2,3-cis-Cyclization of 4-pentenoxyl radicals

Article abstract of DOI:10.1016/j.tet.2016.07.001

4-Pentenoxyl radicals cyclize 2,3-cis-selectively, when substituted by an allylic hydroxy, acetyloxy, or benzoyloxy group. Additional substituents increase or decrease the fraction of 2,3-cis-cyclized product, depending on relative configuration, position

Full text of DOI:10.1016/j.tet.2016.07.001

Accepted Manuscript
2,3-cis-Cyclization of 4-pentenoxyl radicals
Irina Kempter, Christine Schur, Katharina Huttenlochner, Ruth-Maria Bergsträßer,
Benjamin Wolff, Thomas Kopf, Jens Hartung
PII:
S0040-4020(16)30615-9
10.1016/j.tet.2016.07.001
TET 27897
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 25 February 2016
Revised Date: 28 June 2016
Accepted Date: 1 July 2016
Please cite this article as: Kempter I, Schur C, Huttenlochner K, Bergsträßer R-M, Wolff B, Kopf
T, Hartung J, 2,3-cis-Cyclization of 4-pentenoxyl radicals, Tetrahedron (2016), doi: 10.1016/
j.tet.2016.07.001.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.

ACCEPTED MANUSCRIPT
2
,3-cis-Cyclization of 4-Pentenoxyl Radicals
Irina Kempter, Christine Schur, Katharina Huttenlochner, Ruth-Maria Bergsträßer, Benjamin
Wolff, Thomas Kopf, and Jens Hartung*
Fachbereich Chemie, Organische Chemie, Technische Universität Kaiserslautern,
Erwin-Schrödinger-Straße, D-67663 Kaiserslautern, Germany
H
•
O
•
O
1
O•
5-exo-trig
CH
2
5
2
≡
3
H
H
H
H
H
favored
up to 80:20 dr
=
OH, OAc, or OBz
1

ACCEPTED MANUSCRIPT
2
,3-cis-Cyclization of 4-Pentenoxyl Radicals
Irina Kempter, Christine Schur, Katharina Huttenlochner, Ruth-Maria Bergsträßer,
Benjamin Wolff, Thomas Kopf, and Jens Hartung*
Fachbereich Chemie, Organische Chemie, Technische Universität Kaiserslautern,
Erwin-Schrödinger-Straße, D-67663 Kaiserslautern, Germany
*
Corresponding author. Tel.: +49-631-205-2431, Fax: +49-631-205-3921, e-mail:
hartung@chemie.uni-kl.de
Keywords: Alkoxyl radical; Polar effect; Reactive conformer; Stereoselective synthesis;
Tetrahydrofuran.
2

ACCEPTED MANUSCRIPT
Abstract: 4-Pentenoxyl radicals cyclize 2,3-cis-selectively, when substituted by an allylic
hydroxy, acetyloxy, or benzoyloxy group. Additional substituents increase or decrease the
fraction of 2,3-cis-cyclized product, depending on relative configuration, positioning, and their
chemical nature. The preference for 3-acceptor-subsituted pentenoxyl radicals to furnish
products of 2,3-cis-ring closure arises from a secondary orbital interaction between the allylic
oxygen substituent and the alkene entity, kinetically disfavoring the competing 2,3-trans-mode
of 5-exo-cyclization. Aligning the
alkene, which is the preferred conformation for transition structures for 2,3-trans-cyclization,
stabilizes the double bond by delocalizing -electrons into the *(C,O)-orbital. Along with
energy decreases the affinity of -electrons for adding the oxygen radical. In 2,3-cis-
β-C,O-bond in anticline orientation to the plane of the
π
σ
π
cyclization, a similar stabilizing effect cannot occur, because the allylic oxygen substituent and
the alkene align synperiplanar. The kinetic effect of an allylic oxygen substituent becomes
furthermore apparent in cyclization of the 3-hydroxynona-1,8-dien-5-oxyl radical, favoring
intramolecular addition to the unsubstituted allylic double bond by a factor three.
1
.
Introduction
4-Pentenoxyl radicals cyclize 2,3-trans-selectively when substituted with an allylic alkyl
1
or a phenyl group (Scheme 1). The fraction of 2,3-trans-product increases with the size of the
2
,3
allylic substituent, from 80/20 for methyl to above 99/1 for tert-butyl. Transition state
4
theory explains 2,3-trans-selectivity on the basis of cumulative 1,2- and 1,3-repulsion,
5
,6
progressively disfavoring 2,3-cis-addition as steric demand of the allylic substituent grows.
3

ACCEPTED MANUSCRIPT
X = Ph
R = H
X = OH
H
H
H
H
•
•
R = CH3
O
R
O •
X
O
CH
CH
2
2
H
H
OH
Ph
,3-trans
dr = 5:95
2
2,3-cis
dr = 70:30
2
,3,7
Scheme 1. Stereoselectivity in 5-exo-cyclization of allyl substituted 4-pentenoxyl radicals.
F
+
N(CH )
OH
3
3
F
H
H
H
H
O
O
2
3
H
H
OH
HO
allo-isomuscarine
thymidine bioisoster
–
O
P
O
H
O
–
O
HO
H
P
O
O
O
H
N
HO
H
O
O
3
2
H
O
H
N
HO
OH
H
H
HO
OH
UDP-C-D-galactofuranose
Figure 1. Structure formulas of bio-inspired 2- and 3-hydroxy-substituted tetrahydrofurans as
8
,9,10
targets in organic synthesis.
Steric effects controlling selectivity in synthesis of 2,3-trans-substituted heterocycles by
1
1,12
13,14,15,16,17
intramolecularly adding polar reactands
or radicals
to double bonds are well
documented in the scientific literature. In the past decades, however, more and more reports
4

ACCEPTED MANUSCRIPT
appeared describing selectivity not fitting into this stereochemical scheme. Alkenols bearing
an allylic oxygen substituent, for example, show a marked propensity for cyclizing 2,3-cis-
selectively, when treated with molecular iodine. A theory explaining this phenomenon starts to
1
8,19,20,21
evolve, but is not yet consistent.
The affinity of the allylic hydroxy group to direct cyclizations 2,3-cis-selectively also
extends to oxygen radical additions, as recently outlined in synthesis of allo-isomuscarine
8
(
Scheme 1, Figure 1). The alkene in this example poses the nucleophilic component and the
radical oxygen the electrophilic, which is exactly opposed to the situation in electrophile-
induced alkenol cyclization. A theory explaining 2,3-cis-selectivity in radical cyclization so far
does not exist.
For uncovering the principles leading to 2,3-cis-selective ring closures, we investigated
in the study summarized below, reactivity and selectivity of seven 4-pentenoxyl radicals,
differing in substitution at the allylic carbon and at proximal positions. The results from this
effort show that an allylic oxygen substituent reduces the rate of intramolecular 2,3-trans-
addition, and leave the rate of the 2,3-cis-pathway largely unaffected. The rate effect of the
allylic oxygen substituent is not restricted to stereocontrol but also controls selectivity in
intramolecular addition of an oxyl radical to two chemically different C,C-double bonds. The
3
-hydroxynona-1,8-dien-5-oxyl radical thus prefers adding to the unsubstituted allylic double
bond by a partial rate factor three.
2
.
Results and Interpretation
5

ACCEPTED MANUSCRIPT
2
.1 Alkenoxyl radical generation, chain reaction, and design of alkenoxyl radicals for
conducting the study
8
,22
Based on the expertise from preceding mechanistic studies, we used O-pentenyl esters
2
3
of 1,3-thiazole-derived heterocyclic thiohydroxamic acids as progenitors for generating
oxygen radicals. Heating compounds of this kind, for example the 3-alkoxy-1,3-thiazole-
1
2
3
(3H)-thiones, in the presence of a chemical initiator, or photoexciting the molecules with
50 nm-light, almost specifically breaks the N,O-bond, allowing to liberate oxygen radicals
under pH-neutral and non oxidative conditions from otherwise stable compounds.
The mechanism operating for converting 3-alkoxy-1,3-thiazole-2(3H)-thiones into
alkoxyl radicals is a chain reaction. The sequence starts by adding a chain propagating radical
to the thione sulfur of the thiohydroxamate used as progenitor, providing oxygen radical I and
2
4
2,25,26
substituted thiazole 2 (Scheme 2). By extrapolating known kinetic data
substituted derivatives to intramolecularly add with a rate constant between 10 s and 10 s
to the C,C-double bond. In terms of regioselectivity the intramolecular addition should lead to
, we expect allyl-
8
–1
9
–1
2
7
a 5-exo/6-endo-distribution ranging from 98:2 (for X = CH ) to 90:10 (for X = OH). For
3
terminating the sequence, cyclized radical II needs to be trapped by a mediator. In the present
study,
we
used
bromotrichloromethane
for
this
purpose,
providing
2-
8
bromomethyltetrahydrofuran 3 as major oxygen radical-derived product. The requested
information on stereocontrolling effects of oxygen substituents in cyclizations then is stored in
the cis/trans-ratio of bromocyclization product 3. The second product of bromine atom transfer
from bromotrichloromethane to carbon radical II is the trichloromethylradical, being essential
for resuming the chain reaction.
6

ACCEPTED MANUSCRIPT
S
S
N
Br
O
O
R2
R2
OR1
• CCl3
OR1
3
1
S
BrCCl3
exo
•
SCCl3
N
O
5
O•
CH
2
R2
2
R2
trig(onal)
OR1
OR1
II
I
Scheme 2. Chain reaction for bromomethyltetrahydrofuran synthesis from 3-alkenoxy-4-
1
2
8,24
methylthiazole-2(3H)-thione 1 and BrCCl (R = H, Ac, Bz; R = H, OAc, CH ).
3
3
For clarifying the role of the allylic oxygen substituent in stereoselectivity control for
alkenoxyl radical 5-exo-cyclization, we conceived a set of six alkenoxyl radicals: three
monosubstituted 4-pentenoxyl radicals (Ia–c), two homologues bearing acetyloxy groups in
different relative configuration (erythro/threo-Id), and a third (Ie) having an allylic acyloxy
group positioned next to two additional substituents (Figure 2). As we became aware of the
kinetic role of the allylic oxygen substituent, we included nonadienyloxyl radical If into the
study, for exploring regioselectivity effects.
7

ACCEPTED MANUSCRIPT
H
H
O •
O •
O•
AcO
AcO
AcO
OR
OAc
OAc
Ia–c
erythro-Id
Ie
O •
O•
R
I
H
Ia
Ib
Ic
Ac
Bz
OAc
OH
threo-Id
rel-(3R,5S)-If
Figure 2. Structure formulas of 4-pentenoxyl radicals Ia–f conceived for uncovering the origin
of 2,3-cis-selectivity in homolytic 5-exo-cyclization.
2
.2 Convention used for numbering atom positions
Oxygen and carbon differ in priority for systematically naming aliphatic and heterocyclic
compounds. For carbons in the aliphatic side chain of O-alkenyl esters 1, and for the carbons
2
8
of radical I, we adhered to the IUPAC-recommendation for aliphatic compounds. A
transition structure TS-I associated with 5-exo-cyclization, in the chosen stereochemical
model, derives from tetrahydrofuran (cf. section 2.5.3). Numbering atoms in intermediate TS-
I, similar to tetrahydrofuran core of cyclized carbon radical II, and bromocyclization product 3
2
9,30
thus follows the Hantzsch and Widman convention.
For assigning configuration of vicinal stereocenters, we adhered in this article to
descriptors erythro and threo. By this approach we feel, that the reader may easier follow
differences in chemical selectivity, as configuration at one of the stereocenters changes.
8

ACCEPTED MANUSCRIPT
1
O
1
O
1
O
Br
OMTT
O •
1
2
1
2
•
5
•
5
5
2
CH
2
2
4
3
4
2
5
3
5
3
4
3
4
3
4
OH
OH
OH
OH
OH
1a
Ia
TS-Ia
IIa
3a
Figure 3. Convention used in this article for numbering atoms in product classes associated
with alkenoxyl radical cyclization, as exemplified for alkenoxythiazolethione 1a, derived 4-
alkenoxyl radical Ia, transition structure for 5-exo-cyclization TS-Ia, 5-exo-cyclized carbon
radical IIa, and bromocyclization product 3a (MTT = 4-methyl-2-thiooxo-1,3-thiaz-3-yl).
2
.3 3-Alkenoxy-4-methylthiazole-2(3H)-thiones
2
.3.1 Monosubstituted O-pentenyl thiohydroxamates
For constructing the alkenoxy group of thiohydroxamate 1a, we connected in the initial
3
1,32
step ethyl acetate to acrolein by a mixed aldol addition (Scheme 3).
The product formed in
the addition, ethyl 3-hydroxypentenoate, eliminates water on standing. For preventing this
elimination to occur, we converted the reactive hydroxy group into a methoxymethyloxy
3
2,33
(MOMO)-group.
Subsequently, we transformed the ethoxycarbonyl group in ester 4 into
3
2,34
primary alkyl sulfonate group in tosylate 5 by standard prodecures.
Displacing the
sulfonate group in tosylate 5 by the 4-methyl-2-thiooxo-(3H)-1,3-thiazyl-1-oxide anion
furnishes a MOM-protected O-alkenyl thiohydroxamate, which was treated with hot acidic
methanol for solvolytically cleaving off the acetal-protecting group. Esterifying the hydroxy
group in alcohol 1a with acetic anhydride gives acetate 1b. Using benzoyl chloride as
acylation reagent provides benzoate 1c (Scheme 4).
9

ACCEPTED MANUSCRIPT
1
. LDA / THF
acrolein
OEt
O
O
OEt
O
O
O
2.
P2O5
O
4
(43%)
3. LiAlH / Et O
4 2
4
. TsCl / DABCO
–
+
5
. MTTO NEt4
OMTT
OTs
DMF
3
6
. HCl / CH OH
3
OH
O
O
1a (72%)
5 (77%)
Scheme 3. Summary of functional group transformation for preparing 3-(3-hydroxypent-4-
enyloxy)thiazolethione 1a from ethyl acetate (MTTO = 4-methyl-2-thiooxo-1,3-thiazyl-3-oxy;
cf. Figure 3).
Ac O / NEt
BzCl / NEt3
DMAP
2
3
OMTT
DMAP
OMTT
OMTT
3
3
3
CH Cl
CH Cl
2 2
2
2
OAc
OH
OBz
1
b (82%)
1a
1c (52%)
Scheme 4. O-Acylation of 3-(3-hydroxypent-4-enyloxy)thiazolethione 1a.
2
.3.2 Disubstituted O-pentenyl thiohydroxamates
For preparing the erythro-configured side chain of O-(2,3-erythro-bisacetyloxypentenyl)
thiohydroxamate erythro-1d, we started from
available from -isoascorbic acid by oxidatively removing two carbon atoms (Scheme 5). For
preventing oxidative side reactions, we protected the two hydroxy groups of the lactone with
D-erythronolactone. The lactone is synthetically
D
10

ACCEPTED MANUSCRIPT
2
,2-dimethoxypropane in acidic solution under kinetic control, leading to the derived
3
5
acetonide. Reducing the protected
lactol, which was treated with methanediyltriphenylphosphorous ylide to furnish alkenol
D-erythronolactone with sodium borohydride affords a
3
5,36,37
erythro-6.
we adhered to methods already summarized in Scheme 3.
changed protecting groups at oxygen from acetonide, required for synthesis, to acyl required
For converting alkenol erythro-6 into O-alkenyl thiohydroxamate erythro-7,
2
4,38,39
At the final stage, we again
4
0
for alkenoxyl radical cyclization.
1
. Na CO / H O
2 3 2 2
2. MeO OMe
HO
H
O
OH
O
TsOH
HO
H
O
3
3
. NaBH / H O
2
4
2
HO
OH
H
+
–
O
4
. Ph PCH Br
3
3
KOtBu / THF
D-isoascorbic acid
erythro-6 (32%)
5
. TsCl / DABCO
–
+
6. MTTO NEt4
7
. HCl / MeOH
OMTT
OMTT
8
. Ac O / NEt3
2
H
O
3
3
AcO
2
2
H
O
OAc
erythro-1d (74%)
erythro-7 (48%)
Scheme 5. Preparing of 3-[2,3-erythro-2,3-bis-(acetyloxy)pent-4-en-1-oxy]thiazolethione
erythro-1d from -isoascorbic acid.
D
For preparing O-(threo-2,3-bisacetyloxypentenyl) thiohydroxamate threo-1d, we used
dimethyl (2R,3R)-tartrate as starting material. The ester comprises two stereocenters in the
11

ACCEPTED MANUSCRIPT
required configuration and two functional groups for building the carbon skeleton of target
compound threo-1d. One of the ester groups served in a sequence adapted from synthesis of
diastereomer erythro-1d for introducing the thiohydroxamate functional group at a primary
carbon. The second ester group was used for constructing the alkene entity needed for
conducting the alkenoxyl radical cyclization. By this strategy, we obtained monosilyl-protected
diol threo-8 as first key intermediate, acetonide-protected 4-pentenetriol threo-6 as second and
4
1,42,43,44,45
O-alkenyl thiohydroxamate threo-1d as final product.
1
. MeO OMe
TsOH
MeO
HO
O
OTBDMS
O
H
O
3
OMe
2
3
. LiAlH / THF
2
OH
4
H
OH
O
. NaH / THF
TBDMSCl
dimethyl (2R,3R)-tatrate
threo-8 (53%)
4
. (COCl) / DMSO / NEt
2
3
+
–
5
. Ph PCH Br
3 3
KOtBu / THF
6
. TBAF / THF
7
. TsCl / DABCO
–
+
OMTT
8. MTTO NEt4
OH
H
O
3
3
9
. PPTS / MeOH
AcO
2
2
H
OAc
10. Ac O / NEt3
O
2
threo-1d (60%)
threo-6 (49%)
Scheme 6. Synthesis of 3-[2,3-threo-2,3-bis-(acetyloxy)pent-4-en-1-oxy]thiazolethione threo-
1
d from dimethyl (2R,3R)-tartrate.
2
.3.3 Trisubstituted O-pentenyl thiohydroxamate 1e
12

ACCEPTED MANUSCRIPT
For accomplishing synthesis of
D-arabino-configured thiohydroxamate 1e we extended
the five carbon chain from
D
-ribose by one carbon in a Wittig-alkenylation, and inverted
configuration at carbon C2 in the thiohydroxamate O-alkylation step (Scheme 7). For
simplifying work-up procedures we, again, used the acetonide protecting group for the glycol
segment. We also changed the thiohydroxamic acid to 3-hydroxy-5-(p-methoxyphenyl)-4-
methylthiazole-2(3H)-thione, allowing to crystalline target compound 1e, for additional
4
6
purification in the final step.
R
H
S
S
N
O
H
HO
H
OMAnTT
O
1.–5.
2
6.–7.
2
OH
OH
D-ribose
H
O
4
4
3
AcO
3
H
HO
O
OAc
9 (15%)
1e (43%)
Scheme 7. Synthesis of 2,3,4-arabino-configured 3-(hexenoxy)thiazolethione 1e from
D-
ribose [R = 4-methoxyphenyl; reagents and conditions – 1. step: acetone, concentrated sulfuric
4
7
acid (81% yield); 2. step: iodine, imidazole, triphenylphosphine, toluene/acetonitrile (54%
4
8,49
50,51
yield);
triphenylphosphonium bromide, potassium tert-butoxide, tetrahydrofuran (67% yield); 5. step:
-hydroxy-4-methyl-5-(4-methoxyphenyl)-1,3-thiazole-2(3H)-thione (MAnTTOH), diethyl
3. step: palladium on charcoal, hydrogen, ethanol (97% yield);
4. step: methyl
3
azodicarboxylate, triphenylphosphine, benzene (54% yield); 6. step: hydrogen chloride,
methanol (96% yield); 7. step: acetic anhydride, 4-(N,N-dimethylamino)pyridine,
dichloromethane (45% yield)].
13

ACCEPTED MANUSCRIPT
2
.3.4 O-(3-Hydroxy-5-nona-1,8-dienyl) thiohydroxamate 1f
For preparing the C9-carbon skeleton of dienyl tosylate 10 we added but-3-en-1-yl
magnesium bromide to 3-(methoxymethyloxy)pent-4-enal and esterified the resulting product,
-(methoxymethyloxy)nona-1,8-dien-5-ol, with p-toluenesulfonyl chloride. Heating tosylate 10
3
with 3-hydroxy-4-methyl-1,3-thiazole-2(3H)-thione tetraethylammonium salt in DMF yields a
MOM-protected O-dienyl thiohydroxamate as 50/50-mixture of stereoisomers. The
diastereomers separate on the silica gel column, used for purifying the 3-(3-hydroxynona-1,8-
dien-5-yl)thiazolethione 1f after having solvolytically cleaved off the MOM-protecting group
(Scheme 8). Only isomer rel-(3R,5S)-1f provided a correct combustion analysis and was
therefore considered in the subsequent alkenoxyl radical study.
H
OTs
OMTT
OH
1
.–2.
5
3
O
OCH3
1
0
rel-(3R,5S)-1f (23%)
Scheme 8. Preparing 3-(3-hydroxynona-1,8-dien-5-yl)thiazolethione rel-(3R,5S)-1f from
dienyl tosylate 10 (Supplementary data) [reagents and conditions – (1.): 3-hydroxy-4-methyl-
1
,3-thiazole-2(3H)-thione tetraethylammonium salt, DMF, 40 °C (67% yield); (2.): hydrogen
chloride, methanol, and column chromatography for separating diastereomers {rel-(3R,5S)-1f:
3
5% yield}].
2
.3.5 Stereochemical analysis
All stereocenters of O-alkenyl thiohydroxamates erythro-1d, threo-1d, and 1e derive
14

ACCEPTED MANUSCRIPT
from enantiopure natural products. All synthetic manipulations used for transforming the
starting materials into the target compounds occurred stereospecifically, as concluded from
single sets of resonances in proton- and carbon-13 NMR-spectra (Experimental). For unknown
reasons, optical rotations recorded for erythro-1d and threo-1d scattered from batch to batch,
and in one instance even changed sign. Since stereochemical information obtained by
transforming single diastereomers suffice to answer all relevant question associated with the
study, we refrained from emphasizing absolute configuration of O-alkenyl thiohydroxamates
erythro/threo-1d and 1e hereafter.
Table 1. Chemical shifts of oxygen-bound carbons and the inner alkene proton of O-alkenyl
thiohydroxamates 1a–f (in CDCl , ambient temperature)
3
H
H
OMTT
OMTT
OMAnTT
OMTT
α
β
α
δ
γ
δ
γ
δ
γ
δ
γ
β
AcO
AcO
OR
OAc
OAc
OH
1
a–c
erythro/threo-1d
1e
(±)-1f
α
β
γ
δ
entry
1 / R
C / ppm
C / ppm
C / ppm
H / ppm
1
2
3
4
5
6
7
1a / H
1b / Ac
1c / Bz
erythro-1d
threo-1d
1e
73.1
72.0
72.0
72.5
73.8
74.0
84.5
35.1
32.5
32.6
71.2
70.8
72.1
68.5
71.1
71.7
73.3
72.0
76.9
71.0
5.94 (ddd)
5.85 (ddd)
5.97 (ddd)
5.83 (ddd)
5.90 (ddd)
5.79 (ddd)
5.94 (ddd)
a
(±)-1f
–
a
Not assigned; for entire sets of chemical shifts, see the Experimental.
15

ACCEPTED MANUSCRIPT
2
.4 Homolytic bromocyclization
2
.4.1 Parameters for conducting alkenoxyl radical reactions
For studying the chemistry of 4-pentenoxyl radicals Ia–f, we adhered to a standardized
8
,22,24
protocoll derived from preceding mechanistic investigations.
Operational standards
thereby relate to (i) the experimental set-up, (ii) methods for quantifying products, and (iii)
structure analysis.
(i) Experimental set-up. For securing that yields and selectivity are not flawed by
technical details associated with the initiating step, we used photochemical and thermal
activation for initiating radical chain reactions. In the following, we assessed yields and
product manifolds obtained from O-esters 1a and 1b under both conditions, for deciding which
of the methods to use as standard.
The most effective way for initiating a radical reaction photochemically started from a
8
3-millimolar solution of 1a in perdeuterobenzene containing bromotrichloromethane. The
®
solution was irradiated in a Rayonet -chamber reactor equipped with twelve 350-nm light
bulbs. This set-up gave 2-methylsulfanylthiazole 2 in 59% and bromoethers 3a/11a in 71%
yield. For initiating the radical reaction thermally, we heated a likewise prepared solution of 1a
in the presence of azobisisobutyronitrile (AIBN) as chemical initiator. The therml version of
the radical reaction gave 88% of bromoethers 3a/11a in total, and co-product 2 in 81% yield.
The same trend for providing elevated yields under thermal conditions was seen for
reactions between O-acetyl-substituted ester 1b and bromotrichloronmethane. For example,
heating 1b in a 83-millimolar solution of perdeuterobenzene and bromotrichlormethane gave
16

ACCEPTED MANUSCRIPT
O-acetyloxybromoethers 3b/11b in 89% combined yield, and 90% of thiazole 2. The
photochemical reaction provided 77% of bromocyclization products 3b/11b and 71% of co-
product 2. By comparing yields we finally decided to use the thermal method as standard for
conducting alkenoxyl radical bromocyclizations of O-alkenylthio-hydroxamates 1a–f.
Benzoate 1c, under such conditions, gave 95% of thiazole 2 and a combined yield of 94% for
brominated ethers 3c/11c.
Regarding reaction times, thermally initiated conversions of O-alkenyl esters 1a–c took
no longer than two hours for completion. Extending reaction times diminishes yields and
causes side-products to appear.
Photoreactions of standardized solutions of esters 1a–b were complete within one hour.
In some of the experiments the reaction mixture turned in an unpredictable manner yellowish
or turbid with some of the starting material still present. Extending the reaction time
quantitatively turns over substrate 1, however, by also leading to a higher product
2
2,52
manifold.
(ii) Methods for quantifying products. For obtaining the original stereochemical
information of a bromocyclization product 3, chemoselectivity for the reaction between
thiohydroxamate 1 and BrCCl , and maximum yields, we injected samples taken directly from
3
solutions of products in perdeuterobenzene into a gas chromatograph equipped with a flame
ionization detector and coupled to a mass spectrometer. For cross-checking structural
assignments, we analyzed samples from the same mixture by NMR-spectroscopy (proton and
carbon-13).
(iii) Structure analysis. All products described in this article were characterized by high
17

ACCEPTED MANUSCRIPT
resolution mass spectrometry, combustion analysis, and NMR-spectroscopy for constitution
analysis. For assigning relative configuration of bromocyclization products (i.e. 3), we relied
on shift differences of carbon-13 resonances between cis/trans-isomers (vide infra). For
deducing constitution and, wherever possible, relative configuration of brominated
tetrahydropyrans, being formed in 13% yield and less (cf. section 2.5.2), we used mass
spectrometry in combination with NMR-spectroscopy. In one instance (for 11a), we
independently prepared a bromotetrahydropyran for securing our interpretation.
2
.4.2 Bromocyclization of monosubstituted O-(4-pentenyl) thiohydroxamates 1a−c
i) Products. In a larger scale-up experiment, conducted under thermal conditions, 1.12
(
mmol monosubstituted O-alkenyl thiohydroxamate 1a gave 56% of 2-bromomethyl-
tetrahydrofuran-3-ol (3a), 4% of 4-bromotetrahydropyran-3-ol (11a), and 84% of disubstituted
thiazole 2 (Table 2, entry 1). Tetrahydrofuran 3a was formed under such conditions as 74/26-
mixture of 2,3-cis/trans-isomers.
O-Acetyl-substituted ester 1b yielded 71% of tetrahydrofuran 3b as 68/32-mixture of
2
,3-cis/trans-isomers, 11% of tetrahydropyran 11b and 94% of thiazole 2 under standard
conditions (Table 2, entry 2).
O-Benzoyl-substituted thiohydroxamate 1c provided 69% of disubstituted
tetrahydrofuran 3c as 68/32-mixture of 2,3-cis/trans-isomers, 12% of tetrahydropyran isomer
1
1c and 75% of thiazole 2, when heated with bromotrichloromethane (Table 2, entry 3).
18

ACCEPTED MANUSCRIPT
Table 2. Products formed from O-(4-pentenoxy)thiazolethiones 1a–c and
bromotrichloromethane
Br
OMTT
S
O
BrCCl / AIBN
O
2
3
+
+
3
3
SCCl3
3
4
N
C H / 80 °C
6
6
Br
OR
OR
OR
1a–c
2
3a–c
11a–c
a
b
entry
1 / R
2 / %
84
3 / % (cis:trans)
3a: 56 (74:26)
3b: 71 (68:32)
3c: 69 (68:32)
b
11 / % (cis:trans)
c
1
2
3
1a / H
1b / Ac
1c / Bz
11a: 4
94
11b: 11 (57:43)
11c: 12 (50:50)
75
a
Stereodescriptors refer to configuration at carbons C2 and C3. Stereodescriptors refer to
c
configuration at carbons C3 and C4. 65/35-Mixture of stereoisomers.
(ii) Stereochemical analysis. For assigning relative configuration of 5-exo-
bromocyclization product 3 we relied on a systematic highfield shift of resonances for carbons
C2, C3, and for exocyclic bromomethyl carbon C1’ in 2,3-cis-isomers. This guideline derives
from a combined NMR-spectroscopy/X-ray-diffraction study on di- and trisubstituted
8
,12
tetrahydrofurans. This phenomenon arises from steric deshielding of interacting nuclei
5
3
(Table 3).
Table 3. Carbon-13 NMR-chemical shifts for cis/trans-isomers of 2,3-substituted
tetrahydrofurans 3a–c (in CDCl , ambient temperature)
3
19

ACCEPTED MANUSCRIPT
Br
1'
Br
1'
H
H
O
O
2
H
2
H
3
3
OR
OR
(
±)-cis-3a–c
(±)-trans-3a–c
entry
3 / R
C2 / ppm
C3 / ppm
C1’ / ppm
1
2
3
4
5
6
cis-3a / H
cis-3b / Ac
cis-3c / Bz
82.2
80.8
81.2
85.1
83.2
83.3
71.9
73.7
74.5
75.1
77.4
78.0
29.0
28.4
28.7
32.8
33.1
33.2
trans-3a / H
trans-3b / Ac
trans-3c / Bz
2
1
.4.3 Products from erythro/threo-isomers of 3-[2,3-bis-(acetyloxy)pent-4-enyloxy]-4-methyl-
,3-thiazole-2(3H)-thione 1d
(
i) Products. O-(2,3-Bisacetyloxypentenyl) thiohydroxamate erythro-1d furnished 67%
of bromoethers erythro-3d and erythro-11d taken together, and 49% of disubstituted thiazole
, when heated in a laboratory microwave to 80 °C in a solution of α,α,α-trifluorotoluene
2
under otherwise standard conditions (Scheme 9, top).
All attempts to chromatographically separate 2,3-cis/trans-isomers of bromoethers 3d
and 11d always gave in our hands product mixtures, however in differing ratios.
Superimposing information gained from such mixtures allowed to characterize all products by
chemical shifts and fine structures of proton resonances. In high resolution mass spectrometra
bisacetates erythro-3d/11d showed the peculiar phenomen of being by one proton heavier than
9 13 5
calculated for the molecular formula C H O Br. This excess mass-per-charge unit is retained
20

ACCEPTED MANUSCRIPT
throughout the fragmentation of erythro-3d/11d, until one of the acetyl groups dissociates off.
Given consistent proton integrals and fine structures in underlying NMR-spectra, we assigned
+
molecular ions C H O Br to derivatives of erythro-3d/11d, generated in the mass
9
14
5
spectrometer, having one proton chelated by two vicinal acetyloxy groups.
Heating stereoisomer threo-1d in a solution of α,α,α-trifluorotoluene (80 °C) containing
bromotrichloromethane and AIBN gave 65% of thiazole 2, and a combined yield of 51% for
bromocyclization products threo-3d and threo-11d (Scheme 9, bottom). For assigning
constitution and configuration, we, again, relied on information gained from analyzing sets of
threo-3d/11d-mixtures containing differing isomer ratios.
Br
H
OMTT
S
O
BrCCl / AIBN
O
3
2
3
+
+
3
SCCl3
4
4
MW / 80 °C
N
H
AcO
H
OAc
AcO
2
AcO
5
Br
3
C H CF
3
6
5
OAc
OR
erythro-11d (13%)
erythro-1d
2 (49%)
erythro-3d (54%)
2,3-cis:trans = 70:30 4,5-cis:trans = 69:31
Br
H
OMTT
S
O
BrCCl / AIBN
O
3
2
3
+
+
3
2
SCCl3
4
4
MW / 80 °C
N
H
AcO
H
OAc
AcO
AcO
3
5
Br
C H CF
3
6
5
OAc
OR
threo-1d
2 (65%)
threo-3d (44%)
threo-11d (7%)
2,3-cis:trans = 50:50 4,5-cis:trans = 71:29
Scheme 9. Products formed from radical reactions between bromotrichloromethane and 3-
erythro-2,3-bis-(acetyloxy)pent-4-enoxy]thiazolethione erythro-1d [top; yields for experiment
conducted via photochemical activation ( = 350 nm, 22 °C, 30 minutes) on a 196-millimolar
scale of 1d in C D containing 9.6 equivalents of BrCCl : 77% of 2, 68% of erythro-3d and
[
λ
6
6
3
21

ACCEPTED MANUSCRIPT
erythro-11d taken together] and stereoisomer threo-1d [bottom; yields for experiment
conducted via photochemical activation (
scale in C , containing 9.0 equivalents of BrCCl
taken together]; MW = microwave.
λ
= 350 nm, 22 °C, 30 minutes) on a 196-millimolar
6
D
6
3
: 53% of 2, 72% of threo-3d and threo-11d
(ii) Stereochemical analysis of tetrahydrofurans. According to the stereochemical
guideline for interpreting carbon-13 NMR-spectra described in section 2.4.2, the major
stereoisomer formed from O-pentenyl ester erythro-1d is the all-cis-configured tetrahydrofuran
ciserythro-3d. This guideline in a similar manner allows to assigning spectral data of cisthreo-3d
from the equimolar mixture of diastereoisomers.
Table 4. Selected carbon-13 NMR-chemical shifts of 3,4-bisacetyloxy-2-bromomethyl-
tetrahydrofurans erythro/threo-3d (in C D , ambient temperature)
6
6
Br
1'
Br
1'
Br
1'
Br
1'
H
H
H
H
O
O
O
O
2
3
2
2
2
3
4
4
4
4
H
H
OAc
H
AcO
H
OAc
H
AcO
H
OAc
H
AcO
H
3
3
AcO
OAc
ciserythro-3d
cisthreo-3d
transerythro-3d
transthreo-3d
entry
3d
C2 / ppm
C3 / ppm
C4 / ppm
C1’ / ppm
1
2
3
4
ciserythro
cisthreo
79.2
80.2
79.6
83.9
71.9
76.4
74.5
80.2
72.2
77.9
72.3
78.4
29.2
28.0
33.3
32.2
transerythro
transthreo
22

ACCEPTED MANUSCRIPT
2
.4.4 Products from 3-[arabino-3,4-bis-(acetyloxy)hex-5-enoxy]-1,3-thiazole-2(3H)-thione 1e
(i) Products. Arabino-configured O-(3,4-bisacetyloxyhexenyl) thiohydroxamate 1e
reacted with bromotrichloromethane in boiling benzene containing AIBN to give 50%
bromomethyltetrahydrofuran 3e, and 10% of bromotetrahydropyran 11e, and 55% of
trisubstituted thiazole 12 (Scheme 10). Chromatography on silica gel allows to separate
tetrahydropyran 4,5-trans-11e from all-cis-configured tetrahydropyran 4,5-cis-11e, and a 2,3-
cis/trans-mixture of bromomethyltetrahydrofuran 3e.
An
Br
H
H
H
H
OMAnTT
S
O
BrCCl / AIBN
O
3
2
+
+
3
SCCl3
4
4
C H / 80 °C
N
H
AcO
H
OAc
6
6
AcO
2
AcO
5
Br
3
3
OAc
OAc
11e (10%)
,3-cis:trans = 30:70 4,5-cis:trans = 20:80
1
e
12 (55%)
3e (50%)
2
Scheme 10. Products formed from 3-[arabino-3,4-bis-(acetyloxy)hex-5-enoxy]-1,3-thiazole-
(3H)-thione 1e and bromotrichloromethane.
2
(ii) Stereochemical analysis. From low field-shifted resonances of carbons C2–C4 and
carbon C1’ we concluded that the major 5-exo-bromocyclization product obtained from
arabino-configured O-hexenoxyl radical Ie is the 2,3-trans-stereoisomer of tetrahydrofuran 3e.
Table 5. Proton and carbon-13 chemical shift values used for distinguishing tetrasubstituted
tetrahydrofuran cis-3e from isomer trans-3e (CDCl , ambient temperature)
3
23

ACCEPTED MANUSCRIPT
Br
1'
Br
1'
H
H
H
H
O
O
2
3
2
4
4
H
AcO
H
OAc
H
AcO
H
OAc
3
cis-3e
trans-3e
entry
3
C2 / ppm
C3 / ppm
C4 / ppm
C1’ / ppm
1
2
cis-3e
77.9
78.1
72.9
74.9
72.4
73.6
29.4
33.7
trans-3e
2
2
.4.5 Products from 3-[rel-(3R,5S)-3-hydroxynona-1,8-dien-5-oxy]-4-methyl-1,3-thiazole-
(3H)-thione 1f
(i) Products. Heating or photolyzing O-nonadienyl thiohydroxamate 1f in
benzene/bromotrichloromethane-solutions gives products of sequential alkoxyl radical
cyclization and bromotrichloromethane addition across the remaining carbon-carbon double
5
2,54
bond.
For preventing such side reactions to occur, we changed the mediator from
bromotrichloromethane to tributylstannane. This modification, in all instances studied so far,
1
,8
never changed stereo- and regioselectivity of alkenoxyl radical additions.
Boiling a solution of O-nonadienyl ester 1f in benzene in the presence of a 3.7-fold
excess of tributylstannane, provided a 26/74-mixture of tetrahydrofurans 14 and 15 in a total
yield of 76%, besides 79% of thiazole 13 (Scheme 11). Tin compound 13 fragments into 2-
methylthiazole-2(3H)-thione and yet unidentified organotin derivatives when being contacted
with silica gel, and therefore was characterized and quantified from the reaction mixture by
2
4
proton-NMR spectroscopy.
2
,5-Substituted tetrahydrofuran 15 formed from O-nonadienyl ester 1f and
24

ACCEPTED MANUSCRIPT
tributylstannane as 27/73-mixture of cis/trans-isomers, which we used for stereochemical
analysis. Based on a model developed from theory for predicting selectivity in 5-exo-
cyclization of monosubstituted 4-pentenoxyl radicals, we think that the major stereoisomer is
1
disubstituted tetrahydrofuran trans-15. Since most resonances of cis/trans-15 overlap, we
unfortunately could not conduct a more sophisticated stereochemical analysis at this point.
From the reaction mixture we furthermore separated trisubstituted tetrahydrofuran 14 as
2
0/80-mixture of 2,3-cis/trans-isomers in a total yield of 20%. This mixture, however, was
contaminated by 13 percent by weight with a third product, showing in high resolution mass
spectra identical molecular masses to tetrahydrofurans 14 and 15. From supplementary NMR-
spectroscopic information we concluded that the third product is a tetrahydropyran formed
from 6-endo-cyclization. Since the yield of the product was rather low (3%), we were not able
to provide at this point a complete structure analysis (see also section 2.5.2). Given the
propensity of hydroxyl-substituted alkoxyl radicals to afford a higher fraction of 6-endo-
cyclized products, we propose that the third product arises from 6-endo-addition of If to the
allylic alcohol segment.
HO
Bu SnH
3
H
H
H
H
H
OMTT
OH
S
AIBN
O
O
+
2
+
5
2
SSnBu3
C H / 80 °C
N
6
6
H
OH
3
rel-(3R,5S)-1f
13 (79%)
14 (20%)
15 (56%)
dr = 20:80
dr = 27:73
Scheme 11. Products formed from 3-[rel-(3R,5S)-3-hydroxynona-1,8-dien-5-oxy]-4-methyl-
,3-thiazole-2(3H)-thione 1f and tributylstannane.
1
25

ACCEPTED MANUSCRIPT
(ii) Analyzing regioselectivity in 5-exo-cyclization. From a doublet of double-doublet
fine structure for the inner alkene proton and larger shift dispersion for the two terminal
protons, we concluded that the major cyclization product obtained from nonadien-5-yl ester 1f
is disubstituted tetrahydrofuran 15 (Figure 4, top and bottom). In tetrahydrofuranol 14, the
resonance of the inner alkene proton is split into a triplet of double-doublet and the terminal
protons are less shift-dispersed (Figure 4, center).
Figure 4. Section of proton NMR-spectra displaying resonances of inner and terminal alkene
protons of tetrahydrofurans 14 and 15, and O-[rel-(3R,5S)-3-hydroxynona-1,8-dien-5-yl]
thiohydroxamate 1f for comparison.
26

ACCEPTED MANUSCRIPT
2
.5 Mechanistic and stereochemical aspects
To find out what causes 3-substitued-4-pentenoxyl radicals Ia–c to cyclize 2,3-cis-
selectively, we at first verified that the intramolecular addition follows kinetic control. For this
purpose, we monitored stereochemical integrity of diastereomerically pure cyclized radicals
cis-IIb and trans-IIb under conditions that extended their life-times by a factor 100.
2
.5.1 On reversibility of 3-acyloxypent-4-en-1-oxyl radical 5-exo-cyclization
i) Kinetic approach. In a kinetically controlled reaction a notable barrier would prevent
(
tetrahydrofuranyl-2-methyl radical cis-IIb or stereoisomer trans-IIb from ring opening to
alkenoxyl radical I in a β-fragmentation. If the oxygen radical formed, a de-novo-cyclization
would become apparent by altering stereochemical integrity of previously pure cis-IIb and
trans-IIb towards an equilibrium mixture presumably of 68/32 (Scheme 12 and eq. 1). By
extending the life-time of radicals cis-IIb and trans-IIb by, for example, a factor 100
compared to conditions applied in bromocyclizations, the stereochemical fingerprint from a
possible ring-opening should become evident.
The kinetic life-time of carbon radical IIb in bromocyclization is defined as inverse
pseudo-first order rate constant for homolytic substitution for brominative trapping. This rate
Br
constant numerically derives from the product between the rate constant k and concentration
5
5
of bromotrichloromethane (eq. 2). In bromocyclization starting from a 0.2 molar solution of
O-alkyl thiohydroxamate 1b, as used in the standard operational procedure, the concentration
27

ACCEPTED MANUSCRIPT
of bromotrichloromethane declines from 1.49±0.01
M
at the beginning towards 0.91±0.01
M
Br
by the end. The rate constant k for trapping primary carbon radicals cis-IIb and trans-IIb, is
7
–1 –1
approximately 3.9×10
M s . This value derives from the experimental rate constant for hept-
Br
8
–1 –1
6
-en-2-yl radical-trapping with bromotrichloromethane (k = 1.2×10
M
s at 26 °C in
5
6
benzene) divided by three. The factor three corrects for a polar rate effect arising from alkyl
5
7
substitution at the radical center. Since rate effects of
β
-substituents are in general small, we
Br
Br
58
assumed that k cis and k trans are identical (approximation 1). Based on these
–
8
approximations, life-times of cis/trans-IIb in standard bromocyclizations are 1.7×10 s at the
–
8
beginning of the reaction, and 2.8×10 s by the end.
H •
H
•
–
cis
1
–1
trans
O
O•
O
k
k
CH2
CH
2
H
H
OAc
OAc
OAc
cis-IIb
Ib
trans-IIb
Br
cis
1
Br
trans
k
BrCCl
3
k
BrCCl
3
k
cis/trans
Br
Br
H
H
H
•
O
O
O
CH
2
H
H
H
OAc
OAc
OAc
cis-3b
IIb
trans-3b
cis:trans ≈ 70:30
d [3b]
d t
=
=
=
kBr [BrCCl ][IIb]
(eq. 1)
3
1
life-time for IIb:
τ
(eq. 2)
kBr [BrCCl3]
Br
cis
Br
trans
k
k
(app. 1)
Scheme 12. Scheme of elementary reactions, kinetic equations (eqs. 1–2) and approximation
28

ACCEPTED MANUSCRIPT
(app. 1) set up for identifying thermochemical contributions to 2,3-cis-selectivity in 5-exo-
cyclization of 3-acetyloxypentenoxyl radical Ib (see also text).
For extending tetrahydrofuran-2-methyl radicals life-times, we changed the mediator to
tributylstannane. The tin compound traps, primary carbon radicals, for example the 1-butyl
radical with a rate constant of k = 2.4×10 M–1 s at 300 Kelvin in benzene, which is 16-
H
6
–1
Br
H
times slower than the reaction with bromotrichloromethane. The ratio k /k remains constant,
if activation parameters for the two processes respond similar to a change in reaction
temperature. On the basis of this assumption, we extrapolated that the life-time of primary
alkyl radicals cis/trans-IIb in a 0.17 molar solution of tributylstannane, as used for conducting
–
6
the controls, extends to 2.5×10 s.
Br
H
H
O
Bu Sn•
Bu SnH
O
3
3
IIb
+ Bu Sn
3
•
H
OAc
H
OAc
–
Bu SnBr
kH
3
3
b
16
d [16]
d t
H
=
k [IIb][Bu SnH]
(eq. 3)
3
Scheme 13. Elementary reactions (top) and differential equation (bottom) for comparing rates
and life-time of alkyl radical trapping under reductive conditions (eq. is short for equation).
(ii) Stereochemical study. For putting our theoretical considerations on radical life-times
29