Glycosidase Inhibitors and Fluorous Tags

Posted on August 24, 2010 by Marv

The use of fluorine to modulate physical and biological properties of molecules is well established in drug discovery, either for therapeutics or diagnostics.  Usually this means a single fluorine substitution or a trifluoromethyl for a methyl group.  Occasionally, one might see multiple trifluoromethyl groups incorporated such as using hexafluoroleucine in place of leucine in peptides.  Using more densely fluorinated moieties, however, is quite rare, although it has been done recently to produce 19F MRI imaging agents.  An added benefit of using fluorous, as opposed to just fluorine, modifications is that synthesis and purification can be facilitated through the use of fluorous separations.

A recent paper in ChemBioChem co-authored by several groups describes their use of fluorous moieties to produce reversible inhibitors of glycosidases.  In this instance, the fluorous modified iminoalditols, sugars where the ring oxygen has been replaced by a nitrogen, which are known to be glycosidase inhibitors.  Shown above are some representative structures.  The authors write that alkyl substitution leading to lipophilic iminoalditols has already been shown to provide potent inhibitors of various glycosidases.  They hypothesized that by using highly hydrophobic perfluoro groups that they could achieve even better results.  With that in mind they synthesized eight different flourous modified iminoalditols, three of which are shown.  When appropriate, they reported using fluorous solid-phase extraction (FSPE) as a purification method for intermediates and products.  They then screened the compounds against various glycosidases and found that many of the fluorous iminoalditols had smaller Ki’s and greater selectivity than the corresponding unsubstituted iminoalditol.

Rather than go into detail on the assay results, I’d like to spend the rest of this post describing something that I found interesting.  As seen in the scheme below, one of the fluorous modified compounds they made started from hexafluoroisopropanol (HFIP).  Mitsunobu reaction with the protected aminohexanol shown actually provided a fluorous acetal from two Mitsunobu reactions in 82% yield rather than just the single substitution that they desired (Somewhat surprising). They used FSPE to purify the product (Interesting).  While the authors did not describe the exact FSPE wash conditions they used, it seems that the fluorous acetal is retained on the fluorous silica gel.  The stability of the acetal is seen in the last step where they are able to selectively hydrolyze the acetonide under acid conditions.  Not real surprising since cation formation at that carbon is going to more difficult.

One of the concerns of fluorous chemistry is the persistance and bioaccumulative properties of perfluorocarbon chains.  This has certainly dampened the prospects of fluorous chemistry as a green chemistry strategy as was originally envisioned.  What’s needed are alternative fluorous chains which are shorter and more easily degradeable, yet retain the “fluorousness” necessary to partition into fluorous environments.  So here we have a fluorous moiety which is relatively easy to obtain and presumably could be degraded through more stringent acid treatment.  Unfortunately, the MSDS of HFIP does not provide any information on the persistence, degradability, or bioaccumulative potential of HFIP.  Of course, even if those are non-issues there are other things that have to be sorted out.  For instance, how fluorous is the acetal?  Is it equivalent to a C6F13 chain?  How stable is it to a variety of reaction conditions?  I could see that proton on the end being a problem.  Can other related acetals be made?  All interesting questions which might deserve some investigation.

Posted in Life Science Applications | Leave a comment

Evaluating FTI Chemical Libraries

Posted on August 12, 2010 by Marv

To date fluorous methods have been used for small molecule synthesis more than any other application.  The value of using fluorous techniques is that it provides a versatile, easy, and reliable and method for the purification of synthetic intermediates and products.  The generality of fluorous separations makes it particularly well-suited for library synthesis, either using fluorous tags or fluorous reagents and scavengers.

Does it really though?

There have been a few papers comparing fluorous methods to others in library production and they have validated the value of fluorous methods . Another way perhaps to judge is to look at FTI’s participation in the NIH Pilot-Scale Libraries (PSL) intiative.  Fluorous Technologies, Inc. (FTI) is one of only two industrial labs to have been awarded by NIH a Pilot-Scale Libraries grant.  The purpose of the grant is to augment the NIH’s compound collection, the Molecular Libraries Small Molecule Repository (MLSMR).  The MLSMR is part of NIH’s Molecular Library Program and is the source of compounds for another element, the Molecular Libraries Screening Center Network (MLSCN) which is tasked with developing assays and identifying small molecule probes which can be used to study various gene and cell functions to better understand disease states.

Over the last two and half years FTI has submitted over 1600 compounds into the MLSMR of which 1433 have been entered into PubChem.  All of these compounds are de novo compounds made by methodologies developed at FTI.  A PubChem search found that this was more compounds on a per annum basis than any other PSL grant awardee.  Out of the 30 or so awardees only two have submitted more compounds than FTI and both have had PSL grants for almost six years.   We’ve have never had a compound not accepted into the MLSMR due to purity or other issues.  So in terms of absolute number and purity of compounds FTI, and fluorous methodology, is doing pretty well.

Of course that’s only part of the story.  Anyone could make large number of compounds if they have little or no diversity, but that wouldn’t be of much value in finding new chemical probes for various assays.  Nor would libraries of molecules with very little chance of exhibiting biological activity.  Using PubChem’s search and structure clustering functions , I compared FTI’s submitted compounds with the group that submitted the most compounds over the lifetime of the PSL initiative, Prof. Stuart Schreiber’s group at the Broad Institute.

As seen in the table above, the numbers are actually quite similar.  The Broad Institute’s % of active compounds is somewhat higher, but that could just reflect the extra time which translates into more assays being run on their compounds.  What’s interesting is that the diversity as measured by compounds/similarity score cluster is almost identical at the 0.9 level; 46 vs. 47.  FTI’s diversity amongst the active compounds was actually little better, although how meaningful that difference may be is not clear.  At the 0.8 similarity score level (data not shown), however, the Broad Institute’s diversity was clearly higher than FTI’s for all compounds, but about the same for active compounds.

Purely based on number of compounds submitted, Prof. Schreiber’s group is one of the top performers amongst all PSL grant awardees.  He is probably the biggest proponent of diversity-oriented synthesis (DOS), so you would expect a high degree of diversity.  In addition, his labs are one of the best equipped in the world, far beyond FTI’s resources.  Yet, FTI has managed to deposit more compounds per year, achieve comparable activity rates and diversity.  There is the caveat that I have no idea of the budget or man-hours committed by the Broad Institute to the PSL grant, so perhaps the comparison isn’t valid.  On the face of it at least though, FTI has been able to produce almost as much in half the time.  Only through the use of fluorous techniques could we achieve this level of productivity with our resources.

Posted in Assorted Fluorous Items, Small Molecule Synthesis | Tagged , , | Leave a comment

Phase Vanishing Reactions

Posted on August 10, 2010 by Marv

Phase vanishing (PV) reactions are a specific subgroup of reactions which use a fluorous phase, generally fluorous solvents, as a liquid membrane.  The liquid membrane keeps substrates separated from reactants.  The reactant can then diffuse through the fluorous phase to react with the substrate at the interface of the fluorous phase and the substrate phase.  All of the reactant phase eventually diffuses across the fluorous phase and disappears giving the name phase vanishing.  Since the reactants diffuse slowly through the fluorous membrane, PV reactions are especially useful as an alternative to addition controlled reactions where highly reactive or exothermic reagents are used.  These reactions are related to fluorous biphasic reactions and fluorous liquid-liquid extraction (FLLE) since they are based on the immiscibility of the fluorous phase with other other liquid phases.

PV reactions were first described in 2002 by Ryu and Curran.  Since that initial report, a wide variety of creative variations on the PV theme have been reported.  These include variations in the reagents, the number of phases, the fluorous solvents used, reaction apparatus, etc.  For example, while halogenations have been the most prevalent type of reaction, Simmons-Smith cyclopropanations, halolactonizations, and photochemical reactions have all been conducted in PV mode.  A recent short “Concept” paper from Dragojlovic in Chemistry – A European Journal describes many of the variations of PV reactions that have been developed.  Some of the more interesting ones include tetraphasic reactions, multi-reaction reaction vessels, and polytetrafluoroethylene (PTFE) screen reactions.

Prof. Ryu meanwhile has continued working in this area and has just published a photoirradiative PV reaction using bromine for the anti-Markovnikov addition of HBr to terminal alkenes through a radical addition process.   They found that the PV reaction worked quite well for monosubstituted terminal olefins providing 1-bormo-alkanes in >90% yield.  Just add this to the many different types of PV reactions.

Posted in Assorted Fluorous Items, Small Molecule Synthesis | Tagged | Leave a comment

Carbohydrate Research in Action

Posted on August 6, 2010 by Marv

We’ve published quite a number of posts in F-Blog regarding the synthesis and purification of oligosaccharides using fluorous methods.  The justification for the development of many synthetic methods, including fluorous carbohydrate synthesis, is that the molecules produced can then be used to probe biological functions which would ultimately lead to the betterment of human health.

In a recent ASAP paper in J. Am Chem. Soc. Prof Nicola Pohl and co-workers have used synthetic oligosaccharides produced using fluorous means to work in modulating immune response to pathogens.  Oligosaccharides are very common molecules on the surface of bacteria and viruses and are important in recognition and response by immune systems.  The researchers’ goal was to investigate the role of these oligosaccharides in eliciting an immune response.  Their approach was to produce what they call an “artificial pathogen”; something of similar size to a real pathogen with the appropriate oligosaccharides on the surface.  The oligosaccharide chosen was the trimannose molecule labeled 4.  This trisaccharide is the terminal cap for glycolipids in both Mycobacterium tuberculosis and Leishmania parasites.  The artificial pathogen would then be incubated with stimulated macrophages and the levels of cytokine ineterleukin-12 (IL-12) produced measured.  IL-12 is critical for an appropriate immune response.

The trisaccharide and a control lactososide were both prepared using fluorous methods previously developed within Prof. Pohl’s labs. Fluorous solid phase extraction (FSPE) purification is a prime component of this methodology.  The fluorous tag was removed and the sugars attached to 1 micron diameter latex beads to provide the pathogen mimics.  Incubation with the stimulated macrophages resulted in a dampened immune response as measured by IL-12 produced for macrophages exposed to the trimannose coated beads vs. the control lactososide beads.  The researchers, therefore, demonstrated that the carbohydrate cap alone can influence immune response providing a path to new immunomodulators or vaccines.

Posted in Biomolecule Synthesis, Carbohydrates, Life Science Applications | Tagged , , | Comments Off

Synthetic Oligonucleotide Purification by Polymerization

Posted on August 3, 2010 by Marv

An ASAP Organic Letters paper from Fang and Fueangfung at Michigan Technological University describes their work in using phase tags for the purification of synthetic oligonucleotides.  In this case the phase tag is a polymerizable tag attached to the terminal phosphoramidite in a solid phase synthesis.  So the general strategy is very similar to that used with other phase tags, including fluorous.  The oligomer is synthesized on a solid-support with capping reactions in between couplings.  The capping truncates coupling failures so as to prevent n-1 deletion sequences from accumulating.  The final coupling is conducted with a phase tagged monomer.  Assuming all the cappings worked well the only phase tagged final oligomer will be the desired full sequence.  After removal from the solid-support, the phase tag is then used to separate the desired sequence from the undesired capped sequences.  In the case of fluorous methods that phase separation is usually a fluorous solid phase extraction (FSPE).

In the Fang paper a polymerizable acrylamide tag which is attached to the 5′-OH group through a silane is used.  Upon attachment to the final oligonucleotide and cleavage from the support, this tag is polymerized to form a new solid containing only the desired full sequence.  The capped sequences are then washed away.  The desired sequence can then be cleaved from the polymer with HF and the polymer filtered off.  Excess HF and silyl by-products can be removed during lyophilization.  The use of a polymerizable phase tag is certainly not new and has been used for various catalysts and reagents with ROMP based methods being particularly popular.

The method does have some nice advantages, since the purification is a simple filtration.  The disadvantage is, of course, you have to run one more reaction, the polymerization, which may or may not cause problems.  In this case, the authors tested that by subjecting each of the four nucleosides to the radical polymerization conditions and found no change.  In practice, they synthesized a 20-mer and reported 72% recovery of the tagged oligo based on UV absorbance at 260 nm.

So how does this compare with fluorous methods, specifically the use of F-DMT phosphoramidite as reported by Berry and Assoc. in 2005?  Well, to read this latest report, the polymerization approach is superior.  It leads to higher recoveries and doesn’t use expensive materials such as fluorous affinity columns or avidin-coated beads.  So the authors specifically tout the advantages of polymer-method over fluorous techniques.  As one might guess, I’m not as sure.

Let’s compare the Berry and Assoc. 2005 J. Org. Chem. paper using F-DMT with the Fang paper.  To be fair the Berry and Assoc. publication is a full paper while this latest on is a communication, so whatever criticisms I point out may very well be addressed in a forthcoming full paper.  The JOC paper synthesized a 30-mer, a 50-mer, a 75-mer, and a 100-mer multiple times.  Recoveries, also using absorbance at 260 nm ranged from 70-100% with even the 100-mer at quantitative recovery.  The Fang paper reported a single synthesis of a 20-mer with 72% recovery.  So not only were the recoveries higher, but the oligonucleotides produced more challenging in the fluorous paper. The longer oligos are an important distinction since as Fang mentions the polymerization could be hindered by steric bulkiness of the oligonucleotide.  In addition, Berry and Associates notes that their reported recoveries are ranges based on multiple runs rather than a single run of 72% recovery for Fang.

As for cost of materials, there is a cost for the F-DMT phosphoramidite and the FluoroPak cartridge for purification, but to lump it in with avidin-coated beads is not really fair.  Our own cost analysis revealed between fluorous methods and biotin based separations resulted in fluorous methods being 25-50x less expensive.  There is no way, fluorous methods can be grouped with biotin based methods in terms of cost.  So is this polymerization method less expensive than biotin-avidin based methods?  Probably.  Less then fluorous?  Unclear, especially when you consider that F-DMT seems to provide greater recovery.

There is no doubt some appeal to the simple polymerization method, but until some further validation it’s not so easy to say one method is superior.  If I had to guess, I’d say the longer the oligo, the better fluorous comes out.  Of course, the shorter the sequence the less need for any type of tagging strategy.  As pointed out in the Berry and Associates paper, regardless of the tag, any 5′-OH tagging method is going to suffer from certain shortcomings related to the synthesis and apart from the tag itself.

The bottom line is that there is probably room for polymerizable tags, fluorous tags, and other methods.  The best method will probably depend on the exact length and composition of the oligonucleotide.

Posted in Biomolecule Synthesis, Oligonucleotides | Tagged , | Comments Off