A convergent approach toward fidaxomicin: Syntheses of the fully glycosylated northern and southern fragments

Article abstract of DOI:10.1016/j.tet.2020.131673

Efficient approaches that enable the synthesis of analogs of natural product antibiotics are needed to keep up with the emergence of multiply-resistant strains of pathogenic organisms. One promising candidate in this area is fidaxomicin, which boasts impressive in vitro anti-tubercular activity but has poor systemic bioavailability. We designed a flexible synthetic route to this target to enable the exploration of new chemical space and the future development of analogs with superior pharmacokinetics. We developed a robust approach to each of the key macrocyclic and sugar fragments, their union via stereoselective glycosylation, and a convergent late-stage macrolide formation with fully glycosylated fragments. Although we were able to demonstrate that the final Suzuki cross-coupling and ring-closing metathesis steps enabled macrocycle formation in the presence of the northern resorcylic rhamnoside and southern novioside sugars, these final steps were hampered by poor yields and the formation of the unwanted Z-macrocycle as the major stereoisomer.

Full text of DOI:10.1016/j.tet.2020.131673

Journal Pre-proof
A convergent approach toward fidaxomicin: syntheses of the fully glycosylated
northern and southern fragments
Ryan Hollibaugh, Xueliang Yu, Jef K. De Brabander
PII:
S0040-4020(20)30886-3
https://doi.org/10.1016/j.tet.2020.131673
TET 131673
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 2 September 2020
Revised Date: 6 October 2020
Accepted Date: 7 October 2020
Please cite this article as: Hollibaugh R, Yu X, De Brabander JK, A convergent approach toward
fidaxomicin: syntheses of the fully glycosylated northern and southern fragments, Tetrahedron, https://
doi.org/10.1016/j.tet.2020.131673.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Elsevier Ltd. All rights reserved.

Graphical Abstract
A convergent approach toward fidaxomicin: Synthesis of
the fully glycosylated northern and southern fragments
Ryan Hollibaugh, Xueliang Yu, and Jef K. De Brabander
Leave this area blank for abstract info.
Department of Biochemistry, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard,
Dallas, Texas 75390-9038, United States

Graphical Abstract
A convergent approach toward fidaxomicin: Synthesis of
the fully glycosylated northern and southern fragments
Ryan Hollibaugh, Xueliang Yu, and Jef K. De Brabander
Leave this area blank for abstract info.
Department of Biochemistry, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard,
Dallas, Texas 75390-9038, United States

1
Tetrahedron
journal homepage: www.elsevier.com
A convergent approach toward fidaxomicin: syntheses of the fully glycosylated
northern and southern fragments
Ryan Hollibaugh, Xueliang Yu, and Jef K. De Brabander
Department of Biochemistry, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9038, United States
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
Efficient approaches that enable the synthesis of analogs of natural product antibiotics are
needed to keep up with the emergence of multiply-resistant strains of pathogenic organisms.
One promising candidate in this area is fidaxomicin, which boasts impressive in vitro anti-
tubercular activity but has poor systemic bioavailability. We designed a flexible synthetic route
to this target to enable the exploration of new chemical space and the future development of
analogs with superior pharmacokinetics. We developed a robust approach to each of the key
macrocyclic and sugar fragments, their union via stereoselective glycosylation, and a
convergent late-stage macrolide formation with fully glycosylated fragments. Although we
were able to demonstrate that the final Suzuki cross-coupling and ring-closing metathesis steps
enabled macrocycle formation in the presence of the northern resorcylic rhamnoside and
southern novioside sugars, these final steps were hampered by poor yields and the formation of
the unwanted Z-macrocycle as the major stereoisomer.
Available online
Dedicated to Prof. Tomas Hudlicky in honor of his
70th birthday
Keywords:
Lipiarmycin
Tiacumicin
Macrolide antibiotics
Ring closing metathesis
2009 Elsevier Ltd. All rights reserved
.
———
Corresponding author. Tel.: +1-214-648-7808; fax: +1-214-648-0320; e-mail: jef.debrabander@utsouthwestern.edu

2
Tetrahedron
years.10,11 Four syntheses of the macrolide aglycone and two
1. Introduction
syntheses of the natural product have been reported since
2015.12,13 Additionally, we and others have completed syntheses
of the structurally related mangrolides A and D, which share the
macrolactone core of fidaxomicin (see scheme 1).14 Here we
report our approach to the synthesis of fully glycosylated
northern and southern fragments of fidaxomicin, and explore
their union via Suzuki cross-coupling and ring-closing
metathesis.
Fidaxomicin (1) is an FDA approved macrolide antibiotic1
with modest in vitro efficacy against Gram-positive bacteria and
notably potent activity against Clostridium difficile (MIC50 = 31
µg/mL)2 and Mycobacterium tuberculosis (MIC50 = 15 µg/mL)
including drug resistant strains (MIC50 = 8–45 µg/mL).3
A
recently obtained cryo-EM structure of fidaxomicin bound to
RNA-polymerase4 has provided insight on the binding mode and
biochemically important structural features of fidaxomicin.
2. Results and Discussion
Despite its promising anti-tubercular activity, approved
clinical use of fidaxomicin is limited to treatment of C. difficile
associated colitis.5 Unfortunately, fidaxomicin has poor oral
bioavailablity and is rapidly cleared following alternative dosing
modes.6 These limitations preclude its use for the systemic
treatment of tuberculosis infections. Thus, a fidaxomicin analog
with improved bioavailability could represent a new class of
antibiotics for the treatment of systemic bacterial infections,
including tuberculosis and multidrug-resistant tuberculosis.
Semisynthetic approaches have been severely hampered by the
instability of the macrolide, which is readily available by
fermentation,7 to standard acid mediated glycolysis methods.8
Several groups have reported the structure and bioactivity of
biosynthetically prepared analogs.9 However, a robust medicinal
chemistry program would enable access to a wider variety of
structural modifications than is generally possible using
bioengineering approaches. In this context, fidaxomicin has
gained significant attention as a synthetic target in recent
Our synthesis was designed to be as convergent and modular
as possible to expedite subsequent syntheses of natural product
analogs. As outlined in Scheme 1, we envisioned unification of
northern and southern fragments 2 and 3 by a sequential Suzuki
cross coupling and ring closing metathesis (RCM) sequence. This
strategy was adapted from our longstanding work on the
mangrolide family.15 The Gademann group also utilized these
disconnections in their synthesis of both fidaxomicin and the
mangrolides.12c,14c,14d However, where they appended the
rhamnoside unit post macrolactone formation, we endeavored to
pursue the unification of fully glycosylated northern and southern
fragments. This would, in principle, allow us to reduce the
number of linear manipulations at the end of the sequence. To
that end we needed to develop a β-selective rhamnosylation of
northern macrolide fragment 5 with a fully functionalized
resorcyl rhamnoside 4. Southern fragment 3 on the other hand,
would
be
Scheme 1: Structures of fidaxomicin, mangrolide A, and mangrolide D and retrosynthetic strategy towards Fidaxomicin
accessible via the previously described Koenigs-Knorr
glycosylation of southern macrolide fragment 6 with novioside
7.13a Our initial objective was to develop robust, scalable routes
towards fragments 4-7 and explore their unification via
stereoselective glycosylation.
D
-mannose was derivatized as a benzyl ether, followed by
masking the 3- and 4-hydroxyl groups as the butanediacetal 8 in
55% yield for the two steps.16
Garegg-Samuelsson
A
iodination/hydrogenolysis sequence removed the 6-hydroxyl
group to provide rhamnoside 9 in good yield. Methylation of the
remaining alcohol proceeded with excellent yield to give 10 and
was followed by final cleavage of the butanediacetal under mild
hydrolytic conditions to provide up to nine grams of key
rhamnoside 11 in a single batch.
Our forward synthesis began with the preparation of
resorcylated D-rhamnoside donor 4 (Scheme 2A). In an adaption
of the procedure reported by Gademann and coworkers for the
corresponding anomeric methyl ether, the anomeric position of

3
We have previously described both base mediated and
photochemical methods for the preparation of hindered salicylate
and resorcylate esters and amides using dioxinones similar to 14a
and 14b.17 In order to explore these approaches for the
introduction of the homoorsellinate residue we prepared both 14a
and 14b starting with trihydroxybenzoic acid derivatives 12a and
12b (Scheme 2B). Selectively etherification of the para-phenol
under Mitsunobu conditions, followed by triflation of the
remaining ortho-phenol under standard conditions efficiently
provided 13a and 13b.17c,18 Subsequent Stille cross-coupling with
tributyl(vinyl)tin proceeded in excellent yields, and the resultant
styrenes could be selectively reduced in the presence of the
benzyl ether using a NiCl2/sodium borohydride system19 to
provide 14a or 14b on multigram scale.
Sodium hydride promoted acylation of 11 with 14a proceeded
with moderate thermodynamic selectivity for the desired 4-acyl
isomer 16 (31% yield of 15, 52% yield of 16, Scheme 2C).17a,20
Resubmission of 15 to the sodium hydride acylation conditions
provided additional 16. By contrast, photochemical acylation of
11 with 14b proceeded in lower yield, and afforded a kinetic
mixture of acylated products 15 and 16 in which the undesired
regioisomer 15 dominated (37% and 20% isolated yield
respectively).17b,c Using the superior sodium hydride mediated
acylation of rhamnoside 11 with 14a, we prepared up to a gram
of 16 in one batch. In preparation for glycosylation of the
northern fragment (vide infra), intermediate 16 was sequentially
debenzylated via hydrogenolysis, chlorinated using sulfuryl
chloride, and carefully tris-acetylated to afford the fully
functionalized rhamnoside 4 in good overall yield.
Our synthesis of northern macrolide fragment 5 initiates with
a stereoselective synthesis of borylated homopropargyl alcohol
19 and tracks with that of previous reports.12a,13a As shown in
Scheme 3A, Sharpless epoxidation and silylation of 1-butene-3-
ol yielded enantiomerically enriched 17 (38% yield for 2 steps).21
Epoxide opening with propynyl lithium occurred as expected to
afford the desired alkyne 18 in high yield.12a We found that regio-
and stereoselective borylation of the alkyne was most efficient
using conditions developed by the Carretero group,22 providing
the needed secondary alcohol 19 with high yield and excellent
selectivity on gram scale.
Scheme 2. Synthesis of orsellinic rhamnoside 4. Reagents and conditions: (a)
BnOH, HCl; (b) (MeO)3CH, CSA, butanedione, MeOH; then Et3N (55%, 2
steps); (c) I2, PPh3, imidazole, THF (84%); (d) H2, Pd/C, MeOH, Et3N (90%);
(e) NaH, MeI, THF (>95%); (f) TFA, DCM, H2O; Et3N (69%); (g) BnOH,
PPh3, DIAD, THF; (h) pyridine, Tf2O (13a: 61%, 2 steps; 13b: 65%, 2 steps);
(i) Pd(PPh3)4, LiCl, dioxane, Et3N, tributyl(vinyl)tin; (j) NiCl2, NaBH4,
MeOH (14a: 66%, 2 steps; 14b: 58%, 2 steps); (k) 14b, 11, 300 nm UV light
(15: 37%; 16: 20%); (l) 14a, 11, NaH, THF (15: 31%; 16: 52%); (m) NaH,
THF (54%); (n) H2, Pd(OH)2, MeOH; (o) SO2Cl2, EtOAc; (p) Ac2O, DMAP,
EtOAc (66%, 3 steps).
To prepare the dienoic acid 23, we adapted the hydrogen
mediated C-C coupling method developed by the Krische group
(Scheme 3B).23 Known enyne 20,24 available in two steps from
propargyl alcohol, was coupled with aldehyde 21 to provide the
allylic alcohol with excellent regio- and stereoselectivity for the
desired E-diene. The resulting crude alcohol was then oxidized to
dienone 22 in 70% yield for the two step sequence. Activation of
the N-methyl imidazole with methyl triflate was followed by
potassium trimethylsilanolate mediated hydrolysis to generate the
dienoic acid 23 in good yield (82%, 2 steps).25 It is worth noting
that this alternative synthesis of dienoic acid 23 (6 steps, E/Z
In the event, ethyl 2-(hydroxymethyl)acrylate was
dibrominated and protected as the THP ether, giving dibromide
25 in excellent yield for the two-step process (78% overall).26,27
Stereoselective
elimination
of
the
α-bromide
with
tetrabutylammonium fluoride in HMPA provided acryloyl
bromide 26 in 85% yield and excellent stereoselectivity (>20:1
E/Z). Stille coupling of vinyl bromide 26 with tributyl(vinyl)tin
yielded dienoate 27 in 93% yield.28 Using this route, we were
able to prepare multigram quantities of 27 in 62% overall yield
over four steps from commercially available ethyl 2-
(hydroxymethyl)acrylate. Hydrolysis of ester 27 afforded the
dienoic acid 28 (>95%), which was esterified as before with
alcohol 19. This time, the THP-protected primary alcohol could
be selectively removed with ethanolic PPTS to yield the target
northern fragment alcohol 5 in 45% yield (2 steps) with no
observable degradation of the TBS or BPin functionalities.
>20:1, 49% overall yield) compares favorably
to those
previously reported for this or related dienoates.12a-c Acid 23 was
esterified with alcohol 19 using previously described Yamaguchi
conditions to generate known bis-silyl ester 24.13a,14b Conditions
to selectively cleave the primary TBS group in the context of an
intact macrocycle have been reported.8b,13a Unfortunately, we
were unable to identify conditions to effect
a selective
deprotection of the primary silyl ether on the borylated ester 24,
and therefore pursued the synthesis of dienoic acid 28 according
to a protocol adapted from Zhu and coworkers (Scheme 3C).12b

4
Tetrahedron
ether in 80% yield.30 Periodinane mediated oxidation of the
primary alcohol gave aldehyde 37. A diastereoselective Brown
allylation was followed by TBS protection of the resultant
alcohol to afford intermediate 38 in 65% yield and excellent
stereoselectivity (>20:1 dr). A final boron trifluoride mediated
cleavage of the methoxymethyl ether afforded target secondary
alcohol 6 in 77% yield.
Scheme 3. Synthesis of northern fragment borylated ester 5. Reagents and
conditions: (a) Ti(OiPr)4,
D-DIPT, DCM, TBHP, 4Å MS; (b) TBSCl,
imidazole (38%, 2 steps); (c) propyne, nBuLi, BF3•OEt2 (90%); (d) CuCl,
PCy3, NaOtBu, B2Pin2, MeOH, PhMe (81%); (e) rac-BINAP, Rh(COD)2OTf,
21, Ph3CCOOH, H2, DCE; then MnO2, DCM (70%); (f) MeOTf, MeCN; then
KOTMS, Et2O (82%); (g) 19, Et3N, DMAP, 2,4,6-trichlorobenzoyl chloride,
PhMe (60%); (h) Br2, CHCl3 (93%); (i) dihydropyran, PPTS, DCM (84%); (j)
TBAF, HMPA (85%, >20:1 dr); (k) Pd2(dba)3, P(2-furyl)3, tributyl(vinyl)tin,
THF (93%); (l) LiOH, H2O, THF, EtOH (>95%); (m) 19, Et3N, DMAP,
2,4,6-trichlorobenzoyl chloride, THF; (n) PPTS, EtOH (45%, 2 steps).
We next turned our focus to the preparation of southern
fragment 6 (Scheme 4). Our work on mangrolide A disclosed a
route to polyketide fragment
6
utilizing
a
2,3-Wittig
rearrangement29 to establish the C10-C11 stereochemistry and a
late stage bis-functionalization of an alkyne to prepare the
required E-vinyl iodide.14b,15
A related rearrangement has
subsequently been utilized by the Roulland group to elegantly set
this same stereochemical diad.12d As shown in scheme 4A, our
route commenced with
L-lactate-derived aldehyde 29. Wittig
Scheme 4. Two alternative syntheses of southern fragment 6. Reagents and
conditions: (a) nPrPPh3Br, nBuLi, THF (83%); (b) pTsOH, MeOH; (c) 31,
tBuOK, THF, (76%, 2 steps); (d) nBuLi, THF; then MOMCl (88%); (e) O3,
DCM; then Me2S (93%); (f) 34, toluene (85%); (g) DIBAL, DCM; (h) TBAF,
THF; (i) TESOTf, lutidine, DCM (76%, 3 steps); (j) CuCN, (Bu3Sn)2, nBuLi,
MeI; then I2, THF (80%); (k) DMP, DCM (91%); (l) (–)-Ipc2B-allyl, Et2O;
then H2O2, NaOH; (m) TBSOTf, lutidine, DCM (65%, 2 steps); (n) BF3•OEt2,
Me2S (77%); (o) [Ir(COD)Cl]2, (R)-Cl-MeO-BIPHEP, Cs2CO3, mNO2BzOH,
allyl acetate, THF (57%, >20:1 dr).
homologation of the aldehyde afforded Z-olefin 30 in 83% yield
(Z/E > 20:1). Methanolysis (pTSOH) of the THP-ether was
followed by alkylation with propargyl bromide 31 (KOtBu, THF)
to afford [2,3]-Wittig precursor 32 in 76% (2 steps).
Deprotonation with nBuLi in THF induced
a
smooth
rearrangement to the corresponding alcohol which was trapped in
situ with methoxymethyl chloride to provide 33 in 88% yield and
with excellent stereoselectivity. The relative stereochemistry was
predicted to result from [2,3]-rearrangement occurring via cyclic
TTS-A. The stereochemistry was unambiguously assigned via
comparison of fragment 6 resulting from this route, to the
identical material reported previously by us and others.12c,14a
Careful selective ozonolysis of the alkene in 33 (Me2S work-up)
was followed by homologation of the resultant aldehyde with
stabilized ylide 34 to afford E-configured α,β-unsaturated ester
35 in 79% yield (2 steps). To set the stage for subsequent alkyne
functionalization, ester 35 was reduced, followed by a TIPS
deprotection, and TES protection sequence. The resultant
terminal alkyne 36, obtained in 76% for this 3-step sequence, was
subjected to a one-pot stannocupration/methylation/iodination to
yield the vinyl iodide with simultaneous cleavage of the TES
During the course of our synthesis of mangrolide D, we
developed an alternative approach to 6 utilizing an Oppolzer
sultam mediated aldol reaction (Scheme 4B).12a,14a For this work,
we exploited primary alcohol 39 to explore a Krische-type
iridium catalyzed allylation from the alcohol oxidation state to
further streamline this synthetic sequence. Gratifyingly, allylic
alcohol 39 directly provided homoallylic alcohol 40 with
excellent stereoselectivity (>20:1 dr) using the reported
conditions.31 A straightforward protection/deprotection sequence
then revealed key intermediate 6, which was identical to material
produced according to Scheme 4A and previous syntheses.12c,14a

5
acetone simultaneously accomplished an isomerization to the
furanose, anomeric methyl ketal formation, and ketalization of
the cis-diol to provide compound 41 in 75%.32 The primary
alcohol was then oxidized to the carboxylic acid and alkylated to
form the corresponding methyl ester (~65% overall).33 Addition
of excess methylmagnesium bromide afforded tertiary alcohol 42
(>95%). By this four-step sequence up to 9.5 g (41% overall
yield) of material was prepared in a single batch with no
purifications. This method compared favorably to the previous
six-step approach from O-methyl mannose reported by
Kaufmann.13a From 42 onwards our route to 3 followed the
sequence reported by Kaufmann et al.13a with some minor
modifications. Isomerization of 42 to the α- and β-pyranoses,
44a and 44b respectively, was performed in a mixture of
trifluoroacetic acid and methanol. All four possible isomers of
deprotected material were individually isolated and characterized
along with recovered starting material 42 in the indicated
quantities. In two subsequent iterations, furanose isomers
43a and 43b and recovered 42 were combined and resubmitted to
the reaction conditions to generate a similar distribution of
products. The total combined yield of isolated 44a and 44b was
70%. We next took advantage of an improved acylation protocol
reported by Kaufmann8b to install the cyclic carbonate and the
isobutyrate, which provided noviosides 7a and 7b from 44a and
44b respectively (~76-82% yield). Glycosylation of 6 with
unstable glycosyl bromide 45 was performed using a slight
modification of the previously reported method.13a Thus
treatment of 7b with HBr in acetic acid afforded the unstable
Scheme 5. Synthesis of protected novioside 7 and fully glycosylated southern
fragment 3. Reagents and conditions: (a) HCl, MeOH, acetone (75%); (b)
TEMPO, KBr, NaHCO3, NaOCl, EtOAc, H2O (>95%); (c) MeI, K2CO3,
DMF (66%); (d) MeMgBr, THF (>95%); (e) TFA, MeOH, H2O; (f) TFA,
MeOH, H2O (70% yield of 44a and 44b after 3 cycles); (g) 44a or 44b, CDI,
THF; (h) (iPrCO)2O, Et3N, DMAP, DCM (76% from 44a, 82% from 44b
over 2 steps); (i) 7b, HBr, AcOH, DCM; (j) 6, HgBr2, HgO, 4Å MS, DCM;
then Et3N (67%, 2 steps, >2:1 β:α).
glycosyl bromide 45, which was isolated but not purified.35
A
Preparation of the southern protected novioside sugar 7
Helferich type Koenigs-Knorr glycosylation of 6 with crude 45
provided the desired β-anomer 3 as the major product and with
commenced with commercially available
D-lyxose (Scheme 5).
Treatment of this material with HCl in a mixture of methanol and
better
than
2:1
stereoselectivity.36
Scheme 6. Synthesis of fully glycosylated northern fragment 2, cross-coupling with fully glycosylated southern fragment 5, and exploration of the final steps.
Reagents and conditions: (a) 4, HBr, AcOH, DCM; then 5, Ag2CO3, 4Å MS, DCM (35% + 44% recovered 5, >3:1 β:α); (b) 46, K2CO3, MeOH (2: 35%; 47:
34%); (c) 47, K2CO3, MeOH (48%); (d) TlOEt, Pd(PPh3)4, THF, H2O (38%); (e) Grubbs II, PhMe; (f) NaH, ethylene glycol, THF; (g) HF•Et3N, THF.
Glycosylation of the northern fragment 5 was best performed
with the crude glycosyl bromide derived from acetate 4 (HBr,
HOAc, DCM) and mediated by silver carbonate to yield the
desired β-anomer 46 in 35% isolated yield (α/β of crude mixture
>3:1) along with recovered alcohol 5 (44%). This approach
differs from previously reported methods for the installation of
this challenging β-mannose type glycosidic bond in the
fidaxomicin literature.37 Methanolysis of the phenolic acetates
(K2CO3, MeOH) provided target fragment 2 (32%) along with an
equal amount of monoacetate 47 (34%). The reaction time had to
be kept short to avoid degradation of the sensitive pinacol ester.
Purified 47 could be resubmitted to the same reaction conditions

6
Tetrahedron
to furnish additional 2. In contrast to previous syntheses, we
methods for the preparation of the corresponding glycosides.
Finally, though we demonstrated that the cross-coupling and
RCM of fully glycosylated fragments was possible, the yields
were adversely impacted because the sugar decoration in the
northern fragment reversed the E/Z-selectivity of the RCM
macrocyclization, and these isomers were unfortunately no
longer separable.
sought to explore macrolide formation using this fully
glycosylated northern fragment 2. Gratifyingly, Suzuki coupling
of this material with glycosylated southern fragment 3 using
conditions described for coupling of a non-glycosylated northern
fragment (cf. 24) provided the desired hindered diene 48, albeit
with lower efficiency (38% versus published 88% for coupling of
3 with 24).13a
Conflicts of interest
All that remained was a penultimate ring closing olefin
metathesis (RCM) and removal of the carbonate and silyl
protecting groups. The ring closing metathesis of a mono-
glycosylated (at C11),13a or a non-glycosylated12b seco-macrolide
diene was reported to yield separable mixtures of E/Z-
macrocycles with the desired natural E-isomer dominating
(~2:1), and the unwanted Z-isomer recyclable after separation
and resubjecting to the RCM conditions. Unfortunately, similar
conditions with fully glycosylated seco-macrolide diene 48 led to
an inseparable 1:2 mixture of double bond isomers with the
unnatural Z-isomer dominating as determined by 1H NMR of the
crude reaction product. Exploration of other RCM catalysts did
not provide a satisfying solution. These results underscore the
unpredictable influence of peripheral protecting groups and
functionality on double bond geometry in the context of ring
closing metathesis leading to larger ring systems.38,39
The authors declare no conflict of interest.
Acknowledgments
This work was supported by the NIH (grant GM111329) and
the Robert A. Welch Foundation (grant I-1422). J. K. De
Brabander holds the Julie and Louis Beecherl, Jr., Chair in
Medical Science. HRMS data were obtained from the Shimadzu
Center for Advanced Analytical Chemistry (SCAAC) at UT
Arlington.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https:///doi.org/
While these results were a setback, we hoped to be able to
separate the E/Z isomers at a later stage. At this point, we had
only small amounts of crude RCM product mixtures available
and subsequent reactions were performed on sub-milligram scale
with product mixtures characterized primarily by crude NMR and
mass spectrometry, but in insufficient quantities to support a
more rigorous structural assignment. Nevertheless, the crude
RCM mixture was subjected to carbonate cleavage and fluoride-
mediated silyl ether cleavage as described.8b,13a Analysis of the
crude reaction product by TLC and HPLC revealed three
separable products. Each of these three products was isolated in
low yield (<15%) by preparative RP-HPLC and independently
References
1. (a) Erb, W.; Zhu, J. Nat. Prod. Rep. 2013, 30, 161-174. (b)
McAlpine, J. B. J. Antibiot. 2017, 70, 492-494.
2. Goldstein, E. J. C.; Babakhani, F.; Citron, D.M. Clin. Infect. Dis.
2012, 55, S143-S148.
3. Kurabachew, K.; Lu, S. H. J.; Krastel, P.; Schmitt, E. K.; Suresh,
B. L.; Goh, A.; Knox, J. E.; Ma, N. L.; Jiricek, J.; Beer, D.;
Cynamon, M.; Petersen, F.; Dartois, V.; Keller, T.; Dick, T.;
Sambandamurthy, V. K. J. Antimicrob. Chemother. 2008, 62, 713-
719.
4. Lin, W.; Das, K.; Degen, D.; Mazumder, A.; Duchi, D.; Wang, D.;
Ebright, Y. W.; Ebright, R. Y.; Sineva, E.; Gigliotti, M.;
Srivastava, A.; Mandal, S.; Jiang, Y.; Liu, Y.; Yin, R.; Zhang, Z.;
Eng, E. T.; Thomas, D.; Donadio, S.; Zhang, H.; Zhang, C.;
Kapanidis, A. N.; Ebright, R. H. Mol. Cell 2018, 70, 60-71.
5. Venugopal, A. A.; Johnson, S. Clin. Infect. Dis. 2012, 54, 568.
6. (a) Australian Therapeutic Goods Administration, Australian
Public Assessment Report for Fidaxomicin, 2013,
1
characterized by H NMR and HRMS. Each consisted of an
inseparable 2:1 Z/E mixture of olefin isomers and had a mass
(m/z) consistent with that of authentic fidaxomicin. One product
coeluted with commercial fidaxomicin in
a co-injection
experiment, and comparison of the 1H NMR of this material with
that of authentic fidaxomicin supported its assignment as a 2:1
mixture of 4,5-Z-fidaxomicin and fidaxomicin. The remaining
products appear to be constitutional isomers of fidaxomicin
(same m/z), presumably isobutyrate migration products in the
novioside sugar.40 However, none of these materials was obtained
in sufficient quantity to support a more rigorous structural
assignment.
https://www.tga.gov.au/node/781 (b) Shue, Y. K.; Sears, P. S.;
Shangle, S.; Walsh, R. B.; Lee, C.; Gorbach, S. L.; Okumu, F.;
Preston, R. A. Antimicrob. Agents Chemother. 2008, 52, 1391-
1395. (c) Sears, P.; Crook, D. W.; Louie, T. J.; Miller, M. A.;
Weiss, K. Clin. Infect. Dis. 2012, 55, S116-S120.
7. A. Malcangi, G. Trione, Tiacumicin Production. US 8728796B2,
2014.
8. (a) Jamison, M. Ph.D. Dissertation, University of Texas
Southwestern Medical Center, 2013. (b) Kaufmann, E. B. Ph.D.
Dissertation, University of Basel, 2017. (c) Hattori, H.; Kaufmann,
E.; Miyatake-Ondozabal, H.; Berg, R.; Gademann, K. J. Org.
Chem. 2018, 83, 7180-7205.
3. Conclusion
We have established robust synthetic routes to key
intermediates 2-7 en route to fidaxomicin. In particular, the
hydrogenative C-C coupling route (scheme 3B) to dienoate 23
represents a favorable route towards this compound. The 2,3-
Wittig rearrangement is a less contemporaneous, yet highly
efficient approach to install the stereodiad embedded within the
southern polyketide fragment 6. The catalytic Krische-type
allylation from the alcohol oxidation stage represents and
efficient alternative to classical reagent controlled allylations to
install the homoallylic alcohol of fragment 6 with high
9. (a) Xiao, Y.; Li, S.; Niu, S.; Ma, L.; Zhang, G.; Zhang, H.; Zhang,
G.; Ju, J.; Zhang, C. J. Am. Chem. Soc. 2011, 33, 1092-1105. (b)
Niu, S.; Hu, T.; Li, S.; Xiao, Y.; Ma, L.; Zhang, G.; Zhang, H.;
Yang, X.; Ju, J.; Zhang, C. ChemBioChem 2011, 12, 1740-1748.
(c) J. E. Hochlowski, M. Jackson, J.B. McAlpine, R. R.
Rasmussen, Bromotiacumicin compounds, US5767096A, 1998.
(d) Zhang, H.; Tian, X.; Pu, X.; Zhang, Q.; Zhang, W.; Zhang, C.
J. Nat. Prod. 2018, 81, 1219-1224. (d) Dorst, A.; Shchelik, I.S.;
Schäfle, D.; Sander, P.; Gademann, K. Helv. Chim. Acta 2020,
Early View, https://doi.org/10.1002/hlca.202000130.
10. For selected reviews, see: (a) Roulland, E. Synthesis 2018, 50,
4189-4200. (b) Dorst, A.; Gademann, K. Helv. Chim. Acta 2020,
103, e2000038.
diastereoselectivity. The preparation of novioside 7 from
D-
11. For selected examples of syntheses of other complex antibiotics,
see: (a) Wright, P. M.; Seiple I. B.; Myers, A. G. Angew. Chem.,
Int. Ed. 2014, 53, 8840-8869. (b) Seiple, I. B.; Zhang, Z.; Jakubec,
P.; Langlois-Mercier, A.; Wright, P. M.; Hog, D. T.; Yabu, K.;
Lyxose represents a substantial improvement in both step count
and efficiency over previous approaches. Koenigs-Knorr
glycosylations towards both 2 and 3 were efficient and selective

7
Allu, S. R.; Fukuzaki, T.; Carlsen, P. N.; Kitamura, Y.; Zhou, X.;
Condakes, M. L.; Szczypiński, F. T.; Green, W. D.; Myers, A. G.
Nature 2016, 533, 338-345. (c) Charest, M.; Lerner, C. D.;
Brubaker, J. D.; Siegel, D. R.; Myers, A. G. Science 2005, 308,
395-398.
37. Roulland used an activated sulfoxide leaving group for their
hydrogen bond directed glycosylation, and Gademann reported the
activation of an N-phenyl trifluoroacetamide.
38. Martí-Centelles, V.; Pandey, M. D.; Burguete, M. I.; Luis, S. V.
Chem. Rev. 2015, 115, 8736-8834.
12. (a) Glaus, F.; Altmann, K.-H. Angew. Chem., Int. Ed. 2015, 54,
1937-1940. (b) Erb, W.; Grassot, J.-M.; Linder, D.; Neuville, L.;
Zhu, J. Angew. Chem., Int. Ed. 2015, 54, 1929-32. (c) Miyatake-
Ondozabal, H.; Kaufmann, E.; Gademann, K. Angew. Chem., Int.
Ed. 2015, 54, 1933-1936. (d) Jeanne-Julien, L.; Masson, G.;
Astier, E.; Genta-Jouve, G.; Servajean, V.; Beau, J. M.; Norsikian,
S.; Roulland, E. Org. Lett. 2017, 19, 4006-4009.
39. Wu, Y.; Liao, X.; Wang, R.; Xie, X.; Brabander, J. K. De J. Am.
Chem. Soc. 2002, 124, 3245-3253.
40. Such acyl migration products have been documented for natural
fidaxomicin-related metabolites. See ref. 1a.
13. (a) Kaufmann, E.; Hattori, H.; Miyatake-Ondozabal, H.;
Gademann, K. Org. Lett. 2015, 17, 3514-3517. (b) Norsikian, S.;
Tresse, C.; François-Eude, M.; Jeanne-Julien, L.; Masson, G.;
Servajean, V.; Genta-Jouve, G.; Beau, J. M.; Roulland, E. Angew.
Chem., Int. Ed. 2020, 59, 6612-6616.
14. (a) Gong, J.; Li, W.; Fu, P.; Macmillan, J.; De Brabander, J. K.
Org. Lett. 2019, 21, 2957-2961. (b) Yu, X. Ph.D. Dissertation,
University of Texas Southwestern Medical Center, 2018. (c)
Hattori, H.; Roesslein, J.; Caspers, P.; Zerbe, K.; Miyatake-
Ondozabal, H.; Ritz, D.; Rueedi, G.; Gademann, K. Angew.
Chem., Int. Ed. 2018, 57, 11020-11024. (d) Hattori, H.; Hoff, L.
V.; Gademann, K. Org. Lett. 2019, 21, 3456-3459.
15. J. De Brabander, Small molecule compounds selective against
gram-negative bacterial infections, US 20150329582, 2015 (Filed
May 13, 2014).
16. Our route differs from that of Gademann (ref. 13a) in that the
deoxygenation sequence was carried out subsequent to masking of
the cis-diol.
17. (a) Bhattacharjee, A.; De Brabander, J. K. Tetrahedron Lett. 2000,
41, 8069-8073. (b) Soltani, O.; De Brabander, J. K. Angew.
Chem., Int. Ed. 2005, 44, 1696-1699. (c) García-Fortanet, J.;
Debergh, J. R.; De Brabander, J. K. Org. Lett. 2005, 7, 685-688.
18. Gaddam, J.; Reddy, G. S.; Marumudi, K.; Kunwar, A. C.; Yadav,
J. S.; Mohapatra, D. K. Org. Biomol. Chem. 2019, 17, 5601-5614.
19. (a) Kumar Dey, S.; Ataur Rahman, M.; Alkhazim Alghamdi, A.;
Reddy, B. V. S.; Yadav, J. S. Eur. J. Org. Chem. 2016, 2016,
1684-1687. (b) Satoh, T.; Nanba, K.; Suzuki, S. Chem. Pharm.
Bull. 1971, 19, 817-820.
20. For a more detailed discussion of factors controlling
thermodynamic equilibria in this system, see reference 8b.
21. Martin, V. S.; Woodard, S. S.; Katsuki, T.; Yamada, Y.; Ikeda,
M.; Sharpless, K. B. J. Am. Chem. Soc. 1981, 103, 6237-6240.
22. (a) Moure, A. L.; Arrayás, R. G.; Cárdenas, D. J.; Alonso, I.;
Carretero, J. C. J. Am. Chem. Soc. 2012, 134, 7219-7222.
23. Komanduri, V.; Krische, M. J. J. Am. Chem. Soc. 2006, 128,
16448-16449.
24. Nishimura, A.; Tamai, E.; Ohashi, M.; Ogoshi, S. Chem. - Eur. J.
2014, 20, 6613-6617.
25. Evans, D. A.; Song, H.; Fandrick, K. R. Org. Lett. 2006, 8, 3351-
3354.
26. Ben Ayed, T.; Amri, H.; & Gaied, M. E. M. Tetrahedron 1991,
41, 9621-9628.
27. Miyashita, M.; Yoshikoshi, A.; & Grieco, P. A. J. Org. Chem.
1977, 42, 3772-3774.
28. Friest, J. A.; Broussy, S.; Chung, W. J.; Berkowitz, D. Angew.
Chem., Int. Ed. 2011, 50, 8895-8899.
29. Nakai, T.; Mikami, K. Chem. Rev. 1986, 86, 885-902.
30. Lipshutz, B. H.; Ellsworth, E. L.; Dimock, S. H.; Reuter, D. C.
Tetrahedron Lett. 1989, 30, 2065-2068.
31. (a) Kim, I. S.; Nagi, M. -Y.; Krische, M. J. J. Am. Chem. Soc.
2008, 130, 14891-14899. (b) Thorat, R. G.; Harned, A. M. Chem.
Commun. 2018, 54, 241-243.
32. N. A. Petasis, X. Yao, J. C. Raber, Nitrogen-containing
heterocycles, US 6927294B1, 2005.
33. Bosch, M. P.; Campos, F.; Niubo, I.; Rosell, G.; Diaz, J. L.; Brea,
J.; Loza, M. I.; Guerrero, A. J. Med. Chem. 2004, 47, 4041-4053.
34. Frank, S. A.; Chen, H.; Kunz, R. K.; Schnaderbeck, M. J.; Roush,
W. R. Org. Lett. 2000, 2, 2691-2694.
35. We used purified 7b for this transformation, but Gademann has
demonstrated that a mixture of 7a and 7b is also effective (ref.
13a).
36. Stereoselectivity was estimated on the basis of crude NMR
integration. The reported yield is of isolated β-3. Gademann
reported an isolated yield of 48% of β-3 and 15% of α-3 (a 3:1
ratio).

HIGHLIGHTS
Efficient stereoselective synthesis of the key polyketide fragments of Fidaxomicin
Efficient stereoselective synthesis of Fidaxomicin’s resorcylic rhamnoside and novioside sugar fragments
Efficient and selective glycosylation of the polyketide fragments with the sugar fragments
Krische hydrogenative C-C coupling towards Fidaxomicin’s branched dienoic acid fragment
[2,3]-Wittig rearrangement tactic to arrive at a polyketide stereodiad fragment embedded within
Fidaxomicin’s southern polyketide fragment

Declaration of interests
☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.
☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests: