Total synthesis of herbicidin C and aureonuclemycin: Impasses and new avenues

Article abstract of DOI:10.1021/jo401706r

The undecose nucleoside antibiotics herbicidin C and aureonuclemycin are biologically highly active and represent challenging targets for total synthesis. Herein, the gradual evolution of our synthetic strategy toward these natural products is described in detail. The initial route encompasses metalate addition chemistry but suffers from poor stereochemical control. In contrast, the ultimately successful strategy benefits from a variety of reagent-controlled stereoselective transformations, including a surprisingly facile and highly diastereoselective N-glycosylation process. The presented work also describes new building blocks that might find further application in carbohydrate chemistry.

Full text of DOI:10.1021/jo401706r

Subscriber access provided by FORDHAM UNIVERSITY
Article
Total Synthesis of Herbicidin C and
Aureonuclemycin: Impasses and New Avenues
Dominik Hager, Christian Paulitz, Jörg Tiebes, Peter Mayer, and Dirk Trauner
J. Org. Chem., Just Accepted Manuscript • Publication Date (Web): 27 Sep 2013
Downloaded from http://pubs.acs.org on September 30, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Organic Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Total Synthesis of Herbicidin C and Aureonuclemycin:
Impasses and New Avenues
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†
‡
§
†
,†
Dominik Hager , Christian Paulitz , Jörg Tiebes , Peter Mayer , and Dirk Trauner*
†
Department of Chemistry, Ludwig-Maximilians-Universität München and Center of Integrated
Protein Science, Butenandtstr. 513, 81377 Munich, Germany
‡
Bayer Aktiengesellschaft BAG-CD-CS, Corporate Strategy, Building W11, 255, 51368
Leverkusen, Germany
§
Bayer CropScience AG Research, Industriepark Höchst, Gebäude G836, 65926 Frankfurt am
Main, Germany
dirk.trauner@lmu.de
ABSTRACT GRAPHIC
ABSTRACT
The undecose nucleoside antibiotics herbicidin C and aureonuclemycin are biologically highly
active and represent challenging targets for total synthesis. Herein, the gradual evolution of our
synthetic strategy toward these natural products is described in detail. The initial route
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
encompasses metallate addition chemistry but suffers from poor stereochemical control. In
contrast, the ultimately successful strategy benefits from a variety of reagent-controlled
stereoselective transformations, including a surprisingly facile and highly diastereoselective N-
glycosylation process. The presented work also describes new building blocks, which might find
further application in carbohydrate chemistry.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Nucleosides and nucleotides are not only the molecular building blocks of nucleic acids, but also
play a major role in many biosynthetic pathways. They function as energy currencies, secondary
messengers, co-factors for a variety of enzymes or post-translational modifiers, e.g. in
methylation or glycosylation reactions. Thus, it is not surprising that the nucleoside motif can
1,2
also be found in different classes of natural products. Most of these bear the canonical purine
and pyrimidine nucleobases, but are highly diversified with respect to their carbohydrate
moieties. The latter range from relatively simple ribose sugars to sophisticated higher-order
mono- or polysaccharides. A case in point are the herbicidins (19), tunicamycins (10) and
hikizimycin (11), all of which are examples of complex undecose nucleoside antibiotics (Figure
1
). These contain linear 11-membered carbon chains folded into different heterocyclic scaffolds.
2
ACS Paragon Plus Environment

Page 3 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. The herbicidin family (1‒9), as well as the tunicamycins (10) and hikizimycin (11).
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The herbicidins have been isolated from different Streptomyces strains and have attracted
3
-11
considerable attention due their structural sophistication and remarkable biological activities.
Within the family, herbicidins A (1) and B (2) interfere with the growth of Xanthomonas oryzae,
a bacterium that causes leaf blight infection in rice crops. In addition, they reduce the growth of
algae and impede seed germination. However, the most promising feature of these compounds is
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
-6,9-11
their selective toxicity towards dicotyledons with practically no activity against animals.
Overall, the herbicidin compound class offers a framework for the discovery of potential lead
structures in search of new herbicides and the development of novel biological tools.
The structures of the herbicidins exhibit a carbohydrate skeleton consisting of eleven carbons that
are connected in a linear fashion. This chain, however, is incorporated into a tricyclic ring system
to feature an unprecedented furano-pyrano-pyran moiety as well as nine stereogenic centers. The
nucleobase adenine is glycosylated at the 1-position, thus residing on the sterically more
encumbered concave face. Furthermore, the pyrano-pyran system includes a hemiketal at the C-7
position that forces all substituents on the terminal pyran ring into the axial orientation. The
individual members of the herbicidin family differ only in three positions, namely C-2, C-8 and
C-11. While the substituent at C-2 is either a hydroxy or methoxy group, and C-11 is the carbonyl
carbon atom of a carboxylic acid or methyl ester, the side chain at C-8 can be either a hydroxy
group or an ester derived from (E)-2-(hydroxymethyl)-2-butenoic acid, isobutyric acid, tiglic acid
or acetic acid.
Due to the potent bioactivity and the unusual structure, the herbicidins have been the focus of
synthetic interest for some time. As a result, several approaches towards the herbicidins have
1
2-25
been described in literature,
and a selection of these investigations is summarized in Scheme
1
.
4
ACS Paragon Plus Environment

Page 5 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Selected approaches toward the herbicidin framework including our strategies.
In general, two contrasting strategies regarding the N-glycosylation were envisioned: the
nucleobase could be introduced either prior to or after the formation of the tricyclic carbohydrate
moiety. Thus, these strategies are referred to as 'early-stage' or 'late-stage' glycosylation
1
7,26-28
approaches, respectively.
The reported studies on herbicidins mainly focused on the
combination of C - and C -building blocks by means of C-glycosylation, a strategy, which
6
5
includes the most obvious and logical disconnection between C-5 and C-6 for the herbicidin
undecose. For example, Whiting and co-workers employed an organolithium species accessed by
2
3,24
transmetallation from stannane 12 in reaction with aldehyde 13 or triflate 14 (Scheme 1).
However, the installation of the ketone at C-7 (herbicidin nomenclature) was not successful.
25
Aldehyde 13 was later also used as a C -building block by Haines and co-workers. Their aim
5
was to combine ketone 15, derived from L-rhamnose, with aldehyde 13 by means of an aldol
reaction. Remarkably, this is the only example that made use of an open-chain C -carbohydrate
6
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
building block to access the undecose moiety. Furthermore, the stereochemistry at C-8, C-9 and
C-10 of compound 15 was opposite to that in the herbicidins. This was attributed to the
availability of the starting materials: 1-deoxy-L-fructose derivative 15, prepared from naturally
occurring L-rhamnose, was more easily accessible than the respective 1-deoxy-D-fructose
derivative, which would be needed for the assembly of the herbicidin undecose with the correct
absolute configuration.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A molecular tethering approach followed by a radical cyclization was investigated by Sinaÿ and
2
2
co-workers. In this case, silaketal 16 was prepared by linking the respective D-glucose- and D-
xylose-derived precursors. This intermediate eventually underwent radical 8-endo-trig cyclization
to complete the 11-membered carbon chain. However, further studies toward the herbicidin
framework have not been reported.
16-21
The investigations by the group of Gallagher date back to 1988,
encompassing two different
strategies, namely a late- and an early-stage approach. First, the same aldehyde 13, which was
later used by Whiting and Haines, was combined with an enolate derived from ketone 17 to link
the C - and C -building blocks. Analogously, -bromoketone 18, prepared from D-glucose, was
5
6
connected in a similar fashion to aldehyde 19, wherein the nucleobase had already been installed.
Although Gallagher and co-workers were able to assemble the glycosylated tricyclic framework,
the C-11 position could not be adjusted to the right oxidation state.
The first total synthesis of a member of the herbicidin family and only synthesis of herbicidin B
13
to date was developed by Ichikawa and co-workers in 1999. A synthetic plan, based on an
early-stage glycosylation, was envisioned and later successfully realized by a samarium-diiodide-
mediated aldol-type C-glycosylation between phenylthioulose 20 and aldehyde 21. Further
transformations finally afforded herbicidin B (2). This route required heavy optimization of the
6
ACS Paragon Plus Environment

Page 7 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
protecting groups in order to favor an all-axial conformation of the substituents at C-8, C-9 and
C-10 in the C -building block 20.
6
In this article, we report the total synthesis of herbicidin C (3) and aureonuclemycin (9), its
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
hydrolyzed congener using a late-stage glycosylation approach. In particular, we describe two
different approaches toward the undecose moiety. Initially, we sought to combine C -building
4
block 22 with C -lactone 23 but this approach stalled due to stereochemical issues. The more
7
successful route involved C -carbohydrate 24, which was readily elongated to afford the
6
herbicidin undecose moiety. Synthetic challenges as well as stereochemical surprises of this
strategy are included in the discussion. Finally, the stereoselective directed glycosylation of the
C -sugar is also described in detail.
11
RESULTS AND DISCUSSION
Initially, we expected that herbicidin C (3) should be accessible through a challenging late-state
glycosylation on the sterically congested concave face of the herbicidin carbohydrate skeleton
20
(
Scheme 2). As a consequence, we envisioned the known undecose sugar 25 to be the desired
precursor.
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Retrosynthetic analysis of the initial approach toward herbicidin C (3).
In contrast to the combination of two separately prepared sugar units, as reported by Gallagher
2
0
and coworkers, we decided to assemble tetracycle 25 in a more straightforward fashion by a
double epoxide opening sequence from compound 27 via intermediate 26.
Alcohol 27 could stem from dihydropyran 28 and the respective ,-epoxyaldehyde by a
stereoselective addition of an organometallic species 28. This compound would be prepared from
lactone 23, which could, in turn, be accessed from diacetone glucose 29.
8
ACS Paragon Plus Environment

Page 9 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Synthesis of key intermediate 23 including an intramolecular Wittig reaction.
Initially, we aimed at preparing lactone 23 from commercially available diacetone glucose 29
(Scheme 3). It had already been reported that compound 29 can be esterified with 2-bromoacetyl
2
9
bromide to yield ester 30, the structure of which was confirmed by X-ray crystallography.
Subsequent regioselective acetonide deprotection with concomitant diol cleavage using periodic
acid afforded aldehyde 31, which underwent one-pot phosphonium salt formation/Wittig reaction
2
9
to yield the known unsaturated lactone 32. The originally reported low yield of 22% was
improved to 40% by using DBU as a hydrogen bromide scavenger instead of propylene oxide.
Subsequent hydrogenation with hydrogen gas and palladium on charcoal gave key lactone 23 in
quantitative yield. In addition, the identity of 23 was confirmed by X-ray analysis.
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 4. Improved route to intermediate 23 through an intermolecular Wittig reaction.
In order to improve the overall yield of lactone 23, we also investigated an alternative approach,
wherein the succession of steps was altered (Scheme 4). Selective deprotection of the primary
acetonide of diacetone glucose 29 followed by diol cleavage with periodic acid, afforded
aldehyde 33. The intermolecular Wittig reaction of aldehyde 33 with ylide 34 resulted in the
isolation of the Z-ester as the unsaturated lactone 32 in 39% yield, as well as E-ester 35 in 23%
3
0
yield. The former was reduced to 23 as shown above. The latter could be transformed into the
desired lactone 23 by a short reduction/saponification/lactonization sequence. Thus, the overall
yield from diacetone glucose 29 to lactone 23 was improved from 24% (intramolecular Wittig
route) to 51% (intermolecular Wittig route).
Next, we investigated the transformation of lactone 23 into a lithiated glycal species 37, which
could potentially be trapped with electrophiles to yield substituted dihydropyrans of type 38
(Scheme 5).
1
0
ACS Paragon Plus Environment

Page 11 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Conceptual transformation of lactone 23 to dihydropyran 38.
Typically, the preparation of organometallic compounds of type 37 has been achieved by direct
3
1-33
deprotonation of glycals.
Thus, lactone 23 was reduced with DIBAL-H to the corresponding
lactol 39 yielding a 1.9:1 mixture of diastereomers (Scheme 6). The major isomer could be
identified by X-ray crystallography. After careful optimization of the reaction conditions, we
found that dehydration of 38 could be achieved with mesyl chloride as the activating agent and
triethylamine as the base to yield dihydropyran 40.
Scheme 6. Preparation of dihydropyran 40 from lactone 23.
When this reaction was carried out at higher concentrations (> 24.5 mM) an undesired
dimerization of lactol 39 was observed (Scheme 7). After 20 h at 50 °C and a concentration of
1
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
85 mM, two of three possible diastereomeric dimers were isolated, namely unsymmetrical dimer
1
4
1 and C -symmetric dimer 42. As expected, the H-NMR spectrum of dimer 42 showed a
2
reduced signal set compared to that of dimer 41. The structure of the major product 42 was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
confirmed by X-ray crystallography. The second possible C -symmetric diastereomer was not
2
observed. Presumably, 41 and 42 are formed through nucleophilic attack of 39 onto its mesylate
or the corresponding oxonium ion.
Scheme 7. Undesired dimerization of lactol 39 in highly concentrated reaction medium.
With glycal 40 in hand, deprotonation with several bases was attempted (Scheme 8).
Disappointingly, only protonated starting material 40 was recovered after treatment with t-BuLi,
KOt-Bu/n-BuLi (Schlosser's base), LDA and HMDS bases and subsequent D O quench.
2
1
2
ACS Paragon Plus Environment

Page 13 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 8. Unsuccessful deprotonation attempts of glycal 40.
At this point, an alternative route to prepare organolithium species 37 from lactone 23 via
stannane 44 was pursued (Scheme 9). Initial attempts to deprotonate tricycle 23 resulted in the
opening of the lactone ring, and the respective carboxylic acid was isolated. However, trapping of
the enolate with N-phenyl-bis(trifluoromethanesulfonimide) gave a sensitive triflyl ketene acetal
that could be converted into stannane 44 without further purification. The trimethyl tin group was
then exchanged with lithium and the organometallic species so obtained reacted further with
aldehyde 45 to afford alcohol 46 as a 1.2:1 mixture of diastereomers. Unfortunately, we were
unable to carry out this reaction with the more highly functionalized, known ethyl ester derivative
3
4-36
of ,-epoxyaldehyde 22,
which was found to be extremely instable in our hands.
1
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 9. Synthesis of stannane 44 and organolithium addition to aldehyde 45.
Compound 46 is an advanced intermediate that represents a significant portion of the herbicidins.
Although all eleven carbon atoms of their undecose skeleton had been assembled and two of
three rings had been installed, stereoselective functionalization of C-6, C-7, C-8, C-9 and C-10 in
4
6 still presented a formidable task. In addition, the preparation of large amounts of material,
especially at the beginning of the sequence, was not practical and the vinyl stannane chemistry
proved to be capricious. As a result, we decided to switch our synthetic strategy.
We reasoned that the stereochemistry at C-6, C-8, C-9 and C-10 should be set at an early stage of
the synthesis since the functionalization of these carbons late in the game appeared to be more
challenging than originally anticipated. In particular, we felt that early establishment of the
stereocenter at C-6 would be beneficial. The problems encountered in Ichikawa’s synthesis of
herbicidin B at thus junction reinforced us in our belief that this would be crucial for the success
of the synthesis.
1
4
ACS Paragon Plus Environment

Page 15 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 10. Second Generation retrosynthetic analysis toward herbicidin C (3).
Similar to our first retrosynthetic strategy, a late-stage glycosylation would introduce the
nucleobase in precursor 47, whereby a neighboring benzoate at C-2 should control the
1
2
stereochemistry of the reaction (Scheme 10). Tricycle 47 could be derived from terminal alkene
48, which already possesses the right oxidation states at C-7 and C-11. This intermediate was
expected to stem from unsaturated ester 49 through asymmetric dihydroxylation and vinyl
Grignard addition, while compound 49 itself could be accessible from glucose 50 via axial C-
glycosylation and Grubbs' olefin cross-metathesis.
In the event, D-glucose (50) was converted to protected anhydro sugar 24 by a two-step protocol
1
2
37
(Scheme 11). Thus, dehydration of D-glucose (50) provided anhydro glucose 51, which was
3
8
directly tribenzylated to yield compound 24. Although this transformation was low yielding, the
reaction could be carried out on large scale to efficiently furnish multi-grams of bicycle 24. The
3
8-43
following Hosomi-Sakurai allylation afforded the known C-glycoside 52 in 66% yield,
1
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
wherein the stereochemical configuration of the C-6, C-8, C-9 and C-10 centers (herbicidin
nomenclature) in 52 had been successfully controlled.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 11. Synthesis of C-glycoside 52.
The next stage of the synthesis was the preparation of bis-silyl ether 55 in order to differentiate
the C-7 and C-11 alcohols, the oxidation states of which would be eventually adjusted (Scheme
44-46
1
2). To this end, compound 52 was selectively debenzylated through a two-step protocol
involving the formation of iodo ether 53 and subsequent elimination to diol 54. After TBS
protection, the desired intermediate 55 was obtained.
During these investigations, we found that permutation of the steps (TBS protection  56
followed by formation of iodo ether 57 and elimination) gives rise to carbohydrate building block
5
8, which could be, for example, protected as an acetate to have three out of four alcohols
orthogonally protected (Scheme 12). This building block might find further application in
carbohydrate chemistry.
1
6
ACS Paragon Plus Environment

Page 17 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 12. Preparation of silyl ether 55 and carbohydrate building block 58.
After modification of the pyran ring of the herbicidin backbone, we sought to establish the
framework of the furan ring and the cyclic hemiacetal. Thus, the allyl side chain of intermediate
4
7,48
55 was elongated by Grubbs' cross-metathesis with methyl acrylate,
which afforded
unsaturated ester 49 in excellent yield (Scheme 13). A subsequent highly diastereoselective
4
9
Sharpless dihydroxylation proved sluggish but could be was forced to completion by increasing
the osmium tetroxide concentration. Subsequent acetonide formation yielded methyl ester 59 as
an inseparable 14:1 mixture with its minor diastereomer. The diol protection under acidic
conditions had to be optimized due to the facile cleavage of the primary TBS ether in acidic
media.
1
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 13. Synthesis of ester 59.
In order to attach the remaining carbon atoms and install an additional stereocenter, ester 59 was
converted to aldehyde 60 using DIBAL-H (Scheme 14). As a side reaction, over-reduction with
DIBAL-H was observed, necessitating the complete reduction of ester 59 with LiAlH . At this
4
stage, it was possible to separate the diastereomeric alcohols 61 and 62. The major isomer 61
could be readily oxidized using Dess-Martin periodinane (DMP) or via Swern oxidation.
Although addition of vinylmagnesium bromide to the resulting aldehyde 60 did take place to
afford a mixture of two separable allyl alcohols 63 and 64, the diastereomeric ratio was low and
yields unreliable.
1
8
ACS Paragon Plus Environment

Page 19 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 14. Attempts to synthesize vinyl alcohols 63 and 64.
In search of a more efficient transformation of ester 59, we found that the most practical method
was the conversion of 59 to vinyl ketone 66, which could then be stereoselectively reduced by
reagent-controlled reactions, such as the Corey-Itsuno reduction (Scheme 15). The double bond
in 66 was intended to serve as a carbonyl equivalent. Thus, saponification of the diastereomeric
mixture of ester 59, provided the corresponding acid, which was immediately converted to
Weinreb amide 65. In order to access isomerically pure material for analysis, the diastereomers
were separated after reduction of 59 with LiAlH at the stage of alcohol 61, as described above.
4
Ley oxidation and subsequent formation of the Weinreb amide provided another practical route to
6
6
5. Reaction of compound 65 with vinylmagnesium bromide then gave the desired vinyl ketone
6 and considerable amounts of ketone 67. This side product, however, could be recycled to by
means of a Hofmann-type elimination, which gave a good overall yield of vinyl ketone 66.
1
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 15. Transformation of ester 59 and alcohol 61 into vinyl ketone 66.
The seemingly straightforward Corey-Itsuno reduction of vinyl ketone 66 proved to be more
1
2
challenging than anticipated. Reduction of the carbonyl group in 66 with BH and the R-
3
configured reagent (R)-68 gave a 4.4:1 mixture of compounds 63 and 64 in favor of the alcohol
with the undesired stereochemistry (Scheme 16). This result was unexpected since Corey-Itsuno
5
0
reductions are typically reagent-controlled and the generally accepted transition-state model
with (R)-68 reagent predicts the opposite outcome.
Scheme 16. Synthesis of the undesired diastereomer 63.
2
0
ACS Paragon Plus Environment

Page 21 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
The undesired stereochemistry of alcohol 63 was revealed only after further synthetic
transformations and establishment of the tricyclic compound 73 (Scheme 17, top). Protection of
the allylic alcohol in 63 and desilylation with TBAF afforded diol 69. In order to oxidize C-7 and
C-11, compound 69 was subjected to a variety of methods and reagents, such as DMP oxidation,
Pinnick oxidation, pyridinium dichromate oxidation (PDC), pyridinium chlorochromate (PCC),
and Ley oxidation (TPAP, NMO). Most of these furnished intractable mixtures containing little
or no desired product. Eventually, we successfully performed the oxidation using 2,2,6,6-
tetramethylpiperidine-1-oxyl (TEMPO) and (diacetoxyiodo)benzene (DAIB) in the presence of
water. Under these conditions, only the primary alcohol was oxidized to the corresponding
dicarboxylic acid, whereas the secondary hydroxy group remained unaffected. Ketone 70 was
then obtained after ester formation with (trimethylsilyl)diazomethane and subsequent DMP
oxidation. Cleavage of the acetonide with trifluoroacetic acid (TFA) resulted in a mixture of
isomeric hemiacetals with compound 71 as the major compound which was identified by X-ray
analysis. When this mixture was subjected to ozonolysis, the entire carbohydrate moiety
rearranged to give the tricyclic core 72, which was subsequently acylated at the sterically more
accessible hemiacetal, yielding undecose 73 as a single diastereomer in 57% yield over three
steps. The undesired stereochemistry at C-2 resulting from the Corey-Itsuno reduction could be
deduced from the NOESY 2D-NMR spectrum due to a cross-peak of protons H-2 and H-4 in
tricycle 73 and was later confirmed by the X-ray structure of the deprotected triol 74 (Scheme 17,
bottom).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 17. Establishment of the tricyclic core 73 (top) and X-ray structure of triol 74
(bottom).
Detailed analysis of the NMR spectra showed that the 6-membered chair in 70 had undergone
ring inversion during the cyclization process to 73. Whereas the coupling constant of protons H-8
and H-9 in 70 were J = 7.1 Hz, the corresponding coupling in tricycle 73 showed a value of
J = 3.4 Hz, an indication that both H-8 and H-9 now adopted equatorial positions.
In order to correct the stereochemistry at the C-2 position, we applied (S)-68, the enantiomer of
the previously used oxazaborolidine, in catalytic amounts (10 mol-%) (Scheme 18). The resulting
2
2
ACS Paragon Plus Environment

Page 23 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
allylic alcohols 64 and 63 were obtained in a diastereomeric ratio of 14:1 (determined by H-
NMR spectroscopy of the crude product mixture) and could be separated from the minor isomer
by column chromatography, affording 64 in 68% isolated yield.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Having established the undecose framework of herbicidin with the desired configuration at C-2 ,
we proceeded with the oxidation state adjustment and the formation of the heterotricyclic ring
system (Scheme 18). At this point, a benzoyl group was introduced at C-2 to later control the
stereochemical outcome of the N-glycosylation. Desilylation using TBAF yielded diol 75,
followed by the previously established oxidation/esterification sequence (TEMPO/DAIB,
TMSCHN , DMP) to furnish ketone 48. Application of our previously optimized cyclization
2
conditions (TFA; O ; Ac O) gave the desired tricycle 76 with the correct stereochemistry at C-2
3
2
and as a 2.8:1 mixture of anomeric acetates.
2
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 18. Establishment of the undecose skeleton 76.
With undecose 76 in hand, we accomplished the crucial late-stage N-glycosylation by using a
modified silyl-Hilbert-Johnson (also Vorbrüggen) protocol. Under these conditions,
glycosylamine 77 was readily obtained and the protected herbicidin C precursor 78 was isolated
in 53% yield over two steps (Scheme 19).
2
4
ACS Paragon Plus Environment

Page 25 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 19. Glycosylation of compound 78.
In order to complete the total synthesis of herbicidin C (3), the benzyl protecting groups of 77 or
8 had to be cleaved. However, this transformation was more problematic than originally
7
anticipated. First, compound 77 was subjected to a variety of debenzylation conditions [such as
H (6 bar), Pd/C, MeOH; H (8 bar), Pd(OH) /C, MeOH; H (8 bar), Pd(OH) /C, AcOH;
2
2
2
2
2
HCOONH , Pd(OH) /C, MeOH; H , Raney-Ni; FeCl ; BBr ; DDQ; Li, NH ], all of which proved
4
2
2
3
3
3
unsuccessful. Attempts to deprotect compound 78 under reductive conditions (H , Pd/C, MeOH;
2
H , Pd/C, EtOAc; H , PtO , MeOH) also failed. When lithium 4,4'-di-tert-butyldiphenyl (LDBB)
2
2
2
was applied to benzyl ether 78, a monobenzylated product was isolated in 20% yield (as indicated
1
by H-NMR spectroscopy and mass spectrometry).
Since final debenzylation could not be effected during the final stretch of the synthesis, we
decided to change the order of debenzylation and glycosylation (Scheme 20). To this end, the
2
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
benzyl groups in 76 were cleaved via hydrogenolysis. Subsequent acetylation then gave
pentaester 79 in 74% over two steps. Notably, the hemiacetal hydroxy group at the pyran ring
juncture remained unaffected under these conditions.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 20. Total synthesis of herbicidin C (3) and aureonuclemycin (9).
To our relief, Vorbrüggen-glycosylation under the previously established conditions also worked
with substrate 79, yielding adenyl nucleotide 80 as the only isolable diastereomer in 55% yield.
Global deprotection using NaOMe/MeOH gave herbicidin C (3) whose structure was confirmed
by detailed spectroscopic analysis, including an NMR titration experiments with an authentic
sample of the natural product. Further saponification of the base-labile natural product 3 under
mild conditions finally yielded aureonuclemycin (9), another member of the herbicidin family.
2
6
ACS Paragon Plus Environment

Page 27 of 70
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
CONCLUSION
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In summary, we have presented two different strategies toward the herbicidins, which are both
based on a late-stage glycosylation as the key step. Our initial yielded lactone 23, which already
includes seven carbon atoms of the undecose chain, from inexpensive carbohydrate starting
materials. However, this approach suffered from weak stereocontrol and the need to employ toxic
organotin reagents. By contrast, the second route turned out to be more robust and large
quantities of the undecose chain could be procured. Although we had to overcome unexpected
stereochemical problems, we eventually established the correct absolute configuration of the
undecose moiety by a stereoselective C-glycosylation and several reagent-controlled
transformations. One further synthetic obstacle regarding the protecting group chemistry needed
to be solved before the highly diastereoselective N-glycosylation completed the total synthesis of
the natural products herbicidin C (3) and aureonuclemycin (9). The realization of a late-stage
glycosylation strategy provides flexibility and new pathways for the fast preparation of herbicidin
derivatives, e.g. for structureactivity relationship studies.
2
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
EXPERIMENTAL SECTION
(3aR,5R,6S,6aR)-5-[(4R)-2,2-dimethyl-1,3-dioxolan-4-yl]-2,2-dimethyl-tetrahydro-2H-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29
furo[2,3-d][1,3]dioxol-6-yl 2-bromoacetate (30): A stirred solution of diacetone-D-glucose 29
(15.0 g, 57.6 mmol, 1.0 eq), DMAP (70.4 mg, 0.576 mmol, 10 mol%), and pyridine (6.96 mL,
8
6.4 mmol, 1.5 eq) in CH Cl (125 mL) was cooled to 0 °C, and 2-bromoacetyl bromide was
2
2
added dropwise. The reaction mixture was stirred for 40 min at 0 °C, then quenched at 0 °C with
water (1.2 mL) and allowed to warm to room temperature. After stirring for an additional 15 min
at room temperature, the resulting solution was diluted with EtOAc (200 mL). The organic phase
was washed with water (2 x 30 mL) and brine (30 mL), dried (MgSO ) and concentrated in
4
vacuo. Flash column chromatography [CH Cl :acetone 99:1] provided bromo ester 30 (21.3 g,
2
2
ꢃꢄ
5
6.0 mmol, 97%) as a white solid: R 0.39 [Petrol:EtOAc 4:1]; ꢀꢁ 39.6 (c = 0.93, MeOH);
ꢂ
f
1
mp 51 – 53 °C; H-NMR (400 MHz, CDCl ) δ 5.90 (d, J = 3.7 Hz, 1H), 5.32 (d, J = 2.8 Hz, 1H),
3
4
4
.51 (d, J = 3.7 Hz, 1H), 4.27 – 4.19 (m, 2H), 4.11 (dd, J = 8.8, 5.6 Hz, 1H), 4.01 (dd, J = 8.7,
.5 Hz, 1H), 3.87 (d, J = 12.3 Hz, 1H), 3.84 (d, J = 12.3 Hz, 1H), 1.52 (s, 3H), 1.41 (s, 3H), 1.32
1
3
(s, 3H), 1.31 (s, 3H); C-NMR (100 MHz, CDCl ) δ 166.1, 112.6, 109.7, 105.2, 83.2, 80.0, 77.8,
3
7
1
2.4, 67.6, 27.0, 26.9, 26.4, 25.4, 25.4; IR (ATR) ṽ 2983, 1769, 1741, 1383, 1268, 1205, 1135,

1
+
069, 1018, 841 cm ; HRMS (ESI) calcd for C H BrNO
398.0809, found 398.0806
14 25
7
+
[
M+NH ] .
4
(3aR,5S,6S,6aR)-5-formyl-2,2-dimethyl-tetrahydro-2H-furo[2,3-d][1,3]dioxol-6-yl
2
9
2-bromoacetate (31): A solution of bromo ester 30 (6.71 g, 17.6 mmol, 1.0 eq) in Et O/MeOH
2
(
9:1, 50 mL) was cooled to 0 °C, and formic acid (25 mL) followed by periodic acid (4.81 g,
1.2 mmol, 1.2 eq) were added. The reaction mixture was stirred at 0 °C for 10 min, then allowed
to warm to room temperature and stirred for an additional 30 min until complete consumption of
2
2
8
ACS Paragon Plus Environment

Page 29 of 70
The Journal of Organic Chemistry
1
2
3
4
5
the starting material (as indicated by TLC analysis). The mixture was diluted with EtOAc
(
100 mL) and the organic phase was washed with water (2 x 20 mL), aq. NaHCO (3 x 30 mL of
3
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a saturated solution), aq. Na S O (3 x 10 mL of a saturated solution), dried (MgSO ) and
2
2
3
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
concentrated in vacuo to provide aldehyde 31 (3.44 g, 11.1 mmol, 63%) as a white solid: R 0.24
f
ꢃꢄ
1
[PE:EtOAc 1:1] (streaking); ꢀꢁ 17.3 (c = 1.0, MeOH); mp 66 – 67 °C; H-NMR (400 MHz,
ꢂ
CDCl ) δ 9.65 (s, 1H), 6.09 (d, J = 3.5 Hz, 1H), 5.53 (dd, J = 3.4, 0.3 Hz, 1H), 4.74 (dd, J = 3.4,
3
0
.8 Hz, 1H), 4.59 (d, J = 3.5 Hz, 1H), 3.78 (d, J = 12.5 Hz, 1H), 3.76 (d, J = 12.9 Hz, 1H), 1.51
1
3
(s, 3H), 1.33 (s, 3H); C-NMR (100 MHz, CDCl ) δ 197.1, 166.1, 113.3, 105.6, 83.2, 82.8, 78.6,
3
1
2
6.9, 26.4, 24.7; IR (ATR) ṽ 3456, 2987, 1742, 1376, 1263, 1160, 1013, 851 cm ; HRMS (ESI)
+
+
calcd for C H BrNO 326.0234, found 326.0234 [M+NH ] .
1
0
17
6
4
2
,6
(
1S,2R,6R,8R)-4,4-dimethyl-3,5,7,12-tetraoxatricyclo[6.4.0.0 ]dodec-9-en-11-one (32): A
solution of aldehyde 31 (200 mg, 0.647 mmol, 1.0 eq) and PPh (170 mg, 0.647 mmol, 1.0 eq) in
3
MeCN (2 mL) was stirred at room temperature for 9 h. 1,8-Diazabicyclo[5.4.0]undec-7-ene
(98.5 mg, 0.647 mmol, 1.0 eq) was added, and the mixture was stirred at room temperature for an
additional 24 h, before it was diluted with Et O (30 mL). The organic phase was washed with
2
water (2 x 20 mL), dried (MgSO ) and concentrated in vacuo. Flash column chromatography
4
[
CH Cl :acetone 99:1] afforded unsaturated lactone 32 (55.5 mg, 0.262 mmol, 40%) as a white
2 2
ꢃꢄ
1
solid: R 0.59 [PE:EtOAc 1:1]; ꢀꢁ +28.9 (c = 0.56, MeOH); mp 70 °C; H-NMR (600 MHz,
f
ꢂ
CDCl ) δ 6.96 (dd, J = 9.8, 5.7 Hz, 1H), 6.23 (d, J = 9.8 Hz, 1H), 6.02 (d, J = 3.7 Hz, 1H), 4.82
3
(d, J = 3.8 Hz, 1H), 4.81 (d, J = 3.2 Hz, 1H), 4.62 (dd, J = 5.7, 3.1 Hz, 1H), 1.53 (s, 3H), 1.35 (s,
1
3
3H); C-NMR (150 MHz, CDCl ) δ 160.9, 138.8, 125.5, 112.7, 105.4, 84.1, 82.6, 67.7, 26.9,
3
1
2
6.3; IR (ATR) ṽ 2924, 1729, 1384, 1212, 1068, 1017, 888, 826 cm ; HRMS (ESI) calcd for
+
+
C H NO 230.1023, found 230.1022 [M+NH ] .
10
16
5
4
2
9
ACS Paragon Plus Environment