Professional Documents
Culture Documents
Information Theory Description of Synthetic Strategies in The Polyquinane Series
Information Theory Description of Synthetic Strategies in The Polyquinane Series
Abstract: Information theory makes it possible to give a semi- counterpart of these reaction-centered approaches: the structural
quantitative graphical representation of the various strategies used entities which make them possible. Such structural entities (holo-
to reach a given synthetic target. Skeletal complexity and similar- synthons) call attention to synthetic strategies where a global part
ity of the precursors with respect to the target structure provide (holos: whole) of the target is looked at, this view complements
figures which monitor the progress made from the starting mate- the more classical bond by bond, disconnection approach.
rial en route toward the target. Examples selected from the Key words: information theory, startegy, synthesis, triquinane,
triquinane family are used to illustrate the benefits but also the silphinene, hirsutene, coriolin, holosynthon
present limits of such an approach. Whereas for silphinene and
hirsutene various synthetic strategies appear in a clear graphical
form when treated within this framework, coriolin shows that the
1 Molecular Complexity and Information Theory cation will also reveal its limits in the analysis to come. If
these limits are clearly identified, this quantitative ap-
Chemists have had intuitive feelings about molecular proach has the advantage of providing a semi-quantitative
complexity. In 1981 Bertz1 developed a quantitative ap- framework to express the beauty of building a complex
proach based on information theory2 and graph represen- molecule out of simple fragments.
tation of molecules.3 In this approach, the molecular com-
plexity is measured as a function of the number and nature
of its constitutive atoms and of the number and nature of
the constitutive bonds. The overall complexity of the mol-
ecule is calculated as being the sum of complexities asso-
ciated with connectivity factors and complexities associ-
ated with the presence of heteroatoms. Hendrickson and
Toczko4 have developed a simple algorithm for calculat-
ing this complexity for any organic compound. The results
converge with a chemist's intuition on many structural
features: a cyclic compound is considered more complex
than its acyclic counterpart; a ramified hydrocarbon is
more complex than its linear counterpart; a molecule with
several carbons replaced by heteroatoms is more complex.
The virtue of this approach is to provide figures allowing
easy graphical comparisons of synthetic strategies. We
have used it here as a convenient tool for such compari-
sons. Figure 1 shows on simple examples the convergence
of its results with chemical intuition. Its systematic appli- Figure 1. Complexity changes in some classical reactions
1560 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
Biographical Sketches
Michel Chanon is Professor in the Faculté de Sciences de Saint-Jérôme. He has held
visiting appointments at the University of London, Indiana University, Kansas State Uni-
versity, Texas University, Brandeis University, Osaka University Hyderabad School of
Chemistry, Universitad Autonoma de Barcelona. Besides Computer-Aided Synthesis,
his research interests include experimental determination of mechanisms and reactivity
concepts, electron transfer and catalysis.
Christophe Baralotto was born in 1968 and studied chemistry at the University of
Marseille (France) where he received his Ph.D. degree under the supervision of Professor
Michel Chanon in 1997. His thesis subject was devoted to the study of the holosynthon
concept and its application in practical organic synthesis. In 1998 he left the academic
world to create his own company.
Rene Barone has been Directeur de Recherche at CNRS since 1986. He has worked in
the field of Computer-Aided Synthesis since 1970. In 1979 he did post-doctoral study in
M.L.H. Green’s group to develop the first program able to predict the products and by-
products of processes catalyzed by transition metal complexes. In 1981 he began to de-
sign programs working on personal computers with the aim of making available to me-
dium sized laboratories tools for use in synthesis design. His recent programs HOLO-
WIN, CONAN, SESAM are developed along this line.
Professor Hendrickson took his doctoral degree from Harvard under Professor R.B.
Woodward and was on the staff at UCLA before coming to Brandeis, developing re-
search programs both in synthetic chemistry and in application of chemistry to the com-
puter. Some 40 of his 150 research publications have been concerned with the latter area.
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1561
In this review, we will analyze the different total synthe- spect to the question “do I aim for A or for B?”. A conve-
ses which have been reported for some natural products of nient way of circumventing this blindness is to define, at
the triquinane family: silphinene; hirsutene and coriolin. every step, the distance6 between the partially synthesized
Fortunately, Mehta’s recent review provides an in-depth fragment of the target and the target itself. This distance is
coverage of the recent synthetic achievements in the field directly connected to the similarity between this fragment
of polyquinane natural products.5 This review will allow and the desired target. Actually, the chemist knows which
us to lighten the weight of chemical considerations in or- structural target he is aiming at; the other advantage of
der to concentrate more on the field as a kind of bench- completing complexity by similarity results from an al-
mark to judge the strengths and weaknesses of informa- most material visualisation of the different strategic ap-
tion theory as applied to the comparison of different build- proaches.
ing strategies.
We have established a rough method of evaluating the
similarity of two skeletons containing only carbon and hy-
drogen atoms (Appendix). For most of the strategies that
2 Similarity, an Indispensable Complement
we will examine the progression of the synthesis will be
of Molecular Complexity
represented in terms of both complexity and similarity
changes.
In 1981, when Bertz1 applied information theory to syn-
thetic strategies, his approach was a breakthrough. It per- Before leaving this mountain metaphor for strategies, let
mitted the description, in a two dimensional way, of the us examine its virtues and shortcomings. Virtues first.
Figure 2.
Synthetic strategies and mountain climbing
1562 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
Another advantage of the representation in Figure 2 is that kind of complexity and we will see, with the coriolin case,
it demonstrates the difficulty of ranking strategies. In- that such considerations can play a determining role in the
deed, depending upon the walker, the selected path may clever synthesis of some families of targets. The third
drastically change. The security-minded walker (industri- component is “stereochemical complexity”. No actual at-
al approach) will rather follow a succession of smooth tempts has been made to evaluate this complexity.15 These
valleys to reach, with a minimum of bad surprises, the se- three components are also reflected in computer-aided
lected summit. The highly competitive sportsman will synthesis programs which are either only skeleton-orient-
certainly select a way which challenges his technical and ed or deal also with functionality16 and stereochemistry.17
physical abilities. Another one would just focus on the
An equivalent state of imperfection characterizes “molec-
pleasure of exploring a given aspect of the available path-
ular similarity”. As the blossoming of recent reviews and
ways (e.g. exploration of a radical reaction in a new struc-
monographs18 shows, this concept is gaining more and
tural context).
more popularity in molecular sciences. The main incen-
Despite its pedagogical attractiveness the shortcomings of tive for this growth is molecular recognition19 and its
this representation follow from its roughness. In the next pharmacological applications.20a,b
section we will examine some of the limits associated
The scale of similarities for a series of natural compounds
with the present treatment of complexity and similarity.
as perceived by the active site of a given enzyme would be
Even if these two concepts were totally satisfactory in
quite inappropriate for ranking the similarities of the same
terms of quantitative treatment one would still have to an-
series from the point of view of synthetic strategy. Some
ed with a change in the carbon skeleton are shown in the Figures 3 and 5 gather the descriptions of eight reported
schemes. When a transformation has several steps the first synthetic strategies (Schemes 1–8). In Figure 3 the select-
is usually the one inducing a skeletal change. The symbols ed strategies share a common feature: the skeletal com-
“+” and “–” are respectively associated with steps which plexity of silphinene (290) is reached before the seventh
are holosynthetic and steps with a yield lower than 60%. step and as soon as the fourth step a value of 250 is
The symbol “+–” designs steps possessing both character- reached. Of the three displayed complexity curves two
istics (strategically interesting but impeded by low yields). show that a precursor on the way to the target is of higher
The bold numbers placed under each structural formula are complexity than the target itself. This corresponds to a
the skeletal complexities calculated by Hendrickson's pro- negative situation in terms of atom economy23 if the at-
gram.4 The italic number stands for a measure of similarity oms bringing the excedent of complexity have to be elim-
of the given structure with respect to the target. For the sake inated in a further step toward the target. In both cases the
of clarity we have gathered the syntheses which, at first intermediate is obtained by a holosynthetic transforma-
sight, seems to be the best. We stress the point on the arbi- tion (see definition in Section 5) so that nothing can be
trariness of this choice. In Section 2 we have specifically done to improve further the strategy concerning this point.
explained that reasons of context may completely upset
Whereas all the strategies gathered in Figure 3 include at
ranking of syntheses based on simplistic considerations
least one step in which ∆C (increase in complexity) is
such as number of steps or overall apparent yields. Con-
higher than 100, Figure 5 describes only two strategies
cerning the question of uncertainties attached to yields the
(Itô and Yamamura) which contain such a step. These
reader is referred to Hudlicky's lucid discussion.21
a) 58%. b) HClO4. c) NaOMe. d) Me2CuLi. e) LiCl, H2O, Me2SO; 89% (4 Steps). f) EtCO2TMS, TBAF; 89%. g) hν; 86%. h) TMSI; 89%. i)
(n-Bu)3SnH; 98%. j) LDA, (EtO)2POCl; 64%. k) Li, CH3NH2 liq.; 99%. (Overall yield : 22%).
Scheme 1. Crimmins’ strategy for silphinene synthesis24
a) Na+Cp-; 75%. b) Li(Me)C=CH2. 71%. c) 160˚C; d) H2O, PyH, OTs; 68% (2 Steps). e) O3. f) KOH; 75% (2 Steps). g) Jones’ reaction; 95%.
h) Pb(OAc)4, Cu(OAc)2; 68%. i) Me2CuLi; 89%. j) N2H4, K2CO3; 93%. (Overall yield : 15%).
Scheme 2. Sternbach’s strategy for silphinene synthesis25
1564 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
a) 1. Li. 2. NH3. 3. NH4Cl; 87%. b) hν; 35%. c) Li, CH3NH2 liq.; 66%. (Overall yield : 20%).
Scheme 3. Wender’s strategy for silphinene synthesis26
a) AlCl3. b) HO(CH2)2OH; 80% (2 Steps). c) NaIO4, OsO4. d) NaBH4. e) HCl, MeOH; 73% (3 Steps). f) Jones’ reaction. g) CH2N2; 76% (2
Steps). h) LDA, MeI; 86%. i) LiAlH4. j) PCC. k) N2H4, KOH; 72% (3 Steps). l) TMS, NaI; 99%. m) DBU; 87%. n) CuI; 70%. o) HCl; 98%.
p) POCl3; 81%. q) MeLi. r) SOCl2. s) m-CPBA. t) BF3-Et2O; 33% (4 Steps). u) N2H4, KOH; 90%. (Overall yield : 4%).
Scheme 5. Itô’s strategy for silphinene synthesis28
a) LDA, TMSCl. b) HC≡C-CO2Et, ZrCl4; 90% (2 Steps). c) CF3SO3TMS; 94%. d) 85%. e) hν; 85%. f) TBAF; 100%. g) H2SO4; 100%. h)
CF3SO3SiMe2Thexyl. i) SeO2; 76% (2 Steps). j) C6H5CH2OCNHCCl3; 80%. k) DIBAH. l) MnO2; 73% (2 Steps). m) 80%. n) MnO2; 85%. o)
BF3-Et2O. p) TBAF; 50%(2 Steps). q) (COCl)2, DMSO; 89%. r) LDA., MeI; 63%. s) Me2CuLi; 78%. t) N2H4, KOH. u) N2H4, KOH; 46% (2
Steps). (Overall yield : 2%).
Scheme 6. Franck-Neumann’s strategy for silphinene synthesis29
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1565
a) NaH; 88%. b) LDA; 90%. c) 1. Li, NH3 liq. 2. NH4Cl; 60%. d) RuCl3, NaIO4; 100%. e) CH2N2; 100%. f) HO(CH2)2OH; 60%. g)
CH3P(O)(OMe)2, n-BuLi; 100%. h) HCl; 100%. i) (n-Bu)4NOH; 70%. j) NaBH4; 72%. k) p-MPTC; 70%. l) (n-Bu)3SnH; 70%. m, n) See
Crimmins; 63% (2 Steps). (Overall yield : 7%).
Scheme 8. Nagarajan’s strategy for silphinene synthesis31
and stereochemical complexities were available it could In the plot describing Franck-Neumann's strategy there is
well reveal strong complexity increases for these three indeed a step in which complexity decreases while simi-
strategies. The second consideration follows from Figures larity increases. This step corresponds to g, h, i, j, k, l in
4 and 6. Within a given strategy, similarity is often corre- Scheme 6. The complexity value indeed decreases from
lated to complexity, as might be expected. In Figure 2, 216 to 157 while the similarity with respect to silphinene
when one progresses from one valley towards the target, increases from 45% to 66%. One does not necessarily
the normal evolution is an increase in the complexity of need information theory to pick out this fact. The com-
the intermediates and an increase in their similarity with plexity decreases on going from the first intermediate to
respect to the target. This trend reveals itself in Figure 4. the second one because there is one ring less in the second
This usual trend led Bertz to state that complexity and intermediate. A quick glance at the two intermediates and
similarity are correlated.32 Figure 6, however, shows that at the structure of the target convinces any chemist that the
this trend is not always observed. second intermediate is more similar to the target. This sit-
1566 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
Complexity
300
250
200
150
100
Similarity vs
Equation 133a
50
silphinene
0
0 20 40 60 80 100 120
Equation 233b
Complexity (Continuous line) Similarity vs
silphinene
(Dotted line)
300
100
250
80
200
60
Equation 333c
150
Silphinene Itô (21 Steps, 4%) In this treatment of silphinene we have omitted Fraser-
Yamamura (16 Steps, 6%) Nagarajan (14 Steps, 7%) Reid's approach.34 The reason is that the selected strategy
Paquette (15 Steps, 10%) Franck-Neumann (21 Steps, 2%)
involves several protection-deprotection steps. We will
Figure 5. Evolution of skeleton complexities and similarities (with see with coriolin synthesis (Section 4.3) the limitations of
respect of the target) in silphinene synthesis (B) our approach for this kind of strategy. Taylor’s transition
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1567
a) hν; 80%. b) TMSI; 95%. c) 1. Li, NH3 liq. 2. MeI; 47%. d) See
Cohen; 96% (1 Steps). (Overall yield: 33%).
Scheme 12. Iyoda’s strategy for hirsutene synthesis40
a) OsO4, NaIO4. b) HOAc; 62% (2 Steps). c) CH2=CHMgBr; 91%. d) CH3C(OEt)3, Hg(OAc)2, EtCO2H. e) KOH, H2O; 82% (2 Steps). f)
(COCl)2. g) CH3CHN2; 89% (2 Steps). h) Cu(acac)2; 94%. i) PbCO3, 580°C; 66%. j) H2/PtO2; 90%. k) See Cohen; 96% (1 Steps). (Overall
yield: 22 %).
Scheme 11. Hudlicky’s strategy for hirsutene synthesis39
1568 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
a) PhSeSO2Ph, hν. b) H2O2; 70% (2 Steps). c) 160°C. d) Jones’ reaction; 72% (2 Steps). e) hν, CH2=CH2. f) NaN(TMS)2, TBSCl. g) m-CPBA;
68% (3 Steps). h) Al(Hg)x; 91%. i) Ph3P=CH2, (i-Pr)2O; 86%. j) I2; 91%. k) See Cohen; 96% (1 Steps). (Overall yield: 26%).
Scheme 13. Paquette’s strategy for hirsutene synthesis41
60 60
150 150
40 40
100 100
20 50 20
50
Synthetic Synthetic
Intermediates Intermediates
0 0 0 0
1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Hirsutene Paquette (11 Steps, 26%) Hirsutene Curran (13 Stepas, 11%)
Iyoda (4 Steps, 33%) Cohen (6 Steps, 37%) Franck-Neumann (10 Steps, 13%) Funk (12 Steps, 13%)
Biosynthesis (5 Steps) Hudlicky (11 Steps, 22%) Sternbach (20 Steps, 20%)
Figure 7. Evolution of skeleton complexities and similarities (with Figure 8. Evolution of skeleton complexities and similarities (with
respect of the target) in hirsutene synthesis (A) respect of the target) in hirsutene synthesis (B)
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1569
In contrast, Curran’s approach44 saves the key step for the Complexity
very end of the synthesis. Sternbach's smooth approach42 300
provides the best efficiency by overall yield. This strate- 250
gy, however, is the longest one in terms of steps and the 200
complexity of the target is almost reached after only seven 150
steps (out of 17 overall), but to reach 100% similarity to 100
Similarity vs
hirsutene requires ten more steps. This part of the strategy 50
hirsutene
has been recently improved by Rawal46 using their ap- 0
0 20 40 60 80 100 120
proach of forming a latent triquinane ((±) endo-hirsutene)
out of a norbornane via a Paterno-Büchi reaction. Franck- Cossy (9 Steps, 5%) Wender (7 Steps, 4%)
Weedon (7 Steps, 9%) Ley (8 Steps, 8%)
Neumann’s strategy stands out in Figure 9 with similarity
plotted against complexity. This corresponds to the same Figure 11. Complexities versus similarity (with respect of the target)
type of approach for two similar targets (compare Figures in hirsutene synthesis (C)
6 and 9). For both targets the key step involves a decrease
in complexity coupled with an increase in similarity.
terms of number of steps. Indeed, here, the shortest strat-
Figures 10 and 11 reveal some new trends. Ley's egy (number of steps) is also the least efficient (overall
approach48 resembles Funk’s discussed above. This re- yield). In the same spirit of skepticism one may note that
semblance is partly fortuitous because our treatment ne- Wender's strategy,47 which was a winner with silphinene
Sternbach (20 Steps, 20%) Funk (12 Steps, 13%) Figure 12 gathers the syntheses with the lowest overall
Franck-Neumann (10 Steps, 13%) Curran (13 Steps, 11%) yields. One must stress again that classification by this cri-
Figure 9. Complexities versus similarity (with respect of the target) terion is just a matter of convenience. The gathering of all
in hirsutene synthesis (B) the strategies on a single graph would have been undeci-
250 250
80 80
200
200
60
60
150
150
40
40 100
100
50 20
50 20
Synthetic
Synthetic Intermediates
Intermediates 0 0
0 0 2 4 6 8 10 12 14 16 18 20 22 24 26
1 2 3 4 5 6 7 8 9 10
Hirsutene Wender (7 Steps, 4%) Hirsutene Fukumoto (25 Steps, 3%)
Cossy (9 Steps, 5%) Ley (8 Steps, 8%) Hewson (17 Steps, 3%) Magnus (16 Steps, 3%)
Weedon (7 Steps, 9%) Mehta (10 Steps, 3%) Little (15 Steps, 2%)
Figure 10. Evolution of skeleton complexities and similarities (with Figure 12. Evolution of skeleton complexities and similarities (with
respect of the target) in hirsutene synthesis (C) respect of the target) in hirsutene synthesis (D)
1570 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
Complexity pherable. The reader must keep in mind that: (1) a syn-
300 thesis of 10 steps with an average 80% yield amounts to
250 an overall yield of 10% whereas if the average yield is
200 75% this value drops to 5.6% (2) our description does
150 not account for the evolution of functional and stere-
100 ochemical complexities. Of the five strategies displayed
Similarity vs
50
0
hirsutene in Figure 12, two (Little52 and Mehta54) involve a holo-
0 20 40 60 80 100 120 synthetic step. Mehta's starts with a double jump in com-
plexity obtained by a sequence of [4+2] and [2+2] cy-
Hewson (17 Steps, 3%) Magnus (16 Steps, 3%)
Mehta (10 Steps, 3%) Little (15 Steps, 2%) cloadditions. This double jump yields a precursor whose
Fukumoto (25 Steps, 3%) skeletal complexity exceeds hirsutene complexity by 60
Figure 13. Complexities versus similarity (with respect of the target) units. This excess has to be compensated in the next step
in hirsutene synthesis (D) in which a 118 loss in complexity is compensated by a
22% gain in similarity.
Returning to Figure 2, this strategy illustrates the situation
wherein the climber follows a valley which leads him fur-
ther off from the selected summit; he then has to go down
into another one to find his original direction. One may
a) NaBH4, CeCl3. b) Ac2O, Pyr. c) CH2=CH(OTBS)2. d) ∆. e) PhSeCl. f) H2O2; 62% (6 Steps). g) 75%. h) DIBAH; 83%. i) (CF3CO)2O; Pyr.;
81%. j) (n-Bu)4NI; 75%. k) LiC≡C-TMS. l) TBAF; 75% (2 Steps). m) (n-Bu)3SnH; 63%. (Overall yield : 11%).
Scheme 16. Curran’s strategy for hirsutene synthesis44
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1571
a) 1. CuBr, Me2S. 2. HCl; 84%. b) PhCOCl, Pyr; 90%. c) Morpholine, TsOH; 91%. d) 57%. e) HI; 85%. f) H2/Pd; 91%. g) NaOMe; 95%. h)
Electrochmical reduction: 45%. i) PCC; 100%. j) See Cohen; 96% (1 Steps). (Overall yield: 13%).
Scheme 17. Franck-Neumann’s strategy for hirsutene synthesis45
4.3 Coriolin (Schemes 27–35) In Figure 14 the graphical evolutions of skeletal complex-
ities and similarities of the three “best” approaches (com-
The main structural characteristic of this triquinane ses- bination of yields and number of steps) are displayed.
quiterpenoid is its considerable functionality: the carbon Only Curran's synthesis63 (Scheme 29) presents a holo-
skeleton is substituted by two hydroxy and one keto synthetic character in its last phase; the strategy used in
group, as well as by two epoxides. We have purposely se- this phase had already met with success for hirsutene syn-
lected this target to show the limits of a “skeleton-only thesis (Section 4.2). Ikegami61 and Ito62 classically built
complexity” approach. the triquinane skeleton starting from a precursor where
two out of the three rings constituting the triquinanic skel-
Let us first consider the specific problem of the double
eton are already present (Schemes 27 and 28).
epoxidation needed to obtain the target. Tatsuta64 per-
formed this transformation by a one-pot reaction (62%). Figures 15 and 16 gather the “average” synthesis. In Fig-
The same year, Danishefsky68 proposed a four-step trans- ure 15, the graph describing Tatsuta’s approach64 unduly
formation, far better in terms of stereochemical control, suggests a highly holosynthetic step in the first phase of the
and of comparable overall yield (58%). For saving space synthesis. Actually, the great increase in complexity is the
in Schemes 27–35 and Figure 14–16 we have retained the result of addition of the complexity of the reagent to the one
“shorter” Tatsuta's transformation. of the substrate (steps f,g,h,i,j in Scheme 30). The same
1572 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
a) NaH, TsSMe. b) HO(CH2)2OH, TsOH. c) NaIO4. d) CaCO3; 64% (4 Steps). e) EtNO2, Tetramethylguanidine; 93%. f) TiCl4; 77%. g) TFA;
79%. h) NaH; 83%. i) LDA, MeI. j) ∆. k) N2H4, K2CO3. l) m-CPBA; 67% (4 Steps). m) NaH; 72%. n) NaHPO3; 87%. o) HCO2H; 56%. p)
RuO2; 42%. q) See Cohen; 96% (1 Step). (Overall yield : 3%).
Scheme 22. Hewson’s strategy for hirsutene synthesis51
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1573
a) I2, KI, NaHCO3. b) DBU. c) LiAlH4; 58% (3 Steps). d) TBSCl. e) Ac2O, Pyr. f) TBAF; 82% (3 Steps). g) O3, Me2S; 93%. h) Ph3P=CHCOMe.
i) PPTS, MeOH; 69% (2 Steps). j) TsOH; 86%. k) LiOH. l) ortho-NO2PhSeCN, (n-Bu)3P. m) H2O2; 66% (3 Steps). n) 1. LDA. 2. TMSCl. 3.
Pd(OAc)2; 99%. o) H2/Pd. p) Ph3P=CH2. q) CH2I2, Et2Zn; 50% (3 Steps). r) HClO4. s) Ph3P=CH2; 61% (2 Steps). t) PCC. u) PdCl2, CuCl, O2.
v) (n-Bu)4NOH, KOH. w) H2/PtO2; 84%; (4 Steps). x) PCC; 74%. y) See Cohen; 96% (1 Step). (Overall yield : 3%).
Scheme 24. Fukumoto’s strategy for hirsutene synthesis53
1574 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
5 Holosynthons
a) K2CO3, TsN3. b) hν; 82% (2 Steps). c) NaBH4. d) SOCl2, Pyr.; 72% (2 Steps). e) DBU; 94%. f) NaOH; 96%. g) (COCl)2; 68%. h) AgBF4;
38%. i) 1. MeLi. 2. HgCl2, HgO; 93%. j) 1. BuLi. 2. CuI. 3. BF3-OEt2; 60%. k) BzEt3N+Cl-, KF; 84%. l) TsCl, Pyr.; 94%. m) LiN(TMS)2; 75%.
n) NaBH4. o) CS2, NaH, MeI. p) (n-Bu)3SnH, AIBN; 70% (3 Steps). (Overall yield : 3%).
Scheme 26. Magnus’ strategy for hirsutene synthesis55
a) H2O2. b) 40˚C; 66% (2 Steps). c) NaH, BzBr; 91%. d) NBS. e) H2/Pd; 80% (2 Steps). f) PCC. g) TBSCl. h) DBU; 80% (3 Steps). i) Me2CuLi;
95%. j) KOt-Bu, MeI; 77%. k) Li, NH3; 63%. l) DHP. m) TBAF. n) PCC; 90% (3 Steps). o) NaH; 94%. p) PdCl2, CuCl; 77%. q) KOt-Bu; 83%.
r) LDA, MeI; 78%. s) 1. LDA, PhSeBr. 2. H2O2. t) HOAc; 50% (2 Steps). u) KOt-Bu. v) m-CPBA. w) DBU; 48% (3 Steps). x) See Danishefsky;
58% (1 Step). (Overall yield : 1.1%).
Scheme 27. Ikegami’s strategy for coriolin synthesis61
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1575
a) NaBH4, CeCl3. b) Ac2O, Pyr; 95% (2 Steps). c) CH2=C(OTBS)2. d) ∆. e) PhSeCl. f) H2O2; 62% (4 Steps). g) Li, CuBr. h) LiAlH4; 90% (2
Steps). i) PCC. j) LiC≡C__TMS. k) PCC. l) HO(CH2)2OH. m) TBAF. n) PCC; 30% (6 Steps). o) SmI2. p) TsOH; 60% (2 Steps). q) LDA,
TMSCl; 89%. r) DDQ; 72%. s, t, u) See Ikegami; 48% (3 Steps). v) See Danishefsky; 58% (1 Step). (Overall yield : 1.7%).
Scheme 29. Curran’s strategy for coriolin synthesis63
1576 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
300
100
250
80
200
60
150
40
100
20
50
Synthetic
Intermediates
0 0
2 4 6 8 10 12 14 16 18 20 22 24 26 28
a) Pd(P(Ph)3)4, DBU; 76%. b) NBS; 89%. c) 1.P(Ph)3. 2. K2CO3. 3. 40˚C; 79%. d) MeSH. e) HO(CH2)2OH; 92% (2 Steps). f) KH, (MeS)2;
79%. g) KH; 72%. h) m-CPBA; 63%. i) TBAF; 87%. j) CH2I2, ZnEt2; 82%. k) H2/PtO2; 94%. l) SOCl2; 92%. m) m-CPBA; 87%. n) HClO4. o)
DBU; 91% (2 Steps). p) Sodium naphthalenide. q) DBU; 52% (2 Steps). r) Li, NH3 liq. s) m-CPBA; 63% (2 Steps). t) CF3C(O)N(TMS)2. u)
LDA, TMSCl. v) Me3N+,I-. w) MeI. x) DBU; 46%. y) HF; 85%. z) See Danishefsky; 58% (1 Step). (Overall yield : 0.7%).
Scheme 31. Trost’s strategy for coriolin synthesis65
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1577
50 20
Synthetic 20
50
Intermediates Synthetic
0 0 Intermediates
2 4 6 8 10 12 14 16 18 20 22 24 26 0 0
2 4 6 8 10 12 14 16 18 20 22
Coriolin Trost (26 Steps, 0,7%)
Schuda (25 Step, 0,7%) Tatsuta (25 Steps, 03%) Coriolin Wender (14 Steps, 0,4%)
Metha (19 Steps, 0,8%) Danishefsky (22 Steps, 0,2%)
Figure 15. Evolution of skeleton complexities and similarities (with
a) HO(CH2)2OH; 70%. b) 1. Br2. 2. NaOMe. 3. HCl; 59%. c) LiAl(OMe)3H, CuBr; 84%. d) KOt-Bu, MeI; 85%. e) Li, NH3 liq.; 70%. f) KH,
BzBr; 89%. g) OsO4; 96%. h) NaIO4. i) NaBH4; 85% (2 Steps). j) t-BuCOCl, Pyr; 48%. k) (CF3SO2)2O, Pyr. l) TBAI. m) Zn; 81% (3 Steps).
n) KOH; 100%. o) RuO2, NaIO4; 93%. p) HCl; 85%. q) CH(OMe)3, TsOH. r) 160˚C; 82% (2 Steps). s) Hg(OAc)2; 89%. t) KOt-Bu. u) TsOH;
80%. v, w, x) See Ikegami; 48% (3 Steps). y) See Danishefsky; 58% (1 Step). (Overall yield : 0.7%).
Scheme 32. Schuda’s strategy for coriolin synthesis66
gathered most of the relevant contributions to this idea for In Section 4.3. we saw that Bertz’s complexity was rather
“making much chemistry in one step”. good at describing carbon-skeleton complexity changes but
This information theory approach is really not needed to rather poor at describing the functional and stereochemical
recognize these reactions in a given strategy. When we aspects of complexity. The unsatisfactory treatment of co-
first coined the term holosynthon we were not even aware riolin strategies above demonstrated the limits of such a
of Bertz’s approach. We believe, however, that the defini- narrow view. On the other hand, if one accepts the idea that
tion of holosynthon given in terms of complexity and or complexity could be sub-divided into three components the
similarity is more closely defined than the intuitive views holosynthon approach touches the heart of efficient synthe-
that we had at first. The graphical representation of strat- sis. Indeed the following examples show that a functional
egies complements the classical approaches in terms of holosynthon could be a structural unit for which large
number of steps and key steps.78 It is particularly appro- changes of functional complexity and/or similarity can
priate for polycyclic targets which are not too heavily be achieved in a one-pot reaction. Equations (4)79 and
functionalized. (5)80 provide examples of such functional holosynthons:
1578 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
a) NaH CO2; 63%. b) NaOMe, MeOH; 85%. c) TsOH; 51%. d) 120˚C. e) PhSeCl. f) H2O2; 57% (3 Steps). g) MeLi; 73%. h) O3. i) CrO3. j)
Ba(OH)2. k) Pb(OAc)4; 46% (4 Steps). l) KOt-Bu. m) TsOH; 71% (2 Steps). n) KOt-Bu. o) HOAc; 63% (2 Steps). p) DIBAH. q) Li, NH3 liq.
r) m-CPBA. s) PCC; 55% (4 Steps). t) 1. LDA. 2. Phenyl(thiophenyl)sulfonate; 40%. u) 1. m-CPBA. 2. 77˚C; 64%. v) H2O2, 0˚C; 58%. (Overall
yield : 0.2%).
Scheme 34. Danishefsky’s strategy for coriolin synthesis68
November 1998 Information Theory Description of Synthetic Strategies in the Polyquinane Series 1579
a-f) See Mehta’s synthesis of hirsutene ; 18% (6 Steps). g) MeMgI; 90%. h) POCl3, Pyr.; 75%. i) Li, NH3 liq. 63%. j) m-CPBA; 100%. k)
BF3-OEt2; 80%. l) LDA, TMSCl. m) Pd(OAc)2; 90% (2 Steps). n) LDA, PhSeBr. o) H2O2; 35% (2 Steps). p, q, r) See Ikegami; 48% (3 Steps).
s) See Danishefsky; 58% (1 Step). (Overall yield: 0.8%).
Scheme 35. Mehta’s strategy for coriolin synthesis54
Figure 18. Overall value of complexities (addition of the complexity of each intermediate) versus overall yield for hirsutene.
1580 M. Chanon, R. Barone, C. Baralotto, M. Julliard, J. B. Hendrickson SYNTHESIS
Equation 681
Equation 883
On the other hand, the critical examination of various holo- Table 2. Similarity of cyclic skeletons
synthons in this work suggests that the strategy yielding the
most impressive overall yield is not necessarily holosyn-
thetic in essence. A critical question before selecting an ef-
ficient strategy remains “would this kind of target be adapt-
ed to a holosynthetic approach?”. The question is still diffi-
cult to answer because new holosynthetic strategies keep
appearing in the litterature with a high frequency.
Appendix
Our approach to similarity is comparable to Bertz’s
complexity4 and adapted to be an efficient monitor of
skeletal differences for compounds containing only car-
bons and hydrogens. Therefore for compounds like corio-
lin the target is simplified by replacing all the functional-
ities by the appropriate number of hydrogens. Four struc-
tural parameters have been selected to describe every
intermediate in the synthetic tree:
Table 4. Similarity of polycyclic skeletons (* see Fig. 19) (4) Hendrickson, J. B.; Huang, P.; Toczko, A. G. J. Chem. Inf.
Comput. Sci. 1987, 27, 63. Available on request.
(5) Mehta, G. Chem. Rev. 1997, 97, 671.
(6) Hendrickson, J. B.; Braun-Keller, E. J. Comput. Chem. 1980,
1(4), 323.
Jochum, C.; Gasteiger, J.; Ugi, I. Angew. Chem. 1980, 92, 503;
Angew. Chem., Int. Ed. Engl. 1980, 19, 495.
Wochner, M.; Brandt, J.; von Scholley, A.; Ugi, I. Chimia 1988,
6, 217.
Ugi, I.; Bauer, J.; Bley, K.; Dengler, A.; Dietz, A.; Fontain, E.;
Gruber, R.; Knauer, M.; Reitsam, K.; Stein, N. Angew. Chem.
1993, 105, 210; Angew. Chem., Int. Ed. Engl. 1993, 32, 201.
(7) Denis, J. D.; Greene, A. E.; Guénard, D.; Guéritte-Voegelein,
F.; Mangatal, L.; Potier, P. J. Am. Chem. Soc. 1988, 110, 5917.
(8) Wipke, W.T.; Rogers, D. J. Chem. Inf. Comput. Sci. 1984, 24, 71.
(9) Chastrette, M.; Rajzmann, M.; Chanon, M.; Purcell, K. F. J. Am.
Chem. Soc. 1985, 107, 1.
(10) Hendrickson, J. B.; Chanon, M.; Chastrette, M. to be published
(11) Atlan, H. Organisation Biologique et Théorie de l'Information,
Hermann, Paris, 1992.
(12) Hendrickson, J. B.; Braun-Keller, E.; Toczko, G. A. Tetrahe-
(23) Trost, B. M.; Science, 1991, 254, 1471. (54) Mehta, G.; Murthy, A. N.; Reddy, D. S.; Reddy, A. V. J. Am.
Trost, B. M. Angew. Chem. 1995, 107, 285; Angew. Chem., Int. Chem. Soc. 1986, 108, 3443.
Ed. Engl. 1995, 34, 259. (55) Magnus, P.; Quagliato, D. J. Org. Chem. 1985, 50, 1621.
(24) Crimmins, M. T.; Mascarella, S. W. J. Am. Chem. Soc. 1986, (56) Majetich, G.; Defauw, J. Tetrahedron 1988, 44, 3833.
108, 3435. (57) Sarkar, T. K.; Ghosh, S. K.; Subba Rao, P. S. V.; Satapathi, T.
(25) Sternbach, D. D.; Hugues, J. W.; Burdi, D. F.; Banks, B. A. K.; Mamdapur, V. R. Tetrahedron 1992, 48, 6897.
J. Am. Chem. Soc. 1985, 107, 2149. (58) Castro, J.; Sorensen, A.; Riera, A.; Morin, C.; Moyano, A.; Per-
(26) Wender, P. A.; Ternansky, R. J. Tetrahedron Lett. 1985, 26, iras, M. A.; Greene, A. E. J. Am. Chem. Soc. 1990, 112, 9388.
2625. (59) Hua, D. H.; Sinai-Zindge, G.; Venkataraman, S. J. Am. Chem.
(27) Paquette, L. A.; Leone-Bay, A. J. Am. Chem. Soc. 1983, 105, Soc. 1985, 107, 4088.
7352. Hua, D. H.; Venkataraman, S.; Ostander, R. A.; Sinai, G.-Z.; Mc-
(28) Tsunoda, T.; Kodama, M.; Itô, S. Tetrahedron Lett. 1983, 24, 83. Cann, P. J.; Coulter, M. J.; Xu, M. R. J. Org. Chem. 1988, 53, 507.
(29) Franck-Neumann, M.; Miesch, M.; Gross, L. Tetrahedron Lett. Inoue, T.; Hosomi, K.; Araki, M.; Nishide, K.; Node, M. Tetra-
1991, 32, 2135. hedron Asymm. 1995, 6, 31.
Miesch, M.; Miesch, L.; Franck-Neumann, M. Tetrahedron Weinges, K.; Reichert, H.; Huber-Patz, U.; Irngartinger, H. Lie-
1997, 53, 2103 bigs Ann. Chem. 1993, 403.
(30) Yamamura, S.; Ohkubo, M.; Shizuri, Y. Tetrahedron Lett. (60) Oppolzer, W.; Robyr, C. Tetrahedron 1994, 35, 415.
1989, 30, 3798. (61) Iseki, K.; Yamazaki, M.; Shibasaki, M.; Ikegami, S. Tetrahe-
(31) Koteswar Rao, Y.; Nagarajan, M. Tetrahedron Lett. 1988, 29, 107. dron 1981, 37, 4411.
(32) Bertz, S. H.; Herndon, W. C. in “Artificial Intelligence. Appli- (62) Ito, T.; Tomiyoshi, N.; Nakamura, N.; Azuma, S.; Izawa, M.
cations in Chemistry", Pierce, T. H.; Hohne, B. A. Ed. ACS Tetrahedron Lett. 1982, 23, 1721.