`
`2091
`
`Linkers and Cleavage Strategies in Solid-Phase Organic Synthesis and
`Combinatorial Chemistry
`
`Fabrice Guillier,† David Orain,‡ and Mark Bradley*,‡
`
`Department of Chemistry, Southampton University, Highfield, Southampton SO17 1BJ, U.K., and Alanex Division of Agouron Pharmaceuticals,
`A Werner Lambert Company, 3550 General Atomic Courts, San Diego, California 92121
`
`Contents
`
`2091
`Introduction
`I.
`2092
`II. Solid Supports for Organic Synthesis
`2093
`III. Linker and Linker Attachment
`2093
`A. Linker Types
`2094
`B. Scaffold Preloading and Direct Loading
`2095
`C. Spacers
`2095
`D. Linker Attachment
`2095
`E. Leaving Groups and Scavengers
`2096
`IV. Linkers and Cleavages in Organic Synthesis
`2096
`A. Electrophilically Cleaved Linkers
`2096
`1. Strong Acid Cleavable Linkers
`2099
`2. Mild Acid Cleavable Linkers
`2105
`3. Silicon Linkers
`2106
`4. Ketal Linkers for Alcohol Immobilization
`5. Linkers for Carbonyl Group Immobilization 2108
`6. Boronate Linkers
`2109
`7.
`tert-Alkoxycarbonyl-Based Linkers
`2109
`8. Aryltriazene Linkers
`2110
`9. Organometallic-Supported Scaffold
`2110
`B. Nucleophilically Cleaved Linkers
`2110
`1. Oxygen Nucleophiles
`2110
`2. Nitrogen Nucleophiles
`2114
`3. Safety-Catch Linkers Sensitive to
`2118
`Nucleophiles upon Activation
`4. Carbon Nucleophiles
`5. Halogen Nucleophiles
`6. Thiol Nucleophiles
`7. Base-Promoted Cleavage
`C. Photocleavable Linkers
`1. o-Nitrobenzyl-Based Linkers
`2. Phenacyl Linkers
`3. Alkoxybenzoin Linkers
`4. NpSSMpact Linker
`5. Pivaloylglycol Linker
`6. Miscellaneous Photolytic Protocols
`D. Metal-Assisted Cleavage Procedures
`1. Palladium Deblocking of Allyloxycarbonyl
`Group
`2. Metal-Catalyzed Release by C-C Bond
`Formation
`3. Metal and Nonmetal Lewis-Acid-Assisted
`Cleavage.
`E. Cleavage under Reductive Conditions
`1. Catalytic Hydrogenation
`
`2121
`2122
`2125
`2125
`2128
`2128
`2130
`2131
`2131
`2131
`2132
`2132
`2132
`
`2133
`
`2135
`
`Received September 30, 1999
`
`2. Reduction of Disulfide Bonds
`3. Reductive Desulfurization and
`Deselenization Techniques
`4. Hydride Nucleophiles
`5. Miscellaneous Reductive Protocols
`F. Cleavage under Oxidative Conditions
`1. Oxidation of p-Alkoxybenzyl Groups
`2. Oxidation of Sulfur and Selenium Linkers
`3. Ozonolysis of Polymeric Olefins
`4. Miscellaneous Oxidative Protocols
`G. Cycloaddition- and Cycloreversion-Based
`Release
`V. Guide for Functional Groups Release
`VI. Multidetachable Linkers (In-Line Linkers)
`VII. Utilization of Linkers To Monitor Solid-Phase
`Organic Reactions
`1. Single-Bead Direct Analysis
`2. Controlled Partial Release
`3. Monitoring with Multidetachable Linkers
`VIII. Utilization of Linkers for Structural Elucidation of
`“Hits” in Combinatorial Chemistry
`A. Encoding Strategies
`1. Haloaromatic Tags for Binary Encoding
`Strategy
`2. Secondary Amine Tags
`3. DNA-Encoding Strategy
`4. Amino-Acid-Encoding Strategy
`B. Multiple-Release
`C. Ladder Scanning Strategy
`IX. Conclusion
`X. Abbreviations
`XI. Acknowledgments
`XII. References
`
`2137
`2137
`
`2138
`2138
`2139
`2139
`2139
`2140
`2140
`2141
`
`2141
`2141
`2144
`
`2144
`2144
`2144
`2145
`
`2145
`2145
`
`2146
`2146
`2146
`2147
`2149
`2150
`2150
`2152
`2152
`
`I. Introduction
`The massive increase in the number of papers
`describing the use of polymeric supports in organic
`synthesis over the past decade is a vivid demonstra-
`tion of its impact in the chemical community. Few
`other changes in synthetic chemistry methodology
`have displayed such a growing passion or had such
`a profound influence on the way synthetic chemistry
`
`* To whom correspondence should be addressed.
`† Agouron Pharmaceuticals.
`‡ Southampton University.
`
`2136
`2137
`
`© 2000 American Chemical Society
`10.1021/cr980040+ CCC: $35.00
`Published on Web 05/06/2000
`
`Illumina Ex. 1106
`IPR Petition - USP 10,435,742
`
`
`
`2092 Chemical Reviews, 2000, Vol. 100, No. 6
`
`Guillier et al.
`
`Fabrice Guillier was born in 1968 in the city of Saint Denis, France. He
`received his diplome d’Inge´nieur chimiste in 1992 from the National Institute
`of Applied Sciences (INSA Rouen) after studying chemical engineering
`and organic chemistry. In 1996 he completed his Ph.D. thesis on natural
`product synthesis using metalation and cross-coupling reactions in the
`(URA1429-IRCOF- Rouen-
`laboratories of Professor Guy Que´guiner
`France). These methodologies using palladium and lithium derivatives
`were suitable for the synthesis of polyheterocyclic marine alkaloids of the
`pyridoacridine family. He spent then one year as a postdoctoral fellow at
`Southampton University, working with Professor Jeremy Kilburn (Depart-
`ment of Chemistry, University of Southampton) studying the synthesis
`and supramolecular interactions of guanidinium-based tweezer receptors
`of C-terminal tripeptides. From 1997 to 1999 he joined the research group
`of Professor Mark Bradley at the University of Southampton, where he
`conducted research on the development of new methods for screening
`on solid supports and the synthesis of libraries of polyamines as inhibitors
`of
`trypanothione reductase and substituted pyrimidines for kinases
`inhibition. Currently he is developing new reactions for lead discovery
`using solution-phase parallel synthesis at Alanex Division of Agouron
`Pharmaceuticals (San Diego, CA) as a research fellow.
`
`David Orain was born in 1971 in Rennes, France. He studied chemistry
`at the University of Rennes and received his DEA (Masters) degree in
`1994 under the supervision of Dr. Paul Mosset at ENSCR (School of
`Chemistry at Rennes) where he studied the synthesis of R,R¢ -bifunction-
`nalized heterocycles. Then he moved downstairs at ENSCR and worked
`with Dr. J. C. Guillemin’s group during his Ph.D. studies. His research
`involved the use of zirconium organometallic complexes for solution and
`support chemistries. He was awarded his Ph.D. degree in December 1998.
`In January 1999, he crossed the Channel and joined Professor Mark
`Bradley’s group (Department Chemistry, Southampton University) as a
`postdoctoral researcher. His current research focuses on the development
`of a new safety-catch linker for amine release under biological conditions
`and new direct screening processes of a library of trypanothione reductase
`inhibitors.
`is carried out. The advantages gained by this meth-
`odology are striking, with four main factors contrib-
`uting to the popularity of the technique. (i) The ease
`of chemistry. Reactions can be accomplished in only
`three steps: addition of reagents, filtering, and
`
`Mark Bradley was born in the United Kingdom in 1962. He received his
`first degree from the University of Oxford in 1986 and his DPhil degree
`from the same institution three years later under the supervision of
`Professor Sir Jack Baldwin in the area of Penicillin biosynthesis. In 1989
`he moved to the United States to work with Professor Chris Walsh at
`Harvard Medical School. He then returned to the United Kingdom and to
`Southampton University in 1992 as a Royal Society University Research
`Fellow. In 1997 he was made a Professor of Combinatorial Chemistry
`and has published in excess of 50 papers in the combinatorial area. In
`January 2000 he became director of
`the Combinatorial Centre of
`Excellence now housed in Southampton. His research interests span
`across the whole area of combinatorial high-throughput synthesis,
`screening, and analysis from an academic viewpoint. This includes interests
`in solid-phase and small-molecule synthesis as well combinatorial activities
`in the area of catalysts, dendrimers, fluorophores, and polymers. He has
`interests in the area of enzyme inhibition including work in the area of
`proteases, antibacterial agents, and antiparastics.
`washing the resin, thus allowing many simple auto-
`mated procedures to be developed. (ii) The elimina-
`tion of purification steps en route. For each step of a
`multiple-step synthesis, the only purification needed
`is a resin-washing step. Only the final product of
`cleavage needs to be purified. (iii) In a solid-phase
`synthesis, high concentrations of reagents can be used
`to drive reactions to completion. (iv) The straight-
`forward nature of parallel solid-phase synthesis.
`However, for a solid-phase synthesis to be practical,
`several important issues need to be addressed, in-
`cluding the correct choice of solid support and the
`mode of attachment and cleavage of materials from
`the resin matrix. Efficiency in anchoring and remov-
`ing a small organic molecule from the polymeric resin
`relies on the correct choice of the linker group. This
`key fragment is crucial in planning a synthetic
`strategy. However, if the objective of a single or
`parallel solid-phase organic synthesis is to produce
`one or several defined products upon release, the
`correct choice of an adequate linker system can
`enable further goals to be attained. Multidetachable
`linkers allow the preparation of different products
`depending on the cleavage conditions selected. Partial
`release can be useful for monitoring reactions or
`screening mixtures for deconvolution. Structural
`elucidation of hits from mixtures of products is
`another application of linkers in combinatorial chem-
`istry, allowing “tags” of a variety of forms to be
`incorporated and read.
`
`II. Solid Supports for Organic Synthesis
`Cross-linked polystyrene resin beads have been
`used for organic synthesis since 1963, when Merri-
`
`
`
`Linkers and Cleavage Strategies
`
`Chemical Reviews, 2000, Vol. 100, No. 6 2093
`
`field first used a chloromethylated-nitrated copoly-
`mer of styrene and divinylbenzene (DVB).1 Resins
`now used in solid-phase organic synthesis have
`changed little since this time. These insoluble sup-
`ports have a gel-type structure which readily allows
`penetration of reagents and solvents into the beads
`to sites where chemistry is taking place. A compro-
`mise has been found between moderately cross-linked
`resins (5% DVB) which are very stable but do not
`swell particularly well, thus reducing site access and
`low cross-linked resins where mechanical stability
`becomes an issue. A general consensus now seems
`to have been reached, and typical supports used for
`solid-phase synthesis consist of polystyrene with a
`1-2% DVB cross-linking. The three dominant poly-
`styrene supports currently in use are the following.
`(i) Chloromethylpolystyrene. Originally prepared by
`resin postderivatization using chloromethylmethyl
`ether and SnCl4, it has been more recently prepared
`by copolymerization using chloromethylstyrene/sty-
`rene/DVB mixtures. This core resin is used widely
`for the attachment of linkers by ether formation. (ii)
`Hydroxymethylpolystyrene. Prepared from Merrifield
`resin by esterification with potassium acetate fol-
`lowed by saponification or reduction of the ester.2 (iii)
`Aminomethylpolystyrene. Mitchell3 prepared this resin
`either by potassium phthalimide substitution of the
`Merrifield resin followed by hydrazinolysis or by
`direct aminomethylation of the polystyrene resin.
`Aminomethyl resin allows a multitude of spacers/
`linkers to be appended to the resin by amide bonds,
`which are stable under strongly acidic conditions.
`This still provides one of the main workhorse resins
`of today.
`A few other materials are used but polystyrene
`resin dominates; among the other materials used,
`TentaGel resin (TG) and ArgoGel (AG), both poly-
`styrene/DVB-poly(ethylene glycol) graft copolymers
`(PS-PEG), developed by Bayer,4 are the most fa-
`vored. They have specific uses, such as when polar
`solvents are needed or when distancing from the
`resin core becomes necessary. Crowns/Pins (CP) are
`another kind of support used in solid-phase synthesis
`and consist of a radiation-grafted polyethylene/
`polypropylene support.5 Kieselguhr/polyacrylamide-
`based resins (KPA)6 and controlled-pore glass (CPG)7
`are used in continuous flow SPPS and oligonucleotide
`synthesis but are usually avoided by the synthetic
`chemist. PEGA,8 a poly(ethylene glycol)/dimethy-
`lacrylamide copolymer, is a very polar material which
`confers unparalleled swelling properties in water and
`possesses a flexible interior enabling access for a
`variety of large macromolecules such as enzymes;
`however, low mechanical stability makes handling
`difficult and expense precludes large-scale use.
`
`III. Linker and Linker Attachment
`The attachment point of the linker to the solid
`support or spacer should be chemically stable during
`the synthesis and cleavage, and as for any solution-
`phase protecting group, yields for its loading and
`cleavage should be as quantitative as possible. In
`combination, the Resin-(Spacer)-Linker unit can
`thus be considered to be an insoluble, immobilizing
`
`protecting group for solid-phase synthesis. A great
`number of linkers (more than 200) have been devel-
`oped over the past 15 years in order to allow many
`multistep organic syntheses to be performed and the
`use of a broad range of reagents, allowing cleavages
`in a very selective manner (see refs 9-15 for reviews).
`Although a linker should ideally enable a selective
`cleavage to take place under a defined set of condi-
`tions, these conditions are in reality not only depend-
`ent on the linker but also on the compound attached
`to the linker, the spacer, and importantly on the resin
`type, its loading, and bead size. Thus, smaller beads
`have a much greater efficiency of cleavage under
`photolysis conditions. Reduced cross-linking dramati-
`cally enhances the rate of cleavage from solid sup-
`ports under acidic conditions.16 Lack of resin pre-
`swelling in CH2Cl2 prior to cleavage with TFA can
`also cause reduction in yields.16 Many different
`parameters are thus involved in the cleavage of
`compounds from the solid support and not just the
`linker needs to be considered.
`
`A. Linker Types
`In this review, to aid clarification and to avoid
`inaccuracies, there will be a clear distinction made
`between resins and linkers: Resins will be considered
`as an inert matrix, passive to chemistry. Linkers will
`be considered simply as immobilized protecting groups
`and will be classified into one of two types:
`(i)
`Integral linkers in which part of the solid support core
`forms part or all of the linker and (ii) Nonintegral
`(or grafted) linkers in which the linker is attached to
`the resin core. A linker which has been prepared in
`solution will be defined as a unloaded linker.
`Many examples of integral linkers exist (Scheme
`1), and certainly they were very popular in the early
`
`Scheme 1
`
`days of solid-phase synthesis. Thus, linkers such as
`the o-nitro-(R-methyl)bromobenzyl linker 1.1 pre-
`pared by Pillai17 is a classic example of an integral
`linker. Its preparation was realized by functionalizing
`polystyrene/DVB resin with acetyl chloride/AlCl3,
`reducing the resulting ketone and bromination of the
`resulting alcohol. The nitro group was then incorpo-
`rated by nitration of the resin. Also included in this
`list are the benzhydrylamine (BHA) linker 1.2 pre-
`
`
`
`Guillier et al.
`
`contrast to the integral linkers described above. The
`p-alkoxybenzyl alcohol Wang linker 2.3 was initially
`prepared by reacting 4-hydroxybenzyl alcohol with
`Merrifield resin in the presence of sodium methox-
`ide.26 The Sasrin linker 2.4 was first described by
`Mergler27,28 and was initially anchored onto the resin
`by etherification. Sheppard29,30 prepared the un-
`loaded linker 2.5, allowing attachment to an ami-
`nomethylpolystyrene resin. Dimethylsilyl chloride
`groups have also been attached to the polystyrene
`core through an ethylene bridge to give 2.6 by
`hydrosilylation of (vinyl)polystyrene.31 Here again the
`resin core is not an integral part of the linker.
`Linkers that are copolymerized into resin beads can
`be either of the integral or nonintegral type. Those
`which are not part of the polymer core can be
`considered as nonintegral (or grafted) in nature.
`Scheme 3 shows the preparation of a trityl (integral)
`linker 3.4 by suspension copolymerization of mono-
`mer 3.1, DVB 3.2, and styrene 3.3.32
`
`Scheme 3
`
`2094 Chemical Reviews, 2000, Vol. 100, No. 6
`pared by Friedel-Crafts acylation of polystyrene with
`benzoyl chloride.18 The resulting benzophenone de-
`rivative was transformed into the desired product by
`either reduction of an oxime, ammonolysis of the
`bromo derivative, or reductive amination with am-
`monium formate. The original chloromethylated poly-
`styrene resin 1.3 used by Merrifield can in many
`respects be considered as an integral linker. This
`allowed Merrifield to anchor N-protected amino acids
`onto solid supports by formation of immobilized
`benzyl esters. The trityl linker 1.4 was developed by
`Leznoff19 by lithiation of polystyrene and reaction
`with benzophenone and by Fre´chet20,21 by treatment
`of benzophenone-based polystyrene with phenylmag-
`nesium bromide. Another light-cleavable bromine-
`derivatized linker 1.5 was obtained by functionaliza-
`tion of 2% polystyrene/DVB with 2-bromopropionyl
`chloride/AlCl3 under Friedel-Crafts conditions.22
`Two percent cross-linked benzene sulfonyl chloride
`1.6 was prepared from Dowex 50W ion-exchange
`resin (-SO3H).23
`The disadvantage with any integral linker is the
`control of synthesis, taking place as it does directly
`on the resin, with the whole range of steric and
`electronic effects having an influence over the syn-
`thetic outcome. The exact degree of loading and
`functionalization can be hard to control.
`The majority of linkers used in solid-phase syn-
`thesis are thus of the nonintegral type (Scheme 2).
`
`Scheme 2
`
`B. Scaffold Preloading and Direct Loading
`The choice of which method, preloading or direct
`loading of the scaffold onto a nonintegral (or grafted)
`linker, is the most suitable for use in solid-phase
`synthesis is not clear-cut. The first method usually
`ensures much higher loading levels and that only
`purified materials are coupled onto the solid support
`and also reduces the number of solid-phase steps. The
`second method is usually less efficient since excess
`materials are often used in the coupling step, a
`problem if valuable scaffolds are being used, but is
`faster since no solution steps or purifications are
`needed (Scheme 4).
`In addition, if all the derivatized sites on the loaded
`linker are not reacted, then undesirable side reac-
`tions can take place. There are certainly cases where
`linkers attached to resins have not been added
`cleanly and have given rise to numerous side reac-
`tions and impure products. Thus, the attachment of
`4-hydroxybenzyl alcohol onto Merrifield resin to form
`the Wang linker using the original procedure de-
`scribed had to be improved in order to limit the side
`reactions which afforded more than five different
`byproducts as originally reported.33
`
`These can be loaded onto the resin and then deriva-
`tized or preloaded prior to attachment. This resin
`attachment is generally realized in one of three
`ways: (i) Ethers, (ii) Amides, and (iii) C-C bonds.
`Thus, derivatives of the trityl linker 2.1 have been
`prepared in the nonintegral manner by attachment
`of 4-carboxy derivatives24 through an amide bond. A
`light-cleavable linker o-nitrobenzyl (ONB) 2.2 was
`prepared25 by coupling 3-nitro-4-bromomethylbenzoic
`acid onto an aminomethylpolystyrene resin, in clear
`
`
`
`Linkers and Cleavage Strategies
`
`Scheme 4
`
`Chemical Reviews, 2000, Vol. 100, No. 6 2095
`
`Scheme 6
`
`C. Spacers
`A group can be attached to the solid support to act
`as a spacer unit. The spacer has a number of roles.
`Principally, it acts to distance chemistry from the
`solid support and tailors the swelling properties of
`the resin materials to give more “solution-like”
`properties and better solvent compatibility. Typical
`examples of spacers are PEG chains (as in PS-PEG-
`based resins such as TentaGel 5.1 or resin 5.234) or
`alkyl chains as shown in 5.335 (Scheme 5). Spacers
`
`Scheme 5
`
`can therefore alter the cleavage properties of the
`linker, affecting resin swelling as well as complicat-
`ing electronic effects. The extra methylene unit
`present in 5.4 compared to the classical hydroxy-
`methylpolystyrene resin confers crucial properties
`such as acid stability,36,37 although possibly sensitiv-
`ity to (cid:226)-elimination once loaded.
`
`D. Linker Attachment
`Although the core structure of the linker may
`remain unchanged, the group placed between the
`linker and the support can modify the cleavage
`conditions and also alter the degree of linker cleavage
`as well. An example has been illustrated with the
`Rink resin 6.1.38 High concentrations of TFA can
`sometimes cleave some of the Rink linker from the
`polystyrene support and introduce colored impurities
`into the cleaved product. The Rink amide AM 6.2
`(RAM) and Rink amide MBHA 6.3 are much more
`stable to TFA (Scheme 6) (these constructs are made
`by coupling the so-called Knorr linker39 to the resin
`through an amide bond).
`
`E. Leaving Groups and Scavengers
`The Rink linker may be employed to attach a range
`of different functional groups to a common solid
`support, for example acids, amides, amines, etc.
`However, each functionality can be cleaved only
`under specific conditions. Thus, 0.1% TFA in CH2-
`Cl2 will release acids38 from linker 6.4 while 5% TFA
`in CH2Cl2 is needed to cleave alcohols from 6.5 and
`amides40 from 6.6 (Scheme 6). Since the dimethoxy-
`benzhydryl cation is generated in all of these cases,
`labilities must depend on the ease of protonation of
`the attached scaffold as well as its leaving ability.
`Thus, clearly, linkers are only one factor in determin-
`ing “cleavability”.
`Another important factor, which can impede cleav-
`age, is the reversibility of the reaction. This is
`certainly well-documented in solid-phase peptide
`synthesis (SPPS) where linker cation alkylation can
`be a serious problem, especially with peptides con-
`taining cysteines (thiols) or tryptophans (indoles),
`with the extent of alkylation being directly related
`to the proximity of specific residues to the linker. This
`is presumably also a problem in solid-phase organic
`synthesis (SPOS), although less well-documented
`with poor yields usually attributed to other factors.
`In the peptide area, scavengers are often used in
`order to trap the cationic linker species and prevent
`re-attachment. Usually scavengers such as ethanedi-
`thiol (EDT) or thiophenol are used, although ir-
`
`
`
`2096 Chemical Reviews, 2000, Vol. 100, No. 6
`
`reversible quenching using trialkylsilanes is common.
`Water is another very commonly used additive in
`cleavage solutions. Reagent K41 (82.5% TFA, 5%
`phenol, 5% H2O, 5% thioanisole, 2.5% EDT) is a
`cleavage mixture which has been found to be efficient
`(if smelly) and is now a fairly general cocktail for
`peptide chemistry. For multiple parallel solid-phase
`synthesis, much less complex scavenger mixes are
`desired/needed and often just water is used.
`
`IV. Linkers and Cleavages in Organic Synthesis
`In the following sections, classification of the link-
`ers has been attempted across seven major classes
`of cleavage reaction; however; overlap between classes
`does occur: (A) Electrophilically cleaved linkers, (B)
`Nucleophilically cleaved linkers, (C) Photocleavable
`linkers, (D) Metal-assisted cleavage procedures, (E)
`Cleavage under reductive conditions, (F) Cleavage
`under oxidative conditions, and (G) Cycloaddition-
`and cycloreversion-based release.
`Particular attention has been paid to linkers such
`as safety-catch linkers, which can achieve a higher
`degree of orthogonality by being uncleavable until
`activated. Purities and yields are considered through-
`out, and cyclorelease methodologies are exemplified
`in this respect. The common and appealing “traceless
`cleavage” terminology has been avoided as it lacks
`definition and is inaccurate.
`
`A. Electrophilically Cleaved Linkers
`Two main modes of electrophilic cleavage are
`used: protons and halogens. The largest one by far
`involves proton sources. Large arrays of compounds
`have been cleaved, including acids, amides, alcohols,
`thiols, amidines, amines, sulfonamides, etc. Although
`the introduction of a halogen group is useful, the
`range of cleaved products in this area is at the
`present time somewhat restricted.
`
`1. Strong Acid Cleavable Linkers
`a. Merrifield Resin. Solid-phase methodology was
`initially developed for the preparation of peptides; the
`classical synthesis involved the anchoring of the free
`carboxylic acid of benzyloxycarbonyl (Cbz) N-protect-
`ed amino acids onto a nitrated chloromethylpolysty-
`rene resin (Scheme 7). Subsequent deprotection of the
`Cbz-amino-protecting group with hydrogen bromide
`in glacial acetic acid gave the amines ready for chain
`extension after neutralization. Under these Cbz-
`deprotection conditions (10% HBr), the ester bond
`linking the peptide to the resin was stable (with only
`3.2% cleaved after 6 h).1 Release of the peptide was
`carried out under basic conditions (principle of or-
`thogonality) by saponification. In the synthesis of
`Bradykinin, Merrifield42 used tert-butyloxycarbonyl
`(Boc) amino-protecting groups which could be depro-
`tected with 1 M HCl without the loss of peptide from
`chloromethylpolystyrene resin (now known as Mer-
`rifield resin) (Scheme 7). Nitration of the resin was
`no longer necessary due to the reduction in acid
`strength used for amine deprotection (i.e., Boc vs
`Cbz). This alternative use of protecting groups also
`allowed the use of hydrogen fluoride for both resin
`
`Scheme 7
`
`Guillier et al.
`
`and side-chain deprotection,43 a cleavage method
`which has dominated Boc/Bn solid-phase peptide
`chemistry since its introduction in 1965.44 The HF
`method gives products of high purity and in good
`yield with excess volatile HF being removed by
`evaporation.
`Nevertheless, low temperatures (0 °C) and short
`reaction times (30-60 min) are advised with HF in
`order to avoid side reactions.45 Techniques such as
`Low-High HF cleavages, where temperature, reac-
`tion time, concentration, and the nature of the
`scavenger are varied, have been created, but in each
`case they are strongly dependent on the side-chain
`protecting groups that are present.46 Using a low
`concentration of HF, the alkyl side chains are trapped
`by scavengers, preventing side-chain alkylation.47,48
`A higher HF concentration enables the cleavage of
`the clean product from the resin. Trifluoromethane
`sulfonic acid (TFMSA), an alternative to HF, was
`introduced in 1974.49 Even though it is a very strong
`acid, it does not require special laboratory glassware.
`However, unlike HF, TFMSA is not volatile and can
`therefore be difficult to remove. Peptides must be
`precipitated from solution using dry solvents such as
`ether and are susceptible to salt and scavenger
`association. They need to be neutralized and desalted
`before further purification. As with HF, TFMSA can
`be used in a Low-High procedure.
`Some benzyl-based linkers are presented in Schemes
`8-11. Hydroxymethylpolystyrene resin 8.1 has been
`used in place of Merrifield resin to anchor carboxylic
`acids by a number of methods, including DIC/DMAP
`activation or Mitsunobu coupling. The sensitivity of
`the resulting ester to cleavage naturally remains
`unchanged.
`
`
`
`Linkers and Cleavage Strategies
`
`Chemical Reviews, 2000, Vol. 100, No. 6 2097
`
`Amines can be easily loaded onto Merrifield resin,
`but their removal is nearly impossible. However, a
`very convenient way to release amines from benzylic
`supports is to use the carbamate linker 8.3. In this
`case the amine can be liberated under acidic condi-
`tions with evolution of CO2. Burdick50 transformed
`hydroxymethyl resin into the chloroformate deriva-
`tive 8.2 using phosgene51,52 and then reacted this
`intermediate resin with anilines. Cleavage with HF
`and anisole provided the desired anilines. Leznoff
`also obtained carbamate 8.3 from p-nitrophenylcar-
`bamate derivative 8.4 for the anchoring of sym-
`metrical diamines. Cleavage was then achieved ei-
`ther using anhydrous HCl in benzene or TFA/TFAA/
`CHCl3 (10:1:10).53,54 A team from Organon recently
`used disuccinimidyl carbonate linker 8.555 to prepare
`a library of 150 pyridine-based compounds obtained
`with purities of between 40% and 60% after cleavage
`with TFA/thioanisole56 (9:1) for 4 h. Alcohols have
`also been anchored and released from Merrifield
`resin using dissucinimidyl carbonate 8.5, Scheme 8.57
`
`Scheme 8
`
`A particular example of nitrogen-based compound
`release from carbamate resin is shown with the
`cleavage of carbamate 9.1 under oxidative conditions
`(TFA/CH2Cl2 (2:1) and O2 bubbling) or reductive
`conditions (TFA/CH2Cl2 (2:1) and 1 equiv of trieth-
`ylsilane) leading to the formation of pyridines 9.2 or
`tetrahydropyridines 9.3, respectively (Scheme 9).58
`
`Scheme 9
`
`Decarboxylation procedures can be used to afford
`products where the carboxyl group forms part of the
`
`linking group involved in attachment to the solid
`support. The interesting feature is that the generated
`compound “loses” its attachment point. In 1970
`Patchornik59 generated ketones 10.2 after treatment
`of a (cid:226)-ketoester anchored onto a 2% cross-linked
`Merrifield resin using dry HBr in TFA (Scheme 10).
`
`Scheme 10
`
`More recently, supported 2-carboxylquinazolines have
`been involved in the synthesis of quinazolines by
`decarboxylative cleavage using TMSI as the Lewis
`acid source for 1-3 days at 75 °C. Postcleavage
`treatment with 1 M HCl in solution phase was
`carried out when decarboxylation was not spontane-
`ous.60
`The relative weakness of the N-O bond has been
`used in the cleavage of 1-hydroxyimidazole 11.2 from
`resin 11.1.61 However, heating at 100 °C for 20 h in
`TFA using a sealed tube was required followed by
`trituration with 37% aqueous HCl to form the hy-
`drochloride salt (Scheme 11).
`
`Scheme 11
`
`b. PAM Linkers. The stability of the benzylic
`ester-based linkages (esterified chloro- or hydroxy-
`methylpolystyrene resins) to TFA is not absolute.62
`Thus, when ribonuclease A was synthesized on this
`resin, an average of 1.4% of the growing peptide chain
`was lost after each deprotection step (using 50% TFA
`in CH2Cl2), which was disastrous for this 124mer
`peptide.45 In 1976, Sparrow63 decided to place a long
`“spacer” between the point of attachment of the first
`residue and the polystyrene support. Two important
`advantages were expected: Higher overall yields
`after HF cleavage and better homogeneity of the
`peptidic product. The author did not invoke the
`acidolysis problem but believed that uncompleted
`coupling at each step occurred due to the steric
`hindrance of the polystyrene backbone. He observed
`a 3-fold improvement in the overall yield of a 19-
`residue peptide using spacer/linker 12.1. Merrifield64
`prepared the PAM linker 12.2 and suggested that,
`without this linker, loss of material at each cycle of
`peptide synthesis was in fact due more to a partial
`cleavage (using TFA 50% in CH2Cl2) rather than to
`uncompleted coupling. The presence of the electron-
`withdrawing phenylacetamidomethyl (PAM) linker
`was shown to increase the stability of the peptide
`ester 100-fold relative to the peptide ester obtained
`from chloro- or hydroxymethylpolystyrene resin when
`submitted to 50% TFA in CH2Cl2.65 These PAM
`
`
`
`2098 Chemical Reviews, 2000, Vol. 100, No. 6
`
`Guillier et al.
`
`linkers (Scheme 12) thus constitute a more stable
`form of the conventional Merrifield benzyl ester
`linkage.
`
`Scheme 14
`
`Scheme 12
`
`c. Benzhydryl Linkers. Preparation of carboxa-
`mides through acidolysis became feasible using a
`benzhydrylamine (BHA) linker designed by Marshall
`(Scheme 13).66 Selective cleavage of the N-C bond
`Scheme 13
`
`then further peptide elongation allowed the prepara-
`tion of a C-terminal thio acid peptide upon HF
`cleavage. These linkers (Scheme 14) are particularly
`useful for the convergent synthesis of peptides by
`chemical ligation.
`Benzhydryl resin 14.6 can also be used to release
`protected peptide alcohols using repetitive treat-
`ments (2-5 min each) of 1-2% TFA in CH2Cl2.74 In
`this case, the mild acid lability is due to the alcohol
`leaving group.
`d. Miscellaneous Strong Acid Cleavable Link-
`ers. Unde´n75 prepared OMPPA (4-(3-hydroxy-4-
`methoxypentyl)) phenylacetic acid linker 15.1 as a
`new linker for SPPS using Boc chemistry (Scheme
`15). The author studied the stability to TFA of the
`
`Scheme 15
`
`occurs, allowing the formation of the carboxamide.
`A charge-stabilizing group (phenyl) was added onto
`the benzylic carbon in order to increase the stability
`of the carbocation, enabling the equilibrium to be
`displaced toward the cleaved product.
`Matsueda67 described the preparation of the 4-me-
`thylbenzhydrylamine linker (MBHA) 14.2. This linker
`is more easily cleaved than the BHA linker 14.1 due
`to the effect of the extra methyl group stabilizing the
`cation formed upon cleavage. However, HF was still
`required. In a comparative study of benzhydrylamine
`linkers (Scheme 14) and chloromethylated resins by
`Hruby,68 the author points out that yields can be
`slightly improved if proper use of the cleavage
`mixture is made. TFMSA and HBF4/thioanisole in
`TFA, a weaker acid than HBr in TFA, were also
`reported to cleave the MBHA linker.69 Houghten,
`surprisingly, achieved the cleavage of some amines
`from MBHA 14.2