Mesyltoe 3o11
3.O11 Organic Synthesis Part II
Lecture 1
The Principles of Retrosynthesis
Retrosynthesis is vital for rational and intelligent molecule design. We work backwards from the desired molecule (target molecule, TM) one step at a time. We disconnect a molecule into 'synthons', or synthetic equivalents. This means that we look at the charges required on the carbon (or other) atoms, and rationalise these into real molecules that will approximate that charge. For example, C- would rationalise to C-Li.
Disconnections often take place immediately adjacent to functional groups in the target molecule. Given functionality inevitably arises from many forward reactions, this is the best way. A good disconnection simplifies the target significantly.
Amine synthesis
The obvious amine synthesis, disconnecting through the C-N bond, has problems in that polyalkylation (undesirable) can occur. This can be avoided if the amine is less nucleophilic than the starting material: adding an amide to an epoxide takes advantage of this, and does not result in polyalkylation.
We can also use FGI reactions. For a primary amine, we use an azide (N3) and an organohalide which can then be reduced to an amino group with LiAlH4 or similar. For amines with an adjacent methylene, an acyl chloride can be used, as an amide is much less nucleophilic than an amine. The carbonyl can then by removed using lithal. However, this only works if there is an adjacent CH2 to hold the carbonyl.
Reductive amination is an alternative route, where there is no methylene to hold the carbonyl. This involves the creation and subsequent reduction of an imine by adding an amine into a carbonyl. This is then reduced using NaBH(OAc)3, milder than lithal, which allows the imine to be formed before all reduction occurs (so the carbonyl isn't instantly reduced, as it might be with lithal).
The Curtius Reaction
This reaction inserts a nitrogen into a C-C=O. A carbonyl with a leaving group is reacted with NaN3. This is then heated. N2 is released, and the nitrogen inserts between the carbonyl and the R group, giving an isocyanate. If worked up with water, an amine results, and if worked up with alcohol, a carbamate (that alcohol added as an ester) results. Carbamate is a useful protecting group. Boc is a carbamate.
Lecture 2
Carbonyl groups are hella important: they have a great deal of synthetic potential. They can be disconnected at the 2 and the 3 position, 2 being straightforward reaction at a carbonyl, and 3 being the reaction of an enolate. 1,3 dicarbonyls are particularly good for synthesis. We can make them by reacting an enolate with a carbonyl+LG. This is particularly easy if the molecule is symmetrical.
The Claisen reaction and the Dieckmann reaction
We all remember the Claisen ester condensation, I'm sure. As described above (give or take) an enolisable carbonyl is reacted with an ester to give a 1,3 dicarbonyl. When this occurs intramolecularly, it is called a Dieckmann reaction, and it works very well for the production of 5 or 6 membered rings. The driving force of this reaction is the removal of the proton between the carbonyls.
β-Hydroxy Carbonyls and the Aldol Reaction
These compounds also have 1,3 oxygenation, but at a lower oxidation level. The Aldol reaction is accomplished by treating carbonyl with a base. The resulting enolate attacks the other carbonyl, to give β-hydroxy carbonyl. This product is easily dehydrated by heating with acid or base, as the proton on the 2 carbon is very acidic. This forms an α,β-unsaturated carbonyl, also useful.
One problem in the aldol reaction is control. The two carbonyl reagents can both enolise, and some of these reagents can enolise at either side, resulting in a lot of undesirable side reactions. The ensure that there is a cross-enolisation (between two different compounds) we must ensure that only one of the compounds will enolise, only on one side, and that the other compound will act as the electrophile.
Intramolecularly, this tends not to be a problem: the most stable ring conformation will form, as the aldol reaction is reversible. This tends to result in six membered rings. For intermolecular reactions, we can use a component that will not enolise, where the α carbon is trisubstituted, part of an aryl system, or part of an ester.
If there is doubt as to which compound will enolise, we can turn one of the carbonyl compounds into a pre-formed enolate equivalent, such as a silyl enol ether (adding TIPS or TMS to the oxygen) or an enamine (attacking the carbonyl with nitrogen). This route means we know which compound will act as the nucleophile.
It is also possible to use an activating group to control the site of deprotonation. The addition of an ester group to the α position makes that proton much more acidic, meaning that it is much more readily lost, and that the enol is more likely to form on that side. The ester group can be removed by decarboxylation afterwards.
The Mannich Reaction
Formaldehyde might seem like a perfect carbonyl to use in these circumstances, as it cannot enolise, but it is too electrophilic, and will continue to react with the desired product. The Mannich reaction avoids this. It uses formaldehyde, secondary amine and acid catalyst. The amine reacts with the formaldehyde to give an iminium ion. This then acts as the electrophile in the reaction. The amine can subsequently be removed using MeI, to give an α,β-unsaturated carbonyl.
1,3-Relationships of O and N
It is also possible to achieve the 1,3-relationship between O and N that the Mannich reaction gives by other means: addition of an amine to an α,β-unsaturated carbonyl!
Lecture 3
1,5 Dicarbonyl Disconnections & The Michael Reaction
One reaction underpins most of the analysis of this connection, and it is the Michael reaction. This is the reaction of an enolate and an enone (Michael acceptor). We must be sure, as in the reactions above, that the correct compound acts as the nucleophile. We must also be sure that the enone reacts at the β position, not the carbonyl carbon. Using an activating group (such as an ester, seen above) on the enolate is the solution to both of these problems. The correct side of the carbonyl is thus enolised, and the soft nature of this stabilised enolate favours 1,4-attack. The result is a 1,5 dicarbonyl.
Related to the Michael reaction is the Robinson annulation. This has the enolate as part of a pre-existing ring. Using the Michael reaction, the 1,5-dicarbonyl that results can be reacted using the aldol reaction, to give a second adjoining ring.
1,4-difunctionality
1,5 and 1,3 difunctionality both correspond to the 'natural' polarity of carbonyl reagents. 1,4-difunctionalised compounds do not. Disconnecting in the middle gives one enolate equivalent (which is normal) and one carbonyl synthon with a positive charge at the α position, which is quite unexpected, and not what carbonyls are supposed to do at all. This is equivalent to a halocarbonyl (amongst other things). Again, it is important to ensure that the correct reagent enolises. This is accomplished, as above, using an ester activating group. This is especially important because the halide increases the acidity of the α protons, meaning they are more easily enolisable.
Allyl halides provide an alternative to α-carbonyl cations. The bromoallyl is reacted with the nucleophile (the enolate in these cases) and is then converted to a carbonyl compound via ozonolysis. This removes some of the problems with enolisation seen above. This technique is called functional group addition, because we are adding a functional group (in this case, the alkene) knowing that it is removable.
1,4 Disconnection at a lower oxidation level
It is also possible to disconnect in the centre of 1,4 difunctionalisation if the functional groups are carbonyl and alcohol. In this case, the carbonyl synthon is still an enol, and the alcohol sython is an epoxide. We make epoxides, as I'm sure we all remember, using mCPBA (meta-chloroperoxybenzoic acid) on an alkene.
Alternative 1,4-disconnection
The disconnection doesn't have to be made in the centre of the 1,4 gap, but even if it is made at the 3-4 position, unnatural polarity is still present in the synthons. One synthon has an α,β unsaturated carbonyl, which is fine, but the other has an acyl anion, with the negative charge located on the carbonyl carbon: this is quite the opposite of what we'd expect.
There are several synthetic equivalents of the acyl anion. 1,3-dithianes can be made from aldehydes, deprotonated with strong bases, and then reacted with electrophiles. The carbonyl can be regenerated by Hg2+ catalysed hydrolysis. Cyanide anions also work. They can be converted to carboxylic acids by hydrolysis, to aldehydes by DIBALH at low temp. or to a ketone by adding a Grignard followed by hydrolysis.
Alternative Strategy: Alkynes
Terminal alkynes can be deprotonated. The resulting anion can be reacted with electrophiles. Alkynes can be partially reduced to alkenes, so we can apply this analysis to 1,4 dioxygenated targets with an alkene in the middle as well. Using an alkyne anion on one side, our other synthon is a carbonyl cation, which is something we're all quite familiar with. We use base to deprotonate both reagents, and can then reduce for alkene functionality.
Lecture 4
Alkenes
Some of the most reliable methods for creating 1,2-difunctionality come from alkenes. They are particularly useful because of their stereospecificity.
We can functionalise alkenes using OsO4 to give syn-dihydroxylation. Epoxide ring opening can be used to give anti-dihydroxlation. The problem with epoxides is knowing which side will be attacked by the nucleophile. With good nucleophiles, the mechanism is SN2 like, and reaction will occur at the less hindered end. With weak nucleophiles, and in the presence of acid, opening occurs with a more SN1 mechanism, meaning that the most stabilised end opens.
Strategies for Alkene Synthesis
Elimination reactions can be stereospecific, such as E2, which requires the H and the leaving group to be app to one another. Single stereoisomers of the starting product, however, are difficult to obtain, and regiocontrol is also difficult: which β-hydrogen is eliminated?
Alkyne reduction is powerful, because the stereospecificity can be chosen: Lindlar catalyst and H2 gives cis, whilst Na/NH3 gives trans. Non-terminal alkynes are easily synthesised, which is good. The other functionalisation in the molecule has to withstand the reduction, however, and alkyne reduction gives only disubstituted alkenes.
Carbonyl Olefination
Carbonyl olefinations (that is, swapping a carbonyl for alkene functionality) are widely used and very useful. They are regiospecific, as the alkene ends up where the carbonyl was in the starting material, and stereocontrol can often be achieved. The reagents are a carbonyl compound and an anion with a stabilising leaving group (such as a phosphinium, phosphonate or sulfone) and the desired R groups.
Wittig
An organohalide is reacted with PPh3, to give a phosphonium yllid (positive charge on the P). This is reacted with the carbonyl to give the alkene product. The driving force of the reaction is the formation of the strong P=O bond. Stereocontrol depends on the R groups on the yllid. If the R groups are alkyl, the Z product predominates. If the R groups are anion stabilising groups (electron withdrawing), E predominates. Otherwise, mixtures happen.
Julia Olefination
To get E-alkenes when the R groups are alkyl, we use the Julia olefination. The stabilising leaving group in this case is a sulfone (sometimes a phenyl sulfone, in the classical reaction, something a heterocyclic sulfone, in the modified reaction). The sulfone and the hydroxyl are then eliminated using Na(Hg) amalgam, to give E selectivity. In the modified reaction, the amalgam is not required. The sulfone-R anion is reacted with the carbonyl compound, and eliminates spontaneously, removing the need for additional steps.
α,β-Unsaturated Esters
α,β-unsaturated esters are common synthetic intermediates. One reason for this is that they can be reduced to give allylic alcohols, the reagents needed for Sharpless asymmetric epoxidation. The unsaturated esters are reduced using DIBALH, and then undergo the asymmetric epoxidation.
We can use the Wittig reaction to make E isomers in this way, but the reaction rate is rather slow. A faster way of doing this is to use phosphonate as the anion-stabilising leaving group rather than phosphine. This is called the Horner-Wadsworth-Emmons reaction. They are more nucleophilic, meaning a higher rate, but still have E-selectivity.
To make Z isomers for this purpose, we can use two special phosphonates that have been designed to give this selectivity. Instead of Et groups on the phosphonate, these have electron withdrawing groups: trifluoroethyl for the Still phosphonate, and phenyl for the Ando phosphonate.
Alkene Metathesis
Alkene metathesis employs metal carbene complexes to switch R groups between alkenes. Ruthenium complexes, due to their air-stability and their tolerance of oxygen functional groups, are most often used.
Ring closing metathesis is a variant of this. This is an intramolecular metathesis reaction, reacting two alkenes in the same molecule to close a ring. This can often be more effective than, for example, intramolecular carbonyl olefination (such at in the Dieckmann reacton) because there is no need for selective deprotonation.
Cross-metathesis is also becoming more common. The R groups must be chosen carefully to avoid side reactions, such as homodimerisations, and control of E/Z geometry can be tricky. Reacting terminal aliphatic alkenes with acrylates (CO2R) is a successful, E-selective process.
Trisubstituted Alkenes
Some of the above methods work well for trisubstituted alkenes, and are widely used, such as the Horner-Wadworth-Emmons phosphonate synthesis of α,β-unsaturated esters.
Lecture 5
Cross Coupling Reactions
Cross coupling reactions are the transition metal catalysed formation of bonds to sp2 carbons. One one hand is an alkene, one substituent of which is either a metal or a leaving group. The other reagent is the group that we seek to add to the alkene, attached either to a metal or a leaving group (there must be one of each, obviously!). The reaction is catalysed by a transition metal, usually palladium. The group attached to the leaving group must not have β-hydride available for elimination (as β-hydride elimination will occur on the catalyst, and the coupling will not occur). This means that sp2 halide aryls are popular substrates.
Pd(0) is the catalytically active species in the reaction, but it's rather expensive. Pd(II) is much cheaper, and thus it needs reducing. A reducing agent, such as Et3N is used in situ to activate the catalyst. The amounts of Pd needed are very small.
Suzuki reaction
The Suzuki reaction, with a boronic ester/acid on one side (such as pinacolborane) and a halide on the other is very popular. The reagents aren't as nasty as in Stille coupling, and boronic acids are readily commercially available.
These reactions are stereospecific. If the reactants are both E, then the product will be E at both centres. The original stereochemistry of the reagents is preserved.
The boronic ester can be added either by metallation (using BuLi to deprotonate and then adding), hydroboration of an alkyne or crosscoupling (pinacolborane dimer with desired adduct-LG and Pd catalyst).
Heck reaction
The Heck reaction is a cross-coupling that involves joining a halogenated reagent with an alkene of no particular functionality. The regioselectivity is strongly influenced by the alkene structure and the Pd/ligand used. Electron poor alkenes give exclusively linear functionality. Electron neutral alkenes give more linear in the presence of triphenylphosphine, and more branched in the presence of dppp. Electron rich alkenes give mixtures in the presence of triphenylphosphine and exclusively branched in the presence of dppp. If there is only one β-hydrogen, the elimination can only go one way. If there is more than one, often the more stable alkene is predominant. Also, for elimination to occur, Pd must be syn to the hydrogen. This can effect chemistry particularly on a ring.
Lecture 6
Pericyclics and Synthetic Strategy
Pericylic reactions (reaction that proceed via a concerted cyclic mechanism) are powerful synthetic tools. They have high levels of regio- and stereoselectvity, and are stereospecific because of the orbital symmetry rules that control them. We'll look mostly at cycloadditions ([4+2], [3+2] and [2+2]) and [3,3] sigmatropic rearrangements.
The Diels-Alder Reaction
The Diels-Alder reaction is a [4+2] cycloaddition. It's a concerted cycloaddition between a diene (which must be able to adopt a cis conformation) and a dieneophile (an alkene or alkyne with an electron withdrawing group). The electron withdrawing group is ortho or para directing, meaning that these isomers are the major products. Because the cycloaddition is concerted, the configuration of diene and dienophile are retained in the product. It is also (often) endo-selective. The double bond in a Diels-Alder product can undergo further reactions, which makes it pretty good (such as syn hydroxylation and ozonolysis). It can also occur intramolecularly, which is a powerful way of generating two rings in a single reaction.
The Diels-Alder reaction is not confined to C-C bond creation. It can also be used to create heterocycles.
[3+2] Cycloadditions
These reactions, involving 1,3-dipole reagents, create five membered rings. We saw these used to create heteroaromatics.
[2+2] Cycloadditions
Cyclobutane can be synthesised via a [2+2] cycloaddition, which can be either thermally or photochemically activated. Ketenes can also undergo thermal [2+2] cycloadditions.
[3,3] Sigmatropic Rearrangements
The Claisen rearrangement is a bond moving from the 3,4 position to the 1,6 position. The 1,2 and 5,6 bonds must both be double bonds in order for this to occur. In addition, the 3 atom is an oxygen that goes from ether to carbonyl. This is usually a one-way arrangement because of the strength of the C=O bond. LDA is used to effect a Claisen rearrangement. A Claisen rearrangement can be recognised by a γ,δ-unsaturated carboxylic acid derivative.
The Cope rearrangement is the same, but there is no oxygen. This makes the reaction more reversible, as there is no strong driving force. The oxy-cope is a variation of this, where the 3 position has a hydroxy group on it. The rearrangement turns this hydroxy into an enol, which then isomerises to a ketone. The Cope rearrangement can be recognised by a δ,ε unsaturated carbonyl.