Roughley 2011
Roughley 2011
Roughley 2011
pubs.acs.org/jmc
r 2011 American Chemical Society 3451 dx.doi.org/10.1021/jm200187y | J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Table 1. Summary of References Analyzed then analyzed manually for each paper and the results aggregated
for analysis following the broad reaction classifications used in
AstraZeneca (AZ) GlaxoSmithKline (GSK) Pfizer total
the analysis of reactions used in process (or development)
total papers 24 65 50 139 chemistry.2 In some instances, papers were abstracted with a
JMC 2 12 6 20 comment along the lines “analogues prepared in this manner”;
BMC 0 1 3 4 here, we referred to the original paper to elucidate how many
BMCL 22 52 41 115
variations on each step were performed.
For the analysis of compound features, compound structures
reactions 1117 3897 2301 7315
were retrieved by searching the Beilstein database for each
test compounds 602 1801 1163 3566
citation and selecting only those compounds where pharmaco-
logical data were abstracted (Beilstein query term “PHARM
reliable, and widely applicable methodology to allow concise and exists”).6 Structures were manually refined to remove synthetic
dependable access to new areas of chemical space. intermediates and reagents that have pharmacological properties
It became apparent to us during the early stages of the reaction reported elsewhere, along with compounds abstracted from
analysis process that the results of the analysis could be poten- figures giving background to the biological target investigated.
tially biased by the common approach of most medicinal In a small number of cases, structures had not been abstracted,
chemistry programs of synthesizing a common core motif and and in these cases they were entered manually. By use of these
then performing multiple derivatizations of this core in order to methods, the 7315 reactions and 3566 test compounds described
generate useful structureactivity relationships (SAR) for the in 139 publications were compiled into a data set upon which the
project team. The early synthetic stages would only be counted required analyses could be performed (Table 1).
once but potentially be used in the synthesis of multiple Numbers of screened compounds were determined by manual
compounds. To address this potential bias, we decided to employ analysis of the original papers. The number of synthetic steps to
in tandem a second analytical approach, in addition to this each compound was determined manually from the text and
reaction-based method, in which we considered the functional schemes in the paper. Where more than one route to a
groups present in the reported molecular structures, in particular, compound was described, only the shortest route was included
choosing those groups representing the final product of the in this part of the analysis. Figures quoted are for the longest
reaction types used (i.e., reaction = amide formation, functional linear sequence from the starting material described in the paper,
group = amide; reaction = Sonogashira coupling, functional regardless of whether the origins of this material are described as
group = arylacetylene; etc). This latter approach will also being commercial sources, described elsewhere by citation, or
under-represent certain reaction types, as their product func- not discussed. As such, they represent a minimum length for the
tional groups are generally not considered to be “druglike” (for synthetic sequence. It is impossible to ascertain without doubt
example, O-sulfonylation is a widely used reaction due to the what materials were available to the chemists performing the
synthetic utility of the sulfonate ester products, but there are no individual work from internal corporate collections, and so
sulfonate esters present in the compounds with pharmacological further elaboration of this is not possible.
data because of this same reactivity). However, we believe that, in “Lipinksi”7,8 properties [MWt, H-bond donors (HBD) and
concert with the reaction-based analysis, it will provide a deeper acceptors (HBA), clogP,9 number of rotatable bonds (NRot),10
insight into the chemistries employed by medicinal chemists. and topological polar surface area (TPSA)10,11], numbers of
chiral centers, and the Fsp3 parameter recently proposed by
Lovering et al.12 were calculated using Python, version 2.6.5,13
2. DATA GATHERING and RDKit.14
In order to identify relevant articles, CAS Scifinder5 searches Functional group counts were generated using SMARTS15
were performed by institution, using the keywords “AstraZeneca”, substructure query strings in RDKit. SMARTS strings were vali-
“GSK”, and “Pfizer”, and refined by year 2008, followed by analysis dated using SMARTSViewer.16,17 Aromatic rings were counted
by journal name and keeping only records from Bioorganic and following the system adopted by Ritchie and MacDonald,18 where-
Medicinal Chemistry (Elsevier, BMC), Bioorganic and Medicinal by when a ring is part of a fused ring system, each ring is counted
Chemistry Letters (Elsevier, BMCL), and Journal of Medicinal separately (e.g., purine = imidazole þ pyrimidine). Individual
Chemistry (American Chemical Society, JMC). Manual analysis aromatic ring types were counted using a set of SMARTS strings;
removed a small number of papers from this set that did not SMARTS strings for all 6,6-, 6,5-, and 5,5-fused aromatic rings
represent medicinal chemistry SAR publications. Several papers containing C, N, O, or S, whether synthetically feasible or not, were
provided details of SAR of a number of compounds without systematically enumerated computationally. Analysis of amine,
providing synthetic details; in these cases, the compounds are amide, ether, and halogen environments was performed similarly,
included in the compound-based analysis but not in the reaction using further SMARTS query strings. Counts were aggregated using
based analysis. The distribution of papers, reactions, and com- Knime Desktop, version 2.2.2.19,20
pounds among target classes is tabulated in the Supporting In-
formation (Tables S1S3); kinases and GPCRs (peptidic and
aminergic) account for approximately half of the papers and 3. REACTION ANALYSIS
compounds, with a broad spread of other target classes represented The reactions in the data set were classified into 10 major
across the remainder. categories based on the overall reaction type, e.g., heteroatom
Reactions were retrieved from SciFinder,5 or in those cases substitutions (see Table 2), and within each category into further
where no reactions were abstracted for a paper, searches were subtypes, e.g., N-alkylation, O-alkylation, etc. In the following
repeated in the Beilstein database6 on a paper-by-paper basis or, sections, we will discuss the overall picture in the major cate-
in a small number of cases, abstracted manually. Reactions were gories and then discuss each category in detail. We conclude this
3452 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Table 2. Continued
reaction type no. of reactions % of total % of subtype
oxidations (continued)
other 18 16.4
functional group interconversion (FGI) 413 5.6
alcohol to halide 68 16.5
amide to imidoyl chloride 7 1.7
acid to acid chloride 39 9.4
nitrile to acid 5 1.2
dehydration 20 4.8
carbonyl to CdN 26 6.3
other 248 60.0
functional group addition (FGA) 78 1.1
halogenation 37 47.4
nitration 3 3.8
sulfonation 2 2.6
other 36 46.2
total 7315
a
Protecting group manipulations are detailed in Table 3.
part of the analysis with a brief discussion of the “top 10” analysis,2 the nitrogen functionalization reactions were further
reactions reported in the data set. subdivided by reaction type.
3.1. Overview of Classes. Table 2 shows the number of reactions 3.2.1. N-Alkylation. Alkylation of nitrogen with alkyl halides,
abstracted in each broad class. At this level of detail, the obtained despite the contingent issues with overalkylation and side reac-
figures correspond well to those outlined in the analysis of reactions tions, accounts for almost a quarter of transformations in this
used in larger-scale drug candidate synthesis.2 Some clear trends class. Reductive aminations,28,29 offering similar products with a
become evident when considering this summarized data set alone. greater degree of control over reactivity and the product thus
The formation of carbonheteroatom bonds account for obtained, are surprisingly only equally popular. These two
almost half of all reactions (45.5%) in the data set, being almost methods together, however, account for half of all heteroatom
equally divided between acylation reactions (amides, ureas, derivatizations. It is worthy of note that in a recent paper of lead
sulfonamides, etc.; see below) and alkylation and arylation optimization and array chemistry at GSK,4 to the surprise of
reactions, perhaps providing an initial indication of the perceived those compiling the array analysis, no examples of reductive
dominance of reductive aminations and amide formations in amination were reported during the survey period.
medicinal chemistry. By contrast, only a little over 1/10 of the 3.2.2. N-Arylation with ArX. Despite the popularity of the
transformations resulted in the formation of carboncarbon bonds. above two transformations, the most common CN bond
This latter figure is perhaps surprising in view of the supposed forming reaction in this section is the aromatic substitution of
preponderance of palladium-catalyzed cross-coupling reactions in an arylhalogen species by a nitrogen, accounting for more than
medicinal chemistry. Despite advances in chemoselective and one-quarter (27%) of the reactions in this category. This
tolerant reaction methodologies, the practicing medicinal chemist displacement of aromatic halogen atoms with amines, both by
still appears to rely heavily upon protecting group strategies to “traditional” nucleophilic displacement (SNAr and related me-
enable the construction of the desired chemical entities. While chanisms, e.g., ANRORC30,31 and SRN132) and by Buchwald
clearly inefficient in terms of time, reagents, and overall yields,2124 Hartwig palladium-catalyzed processes,3340 highlights the uti-
these facilitating processes account for a further 1/5 of all reactions, lity and applicability of this transformation. This utility is no
with this figure being dominated by deprotection reactions. doubt aided by the ease of preparation of the precursor haloge-
Despite the broad utility of heteroaromatic ring systems as core nated heterocyclic moieties and their widespread commercial
scaffolds within medicinal chemistry programs,2527 only 8.2% of availability, itself almost certainly encouraged by their utility in
reactions involve heterocycle formation. This must reflect either CC bond forming processes (see below).
their formation early in a synthetic scheme or their purchase from 3.2.3. Amide and Heteroaryl N-Alkylation. While amide
commercial sources and subsequent diversification. The remainder formation remains ubiquitous (see below), further derivatization
of the reactions are accounted for by oxidation state adjustments and of this functionality by N-alkylation is notably less commonplace,
other functional group manipulations. with few examples of amide N-alkylation appearing in the
3.2. Heteroatom Alkylation and Arylation. Chemists clearly analysis. While it is a common SAR manipulation to investigate
favor reactions involving heteroatom alkylation and arylation for the importance of a hydrogen bond donor in the derivative, this
which robust and reliable methods are readily available and offer transformation does not appear to be routinely employed to
a diverse set of widely applicable protocols (Table 2). Almost explore the SAR around the amide bond. This is most likely a
80% of all heteroatom alkylation and arylation reactions are result of the ease and popularity of reactions such as reductive
derivatizations of nitrogen atoms, with the large majority of the amination to prepare the substituted amine prior to the amide
remainder being oxygen functionalization. In view of this and in formation step. Similarly limited is the N-alkylation of nitrogen
line with the precedent from the earlier process chemistry containing heterocyclic cores, perhaps limited either by the
3454 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
important role such unsubstituted systems play as hydrogen The metrics regarding the acylation reactions discussed in
bond donors in their interactions with biological macromolecules sections 3.3.1 and 3.3.2 recapitulate very closely those described
or by issues of regioselectivity. in the literature for the synthesis of drug candidates on the
3.2.4. Alkylation of Other Heteroatoms. O-Substitution reac- process scale,2 further highlighting their importance in the
tions account for almost 1/5 of all heteroatom substitutions, preparation of biologically relevant molecules. However, it is
although this figure includes O-arylations along with alkylation difficult to ascertain whether this is the case because (i) the
by both the Mitsunobu reaction41,42 and alkyl halide and product functionalities are critically and uniquely important for
sulfonate reagents. the resultant biological interactions or (ii) their inclusion is self-
In terms of N-substitution reactions, the frequencies observed selecting in the early stages of compound evolution because of
largely reflect those reactions employed in the larger scale synthesis their broad applicability and reliability, leading to a readily
of drug candidate molecules,2 where N-alkylation and reductive accessible subset of molecules having desirable profiles.
amination reactions are highly prevalent. In contrast, both S- and 3.4. CC Bond Formation. Data presented in the literature
O-substitutions appear to be less common in the research setting surveyed clearly demonstrate that palladium-mediated CC
compared to the synthesis of molecules selected as potential drug bond formation is the methodology of choice for the construction
candidates. This cannot be readily attributable to the ease of of such bonds in small molecules for biological evaluation, account-
synthesis of such compounds, as in general there are fewer possible ing for almost two-thirds of all transformations in this class. The
side reactions than in the corresponding N-substitution transforma- diversity of available starting materials (a search of the Available
tions, and it was anticipated that such couplings would be more Chemicals Directory47 reveals ∼6600 aromatic boronic acids alone,
commonplace in the small-scale laboratory environment. along with 4800 boronate esters and 750 trifluoroboronate salts and
3.3. Acylation Reactions. The acylation class of reactions ∼660000 chloro-, bromo-, or iodoarenes), chemoselectivity, and
(including N- and O-substitutions with a variety of carbon- and tolerability of these reactions clearly adds to their utility and
sulfur-based electrophiles) accounts for 22% of the reported reac- attractiveness. Additionally, the reactions are, in general, readily
tions in our data set. As such, it occurs at a very similar rate compared amenable to parallel synthesis operations, often in contrast to the
to the heteroatom alkylation and arylation reaction types. However, other reactions in the CC bond forming category. The Suzuki (or
in contrast to that group, the acylation class is dominated by amide SuzukiMiyaura) cross-coupling reaction40,4850 is the single most
formations, which account for over 70% of such processes (Table 2). numerous reaction within the CC bond forming group, account-
3.3.1. Amide Formations. In line with the common perception ing for 40% of all such reactions. This popularity is almost certainly
described in the introduction, amide formation is clearly the most due to the almost unique combination of reagent stability51 and
numerous reaction, both within the acylation reactions class (in safety.52 Alternative Pd-catalyzed processes are generally only
which 7 out of every 10 reactions are amide formations) and in employed when issues of stability or reactivity prevent a successful
absolute reaction count across all classes (see below). Likely outcome in a Suzuki reaction.
explanations for this include the wide range of robust methodol- The Sonogashira reaction,5355 in which a terminal acetylene
ogies available for the synthesis of amide bonds, as a result of the is coupled directly with a suitable haloarene or haloalkene, ranks
efforts made in the area of peptide synthesis,43,44 the ready surprisingly highly in this list at 18%, the second largest within
availability of starting materials by a range of synthetic methods, the category. Again, this is likely to be a combination of the
and the relative ease of purification of the reaction products, relatively benign nature of the starting materials and the utility of
factors that contribute to the amenability of the reaction to high- the acetylene linker, both in its own right and as a precursor to
throughput parallel synthesis.4 We found a wide variety of other functionalities.
reagents and conditions used to accomplish this transformation These couplings are not without limitations and care must be
within this data set, with no clear preference among the com- taken to ensure that residual palladium does not cause spurious
munity for a particular reagent. noncompound related artifacts, particularly as a compound
3.3.2. Other Acylation Reactions. Additional N-acylations progresses into more advanced cellular and in vivo assay systems,
(including the analogous sulfonylations) are also well repre- where palladium toxicity may prove confounding.5658
sented, with the SAR-informative N-sulfonamides and ureas being Outside the scope of these couplings, a variety of other CC
prepared in roughly similar proportions, between them accounting bond forming reactions are widely represented to a lesser extent,
for 2/3 of the remaining reactions in this class. O-Sulfonylation with selected examples of ester condensations, Grignard reactions
reactions are the next largest group, accounting for almost 5% of (including Weinreb amide-style methodology5961), and Friedel-
reactions in this category, almost invariably representing the forma- Crafts acylations comprising the majority of the described transfor-
tion of mesylate (OMs), tosylate (OTs), and more rarely, triflate mations. Wittig reactions6264 and Grignard reactions,5961 both of
(OTf) sulfonate esters to serve as leaving groups in heteratom which we highlighted in our informal survey,1 each account for
alkylations (directly or via conversion to a halide first) and, to a around 5% of CC bond forming reactions (and therefore around
lesser extent, as cross-coupling partners in Pd-catalyzed processes. 1 in 200 of all reactions), making them some of the most common
The remaining reaction types within this class are considerably less non-Pd-mediated CC bond forming reactions but relatively rare
well exemplified, most likely because of a combination of their overall. We have grouped the various other metal-based processes
in vivo lability (which is often deliberately exploited in the design of together in a single “organometallics” category, in contrast to the
prodrugs45,46) or strong basicity. While O-acylation is described review of process chemistry reactions,2 as the individual subtypes
both here and elsewhere, the examples listed here are distinct from were very rare, with only a few examples of each. Among other
those O-acylations where the sole aim is protection of functionality. methodologies not listed specifically in Table 2 are a wide variety of
In the instances listed above, the O-acylation is a deliberate step to transformations, predominantly aldol-type condensations.6570
introduce the ester functionality, either into compounds for biolo- Once again, the common underlying theme with the majority of
gical evaluation or as a synthetic handle for further manipulation, these transformations is the accessibility of a wide range of starting
rather than the addition of a protecting group. materials, some tolerability of other functionality, and the robust,
3455 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Table 3. Detailed Analysis of Protecting Group Manipula- being O-containing heterocycles syntheses. The analysis of
tions within the Reaction Data Set compound substructures later in this manuscript discusses in
more detail the individual heterocyclic systems reported within
protection deprotection this data set.
3.6. Protecting Group Manipulations. Across the protecting
protected functionality/group no. % no. %
groups encountered in the data set (Table 3), all examples
NH total 88 39.0 608 46.2 showed considerably more (up to 20-fold) deprotections than
N-Boc 73 32.4 357 27.1 corresponding protection steps. This trend is in agreement with
N-Bn 0 0.0 29 2.2 that reported for process chemistry.2 There are several explana-
N-Cbz 1 0.4 10 0.8 tions for this trend: (1) use of ready-protected commercial
other NH 14 6.2 212 16.1
building blocks; (2) use of ready-protected noncommercial
building blocks from internal archives; (3) the protection of a
RCO2H 92 40.9 395 29.9
common motif, followed by parallel diversification and subse-
ROH total 41 18.2 279 21.1
quent deprotections of multiple analogues; (4) the introduction
OBn 1 0.4 19 1.4 of protected functionality by other means, with later deprotec-
OSiR3 11 4.9 89 6.7 tion; examples include displacement of a halogen with nitrogen,
OAc 3 1.3 12 0.9 oxygen, or sulfur nucleophiles71 and the Curtius degradation of
other OH 26 11.6 159 12.1 acids to Boc-protected amines (see Scheme 1 for examples).72 It
RSH 21 1.6 is noteworthy that ∼80% of all protecting group manipulations
others 4 1.8 16 1.2 are related to one or the other component of the amide formation
total 225 100 1319 100 reaction (i.e., acid or amine), reflecting the dominance of this
reaction, although it is important to emphasize that this is not the
sole use of such products.
Scheme 1. Introduction of Protected Functionalities by Protection of carboxylic acids and NH groups occur in similar
Non-Protection Routesa numbers, accounting for 80% of all protection reactions. Boc
protection of NH groups dominates, with no benzylations and
only a single example of a Cbz protection in the data set. The
corresponding deprotections show greater diversity, although
unsurprisingly Boc still dominates. This is almost certainly due to
the broad availability of Boc-protected reagents and the asso-
ciated simple protection and deprotection conditions.
Carboxylic acids are invariable protected as simple esters,
predominantly methyl and ethyl, while occasional use is made
of benzyl and tert-butyl esters when alternative deprotection
conditions are required.
For protection of alcohol and phenol groups (which account
for ∼20% of both protection and deprotection reactions), silyl
ethers are the biggest single category accounting for around one-
quarter of all OH protections and around one-third of OH
deprotections. Over half of all OH protections and deprotections
are categorized as “other”, suggesting that beyond the silyl ethers
there is a broad range of protection strategies used in this class, often
reflecting a combination of “personal favorites” and commercial
availability of preprotected starting materials. The corresponding
a
sulfur analogues are more rarely encountered; there are no examples
Top: Introduction of a methyl ether by SNAr in a synthesis of CXCR2 of S-protection in the data set and only 21 deprotections (1.6% of
antagonists (AZ) and subsequent deprotection.71 Bottom: Introduction total deprotections), 19 of which are from a single paper from GSK
of a Boc-protected amine by Curtius degradation of a carboxylic acid in
the synthesis of EphB4 tyrosine kinase inhibitors (AZ).72
describing a series of thiol-based ACE2 inhibitors.73 This is almost
certainly a consequence of the reactive nucleophilic nature of the
reliable nature of the chemistry involved. The large proportion of deprotected thiol groups, which can disrupt disulfide bridges within
reactions in the “other” category reflects the centrality of the CC proteins and undergo covalent conjugation with free cysteine thiol
bond forming reaction category to organic synthesis and the groups (either directly or via breaking of an existing cystine disulfide
diversity of methods available for its execution. bridge), leading medicinal chemists to avoid their incorporation.74
3.5. Heterocycle Formation. A heterocycle synthesis in 3.7. Reduction Reactions. The most common transforma-
which multiple different heteroatoms are present in the synthe- tion within this group is the reduction of the nitro group to an
sized ring was classified under each; i.e., pyrroles and imidazoles amine (Table 2), often employing the precursor both as a
are classified as N-containing, but thiazoles are classified as both regiochemical director of installed functionality and as a masked
N-containing and S-containing, following the same convention amine group, ready to be derivatized via acylation or reductive
used in the analysis of development chemistry reactions amination. Similarly, the reduction of amides, nitriles, and imines
(Table 2).2 From this analysis, it can be seen that N-containing are also commonplace, in this instance generating a homologated
heterocycle synthesis dominates, accounting for almost 90% of amine that can also provide useful structureactivity informa-
this reaction class, with the majority of the remaining reactions tion. Overall, reductions of various functionalities to amino
3456 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
groups correspond to at least 37% of reactions in this group (see leaving groups in SNAr reactions, where they perform better than
below for “other reductions”), again pointing to the centrality of their thioether precursors.76,77
amine-derived products within medicinal chemistry. These data Oxidation of alcohols to acids is the next biggest single
do not include reductive aminations (another 386 reactions), as oxidation type reported (16%), again highlighting the impor-
in general the imine intermediate is not isolated. tance of the acid group within medicinal chemistry as a precursor
The reduction of ester to alcohol occurs with a frequency to amides and heterocyclic compounds in particular. A broad
similar to that of the analogous amide to amine reduction. range of other oxidation processes are also represented, although
Reduction of aldehydes and ketones to alcohols also occurs at all at lower levels than the above. It is clear that in the majority of
a similar level, and in total OH-forming reductions account for cases, oxidation is used as an adjustment prior to further reaction.
almost 25% of reductions. 3.9. Functional Group Interconversions. Given the preva-
Interestingly, the reduction of alkene to alkane (11.3%) is lence of certain key reaction types already noted, it is hardly
much more prevalent than the reduction of alkyne to alkane (2.2%). surprising that the key functional group interconversions identi-
It has been supposed that both the Heck75 and Sonogashira5355 fied by this analysis are the conversion of an alcohol to a halide
palladium-mediated couplings are primarily employed specifically to (16.5%, Table 2) and carboxylic acid to acid chloride (9.4%),
form such alkane linkages postcoupling by simple hydrogenation. given that these functionalities can both be employed in the
However, these figures do not reflect this hypothesis, being inversely highly prevalent derivatization of amines by alkylation or acyla-
proportional to the frequency of such cross-coupling reactions (the tion reactions, respectively. The vast majority of examples within
Sonogashira coupling is 50-fold more prevalent in this data set than this category come in the “other” category, representing the
the Heck reaction (Table 2)). The number of Sonogashira reactions diversity of this group.
(155) far outweighs the instances of alkyne reduction (9), suggest- Within the “other” category, a number of reactions are
ing that the alkynes so-formed are incorporated for inclusion in the featured on a regular basis: (1) diazotization reactions of anilines
final products or for nonreductive derivatization. Conversely, alkene and their subsequent conversion to iodoarenes (the Sandmeyer
reductions (46) outnumber Heck reactions (3). The most likely reaction); (2) halogen displacement with CN; (3) Pd-mediated
explanation for this is that the alkene group is synthesized by alkoxy- and amino- carbonylations to esters and simple amides.
alternative methods, such as the Wittig reaction6264 or aldol/ While these last two transformations could be considered as
elimination reaction.6570 carboncarbon bond forming reactions, they have been in-
Other reductive processes not specifically mentioned in the text, cluded here rather than the alternative section, as their use was
which account for almost 23% of reductions, predominantly primarily to adapt or interchange pendent functionality, as
included examples of nitroso reductions to the corresponding amine opposed to constructing the molecular scaffold.
and heterocyclic scaffold reformation, such as the reductive cleavage By far the most prevalent transformation not specifically listed
of an isoxazole and subsequent recyclization to a pyrimidine or above, however, was the conversion of aromatic halides to
pyrazole, allowing expansion of SAR by facile scaffold-hopping. boronic acids and boronate esters prior to Suzuki coupling
3.8. Oxidation Reactions. Perhaps the most striking initial reactions.48,49 This accounted for almost half of the unlisted
observation regarding oxidation reactions is how much rarer they “other” transformations, predominantly from a single citation
are than reduction reactions (only 110 oxidation reactions within the data set.79
compared with 406 reductions (Table 2)). The reasons for this 3.10. Functional Group Additions. Halogenations dominate
are likely to be complex and depend on the availability of starting this facet of chemistry, and akin to the functional group inter-
materials, along with the hazardous nature of oxidizing reagents conversions discussed above, their occurrence is anticipated
in general, coupled with environmental concerns surrounding given the prevalence of Pd-mediated couplings observed earlier
the disposal and toxicity of the heavy metals employed in many that depend upon the routine access to halogenated precursor
oxidations. materials. They also benefit from a broad range of reagents, many
The most common member of this class (35%) is the of which show mild and selective reactivity, in contrast to the
oxidation of alcohols to aldehydes, useful intermediates for both other reactions in this group. More surprising, given the utility of
CC bond forming reactions and reductive amination reactions. the resultant functionality, the incorporation of a nitro moiety is
It is noteworthy that this oxidation state is accessed entirely by uncommon in the literature surveyed for this analysis. As the
oxidation within this data set, with no reports of the controlled nitro to amine reduction is relatively common (78 reactions in
reduction of nitriles or esters directly to the aldehydes. Many of the data set), it would appear that nitrated materials, in the
the alcohol precursors for this oxidation were obtained by context of the syntheses discussed herein, are generally pur-
reduction of the corresponding ester to the alcohol prior to chased or obtained by prior synthetic manipulation of commer-
reoxidation. It is likely that this two-step procedure is preferred, cially available nitro compounds rather than being prepared by
as the aldehydes are generally more reactive than the correspond- nitration as part of the published reaction sequence. It is likely
ing esters, resulting in difficulty controlling the reduction route, that this is due to a combination of the potentially hazardous
whereas there are many good methods for controlled oxidation nature of the nitration reaction, along with the diversity of
that offer relatively mild selective conditions to achieve the available nitro-containing building blocks (a search of the Symyx
required transformation without significant overoxidation. Available Chemical Directory47 reveals more than of 97 000
Clearly, a direct reduction method to aldehydes with the opera- nitroarene-containing compounds).
tional simplicity and reliability of the reductions to alcohols Other notable functional group additions not specifically listed
would provide an improvement in synthetic efficiency in these are the introduction of nitroso functionality as a masked amine,
cases.78 which may also go some way toward explaining the surprising
Following this are oxidations occurring at sulfur, sulfide to rarity of nitration processes, and the introduction of an enamine
sulfoxide and to sulfone, accounting for 23% of oxidations. Both substituent using dimethylformamide dimethyl acetal as a pre-
of these products are themselves most likely to be employed as cursor to heterocyclic ring formation.
3457 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Table 4. Top 10 Reactions by Frequency in the 2008 Data Set the top 10 list. Our straw poll correctly identified 7 (sulfonylation
and Sonogashira reactions, in addition to those listed above) of
reaction no. of reactions % of all reactions
this revised top 10 list. However, the Heck reaction is among the
N-acylation to amide 1165 16.0 10 least common reactions in our survey (only 3 reactions,
N-containing heterocycle formation 537 7.4 accounting for 0.04% of the total). The remaining 2 reactions in
N-arylation with Ar-X 458 6.3 our straw poll (Grignard and Wittig reactions) have been
RCO2H deprotection 395 5.4 discussed above.
N-subs with alkyl-X 390 5.3
In general, the trends are similar to those observed within the
field of process chemistry,2 with a few variations noted in the text,
reductive amination 386 5.3
most of which can be attributed to the differing goals of medicinal
N-Boc deprotection 357 4.9
and process chemistries.
Suzuki cross-coupling reaction 338 4.6
O-substitution 319 4.4
other NH deprotection 212 2.9 4. COMPOUND ANALYSIS
total 4557 62.4 We have noted throughout the manuscript that some reac-
tions may be over-represented because of their tendency to occur
later in the synthesis. To address the question of whether this use
3.11. Summary of Reaction Types and a “Top 10” Reac- of certain well-developed reactions that are readily amenable to
tions List. The literature data set shows that while medicinal parallel synthesis (e.g., amide formations) is biasing the reaction
chemists rely heavily on a relatively small number of reaction count in favor of these processes in a way not seen for the analysis
types, the remainder of the reactions used covers a wide range of of process development reactions,2 where only a single com-
different types in order to achieve the goal of discovering new pound is normally being made, we also analyzed all the com-
drug candidate molecules. In general terms, heteroatom alkyla- pounds within the published data set by compound structure. In
tion and acylation reactions account for almost half of the this case, we counted various functional groups that were likely to
reported reactions. Protecting group manipulations account for be formed by the synthetic reactions described. While this also
∼20% of reactions, whereas CC bond forming and hetero- has some limitations (for example, substructures being formed as
cycle-forming reactions each account for ∼10% of reactions. reactive functional handles that are subsequently further reacted
However, within this set, a small number of processes dominate not being included), we feel that it is perhaps more representative
(Table 4); notably, ∼1 in 6 of all reactions in the data set was an and, taken with the above reaction-based analysis, provides a
amide formation. While this figure is lower than has been more complete picture.
suggested (various sources quote figures of up to 50% of 4.1. General Points. Table 5 shows the headline data for the
reactions being amide formations80), it is still remarkably high occurrence rates of a range of functional groups within those
for a single transformation. Indeed, amide formations alone compounds for which biological data were reported (3566
account for more reactions than any of the other broad categories compounds). The data shown are the total number of occur-
except for heteroatom alkylation/arylation and deprotections. rences of the functional group or substructure within the data set,
Nitrogen-containing heterocycle formation is the next most the mean number of occurrences per compound (across all
commonly employed synthetic methodology. Notably, the Su- compounds, not just those in which it is found, i.e., the total
zuki coupling, a perceived favorite, does indeed appear in the top number of occurrences divided by 3566 compounds), the
10 but ranks lower than both N-Boc deprotection and carboxylic number and % of compounds containing at least one occurrence
acid deprotection. Seven of the top 10 list are either acylation of the functional group, and the maximum number of occur-
reactions or reactions leading to potential amide-forming pre- rences of that functional group within a single compound within
cursors, accounting for 30% of all reactions. The reactions in the the data set. Perhaps the most obvious point from Table 5 is the
top 10 list themselves account for almost 2/3 of all reactions universality of aromatic rings within medicinal chemistry; 99% of
reported. Taken together, we see that while medicinal chemists compounds contain at least one aromatic ring of some sort,
use a broad range of synthetic methodologies, the suggestion that 94.3% contain at least one benzene ring, and 72.5% contain at
there are a small number of “favorites” seems to hold some truth. lease one heteroaromatic ring, with an average over the entire
Interestingly, of these “top 10” favorites, 5 were highlighted in data set of almost three aromatic rings per compound. This figure
our previous publication (amide formation, N-arylation with compares well with Ritchie and Macdonald’s suggestion of a
ArX, reductive amination, Suzuki reaction, and O-sub- maximum of three aromatic rings per compound.18
stitution), the selection for which was based on an informal Additionally, amides occur in just over 50% of compounds,
“straw poll”.1 It is likely that the omission of nitrogen-containing and both aliphatic amines and biaryl systems occur in ∼40% of
heterocycle formation reflects the somewhat broad brush that compounds, while 35% of compounds contain an alkoxyaryl
this category covers, but it is interesting to speculate whether the ether group. Thioethers are the least prevalent heteroatom
fact that 3 of the remaining 4 reaction types were deprotection linkage, as expected because of their oxidation potential. Within
reactions reflects the mindset of the medicinal chemist, in which the halides, 14% contain an alkyl fluoride, while the other alkyl
such processes are almost overlooked when considering syn- halides are, as is to be expected because of their reactivity, almost
thetic complexity. Removing protecting group manipulations completely absent. Aryl fluorides (∼20%) and chlorides (∼30%)
(and ignoring the “other” categories within each category, as are common, with the heavier and more lipophilic bromo- and
these represent a combination of many differing processes) adds iodoarenes being less common. More detailed discussion of each
sulfonamide formation (163 reactions, 2.2% of total), Sonoga- subgroup is presented below.
shira reaction (155 reactions, 2.1%), urea formation (155 reac- 4.2. Amines. With 42.9% of compounds containing an
tions, 2.1%), and nitro to amine reduction (78 reactions, 1.1%) to aliphatic amine, this group is one of the most popular
3458 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Table 5. Summary of Functional Group Occurrences across the Data Set of Compounds with Biological Data (3566 Compounds)
total no. of occurrences of mean no./compd across no. of compds with g1 % of compds with g1 max. no. of FG in single
subclass FG in data set entire data set occurrence of FG occurrence of FG compd
Class: Amine
aliphatic 1719 0.48 1529 42.9 4
aryl 896 0.25 810 22.7 3
diaryl 561 0.16 503 14.1 2
Class: (Thio)ether
ROR 609 0.17 530 14.9 6
ROAr 1568 0.44 1231 34.5 5
ArOAr 283 0.08 280 7.9 2
C-S-C 234 0.07 228 6.4 2
Table 5. Continued
total no. of occurrences of mean no./compd across no. of compds with g1 % of compds with g1 max. no. of FG in single
subclass FG in data set entire data set occurrence of FG occurrence of FG compd
Class: Miscellaneous (Continued)
sulfone 88 0.02 86 2.4 2
ArNO2 15 0.00 15 0.4 1
a
5-mr and 6-mr refer to five-membered and six-membered rings respectively.
functionalities outside the aromatic rings. In many cases, an detrimental effects on potency. Visual inspection suggests that
amino group has been added to the molecule in order to enhance these are often introduced in benzylic positions, most likely by
solubility and decrease lipophilicity while attempting to avoid reductive amination of a benzaldehyde precursor or via ether
3460 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
(both of which count as two alkoxy aryl ethers in the headline totals)
account for ∼12% of alkoxyaryls, as do trifluoromethyl ethers
(F3COAr). Approximately 20% are in the “solubilizing” category,
with 28% of these containing a β-aminoethyl group. The remainder
(∼25%) are more complex ethers that appear to form part of the
core scaffold of the molecule.
From these figures, it is clear that the increased liability of the
aromatic ring toward oxidative metabolism on incorporation of
electron-donating substituents is offset by the requirement for
such groups, which presumably primarily serve as H-bond
acceptors, to obtain the required levels of potency. Additionally,
the alkoxy aryl ether motif is clearly a useful synthetic handle for
the assembly of scaffolds and the attachment of pendent func- Figure 4. Selected examples of diaryl ether-containing scaffolds con-
tionality (particularly solubilizing groups) to aromatic templates. tained within the data set.9199
This is almost certainly a result of the reliable and clean reactivity
of the phenol precursor under both classical (base, alkyl halide, or remainder are in general actually part of the solubilizing group (as
sulfonate) or Mitsunobu41,42 alkylation conditions. also in 7, Figure 2)87 or form part of the core scaffold (e.g., 8,
4.3.2. Dialkyl Ethers (ROR0 ). Of the dialkyl ethers, the second Figure 2).88 It is worth noting that 8 also shows examples of a
largest group within the (thio)ether subgroup, 36% are simple diaryl ether and an ethylenedioxy motif in addition to the
alkyl ethers as defined above (Table 8). Almost half (45%) of dialkyl ether.
these are methyl ethers. In contrast to the above, while the data Tetrahydropyrans and tetrahydrofurans account for a further 12%
set contains numerous β- or γ-heteroatom-containing ethers, of the dialkyl ethers (present as both sugar-derived units and simpler
only one example (7)87 of these appears to actually function as a systems), most notably the 4-aminopyran solubilizing group, which
point of attachment for a possible solubilizing group. The accounts for almost half of the tetrahydropyrans. While morpholines
3462 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Figure 7. Selected lactam-containing compound series (>10 compounds with motif), showing broad diversity of lactam-containing scaffolds and
targets.102113
equal portions (Table 10). Only 4% of the aliphatic-derived BACE-1 inhibitors bearing an N-linked pyrrolidin-2-one as an
amides are acetamides, while 65% can be considered to be aromatic substituent (e.g., 27)110,111 and by a series of benzo-fused
“solubilizing” (which we define here as containing an oxygen morpholinones (e.g., 25) showing combined 5-HT1A/1B/1D antag-
or nitrogen attached to the R-, β-, or γ-carbon of the aliphatic onism and serotonin (5-HT) reuptake inhibition.112,113 Many of
chain). Many of these are actually substituents, in which it may be these scaffolds can also be seen to contain additional amide groups.
the amide that is the actual uncharged solubilizing group (in The remainder of the lactams are present as isolated examples
contrast to amine solubilizing groups) and the chain heteroatom throughout the data set.
is actually part of a synthetic handle for their attachment. In addition to these “simple” amides, a small proportion
Unlike the case with the amines described above, no strong belongs to more complex functionalities [N-acylamides (imides),
preferences for particular amine derivatives were found in the data set. N-acylsulfonamides (only one example of this relatively acidic group
However, almost exactly two-thirds of the secondary amides were in the data set), and N-acylureas] which provide a complex array of
derived from aliphatic amines (but only ∼5% from methylamine, the H-bond donor and acceptor vectors.
simplest analogue). “Solubilizing” groups (containing a β- or γ- Sulfonamides and ureas are both present in ∼10% of com-
oxygen or nitrogen atom) account for one in five of the secondary pounds, which is in reasonable agreement with the reaction-
amides. Tertiary amides show an even stronger preference for alkyl based figures and reflects their comparison with amides for SAR
substituents (dialkylamides accounted for over 80% of the tertiary purposes. The other functionalities within this group are only
amides), while there are no diarylamides. This preference is most represented in <5% of compounds, which is unsurprising in view
likely due to a desire to avoid the potential for liberation of potentially of their chemical reactivity, biological lability, or permeability
toxic anilines by in vivo cleavage of the amide bond by hydrolase issues (see discussion in section 3.3.2).
enzymes. Around 50% of the tertiary amines bear a solubilizing group. 4.5. sp2sp2 Linkages. Biaryl linkages are present in almost
The increased proportion of tertiary amides bearing a solubilizing 40% of compounds. Again, this reflects the utility of this motif in
group relative to secondary amides suggests that the addition of a providing appropriate vectors within the core scaffold and also the
solubilizing group to the amide nitrogen is a common strategy prevalence of the Suzuki40,4850 and related Pd-mediated cross-
in situations where the NH motif has been shown to be unneces- coupling reactions40,50 (40% of CC bond forming reactions were
sary for biological activity (perhaps by initial synthesis of the N-methyl also Suzuki reactions). Perhaps surprisingly, the products of the
analogue of the parent secondary amine, as suggested by the increased Heck75 and Sonogashira5355 reactions (arylalkenes and acetylenes,
proportion of tertiary amines bearing a methyl substituent). respectively) are present in only low numbers. While the Heck
A surprisingly large proportion of the amides are also cyclic reaction was rare, the Sonogashira reaction was more common
(lactams, ∼12%), arising predominantly from a small number of (18.4% of CC bond-forming reactions), suggesting that the
compound series in which they are present as part of the core products of this reaction were largely used in further synthetic
scaffold, perhaps artificially elevating their standing within the data manipulations
set. Even within this set, however, there is considerable diversity 4.6. Aromatic Ring Systems. Aromatic rings, and in particular
within the lactam functionality represented (see Figure 7),102107 aromatic heterocycles, are almost ubiquitous in medicinal chem-
and two of these scaffolds each contain two lactam substructures (20 istry, as supported by the finding in this study (Table 5) that 99%
and 26).108,109 This motif is also heavily represented by a series of of compounds contain at least one aromatic ring. This is at least
3464 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Table 11. Occurrences of Individual Heteroaromatic Ring Types within the Data Set
aromatic ring type no. of heteroatoms total occurencesa % of heteroaromatics % of subclass nonfused occurencesc % of occurrences nonfusedd
in part explained by their ability to provide readily functionalized The remaining systems found are only seen in small numbers
scaffolds with well-defined vectors for further derivation. The (<50) of compounds. In general, the occurrence rate falls off as
number of heterocycles present (4362 heteroaromatic rings, with the number of ring heteroatoms increases.
72.5% of compounds containing at least one) far outstrips the Considering only those examples where the (hetero)aromatic
number of heterocycle-forming reactions (601 reactions, ring does not form part of a fused ring system shows some marked
Table 2). This implies that heterocyclic cores are synthesized changes (Table 11). Pyridines, pyrimidines, and pyrazoles again
prior to derivatization into a number of analogues, purchased, or dominate this list, with over half of the examples of each of these
are previously available within corporate compound collections rings not forming part of a larger fused system. Notably, pyrrole
and then derivatized, in accord with their role as core scaffolds. has moved from being one of the most common aromatic rings,
We analyzed in more detail the ring systems found within the considering all occurrences, to having only a small number of
data set, starting with the individual ring systems (Table 11). isolated examples (27 examples, representing ∼7% of pyrroles). It
Only 23 of the 35 possible heteroaromatic rings (we have only is likely that this is due to a combination of the reactivity of the
considered N-containing rings in the six-membered series and a pyrrole ring system precluding its incorporation in many circum-
maximum of four heteroatoms in any ring) are reported in the stances where suitable stabilizing substitution is not tolerated by
data set. In agreement with Murcko’s analysis,26 pyridine is the the receptor, and the preponderance of indoles (see below), which
most common heteroaromatic ring overall, accounting for almost accounts for around half of the fused pyrroles. A number of ring
25% of all heteraromatics. Pyrimidine and pyrazole are the next systems, such as the oxadiazole isomers, 1,2,3-triazoles, and
most common (both ∼15%). In the six-membered series, tetrazoles, occur exclusively or near exclusively in nonfused
triazines are less well represented and tetrazines are not repre- settings; in this case the position of heteroatoms precludes the
sented at all. In the five-membered series, pyrazole is the most formation of simple stable fused ring systems containing these
common (∼25% of five-membered heteroaromatics). Following rings. Benzenoid rings, despite being present in many of the most
this, pyrrole and imidazole are observed at similar levels (both commonly occurring fused ring systems (see below), are also
∼10% of all heteroaromatics). The most common (>250 highly prevalent as nonfused systems. Figure 8 shows the most
occurrences) aromatic heterocycles are shown in Figure 8. common (>150 occurrences) nonfused aromatic rings.
Thiophenes, furans, oxazoles, thiazoles, 1,3,4-oxadiazoles, and While the above analysis starts to show some interesting trends, it
1,2,4-triazoles are less common, although still well represented. is perhaps more informative to consider ring systems wherein two
3465 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
definition, BI values range from 0 to 1, with 1 indicating a across the data set; however, the majority of compounds fall in the
compound in which the aromatic rings are entirely benzenoid. range 50100 Å2, which is lower than the originally proposed limit
On the basis of Ritchie et al.’s analysis,81 increasing values are of 140 Å2.10
indicated as more likely to suffer from poor compound devel- If one considers the number of Lipinski’s rules violated, then
opability. As different numbers of aromatic rings give a different 1310 compounds violate at least one rule (a scenario allowed in
set of discrete values for this index (a system with three rings can the original publication7 but often overlooked). Of these, only
have BI of 0, 0.33, 0.67, or 1, whereas four rings can give 0, 0.25, 326 (9% of the compounds) break two (318 compounds) or
0.5, 0.75, and 1), but with some overlap (most notably 0 and 1, more (three violations, four compounds; four violations, four
which will distort the distribution toward these values), we compounds) of the rules and would thus be categorized as likely
decided to plot the distribution for each number of rings to have poor oral bioavailability. Leeson has published similar
separately (an alternative approach, particularly for larger data findings in a more detailed analysis of physicochemical properties
sets, would be to bin the values across all ring counts). We chose of compounds synthesized by leading drug companies.147 More
to ignore those systems with only a single aromatic ring (as this recent reports proposing harsher cut-offs postdate the publica-
gives only 0 or 1) and those systems with five or more rings, as the tion dates for these compounds, which in most cases will have
BI values are distributed over a larger number of values for only a been synthesized several years prior to their publication, and so
relatively small number of compounds in the data set. It can be they are not considered here except to suggest that the medicinal
seen (Figure 13c), however, that for two- and three-ring systems, chemistry community needs to make continued efforts to
higher BI values dominate, while for four-ring systems, there is a restrain the lipophilicity of the compounds synthesized.
much flatter distribution. Indeed there is a trend toward lower BI
as the number of rings increase. It is clear, however, that
medicinal chemists need to be aware of the potential implications 7. SUMMARY AND CONCLUSIONS
for compound development highlighted by Ritchie et al. of both The analysis of this data set has revealed that while the
aromatic ring count and type.18,81 medicinal chemist’s perceived reliance on a small number of
reactions (amide formations and Suzuki cross-couplings being
the most often cited) is generally true (10 reaction types
6. PHYSICOCHEMICAL (“LIPINSKI”) PROPERTIES comprise almost two-thirds of all reactions), these processes
Following Lipinksi’s seminal work,7 linking a number of and their resulting products are not as universal as might be
simple physicochemical properties (MWt, H-bond donors and believed. In addition to these processes, a large diversity of
acceptors and clogP) to human oral absorption of drug molecules reaction types and functionalities are reported by medicinal
(known as the “rule of 5”),8,138 the importance of physicochem- chemists, although the favoring of a small number of reliable
ical properties in drug discovery has received wide recognition. performers is understandable given the pressures of compound
Later papers have added the number of rotatable bonds delivery. The robust, reliable nature of palladium-mediated cross-
(NRot)10 and polar surface area (PSA10 or TPSA11), along with couplings and their wide applicability, chemoselectivity, and
many other more or less esoteric parameters139 to predict oral functional group tolerance make them natural candidates for
bioavailability. Others have attempted to extend this approach to the rapid generation of compound sets to ask specific questions
predict brain penetration,8 crossing of the placenta,140 and around the biological importance of key substituents. Indeed, this
exposure by a plethora of routes including inhalation, opthalmic importance and dependability were recognized most notably by
administration, and transdermal absorption,141 with varying the award of the 2010 Nobel Prize for Chemistry for the
degrees of success. Particular attention has been paid to the pioneering work in this area.151 Furthermore, the reliance upon
importance of lipophilicity (clogP), which has been linked, in a limited subset of reaction methodologies appears to have
addition to solubility and permeability, to cytochrome P450 helped enable and incentivize commercial reagent suppliers to
inhibition,142 hERG binding,143 increased adverse toxicological cater to these specific needs, resulting in large numbers of cost-
outcomes,144 and off-target effects,95 and recent papers have effective, diverse, and readily available starting reagents, facilitat-
suggested tighter constraints on this parameter.144150 While not ing rapid evaluation of SAR without recourse to the in-house
the primary focus of this review, we looked at how the com- synthesis of large numbers of bespoke building blocks. Con-
pounds in this data set were distributed among these physico- versely, it can be argued that lack of commercial availability of
chemical properties (Figure 14). The number of compounds with reagents may also detract from the popularity of some of the less
MWt > 500 rapidly tails off, although there are still a significant commonly used reactions. It is interesting to ponder whether
number of compounds in the 500 < MWt e 600 range (Lipinski’s those processes that are less commonly used are avoided because
original paper allows for one violation of his rules7). The number of of a dearth of reliable methods for their execution or whether
compounds violating the proposed limits of H-bond acceptors such methods have not been developed because of a perceived
(e10) and H-bond donors (e5) is very small; almost all com- lack of desire to incorporate their resultant functionalities.
pounds in the data set fall within these limits and many considerably The reliance on a small number of reaction types with the
below them. Roughly 10% of compounds exceed Veber’s proposed properties described above highlights the need for new methodol-
limit for NRot (e10),10 while most surprisingly, significant num- ogies or improvements to existing transformations, making them
bers (∼20%) of compounds exceed even Lipinski’s proposed clogP more generally applicable and amenable to parallel chemistry, as
limit of 5. It perhaps provides some vindication of Lipinski’s desire to noted previously.4 This would facilitate their use in the later stages of
appeal to the pattern recognition skills of medicinal chemists in medicinal chemistry syntheses, potentially broadening the diversity
formulating his “rule of 5” that the most violated of these rules is the of compounds synthesized. Indeed, the ACS Green Chemistry
clogP constraint, which is also the only one of his parameters not Initiative Pharmaceutical Roundtable recently issued a “call to arms”
readily calculated directly from the molecular structure without to address this very requirement, specifying a number of processes
recourse to computational tools. TPSA has the flattest distribution that were felt, across the industry, to be in need of additional research
3471 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
Figure 14. Distribution of physicochemical properties in the data set: MWt, molecular weight; NRot, number of rotatable bonds; HBA, H-bond
acceptors; HBD, H-bond donors; TPSA, topological polar surface area.
and optimization.152 This suggested not only more aspirational Despite advances in chemo- and regioselective syntheses and
ideals to improve upon areas of chemistry that were clearly not fully transformations with increasing tolerance of a wide range of
developed or widely applicable but also requests for further im- functionalities, one in five of all transformations analyzed in this
provements to tried and tested (and widely employed) methodol- paper are involved in protecting group manipulation. While some
ogies, where clear deficiencies in areas such as atom economy and notable efforts have been made toward protecting-group-free
safety were identified. In addition to allowing exploration of addi- synthetic strategies,2123 this area of synthetic methodology is far
tional chemical space and structural diversity, such methodologies from fully developed and a key benefit of novel transformations
would improve the safety, cost effectiveness, and scalability of the such as those discussed above would be the ability to run such
transformations employed in the medicinal chemistry laboratory. reactions without the need for recourse to protecting groups and
The discovery and development of robust conditions for the resultant deprotection steps, which are clearly wasteful in
existing and novel transformations that are readily practicable terms of time, reagent costs, and overall yield. Until such
in the context of medicinal chemistry (i.e., amenable to parallel methods are more widely developed, the use of protecting groups
synthesis approaches, using readily available reagents, high remains a necessary synthetic burden to facilitate preparation of
yielding, with broad substrate scope, tolerability, and reliability, required molecules, requiring optimization of extra protection
not requiring extremes of temperature, rigorous exclusion of and deprotection steps. Until protecting-group-free methodol-
oxygen and moisture, simple workup and purification) are key ogy becomes more widespread, medicinal chemists must tolerate
areas in which the academic synthetic chemistry community can the use of protecting groups while bearing in mind that their use
provide invaluable input to future medicinal chemistry programs. is not always necessary. In the authors' experience, it is often the
3472 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
case that deprotection can be moved forward in a synthesis prior heterocyclic systems, presumably in order to gain access to clear
to later diversification steps while maintaining chemoselectivity intellectual property areas. A relatively small number of aromatic
by appropriate choice of reaction and workup conditions. heterocycles (three systems of pyridine, pyrimidine, and
While less common in this data set (7% of reactions), redox pyrazole) account for half of all heteroaromatics, while fusion
processes have also been highlighted by Baran78 as a source of increases the diversity of heterocyclic motifs by a remarkably
synthetic inefficiency, in particular, those processes where multi- small amount, albeit into areas with much less overlap between
ple adjustments are made successively to the redox state of a each company in this analysis. Again, our analysis suggests that
single group, an area we highlighted in the sections dealing with medicinal chemists should actively seek a broader range of novel
these transformations. Of particular note in this review was the heterocyclic motifs, and in particular, care needs to be taken
repeated use of a two-step ester to aldehyde conversion, com- regarding both the overall aromatic ring count and the relative
prising over-reduction to the intermediate alcohol, followed by proportions of benzenoid and heteroaromatic rings (indicated by
oxidation to the required aldehydes. While this transformation our proposed BI index). New synthetic methods will need to be
can often be achieved directly using diisobutylaluminum hydride, developed in order to access novel ring systems, but the academic
this is often considered a poorly controlled reaction with synthetic community is perhaps unlikely to rise to the challenge
troublesome workup, requiring inert atmospheres and low unless it can be justified by evidence of biological activity of such
temperatures. New conditions for this process with operational ring systems. The incorporation of cyclic amine solubilizing
simplicity comparable to that of the reduction of ketones with groups, a common approach within medicinal chemistry to
sodium borohydride would be a welcome addition to the improve solubility and bioavailability of compounds (and indeed
medicinal chemists’ toolkit. It is a testament to the creativity of one of the perceptions we sought to investigate), has been
medicinal chemists that within these parameters (constraints of reported to be beneficial for developability,81 as perhaps expected
time, the need for parallel-amenable chemistry, and speed of because of the improved solubility conferred by such modifica-
access to products), such a broad diversity of biologically and tion. However, there is considerable scope for increasing the
therapeutically relevant compounds are produced from the diversity of amine groups used (see section 4.2) for this purpose.
available transformations. Furthermore, recent reports have also suggested that such groups
In recent years, microwave chemistry has found widespread may contribute to potential hERG toxicity and other off-target
use in the medicinal chemistry community, where its ability to promiscuity,147,148 suggesting that, where possible, nonbasic
provide dramatic rate enhancements, particularly in the areas of solubilizing strategies may be preferable.
metal-catalyzed reactions, has allowed for significant increases in Halogens, in particular fluorine and chlorine, are often in-
efficiency.153156 However, alternative complementary technolo- corporated to solve metabolic problems or modulate off-target
gies, such as flow chemistry reactors, are being embraced more effects. Their use, as such, is often to “fix” an otherwise suboptimal
slowly by the medicinal chemistry community. Indeed there were no molecule. The increasing potential for the heavier halogens to
reports in our data set of the use of flow chemistry systems, despite participate in halogen bonding interactions with biological targets
the ability to provide a safe and scalable entrance into otherwise was alluded to in section 4.8.123,127129 This interaction, in parti-
hazardous reactions such as the Curtius degradation.157161 Clearly, cular, its designed use, is gaining slow recognition within drug
medicinal chemists must continue to investigate and embrace new discovery. Indeed there were only two closely related reports in our
enabling technologies, while equipment manufacturers and the data set of such an interaction, and even this was not explicitly
academic community need to make new technologies readily referred to as a halogen bonding interaction. Greater awareness of
accessible to the inexperienced user. this potential interaction, along with the development of improved
Medicinal chemistry is often perceived of as the synthesis of methods for the incorporation of halogens (particularly fluorine, as
planar (“flat”) aromatic systems lacking chirality. However, the data highlighted in section 4.8) within the academic community, may
set reveals a surprisingly large number of chiral compounds, often allow for the design of improved molecules, in which a halogen
containing multiple chiral centers, with an average of ∼1 defined serves in place of an interacting aromatic substituent, often an
stereocenter per molecule, taken across the entire data set. The electron-rich group that introduces metabolic liabilities, reducing
perception, thus, that medicinal chemistry is entirely “flat” appears metabolic liability a priori instead of ameliorating it later with
misguided, although it is true that aromatic systems are almost minimal potency benefit.
universally present, the likely source of this preconception. Further- The problems of drug attrition during clinical trials have been
more, an assessment of the degree of unsaturation using the Fsp3 well documented.162164 While the continued use of molecular
parameter12 shows that between one-quarter and one-half of carbon fragments known to be “safe” (i.e., found in approved drugs) is
atoms are sp3-hybridized. Despite this, Lovering’s 2009 analysis understandable, the exploration of novel areas of chemical space,
suggests that the incorporation of a greater proportion of chiral in which the current reliance in particular on amines, amides, and
compounds with higher degrees of unsaturation may improve aromatic rings, may be biasing the chemical output into areas of
clinical outcomes; Lovering’s Fsp3 parameter provides a simple space prone to development problems.
way of assessing this for any given structure.12 As chiral substituents Finally, having considered some of the common preconcep-
are in the minority, it would seem that the routine inclusion of a tions, it is interesting to consider what this data set suggests is a
greater range of chiral substituents may be a simple and achievable “typical” medicinal chemistry compound and synthesis. In terms
means of addressing this, while consideration should also be given to of physicochemical properties, the analysis suggests a molecular
modification of aromatic core templates to saturated or partially weight in the range 350550, with four to six HBAs and one to
saturated systems, containing chiral centers at the point(s) were two HBDs, NRot in the range 68, and clogP between 3.5 and 5.5.
substituents are attached. TPSA has a broader distribution but typically is 6090 Å2. The
As noted in a recent publication,27 the extent of coverage of compound will possess one to two chiral centers, with 3050% of C
heterocycle space is surprisingly limited, although it is clear in this atoms in the sp3 hybridization state, and contain a biaryl bond
analysis that different companies have favored alternative between a fused aromatic system and a second ring (with one of the
3473 dx.doi.org/10.1021/jm200187y |J. Med. Chem. 2011, 54, 3451–3479
Journal of Medicinal Chemistry PERSPECTIVE
journal/jmcmar; BMC, http://journals.elsevier.com/09680896/bioor- (27) Pitt, W. R.; Parry, D. M.; Perry, B. G.; Groom, C. R. Hetero-
ganic-and-medicinal-chemistry/; BMCL, http://www.elsevier.com/ aromatic rings of the future. J. Med. Chem. 2009, 52, 2952–2963.
wps/find/journaldescription.cws_home/972/description#description) (28) Abdel-Magid, A. F.; Mehrman, S. J. A review on the use of
at the time of data compilation. sodium triacetoxyborohydride in the reductive amination of ketones and
(4) Cooper, T. W. J.; Campbell, I. B.; Macdonald, S. J. F. Factors aldehydes. Org. Process Res. Dev 2006, 10, 971–1031.
determining the selection of organic reactions by medicinal chemists and (29) Baxter, E. W.; Reitz, A. B. Reductive aminations of carbonyl
the use of these reactions in arrays (small focussed libraries). Angew. compounds with borohydride and borane reducing agents. Org. React.
Chem., Int. Ed. 2010, 49, 2–12. 2002, 59, 1–714.
(5) CAS SciFinder. http://www.cas.org/products/scifindr/index. (30) Makosza, M. Nucleophilic substitution of hydrogen in electron-
html (accessed Feb 7, 2011). deficient arenes - a general process of great practical value. Chem. Soc.
(6) Reaxys Elsevier Beilstein Database. http://www.reaxys.com/ Rev. 2010, 39, 2855–2868.
info (accessed Feb 7, 2011). (31) Van der Plas, H. C. The SN(ANRORC) mechanism: a new
(7) Lipinksi, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. mechanism for nucleophilic substitution. Acc. Chem. Res. 1978,
Experimental and computational approaches to estimate solubility and 11, 462–468.
permeability in drug discovery and development settings. Adv. Drug (32) Rossi, R. A.; Pierini, A. B.; Santiago, A. N. Aromatic Substitution
Delivery Rev. 2001, 46, 3–26. by the SRN1 Reaction. Org. React. 2004, 1–271.
(8) Lipinksi, C. A. Lead- and drug-like compounds: the rule-of-five (33) Hartwig, J. F. Carbonheteroatom bond-forming reductive
revolution. Drug Discovery Today: Technol. 2004, 1, 337–341. eliminations of amines, ethers, and sulfides. Acc. Chem. Res. 1998,
(9) log P is calculated in RDKit using the atomic contribution 31, 852–860.
algorithm of Crippen: Wildman, S. A.; Crippen, G. M. Prediction of (34) Hartwig, J. F. Transition metal catalyzed synthesis of arylamines
physicochemical parameters by atomic contributions. J. Chem. Inf. and aryl ethers from aryl halides and triflates: scope and mechanism.
Comput. Sci. 1999, 39, 868–873. Angew. Chem., Int. Ed. 1998, 37, 2046–267.
(10) Veber, D. F.; Johnson, S. R.; Cheng, H.-Y.; Smith, B. R.; Ward, (35) Hartwig, J. F. Carbonheteroatom bond formation catalysed
K. W.; Kopple, K. D. Molecular properties that influence the oral by organometallic complexes. Nature 2008, 455, 314–322.
bioavailability of drug candidates. J. Med. Chem. 2002, 45, 2615–2623. (36) Surry, D. S.; Buchwald, S. L. Biaryl phosphane ligands in palla-
(11) Ertl, P.; Rohde, B.; Selzer, P. Fast calculation of molecular polar dium-catalyzed amination. Angew. Chem., Int. Ed. 2008, 47, 6338–6361.
surface area as a sum of fragment-based contributions and its application (37) Wolfe, J. P.; Wagaw, S.; Marcoux, J.-F.; Buchwald, S. L. Rational
to the prediction of drug transport properties. J. Med. Chem. 2000, development of practical catalysts for aromatic carbonnitrogen bond
43, 3714–3717. formation. Acc. Chem. Res. 1998, 31, 805–818.
(12) Lovering, F.; Bikker, J.; Humblet, C. Escape from Flatland: (38) Wolfe, J. P.; Buchwald, S. L. Palladium-Catalyzed Amination of
increasing saturation as an approach to improving clinical success. J. Med. Aryl Halides and Aryl Triflates: N-Hexyl-2-methyl-4-methoxyaniline
Chem. 2009, 52, 6752–6756. and N-Methyl-N-(4-chlorophenyl)aniline. Organic Syntheses; Wiley &
(13) Python. http://www.python.org (accessed Mar 19, 2010). Sons: New York, 2004; Collect. Vol. 10, p 423.
(14) RDKit: Open-Source Cheminformatics. http://www.rdkit.org (39) Yang, B. H.; Buchwald, S. L. Palladium-catalyzed amination of
(accessed Apr 25, 2010). aryl halides and sulfonates. J. Organomet. Chem. 1999, 576, 125–146.
(15) SMARTS—A Language for Describing Molecular Patterns. (40) Corbet, J.-P.; Mignani, G. Selected patented cross-coupling
http://www.daylight.com/dayhtml/doc/theory/index.pdf (accessed reaction technologies. Chem. Rev. 2006, 106, 2651–2710.
Jan 20, 2011). (41) Hughes, D. L. The Mitsunobu reaction. Org. React. 1992,
(16) Schomburg, K.; Ehrlich, H. C.; Stierand, K.; Rarey, M. From 42, 335–656.
structure diagrams to visual chemical patterns. J. Chem. Inf. Model. 2010, (42) Mitsunobu, O. The use of diethyl azodicarboxylate and triphe-
50, 1529–1535. nylphoshine in synthesis and transformation of natural products. Synth-
(17) SMARTSViewer. http://smartsview.zbh.uni-hamburg.de/ esis 1981, 1–28.
(accessed Jan 20, 2011). (43) Han, S. Y.; Kim, Y.-A. Recent development of peptide couping
(18) Ritchie, T. J.; Macdonald, S. J. F. The impact of aromatic ring reagents in organic synthesis. Tetrahedron 2004, 60, 2447–2467.
count on compound developability: Are too many aromatic rings a (44) Montalbetti, C. A. G. N.; Falque, V. Amide bond formation and
liability in drug design? Drug Discovery Today 2009, 14, 1011–1020. peptide coupling. Tetrahedron 2005, 61, 10827–10852.
(19) Berthold, M. R.; Cebron, N.; Dill, F.; Gabriel, T. R.; K€otter, T.; (45) Ettmayer, P.; Amidon, G. L.; Clement, B.; Testa, B. Lessons
Meinl, T.; Ohl, P.; Sieb, C.; Thiel, K.; Wiswedel, B. KNIME: The learned from marketed and investigational prodrugs. J. Med. Chem. 2004,
Konstanz Information Miner. In Studies in Classification, Data Analysis, 47, 2393–2404.
and Knowledge Organization; Springer: Berlin, Germany, 2007. (46) Rautio, J.; Kumpulainen, H.; Heimbach, T.; Oliyai, R.; Oh, D.;
(20) KNIME: The Konstanz Information Miner. http://www. J€arvinen, T.; Savolainen, J. Prodrugs: design and clinical applications.
knime.org (accessed Jan 20, 2011). Nat. Rev. Drug Discovery 2008, 7, 255–270.
(21) Baran, P. S.; Maimone, T. J.; Richter, J. M. Total synthesis of (47) Accelrys (formerly Symyx) Available Chemicals Directory,
marine natural products without using protecting groups. Nature 2007, version 2010.08. http://accelrys.com/products/databases/sourcing/
446, 404–408. available-chemicals-directory.html (accessed Feb 2, 2011).
(22) Gaich, T.; Baran, P. S. Aiming for the ideal synthesis. J. Org. (48) Kotha, S. Recent applications of the SuzukiMiyaura cross-
Chem. 2010, 75, 4657–4673. coupling reaction in organic synthesis. Tetrahedron 2002, 58, 9633–9695.
(23) Shenvi, R. A.; O’Malley, D. P.; Baran, P. S. Chemoselectivity: (49) Miyaura, N.; Suzuki, A. Palladium-catalyzed cross-coupling
the mother of invention in total synthesis. Acc. Chem. Res. 2009, reactions of organoboron compounds. Chem. Rev. 1995, 95, 2457–2483.
42, 530–541. (50) Stanforth, S. P. Catalytic cross-coupling reactions in biaryl
(24) Young, I. S.; Baran, P. S. Protecting-group-free synthesis as an synthesis. Tetrahedron 1998, 54, 263–303.
opportunity for invention. Nat. Chem. 2009, 1, 193–205. (51) Many boronic acids and related derivatives are stable solids that
(25) Ertl, P.; Jelfs, S.; M€uhlbacher, J.; Schuffenhauer, A.; Selzer, P. can be easily weighed and stored, often for periods of years without
Quest for the rings. In silico exploration of ring universe to identify serious degradation, in contrast with the organozinc and Grignard
novel bioactive heteroaromatic scaffolds. J. Med. Chem. 2006, 49, reagents of the Negishi and Kumada couplings.
4568–4573. (52) Compare with the Stille reaction of organostannanes, which are
(26) Bemis, G. W.; Murcko, M. A. The properties of known drugs. 1. highly toxic, often significantly volatile, and generate toxic tin-based
Molecular frameworks. J. Med. Chem. 1996, 39, 2887–2893. byproducts that can be troublesome to separate from the desired
product and represent a serious environmental hazard in their storage arrayable syntheses to accelerate SAR studies of pyridopyrimidinone and
and disposal. pyrimidopyrimidinone. Tetrahedron Lett. 2009, 50, 370–372.
(53) Negishi, E.-I. A genealogy of Pd-catalyzed cross-coupling. (77) Yan, H.; Boehm, J. C.; Jin, Q.; Kasparec, J.; Li, H.; Zhu, C.;
J. Organomet. Chem. 2002, 653, 34–40. Widdowson, K. L.; Callahan, J. F.; Wan, Z. An improved and highly
(54) Negishi, E.-I. Palladium-catalyzed alkynylation. Chem. Rev. convergent synthesis of 4-substituted-pyrido[2.3-d]pyrimidin-7-ones.
2003, 103, 1979–2017. Tetrahedron Lett. 2007, 48, 1205–1207.
(55) Sonogashira, K. Development of Pd-Cu catalyzed cross-cou- (78) Burns, N. Z.; Baran, P. S.; Hoffman, R. W. Redox economy in
pling of terminal acetylenes with sp2-carbon halides. J. Organomet. Chem. organic synthesis. Angew. Chem., Int. Ed. 2009, 48, 2854–2867.
2002, 653, 46–49. (79) Baldwin, I.; Bamborough, P.; Haslam, C. G.; Hunjan, S. S.;
(56) Brammertz, A.; Augthun, M. Toxicity of palladium in animals. Longstaff, T.; Mooney, C. J.; Patel, S.; Quinn, J.; Somers, D. O. Kinase
Wiss. Umwelt 1992, 285–289. array design, back to front: biaryl amides. Bioorg. Med. Chem. Lett. 2008,
(57) Iavicoli, I.; Bocca, B.; Fontana, L.; Caimi, S.; Bergamaschi, A.; 18, 5285–5289.
Alimonti, A. Distribution and elimination of palladium in rats after (80) The authors have heard various figures mentioned in confer-
90-day oral administration. Toxicol. Ind. Health 2010, 26, 183–189. ence presentations and other discussions but never with supporting data
(58) Kielhorn, J.; Melber, C.; Keller, D.; Mangelsdorf, I. Palladium, a or citations.
review of exposure and effects to human health. Int. J. Hyg. Environ. (81) Ritchie, T. J.; Macdonald, S. J. F.; Young, R. J.; Pickett, S. D. The
Health 2002, 205, 417–432. impact of aromatic ring count on compound developability: further
(59) Nahm, S.; Weinreb, S. M. N-Methoxy-N-methylamides as insights by examining carbo- and hetero-aromatic and -aliphatic ring
effective acylating agents. Tetrahedron Lett. 1981, 22, 3815–3818. types. Drug Discovery Today 2011, 16, 164–171.
(60) Labeeuw, O.; Phansavath, P.; Gen^et, J.-P. Synthesis of modified (82) Jaeckel, H.; Klein, W. Prediction of mammalian toxicity by
Weinreb amides: N-tert-butoxy-N-methylamides as effective acylating quantitative structureactivity relationships: aliphatic amines and ani-
agents. Tetrahedron Lett. 2004, 45, 7107–7110. lines. Quant. Struct.Act. Relat. 1991, 10, 198–204.
(61) Davies, S. G.; Goodwin, C. J.; Hepworth, D.; Roberts, P. M.; (83) Karki, S. B.; Dinnocenzo, J. P.; Jones, J. P.; Korzekwa, K. R.
Thomson, J. E. On the origins of diastereoselectivity in the alkylation of Mechanism of oxidative amine dealkylation of substituted N,N-dimethy-
enolates derived from N-1-(10 -naphthyl)ethyl-O-tert-butylhydroxa- lanilines by cytochrome P-450: application of isotope effect profiles.
mates: chiral Weinreb amide equivalents. J. Org. Chem. 2010, J. Am. Chem. Soc. 1995, 117, 3657–3664.
75, 1214–1227. (84) Zok, S.; G€orge, G.; Kalsch, W.; Nagel, R. Bioconcentration,
(62) Boutagy, J.; Thomas, R. Olefin synthesis with organic phos- metabolism and toxicity of substituted anilines in the zebrafish
phonate carbanions. Chem. Rev. 1974, 74, 87–99. (Brachydanio rerio). Sci. Total Environ. 1991, 109110, 411–421.
(63) Maercker, A. The Wittig reaction. Org. React. 1965, 14, (85) Ghose, A. K.; Herbertz, T.; Salvino, J. M.; Mallamo, J. P.
270–490. Knowledge-based chemoinformatic approaches to drug discovery. Drug
(64) Maryanoff, B. E.; Reitz, A. B. The Wittig olefination reaction Discovery Today 2006, 11, 1107–1114.
and modifications involving phosphoryl-stabilised carbanions. Stereo- (86) Akritopoulou-Zanze, I.; Hajduk, P. J. Kinase-targeted libraries:
chemistry, mechanism and selected synthetic aspects. Chem. Rev. 1989, the design and synthesis of novel, potent, and selective kinase inhibitors.
89, 863–927. Drug Discovery Today 2009, 14, 291–297.
(65) Cowden, C. J.; Paterson, I. Asymmetric aldol reactions using (87) Brown, A.; Brown, L.; Brown, T. B.; Calabrese, A.; Ellis, D.;
boron. Org. React. 1997, 51, 1–200. Puhalo, N.; Smith, C. R.; Wallace, O.; Watson, L. Triazole oxytocin
(66) Guillena, G.; Najera, C.; Ramon, D. J. Enantioselective direct antagonists: identification of aryl ether replacements for a biaryl
aldol reaction: the blossoming of modern organocatalysis. Tetrahedron: substituent. Bioorg. Med. Chem. Lett. 2008, 18, 5242–5244.
Asymmetry 2007, 18, 2249–2293. (88) Patel, S. D.; Habeski, W. M.; Min, H.; Zhang, J.; Roof, R.;
(67) Mukaiyama, T. The directed aldol reaction. Org. React. Snyder, B.; Bora, G.; Campbell, B.; Li, C.; Hidayetoglu, D.; Johnson,
1982, 203–331. D. S.; Chaudhry, A.; Charlton, M. E.; Kablaoui, N. M. Identification
(68) Palomo, C.; Oiarbide, M.; Garcia, J. M. The aldol addition reaction: and SAR around N-{2-[4-(2,3-dihydro-benzo[1,4]dioxin-2-ylmethyl)-
and old transformation at constant rebirth. Chem.—Eur. J. 2002, 8, 36–44. [1,4]diazepan-1-yl]-ethyl}-2-phenoxy-nicotinamide, a selective R2C adre-
(69) Schetter, B.; Mahrwald, R. Modern aldol methods for the total nergic receptor antagonist. Bioorg. Med. Chem. Lett. 2008, 18, 5689–5693.
synthesis of polyketides. Angew. Chem., Int. Ed. 2006, 45, 7506–7525. (89) Fish, P. V.; Deur, C.; Gan, X.; Greene, K.; Hoople, D.;
(70) Trost, B. M.; Brindle, C. S. The direct catalytic asymmetric aldol MacKenny, M.; Para, K. S.; Reeves, K.; Ryckmans, T.; Stiff, C.; Stobie,
reaction. Chem. Soc. Rev. 2010, 39, 1600–1632. A.; Wakenhut, F.; Whitlock, G. A. Design and synthesis of morpholine
(71) Walters, I.; Austin, C.; Austin, R.; Bonnert, R.; Cage, P.; derivatives. SAR for dual serotonin & noradrenaline reuptake inhibition.
Christie, M.; Ebden, M.; Gardiner, S.; Grahames, C.; Hill, S.; Hunt, Bioorg. Med. Chem. Lett. 2008, 18, 2562–2566.
F.; Jewell, R.; Lewis, S.; Martin, I.; Nicholls, D.; Robinson, D. Evaluation (90) Xu, W.; Gray, D. L.; Glase, S. A.; Barta, N. S. Design and
of a series of bicyclic CXCR2 antagonists. Bioorg. Med. Chem. Lett. 2008, synthesis of reboxetine analogs morpholine derivatives as selective
18, 798–803. norepinephrine reuptake inhibitors. Bioorg. Med. Chem. Lett. 2008,
(72) Bardelle, C.; Coleman, T.; Cross, D.; Davenport, S.; Kettle, 18, 5550–5553.
J. G.; Ko, E. J.; Leach, A. G.; Mortlock, A.; Read, J.; Roberts, N. J.; Robins, (91) Whitlock, G. A.; Blagg, J.; Fish, P. V. 1-(2-Phenoxyphe-
P.; Williams, E. J. Inhibitors of the tyrosine kinase EphB4. Part 2. nyl)methanamines: SAR for dual serotonin/noradrenaline reuptake
Structure-based discovery and optimization of 3,5-bis substituted anili- inhibition, metabolic stability and hERG affinity. Bioorg. Med. Chem.
nopyrimidines. Bioorg. Med. Chem. Lett. 2008, 18, 5717–5721. Lett. 2008, 18, 596–599.
(73) Deaton, D. N.; Gao, E. N.; Graham, K. P.; Gross, J. W.; Miller, (92) Middleton, D. S.; Andrews, M.; Glossop, P.; Gymer, G.; Hepworth,
A. B.; Strelow, J. M. Thiol-based angiotensin-converting enzyme 2 D.; Jessiman, A.; Johnson, P. S.; MacKenny, M.; Pitcher, M. J.; Rooker,
inhibitors: P1 modifications for the exploration of the S1 subsite. Bioorg. T.; Stobie, A.; Tang, K.; Morgan, P. Designing rapid onset selective
Med. Chem. Lett. 2008, 18, 732–737. serotonin re-uptake inhibitors. 2: Structureactivity relationships of
(74) Jiang, C.; Chang, J.-Y. Unfolding and breakdown of insulin in the substituted (aryl)benzylamines. Bioorg. Med. Chem. Lett. 2008, 18,
presence of endogenous thiols. FEBS Lett. 2005, 579, 3927–3931. 4018–4021.
(75) Heck, R. F. Palladium-catalyzed vinylation of organic halides. (93) Mente, S.; Gallaschun, R.; Schmidt, A.; Lebel, L.; Vanase-
Org. React. 1982, 27, 345–390. Frawley, M.; Fliri, A. Quantitative structureactivity relationship of
(76) Wan, Z.; Yan, H.; Hall, R. F.; Lin, X.; Livia, S.; Respondek, T.; phenoxyphenyl-methanamine compounds with 5HT2A, SERT and
Widdowson, K. L.; Zhu, C.; Callahan, J. F. Design and development of hERG activities. Bioorg. Med. Chem. Lett. 2008, 18, 6088–6092.
(94) Fish, P. V.; Ryckmans, T.; Stobie, A.; Wakenhut, F. Structureactivity study of 2,3-benzodiazepin-4-ones noncompetitive
[4-(Phenoxy)pyridin-3-yl]methylamines: a new class of selective nora- AMPAR antagonists: identification of the 1-(4-amino-3-methylphenyl)-
drenaline reuptake inhibitors. Bioorg. Med. Chem. Lett. 2008, 3,5-dihydro-7,8-ethylenedioxy-4H-2,3-benzodiazepin-4-one as neuro-
18, 1795–1798. protective agent. Bioorg. Med. Chem. 2008, 16, 2200–2211.
(95) Whitlock, G. A.; Fish, P. V.; Fary, M. J.; Stobie, A.; Wakenhut, F. (106) Tian, X.; Switzer, A. G.; Derose, S. A.; Mishra, R. K.; Solinsky,
Pyridyl-phenyl ether monoamine reuptake inhibitors: impact of lipo- M. G.; Mumin, R. N.; Ebetino, F. H.; Jayasinghe, L. R.; Webster, M. E.;
philicity on dual SNRI pharmacology and off-target promiscuity. Bioorg. Colson, A. O.; Crossdoersen, D.; Pinney, B. B.; Farmer, J. A.; Dowty,
Med. Chem. Lett. 2008, 18, 2896–2899. M. E.; Obringer, C. M.; Cruze, C. A.; Burklow, M. L.; Suchanek, P. M.;
(96) Pfefferkorn, J. A.; Choi, C.; Winters, T.; Kennedy, R.; Chi, L.; Dong, L.; Dirr, M. K.; Sheldon, R. J.; Wos, J. A. Discovery of orally
Perrin, L. A.; Lu, G.; Ping, Y.-W.; McClanahan, T.; Schroeder, R.; bioavailable 1,3,4-trisubstituted 2-oxopiperazine-based melanocortin-4
Leininger, M. T.; Geyer, A.; Schefzick, S.; Atherton, J. P2Y1 receptor receptor agonists as potential antiobesity agents. J. Med. Chem. 2008,
antagonists as novel antithrombotic agents. Bioorg. Med. Chem. Lett. 51, 6055–6066.
2008, 18, 3338–3343. (107) Kazmierski, W. M.; Aquino, C.; Chauder, B. A.; Deanda, F.;
(97) Chen, Y. L.; Braselton, J.; Forman, J.; Gallaschun, R.; Man- Ferris, R.; Jones-Hertzog, D. K.; Kenakin, T.; Koble, C. S.; Watson, C.;
sbach, R.; Schmidt, A. W.; Seeger, T. F.; Sprouse, J. S.; Tingley, F. D.; Wheelan, P.; Yang, H.; Youngman, M. Discovery of bioavailable 4,4-
Winston, E.; Schulz, D. W. Synthesis and SAR of 2-aryloxy-4-alkoxy- disubstituted piperidines as potent ligands of the chemokine receptor 5
pyridines as potent orally active corticotropin-releasing factor 1 receptor and inhibitors of the human immunodeficiency virus-1. J. Med. Chem.
antagonists. J. Med. Chem. 2008, 51, 1377–1384. 2008, 51, 6538–6546.
(98) Barlaam, B.; Ballard, P.; Bradbury, R. H.; Ducray, R.; Germaine, (108) Liddle, J.; Allen, M. J.; Borthwick, A. D.; Brooks, D. P.; Davies,
H.; Hickinson, D. M.; Hudson, K.; Kettle, J. G.; Klinowska, T.; Magnien, D. E.; Edwards, R. M.; Exall, A. M.; Hamlett, C.; Irving, W. R.; Mason,
F.; Ogilvie, D. J.; Olivier, A.; Pearson, S. E.; Scott, J. S.; Suleman, A.; A. M.; McCafferty, G. P.; Nerozzi, F.; Peace, S.; Philp, J.; Pollard, D.;
Trigwell, C. B.; Vautier, M.; Whittaker, R. D.; Wood, R. A new series of Pullen, M. A.; Shabbir, S. S.; Sollis, S. L.; Westfall, T. D.; Woollard, P. M.;
neutral 5-substituted 4-anilinoquinazolines as potent, orally active Wu, C.; Hickey, D. M. B. The discovery of GSK221149A: a potent and
inhibitors of erbB2 receptor tyrosine kinase. Bioorg. Med. Chem. Lett. selective oxytocin antagonist. Bioorg. Med. Chem. Lett. 2008, 18, 90–94.
2008, 18, 674–678. (109) Harriman, G. C.; Brewer, M.; Bennett, R.; Kuhn, C.; Bazin,
(99) Fang, J.; Akwabi-Ameyaw, A.; Britton, J. E.; Katamreddy, S. R.; M.; Larosa, G.; Skerker, P.; Cochran, N.; Gallant, D.; Baxter, D.;
Navas, F.; Miller, A. B.; Williams, S. P.; Gray, D. W.; Orband-Miller, L. A. Picarella, D.; Jaffee, B.; Luly, J. R.; Briskin, M. J. Selective cell adhesion
Synthesis of 3-alkyl naphthalenes as novel estrogen receptor ligands. inhibitors: barbituric acid based R4β7—MAdCAM inhibitors. Bioorg.
Bioorg. Med. Chem. Lett. 2008, 18, 5075–5077. Med. Chem. Lett. 2008, 18, 2509–2512.
(100) Micheli, F.; Bonanomi, G.; Braggio, S.; Capelli, A. M.; Celes- (110) Clarke, B.; Demont, E.; Dingwall, C.; Dunsdon, R.; Faller, A.;
tini, P.; Damiani, F.; Di Fabio, R.; Donati, D.; Gagliardi, S.; Gentile, G.; Hawkins, J.; Hussain, I.; MacPherson, D.; Maile, G.; Matico, R.; Milner,
Hamprecht, D.; Petrone, M.; Radaelli, S.; Tedesco, G.; Terreni, S.; P.; Mosley, J.; Naylor, A.; O’Brien, A.; Redshaw, S.; Riddell, D.; Row-
Worby, A.; Heidbreder, C. New fused benzazepine as selective D3 land, P.; Soleil, V.; Smith, K. J.; Stanway, S.; Stemp, G.; Sweitzer, S.;
receptor antagonists. Synthesis and biological evaluation. Part one: [h]- Theobald, P.; Vesey, D.; Walter, D. S.; Ward, J.; Wayne, G. BACE-1
fused tricyclic systems. Bioorg. Med. Chem. Lett. 2008, 18, 901–907. inhibitors part 1: identification of novel hydroxy ethylamines (HEAs).
(101) Micheli, F.; Bonanomi, G.; Braggio, S.; Capelli, A. M.; Damiani, Bioorg. Med. Chem. Lett. 2008, 18, 1011–1016.
F.; Di Fabio, R.; Donati, D.; Gentile, G.; Hamprecht, D.; Perini, O.; (111) Clarke, B.; Demont, E.; Dingwall, C.; Dunsdon, R.; Faller, A.;
Petrone, M.; Tedesco, G.; Terreni, S.; Worby, A.; Heidbreder, C. New Hawkins, J.; Hussain, I.; MacPherson, D.; Maile, G.; Matico, R.; Milner,
fused benzazepine as selective D3 receptor antagonists. Synthesis and P.; Mosley, J.; Naylor, A.; O’Brien, A.; Redshaw, S.; Riddell, D.; Row-
biological evaluation. Part 2: [g]-Fused and hetero-fused systems. Bioorg. land, P.; Soleil, V.; Smith, K. J.; Stanway, S.; Stemp, G.; Sweitzer, S.;
Med. Chem. Lett. 2008, 18, 908–912. Theobald, P.; Vesey, D.; Walter, D. S.; Ward, J.; Wayne, G. BACE-1
(102) Young, R. J.; Borthwick, A. D.; Brown, D.; Burns-Kurtis, C. L.; inhibitors part 2: identification of hydroxy ethylamines (HEAs)
Campbell, M.; Chan, C.; Charbaut, M.; Chung, C.-W.; Convery, M. A.; with reduced peptidic character. Bioorg. Med. Chem. Lett. 2008,
Kelly, H. A.; King, N. P.; Kleanthous, S.; Mason, A. M.; Pateman, A. J.; 18, 1017–1021.
Patikis, A. N.; Pinto, I. L.; Pollard, D. R.; Senger, S.; Shah, G. P.; Toomey, (112) Serafinowska, H. T.; Blaney, F. E.; Lovell, P. J.; Merlo, G. G.;
J. R.; Watson, N. S.; Weston, H. E. Structure and property based design Scott, C. M.; Smith, P. W.; Starr, K. R.; Watson, J. M. Novel 5-HT1A/
of factor Xa inhibitors: pyrrolidin-2-ones with biaryl P4 motifs. Bioorg. 1B/1D receptors antagonists with potent 5-HT reuptake inhibitory
Med. Chem. Lett. 2008, 18, 23–27. activity. Bioorg. Med. Chem. Lett. 2008, 18, 5581–5585.
(103) Young, R. J.; Borthwick, A. D.; Brown, D.; Burns-Kurtis, C. L.; (113) Bromidge, S. M.; Bertani, B.; Borriello, M.; Faedo, S.; Gordon,
Campbell, M.; Chan, C.; Charbaut, M.; Convery, M. A.; Diallo, H.; L. J.; Granci, E.; Hill, M.; Marshall, H. R.; Stasi, L. P.; Zucchelli, V.;
Hortense, E.; Irving, W. R.; Kelly, H. A.; King, N. P.; Kleanthous, S.; Merlo, G.; Vesentini, A.; Watson, J. M.; Zonzini, L. 6-[2-(4-Aryl-1-
Mason, A. M.; Pateman, A. J.; Patikis, A. N.; Pinto, I. L.; Pollard, D. R.; piperazinyl)ethyl]-2H-1,4-benzoxazin-3(4H)-ones: dual-acting 5-HT1
Senger, S.; Shah, G. P.; Toomey, J. R.; Watson, N. S.; Weston, H. E.; receptor antagonists and serotonin reuptake inhibitors. Bioorg. Med.
Zhou, P. Structure and property based design of factor Xa inhibitors: Chem. Lett. 2008, 18, 5653–5656.
biaryl pyrrolidin-2-ones incorporating basic heterocyclic motifs. Bioorg. (114) With 23 heteroaromatic rings and benzene exemplified within
Med. Chem. Lett. 2008, 18, 28–33. the data set, even without the possibility of different fusion bond isomers,
(104) Walker, D. P.; Bi, C. F.; Kalgutkar, A. S.; Bauman, J. N.; Zhao, there are 24 24 = 576 possible ring types (comprising 64 6,6-fused, 256
S. X.; Soglia, J. R.; Aspnes, G. E.; Kung, D. W.; Klug-McLeod, J.; 5,5-fused, and 128 6,5-fused systems). Fusion isomers increase this figure
Zawistoski, M. P.; McGlynn, M. A.; Oliver, R.; Dunn, M.; Li, J. C.; dramatically (for example, the relatively symmetrical pyridine and pyrrole
Richter, D. T.; Cooper, B. A.; Kath, J. C.; Hulford, C. A.; Autry, C. L.; pairing can generate seven different fusion isomers).
Luzzio, M. J.; Ung, E. J.; Roberts, W. G.; Bonnette, P. C.; Buckbinder, L.; (115) Lewell, X. Q.; Jones, A. C.; Bruce, C. L.; Harper, G.; Jones,
Mistry, A.; Griffor, M. C.; Han, S.; Guzman-Perez, A. Trifluoromethyl- M. M.; Mclay, I. M.; Bradshaw, J. Drug rings database with Web
pyrimidine-based inhibitors of proline-rich tyrosine kinase 2 (PYK2): interface. A tool for identifying alternative chemical rings in lead
structureactivity relationships and strategies for the elimination of discovery programs. J. Med. Chem. 2003, 46, 3257–3274.
reactive metabolite formation. Bioorg. Med. Chem. Lett. 2008, (116) Southall, N. T.; Ajay Kinase patent space visualisation using
18, 6071–6077. chemical replacements. J. Med. Chem. 2006, 49, 2103–2109.
(105) Micale, N.; Colleoni, S.; Postorino, G.; Pellicano, A.; Zappala, (117) Jin, J.; Wang, Y.; Shi, D.; Wang, F.; Davis, R. S.; Jin, Q.; Fu, W.;
M.; Lazzaro, J.; Diana, V.; Cagnotto, A.; Mennini, T.; Grasso, S. Foley, J. J.; Webb, E. F.; Dehaas, C. J.; Berlanga, M.; Burman, M.; Sarau,
H. M.; Morrow, D. M.; Rao, P.; Kallal, L. A.; Moore, M. L.; Rivero, R. A.; MEK kinase inhibitors be utilized to synthesize novel type III NCKIs?
Palovich, M.; Salmon, M.; Belmonte, K. E.; Busch-Petersen, J. Discovery Does the MEK-pocket exist in kinases other than MEK? Bioorg Med.
of novel and long acting muscarinic acetylcholine receptor antagonists. Chem. Lett. 2009, 19, 226–229.
J. Med. Chem. 2008, 51, 4866–4869. (136) Note that this figure, being almost twice the total number of
(118) Kung, P. P.; Funk, L.; Meng, J.; Collins, M.; Zhou, J. Z.; reactions (7315) in our data set, reflects the general approach of building
Johnson, M. C.; Ekker, A.; Wang, J.; Mehta, P.; Yin, M. J.; Rodgers, C.; a common scaffold, which is subsequently used to prepare multiple
Davies, J. F., II; Bayman, E.; Smeal, T.; Maegley, K. A.; Gehring, M. R. analogues.
Dihydroxylphenyl amides as inhibitors of the Hsp90 molecular chaper- (137) Manual inspection revealed a further 339 compounds contain-
one. Bioorg. Med. Chem. Lett. 2008, 18, 6273–6278. ing 1 (315 compounds) or 2 (34 compounds) undefined or only
(119) Deaton, D. N.; Graham, K. P.; Gross, J. W.; Miller, A. B. Thiol- partially defined stereocenters, a total of 383 additional stereocenters.
based angiotensin-converting enzyme 2 inhibitors: P10 modifications for Many of these were the result of the addition of simple substituents (Me,
the exploration of the S10 subsite. Bioorg. Med. Chem. Lett. 2008, OH, OMe, etc.) to prochiral centers or from a small number of scaffolds
18, 1681–1687. with a racemic center. Presumably, the vast majority of these were not
(120) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Fluorine deemed of sufficient interest to invest in asymmetric synthesis or
in medicinal chemistry. Chem. Soc. Rev. 2008, 37, 320–330. resolution.
(121) O’Hagan, D. Understanding organofluorine chemistry. An (138) While Lipinski’s rules relate to oral absorption, in most cases
introduction to the CF bond. Chem. Soc. Rev. 2008, 37, 308–319. no intended dose route was identified in many of the papers examined.
(122) Hodgetts, K. J.; Combs, K. J.; Elder, A. M.; Harriman, G. C. We assume for the purposes of this analysis that oral dosing would be the
The role of fluorine in the discovery and optimization of CNS agents: desired “gold standard”.
modulation of drug-like properties. Annu. Rep. Med. Chem. 2010, (139) Lipinski notes in ref 7 the importance of their goal of
45, 429–448. generating simple parameters that are easily accessible to medicinal
(123) Bissantz, C.; Kuhn, B.; Stahl, M. A medicinal chemist’s guide chemists using their pattern recognition skills.
to molecular interactions. J. Med. Chem. 2010, 53, 5061–5084. (140) Giaginis, C.; Zira, A.; Theocharis, S.; Tsantili-Kakoulidou, A.
(124) Zhu, L.; Li, Y.; Zhao, Y.; Hu, J. Nucleophilic (phenylsulfonyl)- Simple physicochemical properties as effective filters for risk estimation
difluoromethylation of alkyl halides using PhSO2CF2SiMe3: preparation of drug transport across the human placental barrier. Epitheor. Klin.
of gem-difluoroalkenes and trifluoromethyl compounds. Tetrahedron Farmakol. Farmakokinet., Int. Ed. 2008, 22, 146–148.
Lett. 2010, 51, 6150–6152. (141) Choy, Y.-B.; Prausnitz, M. R. The rule of five for non-oral
(125) Baumann, M.; Baxendale, I. R.; Martin, L. J.; Ley, S. V. routes of drug delivery: ophthalmic, inhalation and transdermal. Pharm.
Development of fluorination methods using continuous-flow micro- Res. 2011, 28, 943–948.
reactors. Tetrahedron 2009, 65, 6611–6625. (142) Lewis, D. F. V.; Jacobs, M. N.; Dickins, M. Compound
(126) Hagooly, Y.; Cohen, O.; Rozen, S. A general route for lipophilicity for substrate binding to human P450s in drug metabolism.
constructing difluoromethyl ethers. Tetrahedron Lett. 2009, 50, Drug Discovery Today 2004, 9, 530–537.
392–394. (143) Waring, M. J.; Johnstone, C. A quantitative assessment of
(127) Auffinger, P.; Hays, F. A.; Westhof, E.; Ho, P. S. Halogen hERG liability as a function of lipophilicity. Bioorg. Med. Chem. Lett.
bonds in biological molecules. Proc. Natl. Acad. Sci. U.S.A. 2004, 2007, 17, 1759–1764.
101, 16789–16794. (144) Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo,
(128) Lu, Y.; Ting, S.; Wang, Y.; Yang, H.; Yan, X.; Luo, X.; Jiang, H.; G. A.; Devraj, R. V.; Elsworth, E.; Fobian, Y. M.; Gibbs, M. E.; Giles,
Zhu, W. Halogen bonding: a novel interaction for rational drug design? R. W.; Greene, N.; Huang, E.; Krieger-Burke, T.; Loesel, J.; Wager, T.;
J. Med. Chem. 2009, 52, 2854–2862. Whiteley, L.; Zhang, Y. Physicochemical drug properties associated with
(129) Zhou, P.; Lv, J.; Zou, J.; Tian, F.; Shang, Z. Halogen in vivo toxicological outcomes. Bioorg. Med. Chem. Lett. 2008,
waterhydrogen bridges in biomolecules. J. Struct. Biol. 2010, 18, 4872–4875.
169, 172–182. (145) Gleeson, M. P. Generation of a set of simple, interpretable
(130) Hernandes, M. Z.; Cavalcanti, S. M. T.; Moreira, D. R. M. ADMET rules of thumb. J. Med. Chem. 2008, 51, 817–834.
Halogens atoms in the modern medicinal chemistry: hints for the drug (146) Gleeson, P.; Bravi, G.; Modi, S.; Lowe, D. ADMET rules of
design. Curr. Drug. Targets 2010, 11, 303–314. thumb II: a comparison of the effects of common substituents on a range
(131) O’Dowd, H.; Lewis, J. G.; Trias, J.; Asano, R.; Blais, J.; Lopez, of ADMET parameters. Bioorg. Med. Chem. 2009, 17, 5906–5919.
S. L.; Park, C. K.; Wu, C.; Wang, W.; Gordeev, M. F. Novel antibacterial (147) Leeson, P. D.; Springthorpe, B. The influence of drug-like
azetidine lincosamides. Bioorg. Med. Chem. Lett. 2008, 18, 2645–2648. concepts on decision-making in medicinal chemistry. Nat. Rev. Drug
(132) Barrett, S. D.; Bridges, A. J.; Dudley, D. T.; Saltiel, A. R.; Discovery 2007, 6, 881–890.
Fergus, J. H.; Flamme, C. M.; Delaney, A. M.; Kaufman, M.; LePage, S.; (148) Leeson, P. D.; Empfield, J. R. Reducing the risk of drug
Leopold, W. R.; Przybranowski, S. A.; Sebolt-Leopold, J.; Van Becelaere, attrition associated with physicochemical properties. Annu. Rep. Med.
K.; Doherty, A. M.; Kennedy, R. M.; Marston, D.; Howard, W. A., Jr.; Chem. 2010, 45, 393–407.
Smith, Y.; Warmus, J. S.; Tecle, H. The discovery of the benzhydrox- (149) Price, D. A.; Blagg, J.; Jones, L.; Greene, N.; Wager, T.
amate MEK inhibitors CI-1040 and PD 0325901. Bioorg. Med. Chem. Physicochemical drug properties associated with in vivo toxicological
Lett. 2008, 18, 6501–6504. outcomes: a review. Expert Opin. Drug Metab. Toxicol. 2009, 5, 921–931.
(133) Warmus, J. S.; Flamme, C.; Zhang, L. Y.; Barrett, S.; Bridges, (150) Waring, M. J. Lipophilicity in drug discovery. Expert Opin
A.; Chen, H.; Gowan, R.; Kaufman, M.; Sebolt-Leopold, J.; Leopold, W.; Drug Discovery 2010, 5, 235–248.
Merriman, R.; Ohren, J.; Pavlovsky, A.; Przybranowski, S.; Tecle, H.; (151) The Nobel Prize in Chemistry 2010—Richard F. Heck, Ei-ichi
Valik, H.; Whitehead, C.; Zhang, E. 2-Alkylamino- and alkoxy- Negishi, Akira Suzuki, “For Palladium-Catalyzed Cross Couplings in
substituted 2-amino-1,3,4-oxadiazoles—O-Alkyl benzohydroxamate Organic Synthesis”. http://nobelprize.org/nobel_prizes/chemistry/
esters replacements retain the desired inhibition and selectivity laureates/2010/ (accessed Feb 7, 2011).
against MEK (MAP ERK kinase). Bioorg. Med. Chem. Lett. 2008, (152) Constable, D. J. C.; Dunn, P. J.; Hayler, J. D.; Humphrey,
18, 6171–6174. J. L., Jr.; Linderman, R. J.; Lorenz, K.; Manley, J.; Pearlman, B. A.;
(134) Protein Data Bank (PDB). http://www.rcsb.org/ (accessed Wells, A.; Zaks, A.; Zhang, T. Y. Key green chemistry research areas, a
Feb 13, 2011). perspective from pharmaceutical manufacturers. Green Chem. 2007, 9,
(135) Tecle, H.; Shao, J.; Li, Y.; Kothe, M.; Kazmirski, S.; Penzotti, J.; 411–420.
Ding, Y. H.; Ohren, J.; Moshinsky, D.; Coli, R.; Jhawar, N.; Bora, E.; (153) Alcazar, J.; Oehrlich, D. Recent applications of microwave
Jacques-O’Hagan, S.; Wu, J. Beyond the MEK-pocket: Can current irradiation to medicinal chemistry. Future Med. Chem. 2010, 2, 169–176.