(5-ethylcarbamoyl--Dribofuranosyl)- N6-methyl-2-(4-pyridin-2-yl-1,2,3-triazol-1-yl) adenine as the most potent one (3.17b, Ki (A3AR) = 1.8nM). P2Y2 agonists are promising potential therapeutics for cystic fibrosis, cancer and dry eye syndrome. P2Y2 antagonists exhibit anti-inflammatory and neuroprotective effects and have been suggested as potential treatment for coronary vasospastic disorders. As for the A3AR receptor ligands, the natural, non-selective ligand uridine 5’-triphosphate was modified to gain access to more selective P2Y2 receptor ligands. Previous investigation demonstrated that the 2-thiouracil modification and the 2’-amino-2’-deoxy modification enhanced both P2Y2 receptor affinity and selectivity. Therefore, we combined these modifications, which resulted in the very potent and selective P2Y2 receptor agonist 2’-amino-2’-deoxy-2-thiouridine 5’-triphosphate (4.2, EC50 (P2Y2) = 8 nM) (chapter 4). UTP and its diverse synthesized analogues are sensitive to enzymatic degradation. Replacement of the -phosphate group by an isosteric phosphonomethyl group should enhance the metabolic stability. However, the resulting uridine 5’- phosphonodiphosphate (4.7) suffered from complete chemical hydrolysis to the corresponding 5’-phosphonate (4.50) in the buffer storage solution (chapter 4). In summary, in this thesis we succeeded to identify new selective ligands for both envisaged receptor subtypes: the A3A and the P2Y2 receptors.

This Ph.D. study constitutes part of an extensive project dealing with the design and development of new agents with antimalarial activity. The enzyme 1-deoxy-D-xylulose-5-phosphate reductoisomerase (DXR), which is involved in the DOXP pathway for isoprenoid synthesis, was chosen as drug target. This non-mevalonate pathway is not operating in humans and inhibitors of DXR can therefore be considered as selective for parasites. Fosmidomycin and its methyl homologue FR900098, well-known inhibitors of this enzyme that show antimalarial activity in vivo, were used as lead compounds for inhibitor design. The structure of fosmidomycin is characterized by a retrohydroxamic acid group, which acts as a ligand for the metal ion in DXR, and a phosphonate group, which mimics the phosphate group in the substrate structure. It was our objective to further expand the structure-activity relationship (SAR) of fosmidomycin as DXR inhibitor and to discover analogues showing superior activity against the enzyme and/or better antimalarial activity. Modifications were applied to the retrohydroxamic acid group and the propyl chain of the parent structures. Synthetic pathways to these modified fosmidomycines were elaborated. As initial screening method all new compounds were tested for their ability to inhibit DXR of Escherichia coli. The most promising inhibitors were further tested in an in vitro assay on two different Plasmodium falciparum strains. Our investigations are discussed in 4 chapters. 1. To explore the impact of structural modifications of the hydroxyl group of the retrohydroxamic moiety, analogues were synthesized with methyl, ethyl or hydroxyl ethyl substituents on the nitrogen (2.1-2.9). These modifications did not produce any acceptable inhibitors, which demonstrates the importance of the hydroxyl group in ligating the divalent metal. 2. A synthetic procedure for the preparation of α-aryl-substituted fosmidomycin analogues was developed starting from the corresponding benzylphosphonates or the appropriate cinnamaldehydes. Seven analogues were synthesized, allowing to perform an initial SAR study of this α-aryl series (3.1c,e and 3.2a-e). Although the α-aryl analogues were generally weaker DXR inhibitors than fosmidomycin, these analogues unambiguously surpassed the activity of fosmidomycin in inhibiting P. falciparum growth. Remarkably, the formyl analogues 3.1c and 3.1e consistently outperformed the acetyl derivatives 3.2c and 3.2e, both in the enzyme and the parasite growth inhibition assays. Compound 3.1e emerged as the most promising analogue with an IC50 value of 0.059 μM. Due to their straightforward synthesis these compounds are suitable for upscaling in further drug development programmes. The deoxygenated sideproducts 3.17a,c demonstrated the contribution of the hydroxyl group to the overall affinity. The Michael addition procedure, used to prepare the α-substituted analogues from phenyl-substituted cinnamaldehydes, proved also valuable to prepare the fosmidomycin cyclopentyl analogues 3.3 and 3.4. These compounds, featuring a restricted mobility, showed that a trans orientation of the retrohydroxamic acid moiety and the phosphonate group is optimal for binding the enzyme. 3. An efficient and reliable synthesis was established for β- and γ-oxa fosmidomycin analogues. Analogue 4.4, which combines a β-oxa modification with a hydroxamate moiety was almost as potent in inhibiting DXR as FR900098. In an in vitro assay it proved almost twice as active as FR900098 in inhibiting the growth of a P.falciparum 3D7 strain. 4. Several synthetic pathways towards the transition state analogue 5.1 were explored. A 10-step sequence, involving an asymmetric dihydroxylation, afforded compound 5.1 in a reasonable yield. Unfortunately, this compound did not show any worth mentioning inhibitory activity. In conclusion, this Ph.D. thesis highlights some important structural requirements for DXR inhibitors. This study also provides valuable information for modifications on fosmidomycin toward new and more efficient inhibitors: introducing aryl modifications on the α-position of the phosphonate moiety significantly improves the antiplasmodial activity in vitro and one oxa analogue with a hydroxamic acid proved to be a more potent inhibitor and to show superior antiplasmodial activity than the parent compound, fosmidomycin. Further investigation on analogues 3.1e and 4.4 is desirable since these compounds are significantly more potent than fosmidomycin on P. falciparum strains. Especially the latter one is very interesting, since it can be easily synthesized. Before further development the toxicological properties of these analogues need to be investigated and their pharmacokinetic profile needs to be established. In vivo studies, employing murine and monkeys models, are required. Further drug design may involve other sterically hindered analogues as well as prodrugs of the most promising analogues. VULGARISERENDE SAMENVATTING: Elke 40 seconden sterft er op de wereld een kind aan de gevolgen van malaria. De bestaande medicatie om malaria te behandelen verliest zijn werkzaamheid ten gevolge van resistentie van de parasiet voor de bestaande medicatie. Er is dan ook een urgente nood aan nieuwe, efficiënte, goedkope en veilige geneesmiddelen. In Gent aan het laboratorium voor medicinale chemie van de faculteit farmaceutische wetenschappen wordt onderzoek gedaan naar nieuwe geneesmiddelen. Recent zijn daar twee nieuwe fosmidomycine analogen ontdekt die zeer beloftevol bleken in het remmen van de malariaparasiet. Deze nieuwe antibiotica zouden een doorbraak kunnen betekenen in de behandeling van malaria.

Dit doctoraatswerk maakt deel uit van een omstandig project over de synthese van biologisch relevante sfingoïd- en ceramide-analogen. In dit opzicht werden twee verschillende klassen van verbindingen gesynthetiseerd. In een eerste deel werd er geopteerd om hybride PDMP analogen the synthetiseren, die zowel op PDMP als op E-styreen analogen van natuurlijk ceramide gebaseerd zijn. Aangezien preliminaire synthetische routes die gebaseerd waren op enkelvoudige bescherming van de sfingoïd stikstof resulteerden in ongewenste cyclisatie reakties, werd er geopteerd om een dubbel bescherming in te voeren waarbij de tert-Boc beschermde amine functie en het secundare alcohol van de sfingoïd keten in een oxazolidine ring werden vastgelegd. Bijgevolg werden de gewenste E-styreen analogen uitgaande van D-serine gesynthetiseerd in 16 stappen. De synthetische route werd zodanig opgevat dat verschillende aromaten op eenvoudige wijze zullen ingevoerd kunnen worden bij verdere synthetische modificaties. Er werd eveneens toegang verkregen tot een aantal structuuranalogen die meer inzicht konden verschaffen in de sructuur-aktiviteitsrelatie van deze klasse van verbindingen. In vitro biologische evaluatie op Golgi extracten en in vivo evaluatie op HEK-293 en COS-7 cellen bracht twee gelijkwaardige verbindingen verbindingen in vergelijking met PDMP aan het licht, die verder kunnen uitgewerkt worden tot meer potente inhibitoren. In een tweede luik werd de synthese van een nieuwe klasse van homoceramiden, N-homoceramiden genaamd, uitgewerkt. Deze verbindingen bevatten een extra koolstof atoom tussen de N-acyl keten en C2 van de sfingoïd keten. Om dit te bewerkstelligen werd zowel vanuit een suiker als door middel van regioselectieve epoxide opening getracht om een cyanide in te voeren. Aangezien deze benaderingswijze niet voldeed voor de synthese van de gewenste verbindingen, werd geopteerd om de vertakking in te voeren door middel van 1,3-dithiaan. Bijgevolg werden de gewenste N-homo(dihydro)ceramiden in 20 stappen gesynthetiseerd, uitgaande van D-isoascorbinezuur, een veel gebruikt bewaarmiddel in de voedingsindustrie. De sleutelreaktie in deze benaderingswijze was de volledig regioselectieve epoxide opening met lithium 1,3-dithiaan. Daarenboven stelde een volledig regioselective Grignard reaktie ons in staat om D-ribo-homophytoceramide te synthetiseren. This Ph.D. study constitutes part of an extensive project dealing with the synthesis of biologically relevant sphingoid and ceramide analogues. In this context, we effected the synthesis of two different classes of sphingolipid analogues. In a first part, we aimed at the synthesis of hybrid PDMP analogues, based both on PDMP and styryl analogues of natural ceramide. Since preliminary synthetic approaches based on a mono-protective strategy resulted in undesired cyclisation products, a double-protective strategy was devised in which the secondary alcohol and the tert-Boc protected amine were locked in an oxazolidine ring. Hence, the desired E-styryl PDMP analogues have been synthesised in 16 steps starting from D-serine. The synthetic route was developed such that future introduction of different aryl groups is straightforward. Throughout the synthetic coarse, we gained access to a number of structural analogues, which provided more insight into the structure-activity relationship of this class of compounds. Biological evaluation, both in vitro on rat liver Golgi extracts as in vivo on HEK-293 and COS-7 cells revealed two lead compounds with comparable inhibitory potency as PDMP, which could be elaborated to more potent inhibitors. In a second part, we aimed at the synthesis of a new class of homoceramides, named N-homoceramides, which contain an extra methylene-spacer between the N-acyl chain and C2 of the sphingoid backbone. To this aim, we evaluated the introduction of branching by use of a cyano group, both in a sugar approach as in regioselective epoxide opening. Since these experiments failed to provide a reliable means for introduction of branching, we opted to introduce an extra methylene spacer by use of 1,3-dithiane. Hence, the desired N-homo(dihydro)ceramides have been synthesized in 20 steps starting from D-ascorbic acid, a common food preservative. Key reaction in this synthetic approach was a fully regioselective epoxide opening with lithium 1,3-dithiane. In addition, a fully stereoselective Grignard reaction gave access to D-ribo-N-homophytoceramide.

SUMMARY The aim of this work was the search for selective inhibitors of M. tuberculosis thymidine monophosphate kinase (TMPKmt) as leads for the development of new anti-tuberculosis agents. Based on the X-ray structure of TMPKmt, it was decided to establish a preliminary structure-activity relationship by synthesising a series of nucleosides and nucleotides modified at the 2’-, 3’- and 5-positions of the dTMP-scaffold. These analogues were tested for their affinities for TMPKmt. From this, a 2'-chlorine and a 2'-fluorine substituent emerged as the most promising modifications of the dTMP scaffold. Furthermore, the affinities of a series nucleosides and their corresponding nucleotides were compared. The deletion of the phosphate moiety typically resulted in a modest, in many cases negligible affinity loss. In view of the drug delivery problems of phosphorylated compounds, nucleosides seemed more useful leads for further drug design. The goal of the synthesis of a series 3'-C-branched-chain nucleosides and nucleotides (3'-CH2N3, 3'-CH2NH2, 3'-CH2F, 3'-CH2OH) was to occupy a cavity in the enzyme near the 3'-position. Biological results and modeling confirmed this hypothesis. 3'-Azidomethyl- 3'-deoxy-thymidine (4.9), combining a low Ki-value (40 mM) with a high selectivity index for the tuberculosis enzyme (Ki TMPKh/ Ki TMPKmt = 26) emerged as the most promising lead for further optimisation. Attempts were made to combine these favourable 3'-substitution patterns with 2'-halogen substituents (6.9-6.14). However, introduction of the 2'-halogen led to a drastic decrease in affinity compared to the corresponding 2'-deoxy-nucleosides. Probably, the 2'-halogens compete with the 3'-substituents for the same binding pocket. Furthermore, it was tried to supersede the good affinities of the 3'-C-branched-chain nucleosides, by further exploration of the enzyme cavity near the 3'-position with alternative nitrogen-containing substituents. However, attempted simultaneous reduction of the 6'-azido function and the 2'-hydroxyl of compound 19 (chapter 6) failed. Instead three peculiar nucleoside analogues (6.10, 6.11 and 6.12) were isolated that were tested for their affinities for TMPKmt and TMPKh. Dinucleoside 6.12 unexpectedly showed a high affinity (Ki = 37 mM) and selectivity for the tuberculosis enzyme, indicating an exceptional flexibility of TMPKmt towards the orientation of the sugar ring. Also the bicyclic nucleosides 6.10 and 6.11 showed excellent affinities. With its Ki-value of 3.5 mM (exceeding the affinity of the natural substrate) and its selectivity index of 200, 6.10 represents the highest affinity and most selective inhibitor of TMPKmt found so far. The favourable affinities and selectivities of these three inhibitors will form the basis for further drug design, considerably increasing the variety of nucleoside analogues that may be envisaged for future synthesis. Based on the encouraging affinity of anhydro-hexitol nucleotide 7.1, the effects of sixmembered sugar rings for the affinity for TMPKmt were further explored through the synthesis of 1-[2,4-dideoxy-4-C-hydroxymethyl-a-L-lyxopyranosyl]thymine (7.4). This nucleoside indeed showed the predicted equatorial orientation of the thymine ring (4C1). Affinity results were however disappointing. In conclusion, this work yielded some high affinity and selective inhibitors for TMPKmt that open interesting perspectives in the search for new anti-tuberculosis agents. Based on their high selectivity indices, especially dinucleoside 6.12 and bicyclic nucleosides 6.10 and 6.11 will be used as the starting point for the search of inhibitors with an optimal fitting in the active pocket of TMPKmt.

SAMENVATTING In Deel I (Hoofdstuk 1-2) van dit werk beschreven we de synthese van gemodificeerde adenosine en thymidine analogen (zie Figuur 1). Deze verbindingen werden als moleculaire probes gebruikt voor de biologische evaluatie van adenosine-receptoren (ARn, Deel II, Hoofdstuk 3-6), die behoren tot de G-proteïne gekoppelde receptor (GPCR) superfamilie, en als inhibitoren voor thymidine-monofosfaat kinase (Deel III, Hoofdstuk 7-8), een potentieel antituberculose doelwit. NONNOHOHNNH2OHClNHINHCH3NH2CH2NH2ONHCH3N3CH2N3NHClMeO263'5'A NH2N3NNHOOHOOOHOHOHClF3'2'13B Figuur 1 Overzicht van de verrichte adenosine (A) en thymidine modificaties (B) Deel II richtte zich voornamelijk op de adenosine A receptor (A33AR), de meest recente AR. Reeds vele structurele variaties van het adenosine-skelet werden onderzocht met het oog op AAR affiniteit en selectiviteit, waaronder N6-benzyl-3 157 substituties en kleinere functionele groepen in 2-positie van het purine gedeelte, al of niet in combinatie met de introductie van een 5’-methylcarbamoyl in het ribofuranose deel. Door bovengenoemde veranderingen aan het adenosine-skelet te combineren met 3’-azido/amino(methyl)-modificaties (zie Figuur 1A), wilden we de invloed bestuderen van dergelijk (gecombineerd) substitutiepatroon op de A3AR selectiviteit, affiniteit en intrinsieke activiteit. Vele van de gesynthetiseerde analogen vertoonden matige tot hoge A3AR affiniteit en selectiviteit. Het beste derivaat uit deze serie was analoog 4.19 (K hAi3AR = 27 nM en A1/A3 selectiviteit > 350). Interessant bij deze verbindingen was dat de intrinsieke activiteit gemoduleerd kon worden afhankelijk van de substituent in 3’-positie: - een 3’-aminogroep (zoals bv. 4.14, 4.16 en 4.19) resulteerde in (sterk) partieel agonisme; - de azide precursoren (zoals bv. 4.11, 4.13 en 4.17) waren allen antagonisten; en - introductie van een methylene-spacer zoals bij 4.1-10, 4.12, 4.15, 4.18 en 4.20 ging ten koste van de intrinsieke activiteit. Ook bestudeerden we de rol van deze gemodificeerde adenosine analogen als tools voor het structuurgebaseerd onderzoek van ARn. Door organische synthese en moleculaire modeling te integreren in het “neoceptor-neoligand” concept, konden we (in samenwerking met dr. Kenneth A. Jacobson) meer inzicht verwerven in de moleculaire complementariteit van wild-type en gemuteerde A3 (Hoofdstuk 5) en A2A (Hoofdstuk 6) ARn. In afwezigheid van exacte structurele kennis van verschillende GPCRn, liet deze aanpak toe met behulp van verbindingen zoals 4.7 en 5.14 de (AAR en A2A3AR) homologie modellen van de te valideren. Bovendien, toonden we in Hoofdstuk 5 aan dat de affiniteit van 3’-amino gemodificeerde adenosine analogen (neoliganden) hoger was voor de H272E gemuteerde A3AR (neoceptor) dan voor de wild-type A3AR. Hiermee bewezen we dat de A3AR (of GPCRn in het algemeen) ontworpen kunnen worden om selectief met synthetische liganden te interageren. 158 In Hoofdstuk 6 beschreven we de verschillende A2AAR-bindingskarakteristieken voor agonisten en antagonisten op basis van het ontworpen 3-D homologie model. Een model dat gevalideerd werd door nieuwe neoceptor (T88D)-neoligand-paren te beschrijven. In Deel III (Hoofdstukken 7-8) bestudeerden we het potentieel van 2’,3’-gemodificeerde thymidine analogen als inhibitoren voor Mycobacterium tuberculosis thymidine-monofosfaat kinase, een nieuw antituberculose doelwit. De (onverwacht) lage affiniteit (d.i. Ki tussen 118 en 1260 μM) van deze verbindingen, spoorde ons aan om hun conformationele eigenschappen nader te bestuderen. De conformationele analyse van verbindingen 8.7 en 8.8 (respectievelijk met Ki’s = 118 en 190 μM) toonde een mogelijk verband aan tussen de matige affiniteit en een suboptimale positionering van de 2’- en 3’-substituenten, toe te schrijven aan de uitgesproken voorkeur voor de noord conformatie.