Chemical Synthesis of Staphyloferrin B Affords Insight into the Molecular Structure, Iron Chelation, and Biological Activity of a Polycarboxylate Siderophore Deployed by the Human Pathogen Staphylococcus aureus

Article abstract of DOI:10.1021/jacs.5b04557

Staphyloferrin B (SB) is a citrate-based polycarboxylate siderophore produced and utilized by the human pathogen Staphylococcus aureus for acquiring iron when colonizing the vertebrate host. The first chemical synthesis of SB is reported, which enables further molecular and biological characterization and provides access to structural analogues of the siderophore. Under conditions of iron limitation, addition of synthetic SB to bacterial growth medium recovered the growth of the antibiotic resistant community isolate S. aureus USA300 JE2. Two structural analogues of SB, epiSB and SBimide, were also synthesized and employed to investigate how epimerization of the citric acid moiety or imide formation influence its function as a siderophore. Epimerization of the citric acid stereocenter perturbed the iron-binding properties and siderophore function of SB as evidenced by experimental and computational modeling studies. Although epiSB provided growth recovery to S. aureus USA300 JE2 cultured in iron-deficient medium, the effect was attenuated relative to that of SB. Moreover, SB more effectively sequestered the Fe(III) bound to human holo-transferrin, an iron source of S. aureus, than epiSB. SBimide is an imide analogous to the imide forms of other citric acid siderophores that are often observed when these molecules are isolated from natural sources. Here, SBimide is shown to be unstable, converting to native SB at physiological pH. SB is considered to be a virulence factor of S. aureus, a pathogen that poses a particular threat to public health because of the number of drug-resistant strains emerging in hospital and community settings. Iron acquisition by S. aureus is important for its ability to colonize the human host and cause disease, and new chemical insights into the structure and function of SB will inform the search for new therapeutic strategies for combating S. aureus infections. (Chemical Equation Presented).

Full text of DOI:10.1021/jacs.5b04557

Subscriber access provided by UNIV OF MISSISSIPPI
Article
Chemical Synthesis of Staphyloferrin B Affords Insight into the Molecular
Structure, Iron Chelation, and Biological Activity of a Polycarboxylate
Siderophore Deployed by the Human Pathogen Staphylococcus aureus
Julie L.H. Madsen, Timothy C. Johnstone, and Elizabeth M. Nolan
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b04557 • Publication Date (Web): 01 Jun 2015
Downloaded from http://pubs.acs.org on June 23, 2015
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.
Journal of the American Chemical Society 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 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Chemical Synthesis of Staphyloferrin B Affords Insight into the Molecular Structure, Iron
Chelation, and Biological Activity of a Polycarboxylate Siderophore Deployed by the
Human Pathogen Staphylococcus aureus
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
Julie L. H. Madsen,a Timothy C. Johnstonea and Elizabeth M. Nolana,*
aDepartment of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue,
Cambridge, MA 02139, USA
*Corresponding author: lnolan@mit.edu
Phone: 617-452-2495
Fax: 617-324-0505
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 32
1
2
3
4
5
Abstract
6
7
8
9
Staphyloferrin B (SB) is a citrate-based polycarboxylate siderophore produced and utilized by
the human pathogen Staphylococcus aureus for acquiring iron when colonizing the vertebrate
host. The first chemical synthesis of SB is reported, which enables further molecular and
biological characterization and provides access to structural analogues of the siderophore.
Addition of synthetic SB to bacterial growth medium recovered the growth of the antibiotic
resistant community isolate S. aureus USA300 JE2 cultured under conditions of iron limitation.
Two structural analogues of SB, epiSB and SBimide, were also synthesized and employed to
investigate how epimerization of the citric acid moiety or imide formation influence its function as
a siderophore. Epimerization of the citric acid stereocenter perturbed the iron-binding properties
and siderophore function of SB as evidenced by experimental and computational modeling
studies. Although epiSB provided growth recovery to S. aureus USA300 JE2 cultured in iron-
deficient medium, the effect was attenuated relative to that of SB. Moreover, SB more effectively
sequestered the Fe(III) bound to human holo-transferrin, an iron source of S. aureus, than
epiSB. SBimide is an imide analogous to the imide forms of other citric acid siderophores that
are often observed when these molecules are isolated from natural sources. Here, SBimide is
shown to be unstable, converting to native SB at physiological pH. SB is considered to be a
virulence factor of S. aureus, a pathogen that poses a particular threat to public health because
of the number of drug-resistant strains emerging in hospital and community settings. Iron
acquisition by S. aureus is important for its ability to colonize the human host and cause
disease, and new chemical insights into the structure and function of SB will inform the search
for new therapeutic strategies for combating S. aureus infections.
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
ACS Paragon Plus Environment

Page 3 of 32
Journal of the American Chemical Society
1
2
3
4
5
Introduction
6
7
8
9
The Gram-positive human pathogen Staphylococcus aureus is acknowledged as a serious
threat to public health because of its ability to subvert the human innate and adaptive immune
responses, cause life-threatening disease, and acquire multi-drug resistance.1-3 Deciphering the
numerous strategies that S. aureus employs to colonize a host and cause infection at the
molecular level is important for developing new therapeutics to treat S. aureus infections.2 Iron
is an essential nutrient for almost all organisms,4-6 and S. aureus must acquire this element in
the host to colonize and cause infection.7,8 One strategy that S. aureus employs to acquire iron
involves the biosynthesis and utilization of siderophores,9,10 which are small-molecule iron(III)
chelators produced by bacteria confronted with conditions of iron limitation.11-13 Despite
significant biological interest in how staphylococcal siderophores contribute to pathogenesis, a
paucity of chemical information about these molecules exists. In this work, we combine
experimental chemistry, computational modeling, and microbiology to provide the first chemical
synthesis and allow structural and functional studies of staphyloferrin B, a polycarboxylate
siderophore and virulence factor of S. aureus.
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
Staphyloferrin A (SA) and B (SB) are the two known siderophores produced by S.
aureus.9,14 SB was first identified in cultures of S. aureus in 1990; however, the isolation was
complicated by the instability of the molecule under acidic conditions.9,14 SB was successfully
isolated from culture supernatants of the veterinary pathogen S. hyicus DSM 20459 in 1993,15,16
and was next isolated from culture supernatants of the metal-tolerant Gram-negative soil
bacterium Cupriavidus metallidurans CH34 (formerly Ralstonia eutropha and Alcaligenes
eutrophus)17,18 in 199618,19 and 1999.20 S. aureus biosynthesizes SB from three building blocks:
L-2,3-diaminopropionic acid (L-DAP), citric acid, and α-ketoglutaric acid (αKG). Structural
determination of SB isolated from S. hyicus and C. metallidurans resulted in two different
structures (Figure 1).15,20 SB isolated from S. hyicus was reported to harbor a linearized αKG
moiety,15 whereas a cyclic hemiaminal αKG structure was described for the C. metallidurans
isolate.20
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 32
1
2
3
4
Staphyloferrin B (SB)
5
6
O
O
C
O
O
H
O
H
7
8
9
O
O
OH
HO
HN
NH
H2N
O
HN
NH
H2N
O
HN
O
OH
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
HO
N
O
OH
O
O
O
HO
1, SB
Hemiaminal structure
C. metallidurans 1999, this study
Linear structure
S. hyicus 1993
SB analogues
O
O
OH
(R)
O
O
O
H2N
HN
NH
HN
N
O
N
H2N
O
O
HO
HO
O
HO
OH
OH
N
HO
O
O
OH
Epimer of SB
Imide of SB
3, SBimide
2, epiSB
Figure 1. Chemical structures and nomenclature of staphyloferrin B (SB, 1) and two structural
analogues. The two structures previously reported for SB are shown.15,20 SB and the analogues
epiSB 2 and SBimide 3 were prepared and investigated in this work. The structural differences
of epiSB and SBimide compared to native SB are colored in blue.
The linear SB structure contains two stereocenters (C2 and C6, Figure 2) that arise from
the DAP and citrate building blocks. The hemiaminal structure contains a third stereocenter at
the hemiaminal quaternary carbon atom (C15, Figure 2). Based on feeding studies in which
increased production of SB by S. hyicus was observed when the culture medium was
supplemented with L-DAP but not with D-DAP, the DAP moiety of the siderophore was assigned
to have (S)-stereochemistry.15 The assigned (S)-stereochemistry of the citric acid moiety was
supported by a comparison of the circular dichroism (CD) spectrum of native SB isolated from
C. metallidurans with the spectra obtained for synthetic DAP-citric acid fragments of SB with
either (R)- or (S)-stereochemistry at the citric acid moiety.20 The (S)-configuration of the citric
acid moiety was further supported by the recent crystal structure of SB in complex with its
receptor SirA.21,22 For the hemiaminal structure, the stereocenter of the hemiaminal quaternary
4
ACS Paragon Plus Environment

Page 5 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
carbon center was reported as a mixture of two isomers in rapid equilibrium via the linear SB
structure.20 To date, a structural inconsistency persists in the literature with either the linear or
hemiaminal form of SB depicted in different reports, and the linear SB structure is most often
shown.
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
O
4
O
9
OH
(*)
7
O
13
12
O
5
O
H2N
2
HN
NH
HN
3
14
10
O
15
N
5
8
7
6
4
6
O
N
H2N
10
3
O
11
11
1
OH
HO
16
OH
2
HO
16
1
HO
N
15
9
HO
8
O
12
O
OH
O
14 13
(S): 1, SB
3, SBimide
(R): 2, epiSB
Figure 2. Atom numbering for NMR assignments.
The proteins involved in SB biosynthesis and transport in S. aureus have been identified
and characterized.22-30 The sbn gene locus (sbnA-I) encodes the full biosynthetic machinery
(SbnABCEFGH) for the production of SB from the metabolic precursors oxaloacetate, acetyl-
coenzyme A, L-glutamate, and O-phospho-L-serine. SbnAB are required for the biosynthesis of
L-DAP and αKG from L-glutamate and O-phospho-L-serine.26,27 SbnG is an unusual iron-
regulated citrate synthase that supplies citrate to the SB biosynthetic pathway by catalyzing an
aldol condensation between oxaloacetate and acetyl-CoA.28 SbnCEF are three non-ribosomal
peptide synthetase (NRPS) independent siderophore (NIS)31 synthetases that work in concert
with the decarboxylase SbnH to assemble SB from citrate, αKG, and two equivalents of L-
DAP.25 SbnD is an efflux pump that contributes to SB export, and the role of SbnI is as-yet
undetermined.23,30 The enzymes SbnABG enable SB biosynthesis in the absence of the
tricarboxylic acid (TCA) cycle,32 the latter of which is down-regulated by S. aureus under
conditions of iron limitation.33 In contrast, the biosynthetic machinery required for SA production,
sfaABCD, employs citrate produced by the TCA cycle.32,34 Because S. aureus reduces TCA
cycle activity in response to the iron-limited environment of the vertebrate host, the differing
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 32
1
2
3
4
5
6
7
TCA cycle dependence of SA and SB biosynthesis suggests that SB plays a larger role in S.
aureus virulence than SA.32
8
9
A variety of in vitro and in vivo studies indicate that the SB gene locus contributes to S.
aureus virulence.24,34-37 S. aureus can enter the bloodstream and cause sepsis, and one set of in
vitro experiments demonstrated that deletion of the sbnA-I gene locus resulted in impaired
growth of S. aureus strains Newman and RN6390 in horse serum.34 Moreover, deletion of the
NIS synthetase SbnE resulted in impaired growth of S. aureus Newman in a mouse kidney
abscess model as well as in a rat model of infective endocarditis.24,37 Transcriptional analysis of
S. aureus isolates from a rat model of infective endocarditis revealed that the sbn locus was up-
regulated.37 SB is also biosyntheized by some coagulase-negative (CoN) staphylococci. A
comparison of data obtained from clinical isolates of CoN staphylococci cultured from the
peritoneal dialysis fluid of patients showing signs of infection with the data from commensal skin
isolates obtained from healthy individuals was performed.35,36 The CoN staphylococci isolated
from the dialysis fluid exhibited higher expression levels of siderophores, particularly SB, than
the commensal organisms.35,36 Taken together, these studies indicate that the sbn gene cluster
and SB biosynthesis contribute to the pathogenicity of S. aureus, and that SB production is
linked to the severity of infection at least in animal models. Indeed, strategies to inhibit SB
biosynthesis are of interest for therapeutic purposes. Two natural products, baulamycin A and
B, isolated from a library of marine microbial-derived natural products were recently identified as
inhibitors of SbnE.38 Further chemical and biological studies addressing how SB contributes to
S. aureus virulence and disease progression are needed in light of the public health threat
posed by recalcitrant S. aureus strains.
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 this work, we approach this need from a chemical perspective. We report, to the best
of our knowledge, the first total synthesis of SB 1. This twelve-step synthesis enabled access to
two structural analogues, epiSB 2 and SBimide 3 (Figure 1), as well as chemical and biological
evaluation of the resulting siderophores. We address the inconsistent depictions of the SB
chemical structure in the literature, confirm that SB is the hemiaminal form, and reevaluate the
6
ACS Paragon Plus Environment

Page 7 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
crystal structure of Fe(III)-SB in complex with its receptor SirA.21 Moreover, our studies of epiSB
demonstrate the importance of (S)-stereochemistry at the citric acid moiety for the siderophore
function of SB. This work provides new chemical insights into this important staphylococcal
secondary metabolite and virulence factor, a synthetic platform for the design and preparation of
additional SB analogues, and a foundation for future studies at the chemistry/biology interface of
how SB contributes to iron homeostasis and microbial pathogenesis.
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
Results and Discussion
Chemical Synthesis of Staphyloferrin B and Structural Analogues. To the best of
our knowledge, no total synthesis of SB has been reported, which limits the availability of the
native siderophore for investigations of its coordination chemistry, biological activity, role as a
virulence factor, and potential in therapeutic design. Structural analogues of SB, which can be
readily accessed once a total synthesis has been achieved, are also invaluable in such
investigations. In prior structural and functional studies, SB was isolated from bacterial cultures
or obtained by in vitro reconstitution of the SB biosynthetic pathway.15,16,20,25 An unresolved
issue concerns the chemical structure of SB, which has been inconsistently presented in the
literature (Figure 1). To enable further physiochemical and biological studies of SB, gain
synthetic access to structural analogues of the siderophore, and address the structural
discrepancy, we designed and executed a total synthesis that affords native SB (Scheme 1).
The synthesis was achieved in twelve steps from four commercially available starting materials:
(S)-malic acid, Z-L-asparagine, 1,2-diaminoethane (DAE) and αKG. The major challenges
posed by the synthesis were control of the stereogenic center of the citric acid moiety and
control/prevention of imide formation by the citric acid moiety during the reactions and
purification steps (Supporting Information).
The synthesis is based on three building blocks, 6, 10, and 13. Cbz- and benzyl-protected L-
DAP 6 was prepared in two steps from Z-L-asparagine by initial benzyl protection of the
carboxylic acid followed by a Hofmann rearrangement to generate the primary amine 6.
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 32
1
2
3
4
5
6
7
8
9
Coupling of N-boc-diaminoethane 12 to mono-benzylated α-ketoglutarate 11, the latter of which
was prepared according to a procedure in the patent literature,39 afforded the hemiaminal
product 13. The cyclic hemiaminal structure of 13 was confirmed by HSQC and HMBC 2D NMR
spectroscopy, which revealed a correlation between the internal methylene protons 5 of the
diaminoethane moiety to both the quaternary stereocenter 9 and the carbonyl carbon 6 of the
hemiaminal ring (Figure S1). In order to control the stereochemistry of the citric acid moiety
throughout the synthesis, a citric acid synthon 10 was employed. Masking one of the
enantiotopic carboxylic acids of citric acid as an alkene afforded full control of the
stereochemistry. Starting from (S)-malic acid, reaction with trimethylacetaldehyde afforded
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
dioxolanone
8
stereoselectively.40 Allylation of the dioxolanone using lithium
hexamethyldisilazide (LiHMDS) as the base provided alkene 9, which was ring-opened in a
transesterification reaction to afford benzyl ester 10.41 Intermediates 6, 10, and 13 were
subsequently employed as building blocks for the assembly of SB and the structural analogues.
Prior to any coupling of the hemiaminal building block 13, the carbamate protecting group was
removed immediately prior to the coupling reaction by using trifluoroacetic acid (TFA) in
dichloromethane to yield amine 14.
8
ACS Paragon Plus Environment

Page 9 of 32
Journal of the American Chemical Society
1
2
O
3
4
5
6
O
NH2
Building blocks
NH2
OBn
NH2
OH
a,b
c
OBn
CbzHN
CbzHN
CbzHN
O
O
O
4
5
6
7
8
9
O
O
O
Bn
O
OH
O
HO
O
d
e
f
O
O
OH
O
O
HO
C
OH
HO
O
O
O
HO
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
8
9
10
7
O
O
O
O
g
h
NH2
N
N
O
+
BocHN
BocHN
TFA H2N
HO
OBn
OH
OH
O
OBn
OBn
O
11
12
13
14
Staphyloferrin B (SB)
B n
O
O
O
O
Bn
Bn
C O O H
OH
HO
O
O
C
OH
O
O
Bn
O
O
O
j
NH
k
l
i
O
O
O
OH
HN
NH
C
OH
C
HN
O
O
C
OH
CbzHN
O
CbzHN
O
H2N
CbzHN
OH
N
H
OBn
N
HO
H
HO
HO
N
N
O
BnO
O
OBn
O
OH
O
OBn
O
O
10
15
16
17
1, SB
Staphyloferrin B epimer (epiSB)
Bn
O
O
O
O
OH
HO
O
Bn
Bn
O
OH
C
O
O
Bn
O
O
O
O
O
p
NH
HN
m
n
o
O
O
O
OH
NH
HN
C
OH
C
O
O
C
OH
H2N
O
N
N
CbzHN
O
OBn
OBn
N
HO
N
H
H
HO
HO
HO
O
N
HO
HO
N
O
BnO
O
OH
O
O
OBn
O
10
18
19
20
2, epiSB
Staphyloferrin B imide (SBimide)
O
O
O
O
O
O
O
O
O
O
RHN
O
H2N
O
HN
HN
O
q
r
s
u
O
O
N
N
BocHN
O
O
O
O
N
N
BocHN
O
HO
N
H
OtBu
O
O
O
HO
HO
N
O
HO
HO
OR
OBn
OH
OH
H
O
O
OtBu
23, R1 = Boc, R2 = tBu
24, R1 = H, R2 = H
t
9
21
22
3, SBimide
(a) Cs2CO3, H2O, MeOH; (b) BnBr, DMF, 76% yield over two steps; (c) PhI(OAc)2, pyridine, DMF, H2O, 50% yield; (d) Trimethylacetaldehyde, p-TsOH, H2SO4, pentane,
42 oC, 40% yield; (e) Allyl bromide, LHMDS, THF, -78oC, 51% yield; (f) BnOH, NaH, 0 oC, 98% yield; (g) EDC, HOAt, DIPEA, CH2Cl2, 60% yield; (h) 33% TFA in CH2Cl2,
0 oC, quant. yield; (i) 6, HATU, HOAt, DIPEA, CH2Cl2, 69% yield; (j) NaIO4, RuCl3, MeCN, H2O, 89% yield; (k) 14, HATU, HOAt, DIPEA, CH2Cl2, 43% yield;(l) Pd(OH)2/C,
H2, MeOH, 87% yield; (m) 14, HATU, HOAt, DIPEA, CH2Cl2, 72% yield; (n) NaIO4, RuCl3, MeCN, H2O, 86% yield; (o) 6, HATU, HOAt, DIPEA, CH2Cl2, 38% yield;
(p) Pd(OH)2/C, H2, MeOH, 93% yield; (q) Boc-DAP-OtBu, HATU, HOAt, DIPEA, CH2Cl2, 75% yield; (r) NaIO4, RuCl3, EtOAc, MeCN, H2O, 89% yield; (s) 14, EDC, HOAt,
NMM, CH2Cl2, 45% yield; (t) 10% H2O, 45% TFA, 45% CH2Cl2, 4 oC, 66% yield; (u) Pd(OH)2/C, H2, MeOH, 86% yield.
Scheme 1. Syntheses of SB 1, epiSB 2, and SBimide 3.
SB was prepared in four steps beginning with the coupling of synthon 10 to Z-DAP-OBn
6 in a 30-min amide coupling reaction that employed HATU/HOAt as coupling reagents. The
resulting alkene 15 was oxidized to the corresponding carboxylic acid with in situ-generated
perruthenate as the oxidant,42 and the corresponding carboxylic acid 16 was employed directly
in an amide coupling reaction with hemiaminal 14, affording fully protected SB 17. The product
was obtained as a mixture of stereoisomers as a result of the equilibrating hemiaminal
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 32
1
2
3
4
5
6
7
8
9
quaternary carbon. This mixture of isomers was also reported for SB that was isolated from
natural sources.20 In the final step, the siderophore was globally deprotected in a hydrogenation
reaction, which provided SB 1 with an overall yield of 22% for these four steps.
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
The epimeric analogue of SB, epiSB 2, differs from SB only in the stereochemistry of the
citric acid moiety where it possesses (R)-stereochemistry. This analogue was synthesized using
the same building blocks employed for SB and simply reversing the order of the coupling
reactions. By coupling the citric acid synthon 10 to hemiaminal 14 prior to oxidation of the
alkene, the stereochemistry of the citric acid moiety in the resulting siderophore product was
inverted. Following the same reaction procedures described for SB, epiSB 2 was also obtained
in an overall yield of 22% for these four steps.
The imide analogue of SB, SBimide 3, differs from SB by having the carboxylic acid of
the citric acid moiety closed in an imide ring with the neighboring amide nitrogen of the DAE
moiety. A different synthetic strategy was employed to obtain SBimide in a regioselective
manner. Two regioisomers of the imide can be formed, and both isomers were observed as
byproducts during preliminary studies of the total synthesis of SB when no precautions were
taken to prevent this side reaction (see Supporting Information for further discussion). SBimide
was prepared regioselectively where only the DAE amide nitrogen participates in imide
formation (Scheme 1). Dioxolanone 9 was found to promote imide formation during amide
couplings of extended reaction time when applied as the citric acid synthon. Thus, to afford 21
without imide formation, amide coupling between commercially available Boc-DAP-OtBu and
dioxolanone 9 was performed using HATU/HOAt for 30 min. Following oxidation of alkene 21 to
afford carboxylic acid 22, this intermediate was allowed to react with hemiaminal 14 in a second
amide coupling using EDC/HOAt as coupling reagents. The coupling reaction time was
extended to 20 h, providing the DAE-closed imide 23 as the major product. Following
purification and global deprotection, SBimide 3 was obtained in an overall yield of 17% for these
five steps.
10
ACS Paragon Plus Environment

Page 11 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
SB Displays a Cyclic Hemiaminal Structure. The reported chemical structure of SB is
inconsistent throughout the literature (Figure 1). In 1993, a chemical structure of SB containing
a linear αKG moiety was reported.15 The 13C NMR spectroscopic data presented in this early
study revealed that the chemical shift of one carbonyl carbon atom was shifted from the
expected ≈200 ppm to 91 ppm. This change in chemical shift was rationalized by hydration of
the corresponding ketone. In 1999, a structure of SB containing a cyclic hemiaminal αKG moiety
was reported.20 This study presented a 13C NMR spectrum that exhibited a resonance with a
chemical shift of 94 ppm. This peak was shown by 2D NMR to arise from hemiaminal formation
of the αKG moiety, resulting in a quaternary carbon center. Although the correct structure of SB
seems to be the hemiaminal structure on the basis of the 2D NMR analysis and the similar 13C
NMR spectra obtained for the two isolates, the linear structure of SB is more frequently depicted
in the literature. To help clarify this matter, we studied synthetic SB by NMR spectroscopy and
compared these data to those obtained for SB isolated from natural sources (Table S3 and S4).
The observed chemical shifts of all proton and carbon atoms of synthetic SB correspond well
with the reported data,15,20 and minor differences can be attributed to the different spectroscopic
conditions. We employed HMBC 2D NMR spectroscopy to identify long-range correlations in the
molecule. Clear correlations between methylene protons 11 of the DAE moiety and both the
hemiaminal ring quaternary carbon 15 and carbonyl carbon 12 were observed, confirming that
the structure of SB in aqueous solution contains the cyclic hemiaminal (Figures 2, S2).
The NMR spectra of epiSB are very similar to those of SB (Table S3 and S4, Figure S3).
A comparison of the spectra of SB to those of SBimide reveals a clear difference for the
methylene protons 5 and 8 of the citric acid moiety (Figure 2, Table S3). For SB, these four
protons produce four distinct doublets, whereas the same proton signals of SBimide overlap and
result in two doublets and one multiplet signal. In addition, the HMBC spectrum of SBimide
showed correlations between the two carbonyl carbons 7 and 9 of the imide ring and the
neighboring methylene protons 10 of the DAE moiety, making the distinction of the two
compounds unambiguous (Figures 2,S4).
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
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 32
1
2
3
4
5
6
7
8
9
Re-Evaluation of a Crystal Structure of Fe(III)-SB Bound to SirA. The hemiaminal
structure of SB reported in 1999 is largely forgotten in the siderophore and S. aureus literature
for reasons that are unclear. Recently, a crystal structure of Fe(III)-SB in complex with SirA, the
SB receptor of S. aureus, was reported with the linear structure of Fe(III)-SB modeled into the
binding site (PDB 3MWF).21 In this structural determination, the electron density of the
siderophore backbone spanning the citrate, DAE, and αKG moieties is not well defined, as
noted in the original report.21 Our confirmation of the hemiaminal structure of SB in aqueous
solution motivated us to investigate the electron density map of Fe(III)-SB-SirA and determine
whether the hemiaminal form provides a better fit to the electron density corresponding to the
bound siderophore. In order to perform this analysis, we determined the lowest energy structure
of Fe(III)-SB to be Λ-Fe(III)-SB(SSR) (vide infra, Table S5, Figure S5) by computational
methods; a comparison of the corresponding calculated CD spectrum with that obtained
experimentally for Fe(III)-SB provided support for the validity of the computational result (vide
infra). When we aligned the computationally optimized Λ-Fe(III)-SB(SSR) hemiaminal structure
with the linear structure reported in the crystal structure of Fe(III)-SB-SirA, we found that the
structures of the iron complexes differ significantly only in the vicinity of the DAE and αKG units
(Figure 3). Further analysis of the electron density in the area where the hemiaminal and linear
structures differ indicates that the hemiaminal form of SB provides a better fit to the electron
density than the linear structure (Figure 3). This analysis reveals that the hemiaminal Fe(III)-SB
structure, not the linear structure, is bound to SirA in the crystal structure, and provides further
support for the conclusion that the hemiaminal is the form of SB isolated from S. aureus
cultures.
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
12
ACS Paragon Plus Environment

Page 13 of 32
Journal of the American Chemical Society
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 3. Overlay of the calculated Fe(III)-SB structure based on the hemiaminal form (dark
grey carbon sticks) and the crystallographically determined Fe(III)-SB (light grey carbon sticks).
(A) Chemical structure of Fe(III)-SB. (B) The Fo map, contoured at 2σ and 1.6 Å from the atoms
modeled into the electron density, is shown as a grey mesh. The 2Fo–Fc map, contoured at 3σ
showing (C) positive electron density as a blue mesh and (D) negative electron density as a red
mesh. Color code: C: light or dark grey, O: red, N: blue, and Fe: yellow sphere.
Fe(III)-Binding Properties and Photochemistry of SB. SB chelates Fe(III) in a
hexadentate manner, utilizing the two α-hydroxy carboxylic acids and the α-amino carboxylic
acid as three bidentate ligands (Figure 3). This coordination mode is supported by both the
crystal structure of Fe(III)-SB bound to SirA21 and an NMR study of Ga(III)-SB.20 Solution
studies of Fe(III) chelation by SB are limited, however. Elucidating the Fe(III)-binding properties
of SB in solution is essential for achieving a thorough understanding of the role of SB in iron
homeostasis and as a virulence factor of S. aureus as well as informing experimental design for
studies that utilize the Fe(III)-SB complex. We therefore employed synthetic SB to examine the
Fe(III) coordination complex. In agreement with previous reports, addition of Fe(III) to SB
resulted in formation of a charge-transfer band centered at ≈330 nm (Figure 4A).15,16 Moreover,
the CD spectrum of Fe(III)-SB revealed a positive Cotton effect at the wavelengths of the
charge-transfer band, which is also in agreement with the CD spectrum of Fe(III)-SB isolated
from S. hyicus cultures (Figure 4B).15
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 32
1
2
3
4
5
6
7
8
9
A
B
C
1.50
Fe(III)-SB
Fe(III)-epiSB
SB
epiSB
1.25
1.00
0.75
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
0.50
0.25
0.00
+Fe(III)
250
300
350
λ (nm)
400
450
500
4000
3000
2000
1000
0
Fe(III)-SB
Fe(III)-epiSB
SB
epiSB
250
300
350
400
450
500
-1000
-2000
λ (nm)
30
20
10
0
Fe(III)-SB
Fe(III)-epiSB
350
400
450
500
550
600
-10
-20
λ (nm)
Figure 4. Iron-binding studies of SB and epiSB. (A) Optical absorption spectrum of 200 µM apo
and Fe(III)-bound siderophores in 10 mM HEPES, 100 mM KCl, pH 7.4. (B) CD spectra of 300
µM apo- and Fe(III)-bound siderophores in Milli-Q water. (C) Theoretical CD spectra of Λ-Fe(III)-
SB(SSR) and Δ-Fe(III)-epiSB(SRS) as determined by TD-DFT calculations. SB (black), epiSB
(green), Fe(III)-SB (red), and Fe(III)-epiSB (blue).
14
ACS Paragon Plus Environment

Page 15 of 32
Journal of the American Chemical Society
1
2
O
3
4
5
6
7
8
9
NH
O
OH
O
A
B
O
O
NH
HN
O
sunlight
N
H2N
O
O
NH
O
O
O
HO
O
Fe
N
HO
O
H2N
OH
CO2
Fe(II)
O
O
H
O
Fe(III)-SB
SBphoto
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
1.5
0 min, Fe(III)-SB in dark
30 min
60 min
120 min
180 min
300 min
1.0
0.5
0.0
200 250 300 350 400 450 500
λ (nm)
C
1.5
0 min, Fe(III)-SB in sun
30 min
60 min
120 min
180 min
300 min
1.0
0.5
0.0
200 250 300 350 400 450 500
λ (nm)
Figure 5. Photoreactivity of Fe(III)-SB. (A) Proposed photoreaction of Fe(III)-SB. (B) Optical
absorption spectra for Fe(III)-SB incubated in the dark. (C) Optical absorption spectra for Fe(III)-
SB exposed to sunlight for the indicated times. The photoreactions were performed with 200 µM
Fe(III)-SB in 10 mM HEPES, 100 mM KCl, pH 7.5, rt.
Because Fe(III) complexes of citric acid siderophores, including those of petrobactin43
and aerobactin,44 undergo photochemistry when exposed to sunlight, we questioned whether
Fe(III)-SB exhibits similar photochemistry. The photoreaction of these siderophores involves
decarboxylation of the carboxylic acid at the citric acid moiety, followed by oxidation of the
resulting secondary alcohol to generate a β-hydroxy-α,β-unsaturated amide, and the Fe(III) is
concomitantly reduced to Fe(II) (Figure 5A).43,44 The optical absorption spectrum of Fe(III)-SB in
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 32
1
2
3
4
5
6
7
8
9
aqueous solution remained unchanged when the complex was incubated in the dark (Figure
5B). In contrast, when we exposed an aqueous solution of Fe(III)-SB to sunlight and recorded
the optical absorption spectrum at various time points, formation of a new absorbance feature
centered at 290 nm occurred over the course of 3 h (Figure 5C), revealing that Fe(III)-SB is
photoreactive. This absorbance feature is in agreement with previous reports of photoproducts
from structurally related siderophores.43,44 Light was therefore excluded during all chemical and
biological experiments with the Fe(III)-siderophore complexes.
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
Epimerization of the Citric Acid Stereocenter Perturbs the Fe(III)-binding
Properties of SB. During some exploratory experiments with SB and its structural analogues,
we performed a chrome azurol S agar diffusion (CASAD) assay.45 The chrome azurol S (CAS)
assay provides colorimetric detection of apo siderophores and is routinely used to screen for the
presence of siderophores in bacterial culture supernatants.46 Siderophore detection is based on
the ability of siderophores to sequester Fe(III) from the indicator formed from CAS,
hexadecyltrimethylammonium (HDTMA) bromide, and Fe(III). Siderophore sequestration of
Fe(III) from the CAS results in a color change from blue to orange. When performed using agar
diffusion plates (CASAD assay), siderophores can be detected based on the radius and color of
the orange disks formed following addition of a solution containing siderophores to the plate.
The “siderophore activity” is qualitatively assayed by the color intensity and size of the orange
disks. Siderophores that coordinate Fe(III) with relatively high affinity will afford brighter orange
disks relative to siderophores that bind Fe(III) with relatively low affinity. Moreover, higher
siderophore concentrations will result in larger disks because more siderophores diffuse in the
plate. When we spotted SB on CASAD plates, we observed orange spots that indicated
sequestration of Fe(III) from the indicator and, as expected, the radius of the spots increased
with increasing the concentration of SB (Figure 6). A comparison of the spots obtained for
equimolar quantities of SB and epiSB suggested that epiSB coordinates Fe(III) with lower
affinity than SB. Less orange color evolved for the epiSB spots relative to those for SB. This
16
ACS Paragon Plus Environment

Page 17 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
qualitative, colorimetric assay provided the first evidence that epimerization of the citric acid
stereocenter perturbs the Fe(III)-binding properties of SB.
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
10 mM 5 mM 2.5 mM
SB
epiSB
SBimide
Figure 6. Chrome azurol S agar diffusion (CASAD) assay comparing Fe(III) scavenging by SB,
epiSB, and SBimide. The agar plate was spotted with each siderophore at concentrations of 2.5,
5, and 10 mM (10 µL/spot).
Similarly to Fe(III)-SB, the optical absorption spectrum of Fe(III)-epiSB exhibits a charge-
transfer band centered at ≈330 nm (Figure 4A). An overlay of the optical absorption spectra of
Fe(III)-SB and Fe(III)-epiSB reveals subtle differences, which suggests that the Fe(III)
coordination sphere is perturbed by epimerization of the citric acid moiety. As expected, the CD
spectrum of Fe(III)-epiSB differs from that of SB. The molar ellipticity of the Fe(III)-epiSB
complex shifts from positive to negative across the 275-400 nm region of the spectrum and the
intercept of the Fe(III)-epiSB spectrum with the abscissa coincides with the wavelength of
maximum molar ellipticity (315 nm) of the Fe(III)-SB complex (Figure 4B).
To obtain more quantitative insight into how epimerization of the citric acid stereocenter
perturbs the Fe(III)-binding properties of SB, we determined pFe(III) values for SB and epiSB.
The pFe(III) is defined as the negative logarithm of the unbound iron concentration in the
presence of 10 µM ligand and 1 µM Fe(III) at pH 7.4.47 This value is commonly used to compare
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 32
1
2
3
4
5
6
7
8
9
the Fe(III) affinities of siderophores. We performed EDTA competition titrations where the
Fe(III)/EDTA/siderophore mixtures were incubated for 16 h at room temperature prior to
analysis, and determined the pFe(III) values of SB and epiSB to be 23.6 and 23.2, respectively,
at this time point (Figure S8). These pFe(III) values are similar to the experimentally determined
pFe(III) values of other citric acid siderophores including petrobactin (pFe = 23.0)48 and
aerobactin (pFe = 23.3)47 (Table S7). Although similar in value, the relative pFe(III) values
indicate that epiSB coordinates Fe(III) with somewhat lower affinity than native SB and are in
agreement with the CASAD assay results. One caveat to these pFe(III) experiments and the
data interpretation involves the equilibration time. Literature procedures for pFe(III)
determination typically describe incubation times of 16-24 h, and these protocols provided a
basis for our initial experimental design.48-50 We attempted to determine pFe(III) values at time
points >16 h to probe whether a longer incubation period is required for equilibrium to be
achieved; however, these experiments were compromised because of decomposition of
SB/Fe(III)-SB as evidenced by subtle changes in the optical absorption spectra of Fe(III)-SB
over time (up to seven days). Thus, whether the measurements were performed using samples
that had reached equilibrium is unclear.
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
Computational Modeling Provides Further Insight into the Fe(III)-SB Coordination
Sphere. To gain further insight into the consequences of epimerization of the citric acid
stereocenter on Fe(III) chelation, we investigated the structures of Fe(III)-SB and Fe(III)-epiSB
using computational methods. The six donor atoms of SB and epiSB are chemically distinct,
affording an M(ABCDEF) system that can occur as 30 different geometric isomers. The majority
of these possible structures are not relevant because of steric constraints imposed by the
ligand. Because SB is composed of three bidentate ligands, the skew-lines convention (Δ/Λ) can
be used to describe the helicity at the metal center.51 Four isomers were constructed manually
for each Fe(III)-SB and Fe(III)-epiSB isomer spanning the combinations allowed by Δ or Λ
helicity and R or S chirality at the quaternary hemiaminal carbon center 15 (Figure 2). The
conformations of these hand-built structures were first subjected to a semi-empirical
18
ACS Paragon Plus Environment

Page 19 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
conformational search to determine the energetically feasible conformations. Subsequent
density functional theory (DFT) calculations revealed that the lowest energy Fe(III)-SB structure
(-2946.206 Eh) was Λ-Fe(III)-SB(SSR). The notation used indicates that in this species the
Fe(III) is chelated with Λ helicity and the stereochemical descriptors in parentheses provide the
chirality at the DAP (carbon 2), citrate (carbon 6), and hemiaminal (carbon 15) stereocenters.
(Figure S5). The Δ-Fe(III)-SB(SSS) isomer was found to be of higher energy (-2946.181 Eh),
whereas the configurations of Λ-Fe(III)-SB(SSS) and Δ-Fe(III)-SB(SSR) converted to the
opposite helicity during the DFT geometry optimization (Table S5). The lowest energy structure
for the Fe(III)-epiSB complex, Δ-Fe(III)-epiSB(SRS) (-2946.198 Eh), has the opposite ligand
helicity and quaternary carbon center chirality compared to the lowest energy structure of native
Fe(III)-SB (Figure S5). The energies of the less stable conformations are listed in Table S5.
Next, we employed time-dependent DFT (TD-DFT) to simulate CD spectra corresponding to the
lowest energy structures, Λ-Fe(III)-SB(SSR) and Δ-Fe(III)-epiSB(SRS), and compared these
predictions to the experimentally obtained spectra (Figure 3B,C). In agreement with the
experimental data, the calculated spectra exhibited a positive Cotton effect for Λ-Fe(III)-
SB(SSR) and a shift from positive to negative for Δ-Fe(III)-epiSB(SRS) in the low energy
spectral feature, thereby supporting the different coordination geometries of the two lowest
energy complexes.
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
The geometry optimizations allowed further investigation of the Fe(III) coordination
spheres of SB and epiSB. The two most regular coordination geometries of six-coordinate metal
complexes are octahedral and trigonal prismatic. The twist angle, α, defined as the angle
formed between two opposite metal-ligand bonds when viewing the complex along the C3-axis,
can be used to distinguish these coordination modes (Figure S6B).52 With this definition,
idealized octahedral complexes have α = 60° and trigonal prismatic complexes have α = 0°.
Upon inspection of the coordination polyhedron of Fe(III)-SB (Figure S6A),53 it can be
appreciated that the geometry is distorted from octahedral. When viewed along the pseudo C3-
symmetry axis, it becomes apparent that the α values are intermediate between those of the
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 32
1
2
3
4
5
6
7
8
9
octahedron and the trigonal prism (Figure S6B). The average of twist angles αi, αii, and αiii for Λ-
Fe(III)-SB(SSR) is 36°, indicating a significant distortion towards trigonal prismatic. The average
twist angle for the Δ-Fe(III)-epiSB(SRS) structure is –29°, reflecting a similar trigonal prismatic
distortion of the two complexes with opposite helicity at the metal center.
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
Epimerization of the Citric Acid Stereocenter Perturbs the Siderophore Activity of
SB. We employed a S. aureus growth recovery assay to evaluate the siderophore activity,
which we define as the ability of a siderophore to recover bacterial growth under conditions of
iron limitation, of SB, epiSB, and SBimide. For this assay, we obtained two mutant strains,
hereafter ΔsbnE and ΔsirA, of the multi-drug resistant community isolate S. aureus USA300 JE2
from the Nebraska Transposon Mutant Library (NTML).54 The siderophore-deficient ΔsbnE
mutant cannot biosynthesize SB because it lacks SbnE, the first NIS synthetase in the SB
biosynthetic pathway encoded by the SB gene locus. This mutant is a useful tool for studying
the effect of exogenous SB and its structural analogues on S. aureus growth because there will
be no competition resulting from production of the endogenous siderophore. The ΔsirA mutant is
deficient in the membrane receptor SirA responsible for Fe(III)-SB uptake and therefore cannot
import SB (Table S1).22 The mutant strains were identified using the NTML gene explorer (see
further description in Supporting Information). The strains were cultured under either iron-
sufficient or -deficient conditions. A defined minimal medium, Tris-minimal succinate (TMS)
medium, was employed for all studies, and the medium was supplemented with the chelator
2,2’-dipyridyl (200 µM, DP) to mimic conditions of iron limitation. When cultured in 96-well plates
and in the absence of DP, both ΔsbnE and ΔsirA grew to an OD600 of ≈0.6 following a 10-h
incubation at 37 oC. The growth of both strains was impaired when DP was included in the TMS
medium. Under such iron-deficient conditions, an OD600 of ≈0.2 was observed following the 10-h
incubation period (Figure 7 and S9).
20
ACS Paragon Plus Environment

Page 21 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
A
C
B
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.8
ΔsnbE
ΔsirA
0.7
No DP
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
200 µM DP
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0
0.01 0.1
1
5
10 25 50 100
0
0.01 0.1
1
5
10 25 50 100
[SB] (µM)
[SB] (µM)
D
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
ΔsnbE
ΔsirA
0
0.01 0.1
1
5
10 25 50 100
0
0.01 0.1
1
5
10 25 50 100
[epiSB] (µM)
[epiSB] (µM)
0.6
0.5
0.4
0.3
0.2
0.1
0.0
E
SB ΔSbnE
epiSB ΔSbnE
SB ΔSirA
IC50
EC50
SB 0.7 µM 2 µM
epiSB 8 µM 60 µM
epiSB ΔSirA
0
20
40
60
80
100
[Siderophore] (µM)
Figure 7. Growth recovery assays of the ΔsbnE and ΔsirA mutants of S. aureus in the presence
of increasing concentration of SB or epiSB in the growth medium (37 oC, t = 10 h). Iron sufficient
conditions, TMS medium (gray), Iron deficient conditions, TMS medium + 200 µM DP (black).
(A) SB-mediated growth recovery of S. aureus ΔsbnE; (B) Growth inhibition by SB of S. aureus
ΔsirA; (C) epiSB-mediated growth recovery of S. aureus ΔsbnE; (D) Growth inhibition by epiSB
of S. aureus ΔsirA. (E) The OD600 at t = 10 h plotted as a function of siderophore concentration
and tabulated IC50 and EC50 values determined from the plot The results are the average of
three independent trials, error bars are standard deviations. The growth recovery assay was
monitored for 22 h and the full growth curve plots are shown in Figure S9.
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 32
1
2
3
4
5
6
7
8
9
In the absence of DP, addition of SB to the culture medium had negligible effect on the
growth of the ΔsbnE mutant as ascertained by OD600 values. Under iron-deficient conditions
where the growth of ΔsbnE is impaired, SB afforded growth recovery of the strain in a
concentration-dependent manner. The OD600 of the cultures increased to ≈0.5 when >1 µM SB
was added to the TMS medium containing DP; however, full growth recovery of ΔsbnE was not
observed under these conditions (Figure 7A). Inductively coupled plasma mass spectrometry
established that the TMS medium contains 700 nM of total iron (Table S2). Thus, maximum
growth recovery occurred when the concentration of SB equaled or exceeded the concentration
of total Fe in the medium. To determine whether Fe(III)-SB can fully recover the growth of
ΔsbnE, we performed an experiment where SB was preloaded with Fe(III) prior to the assay,
and observed that the Fe(III)-SB complex (10 µM) afforded full growth recovery of ΔsbnE
(Figure S10).
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
The ΔsirA mutant lacks the SB receptor and cannot import Fe(III)-SB.22 Addition of apo
SB (>1 µM) to cultures of ΔsirA resulted in growth inhibition regardless of the absence or
presence of DP in the medium (Figure 6B). The inhibited growth of ΔsirA can be attributed to
sequestration of iron in the growth medium by SB, which persists as a high-affinity extracellular
iron chelator because it cannot be imported by the ΔsirA mutant. Taken together with the data
for ΔsbnE, these growth recovery assays confirm that synthetic SB exhibits siderophore activity
and delivers iron to S. aureus. Moreover, the siderophore activity depends on iron availability
and requires the membrane receptor SirA.
Next, the growth recovery assay was performed using epiSB to evaluate the importance
of the stereochemistry of the citric acid moiety for siderophore activity. The results indicate that
epiSB possesses siderophore activity that is dependent on the SB receptor SirA (Figure 7C,D).
Thus, the (S)-stereochemistry at the citric acid moiety of SB is not essential for S. aureus to use
this siderophore for acquiring iron. Nevertheless, epiSB exhibits reduced siderophore activity
compared to SB. Under conditions of iron limitation, epiSB recovers the growth of the ΔsbnE
mutant, but maximum growth recovery is observed at 50 µM epiSB, whereas the cultures
22
ACS Paragon Plus Environment

Page 23 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
recover to a similar OD600 value with 1 µM SB (Figure 7A,C). Moreover, higher concentrations of
epiSB relative to SB are required to inhibit growth of the ΔsirA mutant (Figure 7B,D). The
difference in siderophore activity between the two analogues was quantified by comparing EC50
(concentration resulting in 50% growth recovery of ΔsbnE) as well as IC50 (concentration
resulting in 50% growth inhibition of ΔsirA) values obtained from the growth recovery assay data
(Figure 7E). The EC50 and IC50 values of SB are 10-fold and 30-fold lower, respectively, than the
values obtained for epiSB. The reduced siderophore activity of epiSB may result from its lower
Fe(III) affinity, as indicated by the CASAD assay and pFe(III) value determination. Indeed, the
reduced IC50 value of epiSB against ΔsirA supports this notion. Nevertheless, it is also possible
that epimerization of the citric acid moiety affects Fe(III)-binding kinetics, impairs recognition
and/or transport of the ferric complex by SirA, or alters the fate of Fe(III)-SB once inside of the
cell. An investigation of vibrioferrin diastereoisomers differing only in the stereochemistry of the
citric acid moiety revealed that only the natural (R)-isomer possesses siderophore activity,55
which is in contrast to the observed siderophore activity of epiSB. On the basis of the published
work, it is unclear whether the impaired siderophore activity of epi-vibrioferrin results from
impaired Fe(III)-binding properties or because the membrane receptor cannot recognize and/or
transport the non-native epimer.55
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
SBimide Converts to SB Under Simulated Physiological Conditions. We also
examined whether SBimide (Figure 1) is able to chelate Fe(III) and influence the growth of S.
aureus. Imide byproducts of citric acid siderophores are commonly observed when these
siderophores are isolated from natural sources.14 During the course of this work, we observed
imide byproducts during optimization of the total synthesis of SB (Supporting Information). As a
result, we isolated and characterized SBimide. When we spotted SBimide on the CASAD plates,
we observed an orange annular pattern that differed from the orange disks obtained for SB or
epiSB (Figure 6). The orange annulus bounded a concentric blue disk, the latter of which is
where SBimide was positioned as a drop onto the plate. This pattern indicates that no iron was
sequestered from the indicator where SBimide was spotted on the plate, but that iron was
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 32
1
2
3
4
5
6
7
8
9
sequestered after SBimide diffused through the agar, resulting in formation of the orange ring.
We reasoned that SBimide converted to native SB as it diffused through the agar. To investigate
this hypothesis, we incubated SBimide in aqueous buffer at pH 7.4 (10 mM HEPES) and
analyzed the sample by mass spectrometry, and observed a change in m/z value from
430.1336 to 448.1442, which is consistent with ring opening of the imide and formation of SB
(Figure S11C). Moreover, optical absorption spectroscopy revealed that SBimide only chelates
Fe(III) under basic conditions whereas both SB and epiSB chelated Fe(III) at pH < 4 (Figure
S6). In growth recovery assays with the ΔsbnE and ΔsirA mutants, SBimide provided
siderophore activity comparable to that of SB (Figure S11A,B), which we attribute to its
conversion to SB under these assay conditions. The fact that SBimide is ring-opened to
generate SB under physiological conditions indicates that the presence of imide byproducts in
the citric acid siderophore isolates from natural sources is most likely a consequence of the acid
treatments employed during the isolation and purification steps, and that the imide is not
physiologically relevant.
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
Epimerization of the Citric Acid Stereocenter Perturbs Iron Sequestration from
Transferrin. Transferrin is the major Fe(III)-transport protein in serum, and the human protein
binds Fe(III) with a pFe(III) value of 22.56 Various studies have evaluated the ability of
siderophores to sequester Fe(III) from transferrin,57,58 and S. aureus can utilize holo-transferrin
as an iron source.59-61 Moreover, SB is reported to remove Fe(III) from holo-transferrin.60,61
Transferrin has two lobes, and each lobe binds one equivalent of Fe(III) synergistically with
carbonate. Urea PAGE affords separation of the four forms of transferrin – apo, holo (diferric),
monoferric where Fe(III) is bound to the N-terminal lobe (mono-N transferrin), and monoferric
where Fe(III) is bound to the C-terminal lobe (mono-C transferrin).62 We used this technique to
monitor the removal of Fe(III) from holo-transferrin by SB and epiSB (Figure 8). Following a 24-h
incubation of holo-transferrin (40 µM) with 10 to 50 equivalents of SB (50 mM Tris, 150 mM
NaCl, 20 mM NaHCO3, pH 7.4, 37 oC), urea PAGE revealed a mixture of holo- and demetallated
transferrin. Loss of Fe(III) from both the N- and C-terminal lobes occurred and the degree of
24
ACS Paragon Plus Environment

Page 25 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Fe(III) removal was proportional to the SB concentration. At 30 and 50 equivalents of SB, bands
corresponding to apo-transferrin appeared, and these bands were absent from the control
containing only holo-transferrin. Although epiSB was able to remove Fe(III) from holo-
transferrin, as evidenced by the intensities of bands corresponding to the two monoferric
transferrin species, epiSB sequestered markedly less Fe(III) from holo-transferrin than SB. We
also monitored the holo-transferrin/siderophore mixtures over a 70-h time period by urea PAGE,
and observed increased siderophore-dependent loss of Fe(III) from holo-transferrin over time
(Figure S12). Under these experimental conditions, the kinetics of Fe(III) removal from holo-
transferrin is relatively slow, even for SB. Nevertheless, the data support the notion that chirality
at the citric acid stereocenter is important for the siderophore activity of SB.
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
_
_
_
+
_
+
_
+
_
_
_
SB
epiSB
+
+
+
A
_
_
_
_
_
_
apo
mono-C
mono-N
holo
100
B
80
60
40
20
0
apo
mono-C
mono-N
holo
10
30
50
10
30
50
equiv SB
equiv epiSB
Figure 8. Siderophore-mediated iron sequestration from holo-transferrin monitored. (A) Holo-
transferrin (40 µM) was incubated with 0, 10, 30 or 50 equivalents of SB or epiSB and the
o
solutions were incubated for 24 h at 37 C in buffer. The resulting samples were analyzed by 6
M urea PAGE and Coomassie stain, and a representative gel is shown. (B) Quantification of the
band intensities for the gel shown in A. Quantification was performed using ImageJ. Apo-
transferrin, mono-C transferrin, mono-N transferrin and holo-transferrin are depicted in black,
dark gray, light gray, and white, respectively.
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 32
1
2
3
4
5
Summary
6
7
8
9
In this work, we report the first total synthesis of staphyloferrin B, a siderophore and virulence
factor of S. aureus. The total synthesis along with computational modeling studies confirmed
that SB exists as the hemiaminal form with an (S)-configuration of the citric acid moiety.
Synthetic SB provided growth recovery of S. aureus that is dependent on its receptor SirA. The
total synthesis provided a foundation for the preparation of two structural analogues, epiSB and
SBimide. SB and epiSB differ only in the stereochemistry of the citric acid quaternary center.
Fe(III)-binding studies and growth recovery assays provided valuable insight into the importance
of the citric acid stereocenter for iron chelation and siderophore activity. Inversion of this
stereocenter to generate epiSB alters the coordination sphere around the Fe(III) center and
perturbs function. Although epiSB retains the ability to chelate iron and function as siderophore
for S. aureus in vitro, the (S)-configuration at the citric acid moiety of SB appears to be
important because the siderophore activity of epiSB is reduced in all contexts examined in this
work. Our studies of SBimide confirmed that the imide ring is opened to generate the native
siderophore SB under physiological conditions. In closing, S. aureus infections caused by
recalcitrant and drug-resistant strains are a serious problem for public health. This work
provides the necessary basis for further chemical and biological investigations of how SB
contributes to the molecular pathogenesis of S. aureus infections as well as a guide for further
synthetic efforts directed at achieving siderophore-based therapeutics for bacterial infections.63
Efforts along these lines are in progress.
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
Acknowledgements
We thank the Pacific Southwest Research Center of Excellence, the Kinship Foundation (Sloan
Foundation), and the Alfred Benzon Foundation (post-doctoral fellowship to J.L.H.M.) for
financial support. We thank Toshiki Nakashige for purifying apo-transferrin, Prof. Timothy F.
Jamison for use of an IR spectrometer, and Profs. Manuela Raffatellu and Alan G. Barbour for
26
ACS Paragon Plus Environment

Page 27 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
helpful suggestions. NMR instrumentation maintained by the MIT Department of Chemistry
Instrumentation Facility (DCIF) is supported by NSF grants CHE-9808061 and CHE-8915028.
Circular dichroism instrumentation maintained by the MIT Biophysical Instrumentation Facility
(BIF) is supported by NSF-0070319. The single-gene knockout transposon mutants were
obtained from the Network on Antimicrobial Resistance in Staphylococcus aureus (NARSA)
program distributing the Nebraska Transposon Mutant Library (NTML) of S. aureus mutants
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
harboring
single-gene
knockouts,
supported
under
NIAID/NIH
contract
no.
HHSN272200700055C.
Supporting Information
Experimental procedures, supporting tables, supporting figures, and spectroscopic data. This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1)
(2)
McGavin, M. J.; Heinrichs, D. E. Front. Cell. Infect. Microbiol. 2012, 2, 66.
DeDent, A.; Kim, H. K.; Missiakas, D.; Schneewind, O. Semin. Immunopathol. 2012, 34,
317-333.
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Chambers, H. F.; DeLeo, F. R. Nat. Rev. Microbiol. 2009, 7, 629-641.
Cassat, J. E.; Skaar, E. P. Cell Host Microbe 2013, 13, 509-519.
Andrews, N. C.; Schmidt, P. J. Annu. Rev. Physiol. 2007, 69, 69-85.
Sheldon, J. R.; Heinrichs, D. E.. FEMS Microbiol. Rev. 2015, ASAP.
Beasley, F. C.; Heinrichs, D. E. J. Inorg. Biochem. 2010, 104, 282-288.
Hammer, N. D.; Skaar, E. P. Annu. Rev. Microbiol. 2011, 65, 129-147.
Meiwes, J.; Fiedler, H.-P.; Haag, H.; Zähner, H.; Konetschny-Rapp, S.; Jung, G. FEMS
Microbiol. Lett. 1990, 67, 201-205.
(10)
Sebulsky, M. T.; Heinrichs, D. E. J. Bacteriol. 2001, 183, 4994-5000.
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 32
1
2
3
4
5
6
7
8
9
(11)
(12)
Hider, R. C.; Kong, X. Nat. Prod. Rep. 2010, 27, 637-657.
Chu, B. C.; Garcia-Herrero, A.; Johanson, T. H.; Krewulak, K. D.; Lau, C. K.; Peacock,
R. S.; Slavinskaya, Z.; Vogel, H. J. Biometals 2010, 23, 601-611.
Crosa, J. H.; Walsh, C. T. Microbiol. Mol. Biol. Rev. 2002, 66, 223-249.
Konetschny-Rapp, S.; Jung, G.; Meiwes, J.; Zähner, H. Eur. J. Biochem. 1990, 191, 65-
74.
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
(13)
(14)
(15)
(16)
(17)
(18)
Drechsel, H.; Freund, S.; Nicholson, G.; Haag, H.; Jung, O.; Zähner, H.; Jung, G.
Biometals 1993, 6, 185-192.
Haag, H.; Fiedler, H.-P.; Meiwes, J.; Drechsel, H.; Jung, G.; Zähner, H. FEMS Microbiol.
Lett. 1994, 115, 125-130.
Mergeay, M.; Monchy, S.; Vallaeys, T.; Auquier, V.; Benotmane, A.; Bertin, P.; Taghavi,
S.; Dunn, J.; van der Lelie, D.; Wattiez, R. FEMS Microbiol. Rev. 2003, 27, 385-410.
Janssen, P. J.; Van Houdt, R.; Moors, H.; Monsieurs, P.; Morin, N.; Michaux, A.;
Benotmane, M. A.; Leys, N.; Vallaeys, T.; Lapidus, A.; Monchy, S.; Médigue, C.;
Taghavi, S.; McCorkle, S.; Dunn, J.; van der Lelie, D.; Mergeay, M. PLoS One 2010, 5,
e10433.
(19)
Gilis, A.; Khan, M. A.; Cornelis, P.; Meyer, J. M.; Mergeay, M.; van der Lelie, D. J.
Bacteriol. 1996, 178, 5499-5507.
(20)
(21)
Münzinger, M.; Taraz, K.; Budzikiewicz, H. Z. Naturforsch. C 1999, 54, 867-875.
Grigg, J. C.; Cheung, J.; Heinrichs, D. E.; Murphy, M. E. P. J. Biol. Chem. 2010, 285,
34579-34588.
(22)
(23)
(24)
Dale, S. E.; Sebulsky, M. T.; Heinrichs, D. E. J. Bacteriol. 2004, 186, 8356-8362.
Bhatt, G.; Denny, T. P. J. Bacteriol. 2004, 186, 7896-7904.
Dale, S. E.; Doherty-Kirby, A.; Lajoie, G.; Heinrichs, D. E. Infect. Immun. 2004, 72, 29-
37.
(25)
Cheung, J.; Beasley, F. C.; Liu, S.; Lajoie, G. A.; Heinrichs, D. E. Mol. Microbiol. 2009,
74, 594-608.
28
ACS Paragon Plus Environment

Page 29 of 32
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(26)
(27)
Beasley, F. C.; Cheung, J.; Heinrichs, D. E. BMC Microbiol. 2011, 11, 199.
Kobylarz, M. J.; Grigg, J. C.; Takayama, S.-i. J.; Rai, D. K.; Heinrichs, D. E.; Murphy, M.
E. P. Chem. Biol. 2014, 21, 379-388.
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
(28)
(29)
Cheung, J.; Murphy, M. E. P.; Heinrichs, D. E. Chem. Biol. 2012, 19, 1568-1578.
Kobylarz, M. J.; Grigg, J. C.; Sheldon, J. R.; Heinrichs, D. E.; Murphy, M. E. P. J. Biol.
Chem. 2014, 289, 33797-33807.
(30)
(31)
(32)
(33)
Hannauer, M.; Sheldon, J. R.; Heinrichs, D. E. FEBS Lett. 2015, 589, 730-737.
Challis, G. L. ChemBioChem 2005, 6, 601-611.
Sheldon, J. R.; Marolda, C. L.; Heinrichs, D. E. Mol. Microbiol. 2014, 92, 824-839.
Friedman, D. B.; Stauff, D. L.; Pishchany, G.; Whitwell, C. W.; Torres, V. J.; Skaar, E. P.
PLoS Pathog. 2006, 2, e87.
(34)
Beasley, F. C.; Vinés, E. D.; Grigg, J. C.; Zheng, Q.; Liu, S.; Lajoie, G. A.; Murphy, M. E.
P.; Heinrichs, D. E. Mol. Microbiol. 2009, 72, 947-963.
(35)
(36)
Lindsay, J. A.; Riley, T. V. Infect. Immun. 1994, 62, 2309-2314.
Lindsay, J. A.; Riley, T. V.; Mee, B. J. Eur. J. Clin. Microbiol. Infect. Dis. 1994, 13, 1063-
1066.
(37)
(38)
Hanses, F.; Roux, C.; Dunman, P. M.; Salzberger, B.; Lee, J. C. Genome Med. 2014, 6,
93.
Tripathi, A.; Schofield, M. M.; Chlipala, G. E.; Schultz, P. J.; Yim, I.; Newmister, S. A.;
Nusca, T. D.; Scaglione, J. B.; Hanna, P. C.; Tamayo-Castillo, G.; Sherman, D. H. J.
Am. Chem. Soc. 2014, 136, 1579-1586.
(39)
(40)
(41)
Cancer Research Technology Limited. Patent. WO2006/16143 A1 2006.
Seebach, D.; Naef, R.; Calderari, G. Tetrahedron 1984, 40, 1313-1324.
Eckelbarper, J. D.; Wilmot, J. T.; Epperson, M. T.; Thakur, C. S.; Shum, D.; Antczak, C.;
Tarassishin, L.; Djaballah, H.; Gin, D. Y. Chem. Eur. J. 2008, 14, 4293-4306.
Carlsen, P. H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J. Org. Chem. 1981, 46,
3936-3938.
(42)
29
ACS Paragon Plus Environment