Mixing Up the Pieces of the Desferrioxamine B ... - ACS Publications

Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI

Article

Mixing Up the Pieces of the Desferrioxamine B Jigsaw Defines the Biosynthetic Sequence Catalyzed by DesD Thomas J Telfer, Michael P. Gotsbacher, Cho Zin Soe, and Rachel Codd ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.6b00056 • Publication Date (Web): 08 Mar 2016 Downloaded from http://pubs.acs.org on March 10, 2016

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.

ACS Chemical Biology 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 23

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

ACS Chemical Biology

Figure 1 215x139mm (300 x 300 DPI)

ACS Paragon Plus Environment


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 2 203x279mm (300 x 300 DPI)


Page 2 of 23

Page 3 of 23

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 508x228mm (300 x 300 DPI)



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 4 215x177mm (300 x 300 DPI)


Page 4 of 23

Page 5 of 23

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


Chart 1 141x181mm (300 x 300 DPI)



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

Chart 2 136x179mm (300 x 300 DPI)


Page 6 of 23

Page 7 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 1 129x71mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Scheme 2 228x471mm (300 x 300 DPI)


Page 8 of 23

Page 9 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 3 161x109mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Scheme 4 79x51mm (300 x 300 DPI)


Page 10 of 23

Page 11 of 23

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


Table of Contents Graphic



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

Page 12 of 23

Mixing Up the Pieces of the Desferrioxamine B Jigsaw Defines the Biosynthetic Sequence Catalyzed by DesD Thomas J. Telfer, Michael P. Gotsbacher, Cho Zin Soe, and Rachel Codd* School of Medical Sciences (Pharmacology) and Bosch Institute, The University of Sydney, New South Wales 2006, Australia ABSTRACT: Late-stage assembly of the trimeric linear siderophore desferrioxamine B (DFOB) native to Streptomyces pilosus involves two DesD-catalyzed condensation reactions between one N-acetyl-N-hydroxy-1,5-diaminopentane (AHDP) unit and two N-succinyl-N-hydroxy-1,5-diaminopentane (SHDP) units. AHDP and SHDP are products of DesBC-catalyzed reactions of the native diamine substrate 1,5-diaminopentane (DP). The sequence of DesD-catalyzed DFOB biosynthesis was delineated by analyzing the distribution of DFOB analogues and dimeric precursors assembled by S. pilosus in medium containing 1,4-diamino-2(E)-butene (E-DBE). Seven unsaturated DFOB analogues were produced that were partially resolved by liquid chromatography (LC). Mass spectrometry (MS) measurements reported on the combination of E-DBE- and DP-derived substrates in each trimer (uDFOA1 series, 1:2; uDFOA2 series, 2:1; uDFOA3, 3:0). MS/MS fragmentation patterns reported on the absolute position of the substrate derivative at the N-acetylated terminus, the internal region, or the amine terminus of the trimer. The uDFOA1 and uDFOA2 series were each comprised of three constitutional isomers (binary notation (DP-derived substrate ‘0’, E-DBE-derived substrate ‘1’); direction, N-acetylated-internal-amine): uDFOA1[001], uDFOA1[010], uDFOA1[100]; and uDFOA2[011], uDFOA2[110], uDFOA2[101]. E-DBE completely replaced DP in uDFOA3[111]. Relative concentrations of the uDFOA1 series were: uDFOA1[001] >> uDFOA1[100] > uDFOA1[010]; and of the uDFOA2 series: uDFOA2[101] > uDFOA2[011] >> uDFOA2[110]. Dimeric compounds assembled from one N-acetylated and one N-succinylated substrate derivative were detected as trimer precursors: dDFX[00] >> udDFX[10-] > udDFX[01-] (d = dimer, vacant position, ‘-‘). Relative concentrations of all species were consistent with the biosynthetic sequence: (i) SHDP activation; (ii) condensation with AHDP to form AHDP-SHDP; (iii) SHDP activation; (iv) condensation with AHDP-SHDP to form DFOB.

Desferrioxamine B (DFOB) is a linear trihydroxamic acid siderophore produced by several species of Streptomyces.1-3 This siderophore (Gk, ‘iron carrier’) has been evolved as a high affinity chelator of Fe(III) and provides a mechanism for bacteria to acquire sparingly soluble environmental Fe(III) as essential for growth.4-8 DFOB has a long clinical history for the removal of excess iron that accumulates in patients who receive regular blood transfusions to treat genetic hemoglobin disorders, such as beta-thalassaemia, sickle cell anaemia and myelodysplastic syndromes.9 The chemical synthesis of DFOB is complex and multi-step10-12 and its production as a pharmaceutic relies on industrial-scale fermentation of an overproducing strain of S. pilosus.13 Its importance in medicine has prompted interest in the biosynthesis of DFOB and related non-peptide siderophores.14-16 Early studies established 1,5diaminopentane (DP) or common name, cadaverine, as the primary diamine substrate for DFOB assembly, which is produced from the decarboxylation of L-lysine.14 Mono-Nhydroxylation of DP was determined as the next step in DFOB biosynthesis.15 Further understanding was enabled by the genome sequence of S. coelicolor A3(2), which contained a gene cluster SCO2782-2785 predicted to encode for the biosynthesis of linear desferrioxamine G1 (DFOG1) and its macrocyclic product desferrioxamine E (DFOE).17 The cognate enzyme cluster DesABCD was subsequently shown by experiment to be functional via pathways independent of non-ribosomal peptide synthetases in the biosynthesis of DFOG1 and DFOE, as siderophores native to S. coelicolor M145, and in DFOB biosynthesis.16,18,19

The current understanding of the biosynthesis of DFOB involves the decarboxylation of L-lysine to generate DP (DesA), mono-N-hydroxylation of DP to produce N-hydroxy-DP (DesB), succinylation or acetylation of N-hydroxy-DP to produce N-succinyl-N-hydroxy-DP (SHDP) or N-acetyl-Nhydroxy-DP (AHDP), respectively (DesC), and the condensation of two SHDP and one AHDP fragments (DesD) to form DFOB (Scheme 1a).19 The current work aimed to delineate the sequence of steps catalyzed by DesD in DFOB biosynthesis. The existing proposition was that DesD first catalyzed the condensation of two SHDP units and in the second step appended the AHDP unit to the SHDP homodimer.19 While this was a reasonable supposition based on the biosynthesis of DFOG1 and DFOE as SHDP homotrimers, the sequence was less clear as applied to heterotrimeric DFOB, since the dimer precursor could be either a SHDP-SHDP homodimer or an AHDP-SHDP heterodimer. To dissect the sequence of steps of DesD-catalyzed DFOB biosynthesis, a precursor-directed biosynthesis approach was used to measure the distribution of dimer and trimer products assembled in situ from combinations of fragments derived from native DP and non-native diamine precursors. Reconstituting the native assembly of DFOB analogues from different jigsaw pieces has established the directional preference of DFOB biosynthesis and identified the sequence of steps catalyzed by DesD.


Page 13 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 1. Biosynthesis of (a) Desferrioxamine B (DFOB) From Three Units of Native Substrate DP; or (b) the Unsaturated Analogue uDFOA3[111] From Three Units of Non-Native Substrate E-DBE

RESULTS AND DISCUSSION Rationale. Analysis of the structures and relative concentrations of DFOB analogues produced in cultures of S. pilosus supplemented with exogenous diamine substrates could provide new insight into the native DFOB biosynthetic pathway as a complement to knowledge built from genetics and molecular biology. Precursor-directed biosynthesis with Streptomyces olivaceus Tü 2718 has been used to produce analogues of macrocyclic DFOE.20,21 This work examined analogues of DFOB produced in medium inoculated with S. pilosus supplemented with: 1,4-diamino-2(E)-butene (E-DBE) or 1,4diaminobutane (DB), as the saturated analogue of E-DBE. The E-DBE substrate has been used in cultures of Shewanella putrefaciens to produce unsaturated macrocycles of the dihydroxamic acid putrebactin,22 which is one member of this small class of siderophores.23-26 The similarity of siderophore biosynthesis gene clusters between S. putrefaciens and S. coelicolor,16,27-29 suggested that E-DBE would also be competent as a substrate in the biosynthesis of linear DFOB by closely related S. pilosus. Unsaturated Analogues of DFOB from E-DBE. The LC trace from the semi-purified extract of S. pilosus cultured in medium optimized for DFOB production30 showed two peaks (Figure 1a). MS analysis of the major peak gave signals at m/z 561.4 and 281.2, consistent with DFOB (Chart 1, 1) as [M+H]+ and [M+2H]2+ adducts, respectively. The minor peak analyzed as DFOA1 (m/z 547.3, 274.2), which is a known DFO-type siderophore7,31,32 assembled from two DP- and one DB-derived substrates, with the latter diamine available as an endogenous substrate at low levels.

Figure 1. LC-MS-Q trace with total ion current (TIC) detection mode from semi-purified supernatant of S. pilosus cultured in: (a) base medium, or (b) medium supplemented with E-DBE, analyzed as isolated (column at left) or as Fe(III)-loaded (column at right) solutions. The gradient in (a) or (c) was the same in (b) or (d), respectively. Peak numbering refers to the compounds in Chart 1, Chart 2 and Supplementary Chart 1, or the equivalent 1:1 complexes formed with Fe(III) (number underlined).

The LC profile from the extract purified from the medium supplemented with E-DBE showed five well-resolved peaks (Figure 1b). The LC peak at tR 34.55 min was ascribed to DFOB (Figure 2a, lower x axis) and agreed well with calculated data (upper x axis). MS analysis of the peaks at tR 33.12 (Figure 2b) and 32.04 min (Figure 2c) each gave m/z values consistent with a trimer assembled from two DP- and one EDBE-derived substrates (m/z 545.3, 273.2) attributable to isomers of uDFOA1, as unsaturated analogues of DFOA1. uDFOA1 could be formulated as three constitutional isomers, in



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

which the E-DBE-derived substrate was inserted at the Nacetylated region, the internal region, or the amine region of the trimer, with the structural balance met by two units of the DP-derived substrate. A binary nomenclature system (DPderived substrate as ‘0’, E-DBE-derived substrate as ‘1’; analogue direction from left to right: N-acetylated-internal-amine) would describe these constitutional isomers as: uDFOA1[001] (2), uDFOA1[010] (3) and uDFOA1[100] (4). The LC results supported the presence of at least two of the three isomers, similar to previous results for DFOA1.32 Chart 1. DFOB (1) and Analogues (2-8) Assembled from Substrates DP and E-DBE

The peak at tR 30.35 min gave m/z signals (m/z 529.3, 265.2) (Figure 2d) consistent with uDFOA2. Similar to the uDFOA1 series, uDFOA2 could be formulated as three constitutional isomers: uDFOA2[011] (5), uDFOA2[110] (6) or uDFOA2[101] (7). The production of more than one uDFOA2 isomer was unclear, based on the observation of only a single peak. The complete exchange of the native DP-derived fragments for the non-native E-DBE-derived fragments was evident from the m/z signals from the peak at tR 27.37 min (m/z 513.2, 257.1), consistent with uDFOA3[111] (8) (Figure 2e). Relative to DFOB ([M+H]+, 561.4), the m/z values decreased in a systematic fashion by 16 units for uDFOA1, uDFOA2 and uDFOA3 as a result of the incremental exchange of one DPderived fragment for one E-DBE-derived fragment. The LC traces from the native and the E-DBE supplemented extracts were re-acquired following the addition of Fe(III) (Figure 1c and Figure 1d, respectively). This provided an additional level of characterisation and demonstrated the functional integrity of the analogues as siderophores. In the native system, observed MS signals were consistent with the [M+H]+ and [M+2H]2+ adducts (where [M] = [Fe(III)-DFOB-H3] or equivalent) of Fe(III)-DFOB and Fe(III)-DFOA1. MS signals from the LC peak at tR 28.96 min in the Fe(III)-loaded E-DBE supplemented system were similarly characteristic of Fe(III)-

Page 14 of 23

DFOB (m/z 614.3, 307.7) (Figure 2f). Two LC peaks at tR 28.52 min (Figure 2g) and 26.62 min (Figure 2h) gave MS signals with identical m/z values (m/z 598.3, 299.6) consistent with the formation of Fe(III) complexes with the two constitutional isomers of uDFOA1. The single LC peak attributed to uDFOA2 was resolved in the Fe(III)-loaded extract into three peaks at tR 26.09 (Figure 2i), 25.35 min (Figure 2j) and 24.24 min (Figure 2k), which each gave MS signals (m/z 582.2, 291.6) characteristic of Fe(III)-uDFOA2 isomers. Signals for the peak at tR 24.24 min were of low intensity. Fe(III)uDFOA3 was present in the LC at tR 23.05 min (m/z 566.2, 283.6), together with other species (Figure 2l).

Figure 2. MS signals (LC-MS-Q) from the as isolated extract (column at left) consistent with: (a) DFOB at tR 34.55 min; or isomers of the uDFOA1 series at (b) tR 33.12 min, or (c) tR 32.04 min; or isomers of the uDFOA2 series at (d) tR 30.35 min; or uDFOA3 at (e) tR 27.37 min; or from the Fe(III)-loaded extract (column at right) consistent with: (f) Fe(III)-DFOB at tR 28.96 min; or isomers of the Fe(III)-uDFOA1 series at (g) tR 28.52 min, or (h) tR 26.62 min; or isomers of the uDFOA2 series at (i) tR 26.09 min, (j) tR 25.35 min or (k) tR 24.24 min; or Fe(III)-uDFOA3 at (l) tR 23.05 min. Data shown as experiment (lower x axis) or calculated (upper x axis: [M+H]+ and [M+2H]2+ adducts).

Analogues of DFOB from DB. The LC profile from the extract from S. pilosus cultured in medium supplemented with DB was similar to the E-DBE system (Supporting Information). Five major peaks were resolved in the LC (Supplementary Figure 1a) that from the MS signals (Supplementary Figure 2a-e) characterized as: DFOB (1); DFOA1 (two LC peaks, with three possible isomers (DP-derived substrate ‘0’, DB-derived substrate ‘1’): DFOA1[001] (9), DFOA1[010] (10), DFOA1[100] (11)); DFOA2 (one LC peak, with three possible isomers: DFOA2[011] (12), DFOA2[110] (13), DFOA2[101] (14)); and DFOA3[111] (15) (Supplementary


Page 15 of 23

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


Chart 1). The LC peaks from the Fe(III)-loaded extract were less well resolved than the E-DBE system (Supplementary Figure 1b), with MS signals (Supplementary Figure 2g-k) attributable to Fe(III)-DFOB, two isomers of Fe(III)-DFOA1, one isomer of Fe(III)-DFOA2 and Fe(III)-DFOA3. Scheme 2. MS/MS Fragmentation of DFOB and Analogues Assembled From Substrates DP and E-DBE

Constitutional Isomers of uDFOA1 and uDFOA2. The EDBE supplemented system was analyzed by LC-MS-QQQ using selected ion monitoring (SIM) as the detection mode, with values set for the relevant [M+H]+ ion (Figure 3a-e, column at left). Well resolved peaks were consistent with DFOB (SIM 561), uDFOA1 (SIM 545) (major and minor peaks), uDFOA2 (SIM 529) and uDFOA3 (SIM 513). The peak areas ascribed to isomers of uDFOA1 integrated in a ratio of 3:1. Overall relative concentrations were estimated from peak integration as: DFOB (21.2%), uDFOA1 (major) (34.0%), uDFOA1 (minor) (11.3%), uDFOA2 (29.1%) and uDFOA3 (4.4%). This assumed that the extent of ionization was similar among the structural isomers, which was supported by the similarity in the distribution of signals for each species be-

tween different MS systems (LC-MS-Q, LC-MS-QQQ) using different detection modes (extracted ion chromatograms, EIC; SIM) (Supplementary Figure 3). The relative concentrations broadly reflected the yields of compounds isolated using semipreparative HPLC from one 50-mL culture: DFOB (4.1 mg, 17.0%), uDFOA1 (major, 5.5 mg, 22.8%), uDFOA1 (minor, 3.7 mg, 15.3%), uDFOA2 (6.4 mg, 26.6%) and uDFOA3 (4.4 mg, 18.3%). Peak integration showed the ratio of incorporation of DP:E-DBE into the library of analogues was 1.6:1. MS/MS fragmentation patterns were analyzed to identify and establish the distribution of constitutional isomers within the peak(s) that corresponded with the uDFOA1 or the uDFOA2 series (Figure 3a-e, column at right). The fragmentation of DFOB as the parent compound was similar to previous studies of DFOB or related siderophores.33-36 MS/MS data showed that in the uDFOA1 series, the major signal corresponded with uDFOA1[001] (2). The presence of MS/MS signals at m/z 361.1 and 443.2 specific to fragments of uDFOA1[001] (2), together with the absence of signals at m/z 345.2 and 427.2 for the analogous fragments common to uDFOA1[010] (3) and uDFOA1[100] (4), made this assignment unambiguous (Table 1). The LC peak intensity showed that uDFOA1[001] (2) was produced from the uDFOA1 series as the major isomer, which correlated with its isolation in higher yield. The minor signal in the uDFOA1 series gave a fragmentation pattern consistent with the presence of a mixture of uDFOA1[010] (3) and uDFOA1[100] (4). Signals at m/z 128.1, 227.1 and 319.2 were attributable to fragments unique to uDFOA1[100] (4). Signals at m/z 243.2 and 303.2 were attributable to fragments of uDFOA1[010] (3). These latter signals would also arise from fragments of uDFOA1[001] (2), however, this isomer was not present in the minor peak, based on the absence of characteristic signals at m/z 361.2 and 443.2. The relative intensity of the signals due to pairs of fragment analogues of uDFOA1[010] (3) at m/z 243.2 and uDFOA1[100] (4) at m/z 227.1; or m/z 303.3 (3) and m/z 319.2 (4), suggested a higher relative concentration of uDFOA1[100] (4) than uDFOA1[010] (3), in a ratio of about 2:1. In the uDFOA2 series, MS/MS data showed signals that correlated with the presence of a mixture of isomers. Signals of fragments characteristic of both uDFOA2[011] (5) and uDFOA2[101] (7) were present at m/z 345.2 and 427.2, confirming the presence of at least one of these isomers. Fragment signals at m/z 329.1 and 411.1 which within the uDFOA2 series were unique to uDFOA2[110] (6), were detected at very low levels, indicating that this isomer was present in trace amounts. MS analysis across the peak indicated low levels of uDFOA2[110] in the peak front. The minor peak at tR 24.66 min did not characterize as a DFOB analogue. The relative intensity of the signals due to pairs of fragment analogues of uDFOA2[011] (5) at m/z 243.1 and uDFOA2[101] (7) at m/z 227.1; or m/z 287.2 (5) and m/z 303.2 (7), suggested a higher relative concentration of uDFOA2[101] (7) than uDFOA2[011] (5), in a ratio of about 2:1. A signal at m/z 369.2 from the MS/MS fragmentation of the LC peak at tR 23.11 min was characteristic of uDFOA3[111].



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

Page 16 of 23

Figure 3. MS/MS fragmentation spectra (column at right) from LC-MS-QQQ signals (column at left) detected from the E-DBE supplemented system at SIM values: (a) 561, (b) 545 (major), (c) 545 (minor), (d) 529 (with insets of low-intensity signals), and (e) 513.

Table 1. MS/MS Fragmentation Patterns From DFOB and Analogues Assembled From Substrates DP- and E-DBE F1 Analogue

DFOB

F2

F3

F4

F5

F6

F7

F8

F9

[M+H]+ Conca 8.9Cb 9.10N 13.14C 13.14N 20.21C 20.21N 24.25C 24.25N 3.8I (calc) (%)

F10

F11

F12

F13

F14

F15

14.19I 25.30I 10.20I 3.24I 3.13I 14.24I

1 561.3

21.2

144.1 401.2 243.1 319.2 361.2 201.1 443.3 119.1 102.1 102.1 102.1 201.1 401.2 201.1 201.1

uDFOA1[001] 2 545.3

34.0

144.1 385.2 243.1 303.2 361.2 185.1 443.3 103.1 102.1 102.1 86.1

uDFOA1[010] 3 545.3

3.6

144.1 385.2 243.1 303.2 345.2 201.1 427.2 119.1 102.1 86.1 102.1 185.1 385.2 201.1 185.1

uDFOA1[100] 4 545.3

7.7

128.1 401.2 227.1 319.2 345.2 201.1 427.2 119.1 86.1 102.1 102.1 201.1 385.2 185.1 201.1

uDFOA2[011] 5 529.3

10.2

144.1 369.2 243.1 287.2 345.2 185.1 427.2 103.1 102.1 86.1 86.1

uDFOA2[110] 6 529.3

trace

128.1 385.2 227.1 303.2 329.2 201.1 411.2 119.1 86.1 86.1 102.1 185.1 369.2 185.1 185.1

uDFOA2[101] 7 529.3

18.9

128.1 385.2 227.2 303.2 345.2 185.1 427.2 103.1 86.1 102.1 86.1

201.1 385.2 185.1 201.1

uDFOA3[111] 8 513.3

4.4

128.1 369.2 227.1 287.2 329.2 185.1 411.2 103.1 86.1 86.1 86.1

185.1 369.2 185.1 185.1

201.1 401.2 201.1 201.1

185.1 385.2 201.1 185.1

a

Determined from peak integrals and the relative intensity of MS/MS signals resolved for isomers of the uDFOA1 or uDFOA2 series. b The term 8.9C describes breaking the bond between atoms 8 and 9 of the main chain of DFOB (refer Scheme 2 for numbering) in which the relevant fragment contains the terminal acetyl group (C terminus). Where the subscript C is replaced with N, the relevant fragment contains the terminal amine group (N terminus). The subscript I describes an internal fragment. Atom numbers are referenced to DFOB and may differ from the absolute atom number in a given analogue.

This LC-MS/MS analysis allowed the identification of isomers and an estimate of the relative concentrations within the uDFOA1 and the uDFOA2 series as: uDFOA1[001] >> uDFOA1[100] > uDFOA1[010]; and uDFOA2[101] > uDFOA2[011] >> uDFOA2[110]. The MS/MS fragmentation patterns from the species produced in the DB-supplemented system (Supplementary Figure 4, Supplementary Scheme 1) showed the same trend in isomer distribution within the DFOA1 and DFOA2 series: DFOA1[001] >> DFOA1[100] > DFOA1[010]; and DFOA2[101] > DFOA2[011] >> DFOA2[110] (Supplementary Table 1). NMR Spectroscopy. The two peaks that were resolved for the uDFOA1 series (major and minor), the single peak of the uDFOA2 series and the single peak containing uDFOA3 were isolated from the LC and these purified fractions were analyzed by HRMS and by 1D- (1H, 13C) and 2D- (COSY,

HMBC, HSQC) NMR spectroscopy. The 1H-1H COSY experiment confirmed that the major peak from the uDFOA1 series contained uDFOA1[001] (Supplementary Figures 5-6). The methine protons in this sample gave a signal at δH 5.66 ppm, with 1H-1H correlations to δ 4.06 ppm and δ 3.30 ppm. The latter correlation was evidence for the incorporation of the EDBE-derived substrate in the amine region of the trimer. The 1 H-1H COSY data was in agreement with the conclusions based on MS/MS fragmentation that the minor peak from the uDFOA1 series contained a mixture of uDFOA1[010] and uDFOA1[100], although these isomers were unable to be distinguished. The 1H-1H COSY correlation data was consistent with the incorporation of two or three E-DBE-derived substrates in the uDFOA2 or uDFOA3 samples, respectively (Supplementary Figures 7-12).


Page 17 of 23

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


Table 2. MS/MS Fragmentation Patterns From Dimeric Precursors of DFOB and Analogues Assembled From the NAcetylated or the Amine Region From Substrates DP and E-DBE F16 Dimer

[M+H] (calc)

+

Conc (%)

a

8.9C

F17 9.10N

F18

F19

F20

F21

F22

13.14C

13.14Nb

F23

F24

F25

F26

F27

3.9I

3.13I

14.19I 20.21N 24.25N 25.30I 24.25C 14.24I

dDFX[00-]

16 361.2

62.0

144.1 201.1 243.1 119.2

102.1 201.1 102.1 NRc

NR

NR

NR

NR

udDFX[01-]

17 345.2

4.4

144.1 185.1 243.1 103.2

102.1 201.1 86.1

NR

NR

NR

NR

NR

udDFX[10-]

18 345.2

22.4

128.1 201.1 227.1 119.2

86.1

185.1 102.1 NR

NR

NR

NR

NR

NR

NR

NR

NR

119.1

102.1 301.1

udDFX[11-]

19 329.2

2.5

128.1 185.1 227.1 103.2

86.1

185.1 86.1

dDFX[-00]

20 419.3

0.9

NR

NR

NR

319.2

NR

NR

102.1 201.1

udDFX[-01]

21 403.2

1.9

NR

NR

NR

303.1

NR

NR

102.1 185.1

103.1

86.1

301.1

201.1

udDFX[-10]

22 403.2

0.1

NR

NR

NR

303.1

NR

NR

86.1

201.1

119.1

102.1 285.1

185.1

udDFX[-11]

23 387.2

5.8

NR

NR

NR

287.1

NR

NR

86.1

185.1

103.1

86.1

185.1

a

NR

285.1

201.1

b

Determined from peak integration. The term 13.14N describes breaking the bond between atoms 13 and 14 of the main chain of dDFX[00-] or dDFX[-00] in which the relevant fragment contains the terminal amine group (N terminus). Where the subscript N is replaced with C, the relevant fragment contains the terminal acetyl group (C terminus). The subscript I describes an internal fragment. Atom numbers are referenced to dDFX[00-] or dDFX[-00] and may differ from the absolute atom number in a given analogue. c NR, not relevant for this compound.

Dimeric Precursors of DFOB and Analogues. A wellresolved signal in the extract from the DB-supplemented system at tR 28.79 min (Supplementary Figure 1a) analyzed as m/z 361.2 (Supplementary Figure 2f) and in the presence of Fe(III) as m/z 414.1 (Supplementary Figure 2l). These data were consistent with the assignment of this species as the dimer formed from one AHDP and one SHDP fragment: dDFX[00-], where ‘d’ = dimer; and ‘-‘ = vacant position at the amine region (Chart 2, 16). This species was also observed at low levels in the extracts from the native and the E-DBEsupplemented systems (Figure 1). LC-MS-QQQ analysis using SIM detection from the E-DBE extract showed the presence of dDFX[00-] (16) and signals in accord with the presence of the other dimers that could be assembled from the N-acetylated region: udDFX[01-] (17), udDFX[10-] (18) and udDFX[11-] (19) (Figure 4a-c). The identity of each dimer was supported from characteristic MS/MS fragmentation patterns (Table 2, Supplementary Figure 13, Supplementary Scheme 2). The relative concentration of the dimers from SIM peak integration was: dDFX[00-] (16) >> udDFX[10-] (18) >> udDFX[01-] (17) > udDFX[11-] (19). The dimeric fragments assembled from the amine region from the E-DBE extract were identified using SIM detection (Figure 4d-f) and MS/MS fragmentation patterns (Table 2, Supplementary Figure 14, Supplementary Scheme 2). The concentration of the dimer species dDFX[-00] (20), where ‘-‘ = vacant position at the N-acetylated region, was about 70-fold less than dDFX[00-] (16) (Figure 4a, d). It was possible that this difference in concentration was a result of distinct ionization properties of the differentially-terminated fragments (Nacetyl/amine, carboxylic acid/amine). MS analysis of a mixture of two standard compounds that modeled each fragment type (N-(5-aminopentyl)-N-hydroxyacetamide and 4-[(5aminopentyl)(hydroxy)amino]-4-oxobutanoic acid)19,37 gave similar signal intensities (Supplementary Figure 15), which supported that the concentration differences were real and not an artefact of the analysis. Dimers assembled from the amine region from one DP and one E-DBE-derived fragment (SIM 403) were present at similarly low levels as dDFX[-00] (20), with udDFX[-01] (21) > udDFX[-10] (22). From this dimer

series udDFX[-11] (23), assembled from two E-DBE-derived fragments, was present in the highest relative concentration. Chart 2. Dimeric Precursors of DFOB and Analogues From the N-Acetylated (16-19) or Amine (20-23) Region Assembled From Substrates DP and E-DBE

Trends in the distribution of the dimeric precursors of DFOB and analogues were similar between the E-DBE- and the DB-supplemented systems (Supplementary Table 2). Dimer dDFX[00-] (16) was the dominant dimer, with dDFX[-00] (20) present in about 30-fold less concentration (Figure 4g, j). The concentration of dDFX[10-] (25) >> dDFX[01-] (24) (Figure 4h), with the identity of the dimers confirmed from



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

MS/MS fragmentation (Supplementary Figure 16, Supplementary Scheme 3). Aside from dDFX[-00] (20), the other dimers that would require the initiation of assembly from the amine region: dDFX[-01] (27), dDFX[-10] (28) or dDFX[-11] (29) (Supplementary Chart 2); were not detected (Supplementary Figure 17).

Figure 4. LC-MS-QQQ signals detected from the E-DBE supplemented system (a-f) at SIM values: (a) 361; (b) 345; (c) 329; (d) 419; (e) 403; or (f) 387; or from the DB supplemented system (g-l) at SIM values: (g) 361; (h) 347; (i) 333; (j) 419; (k) 405; or (l) 391.

Preference in the Direction of DFOB Biosynthesis. Although the uDFOA1 isomers uDFOA1[001] (2), uDFOA1[010] (3) and uDFOA1[100] (4) had the same 2:1 stoichiometry of the DP- or E-DBE-derived fragments, the relative concentration of the isomers was different, which indicated that the assembly was occurring in a selective fashion. This notion was also supported by the isomers in the uDFOA2 series, which despite the common 1:2 stoichiometry of the DP- or E-DBEderived fragments, were present in different relative concentrations. The trends in the relative concentrations of isomers in the uDFOA1 series: uDFOA1[001] (2) >> uDFOA1[100] (4) > uDFOA1[010] (3); and in the uDFOA2 series: uDFOA2[101] (7) > uDFOA2[011] (5) >> uDFOA2[110] (6), allowed for insight into two aspects of DFOB biosynthesis: the directional preference of DesD-catalyzed assembly; and the specific sequence of steps. Late-stage DFOB assembly involves the formation of two amide bonds, as catalyzed by DesD (Scheme 1). The direction of these two condensation reactions could occur from the Nacetylated region to the amine region (Scheme 3, sequence at left: Camide to Namide) or from the amine region to the Nacetylated region (Scheme 3, sequence at right: Namide to C amide). A directional preference from the N-acetylated region to the amine region would involve as the first condensation step the reaction between AHDP and SHDP to form heterodimer 16. In the second condensation step, dimer 16 would react with a second unit of SHDP to form DFOB (1). In the reverse direction, the first step would yield homodimer 20 as a condensation product from the reaction between two units of SHDP. The second-step condensation between dimer 20 and one unit of AHDP would yield DFOB (1).

Page 18 of 23

Results from this work were consistent with DFOB biosynthesis being preferentially directed from the N-acetylated region to the amine region. This was supported by the dominance of uDFOA1[001] (2) above uDFOA1[100] (4). In this direction, the biosynthesis of uDFOA1[001] (2) would require dimer 16 as the precursor of the final trimer. Since dimer 16 is common to DFOB biosynthesis, it would be expected, as the native substrate, to be available in a relatively high concentration. Dimer dDFX[00-] (16) was detected in the native, the EDBE and the DB-supplemented systems. In this direction, uDFOA1[010] (3) and uDFOA1[100] (4) would be assembled from dimer precursors 17 or 18, respectively, that featured one native DP and one non-native E-DBE-derived substrate, which would likely be available in concentrations below native 16. If the preference in the biosynthetic direction occurred from the amine region to the N-acetylated region then uDFOA1[100] (4) would be expected to dominate above uDFOA1[001] (2). In this opposing direction, uDFOA1[100] (4) would require homodimer 20 as the precursor substrate of the final trimer. As the native dimer precursor of DFOB (1), the concentration of dimer 20 would be expected to be relatively high and would predict for uDFOA1[100] (4) as the dominant isomer, which was not the case. Dimer dDFX[-00] (20) was detected in concentrations between 30-70-fold less than dimer dDFX[00-] (16). Identification of a preference in the biosynthetic direction of DFOB and analogues was only possible by virtue of its assembly from two different N-acylated substrates. Linear desferrioxamine G1 (DFOG1) and its macrocyclic product desferrioxamine E (DFOE) are assembled from SHDP as the unique N-acylated substrate.35,38 In these cases, it would not be possible to use a precursor-directed biosynthesis approach to determine a preference in the biosynthetic direction of the finalstage condensation steps, since the dimer precursors would be identical. Sequence of DesD-Catalyzed DFOB Biosynthesis. Within each of the uDFOA1 and the uDFOA2 series, two trimers were built following the second condensation step from a dimer precursor comprised of one native DP-derived substrate (AHDP or SHDP) and one E-DBE-derived substrate (AH-EDBE or SH-E-DBE). For the uDFOA1 isomer subset, uDFOA1[100] (4) > uDFOA1[010] (3); and for the uDFOA2 isomer subset, uDFOA2[101] (7) > uDFOA2[011] (5). As paired by relative concentration, the high concentration isomers uDFOA1[100] (4) and uDFOA2[101] (7) used 18 as the common heterodimer precursor; and the low concentration isomers uDFOA1[010] (3) and uDFOA2[011] (5) used heterodimer 17 (Scheme 3). This suggested that the dimer precursor udDFX[10-] (18) was available at higher concentration than udDFX[01-] (17). This was supported by the SIM measurements (Figure 4b), with the dimers identified from characteristic MS/MS fragmentation patterns (Supplementary Figure 13). The dominance of udDFX[10-] (18) above udDFX[01-] (17) allowed for subtle insight into the sequence of steps catalyzed by DesD in DFOB biosynthesis. Since the relative concentration of products from reactions between combinations of native AHDP and SHDP would be expected to be higher than those involving non-native AH-E-DBE and SH-E-DBE, the trend of udDFX[10-] (18) > udDFX[01-] (17) was consistent with the notion that the activation of native SHDP occurred more readily than the activation of SH-E-DBE as the first step of the biosynthesis, and that the availability of activated SHDP


Page 19 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 3. Biosynthesis of (a) DFOB (1); and (b-h) Analogues (2-8) Shown in the Direction From CAmide to NAmide (Pathway at Left) or NAmide to CAmide (Pathway at Right).

Refer Chart 1 and Chart 2 for full structures. Fragments derived from native DP or non-native E-DBE are depicted as solid lines in black or grey, respectively. Amide bonds are represented as narrow lines perpendicular to the solid lines.

was the major determinant of the final concentration of a given dimer precursor. It has been established that the activation of the SHDP unit occurs as part of the DesD-catalytic cycle via the formation of an acyl adenylate intermediate.19 Levels of activated SHDP were sufficient to allow the condensation of non-native AH-E-DBE to form udDFX[10-] (18). Levels of activated SH-E-DBE were relatively low, which reduced the concentration of udDFX[01-] (17) formed from the condensation with native AHDP. This notion was also in accord with dDFX[00-] (16) >> udDFX[10-] (18), whereby activated SHDP was condensed in the first step with either native AHDP (high concentration) or non-native AH-E-DBE (low concentration). The trimer products would ultimately be formed following the second condensation step between the available dimer precursors and the second unit of activated SHDP or activated non-native SH-E-DBE. Consideration of the relative concentration of the trimeric DFOB analogues and the dimer precursors has allowed the sequence of steps in the two DesD-catalyzed condensation reactions in DFOB biosynthesis to be defined (Scheme 4). The first step involves the activation of SHDP, which is then condensed with AHDP to form an AHDP-SHDP heterodimer. A second aliquot of activated SHDP would next be formed which would undergo condensation with the heterodimer to

form the final DFOB trimer. The condensation between AHDP and activated SHDP would effectively cap one end of the dimer product, thereby reducing the propensity of side reactions. It would be likely that the activation of SHDP would occur as two discrete steps (step (i) and step (iii)), since a pool of activated SHDP produced in a single step in excess of AHDP could promote the formation of SHDP oligomers. Scheme 4. Sequence of DesD-Catalyzed DFOB Biosynthesis



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

Sequence of DesD-Catalyzed Biosynthesis: DFOB, DFOG1, DFOE. The biosynthetic sequence for DFOB can be considered in terms of the effectiveness of monomeric SHDP or dimeric SHDP-SHDP as a substrate in the DesD-catalyzed activation step. Results from this work suggest that SHDP is the optimal substrate for activation, with SHDP-SHDP significantly less effective. DFOB biosynthesis from the Nacetylated region to the amine region would only ever require the activation of a SHDP monomer. In the reverse direction, the biosynthesis would require activation of the SHDP-SHDP dimer as the step prior to AHDP condensation. In the case of DFOB biosynthesis, this necessarily prescribes a preference in the biosynthetic direction in appending activated SHDP to AHDP-SHDP rather than AHDP to poorly activated SHDPSHDP. In the case of siderophores DFOG1 and DFOE assembled only from SHDP, a directional preference in the biosynthesis would not be prescribed, since the second-step condensation could occur between an activated SHDP monomer and SHDP-SHDP. This rationale is consistent with a study that in DB-supplemented cultures of Erwinia amylovora identified constitutional isomers of DFOG1 assembled from two units of SHDP and one unit of SHDB that were produced in approximately equal concentration.35

CONCLUSION The DesBCD enzyme cluster in S. pilosus was competent in the use of E-DBE as a competitive non-native substrate against native DP to produce a library of unsaturated DFOB analogues that retained function as Fe(III) chelators in vitro. The distribution of dimer precursors and trimer products assembled from a mixture of DP- and E-DBE-derived substrates has enabled the steps in DesD-catalyzed DFOB biosynthesis to be defined. DFOB biosynthesis involves: (i) activation of SHDP; (ii) condensation with AHDP to form AHDP-SHDP; (iii) SHDP activation; (iv) condensation with AHDP-SHDP to form DFOB. This directional preference for DFOB biosynthesis is prescribed by the substrate selectivity of DesD towards activating monomeric SHDP but not dimeric SHDP-SHDP.

METHODS Reagents. Difco yeast mold (YM) broth was sourced from BD Biosciences. Potassium phosphate monobasic (≥99%), sodium phosphate dibasic (≥99%), trizma base (≥99%), zinc sulphate heptahydrate (≥99%), L-threonine (≥99.5%), chelex 100 sodium form, 1,4-diaminobutane dihydrochloride (≥98%), hydrochloric acid (37%), sodium hydroxide (≥98%), iron(III) perchlorate hydrate (≥99%), perchloric acid (70%), Amberlite XAD-2, chrome azurol S (CAS, 65%), piperazine (99%), 5sulfosalicylic acid dihydrate (≥99%), and sodium chloride (≥99%) were sourced from Sigma-Aldrich. Calcium chloride (≥94%), magnesium sulphate heptahydrate (≥98%), methanol (≥99%) ethanol (≥99.5%), acetonitrile (190 grade), and iron(III) chloride hexahydrate (≥98%) were sourced from Ajax Finechem. 1,4-diamino-2(E)-butene dihydrochloride (>96%) was from Small Molecules Inc., Ni(II) Sepharose 6 Fast Flow from GE healthcare, and 4-(2-hydroxyethyl)-1piperazineethanesulfonic acid (HEPES) from AMResco. N-(5Aminopentyl)-N-hydroxyacetamide was from ApexBio and 4[(5-aminopentyl)(hydroxy)amino]-4-oxobutanoic acid was prepared from literature.19,37 MilliQ water was used in all experiments, as required.

Page 20 of 23

Instrumentation. Liquid chromatography-mass spectrometry (LC-MS-Q) was conducted using an Agilent system (Santa Clara, CA, USA) consisting of a 1260 series quaternary pump with an inbuilt degasser, 1200 series autosampler, thermostated column compartment, diode array detector, and fraction collector, and a 6120 series single quadrupole mass spectrometer. The drying gas flow, temperature, and nebuliser were set to 12 L min–1, 350 °C, and 35 psi. Agilent OpenLAB Chromatography Data System (CDS) ChemStation Edition (B.04.02) was used for data acquisition and processing. LCMS/MS-QQQ was conducted using an Agilent system consisting of 1290 series quaternary pump with inbuilt degasser, autosampler, thermostated column compartment, and diode array detector, and a 6460 series triple quadrupole mass spectrometer with jet stream technology. Agilent MassHunter Workstation (B.07.01) was used for data acquisition and processing. Product ion mode was used for fragmentation using collision energy voltages optimized for individual precursor ions, ranging from 14–30 V. The fragmentor voltage was set to 150 V, and drying gas flow, temperature, and nebuliser were set to 10 L min–1, 300 °C, and 25 psi. An Agilent Zorbax Eclipse XDBC18 column (5 µm particle size, 150 mm length, 2.1 mm internal diameter) was used for analytical runs on both machines at a 0.2 mL min–1 flow rate with a gradient of 0-30% ACN (0.1% formic acid) in water (0.1% formic acid) over 40 min at 30 °C. Electrospray ionisation (ESI) was used on both machines to analyse aliquots (10 µL) in positive ion mode with a 4000 V capillary voltage. An Agilent Zorbax Eclipse XDBC18 (5 µm particle size, 250 mm length, 9.4 mm internal diameter) was used for semi-preparative runs at a 1 mL min–1 flow rate with a gradient of 0-30% ACN (0.1% formic acid) in water (0.1% formic acid) over 40 min at 30 °C (100 µL aliquots). 1H NMR and 13C NMR spectra were recorded in 5 mm Pyrex tubes, using a Varian 400-MR NMR spectrometer at a frequency of 399.73 MHz at 24 °C operated with VnmrJ 3.1 software. The spectral data are reported in ppm (δ) and referenced to residual solvent (DMSO-d6 2.50/39.52 ppm). Bacterial Culturing. Streptomyces pilosus (ATCC 19797) permanents were stored at –80 °C in DMSO (10% v/v) in YM broth (2.1% w/v). Precultures of S. pilosus (100 mL) were grown in YM broth (2.1% w/v) for 4 d at 28 °C in an Eppendorf New Brunswick Innova 42 Shaking Incubator (Hamburg, Germany) at 160 rpm. The base medium was made by combining a phosphate buffer with solutions of YM broth and other components. The final composition was YM broth (2.1% w/v), potassium phosphate monobasic (235 mM), trizma base (35.0 mM), zinc sulphate heptahydrate (13.9 µM), calcium chloride (13.6 mM), sodium phosphate dibasic (11.2 mM), magnesium sulphate heptahydrate (2.43 mM), and threonine (0.84 mM).30 All YM broth solutions and the phosphate buffer were stirred with chelex 100 for 2 h to remove residual iron, were pH adjusted to 6.00 ± 0.05 using hydrochloric acid or sodium hydroxide, and were autoclaved (121 °C, 20 min) for sterilisation. All work was completed in a biosafety cabinet using sterilised consumables and iron deplete glassware and plasticware. Cells were withdrawn from the precultures, centrifuged (5000 rpm), resuspended in fresh medium, and used to inoculate 50 mL cultures of the base medium. Individual conditions were established by addition of DB or E-DBE solutions (pH 6.00 ± 0.05) to a concentration of 20 mM with respect to the dihydrochloride salt. The other component mixtures and substrate solutions were filtered with Minisart 0.2 µm syringe filters before use. The bacteria was grown at 28 °C


Page 21 of 23

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


and 160 rpm for 8 d. Siderophore production was tracked daily by adding supernatant (200 µL) to ferric perchlorate (10 mM) in perchloric acid (200 mM) (100 µL) and measuring absorbance at 465 nm using a BMG Labtech FLUOstar Omega microplate reader. The supernatant was collected for purification by centrifugation (5000 rpm), and the bacteria quantity in each culture was monitored by lyophilisation. Purification. XAD-2 purification was modified from previous methodology.23,39 Amberlite XAD-2 resin (100 mL) was activated in methanol, suspended in a column in water, and washed with 2 column volumes (CVs) of water. The supernatant was adsorbed on the resin which was then washed with 2 CVs of water. The sample was eluted using 2-4 CVs of aqueous methanol (50% v/v). Siderophore presence was tracked using a CAS assay.40,41 The CAS positive fractions from the elution step were taken to dryness in vacuo using a Buchi Rotavapor R-300. Ni(II)-IMAC was adapted from previous methodology.41 Ni(II) Sepharose (40 mL) was washed with water (5 CV) and the resin was charged with binding buffer (5 CV; 10 mM HEPES, 0.2 M sodium chloride, pH 9.0). The sample was dissolved in binding buffer (2 mL) and adsorbed onto the resin, which was then washed with binding buffer (5 CV), and the sample eluted with elution buffer (5 CV; 10 mM HEPES, 0.2 M sodium chloride, pH 5.5). The CAS assay positive fractions from the elution step were taken to dryness in vacuo and were desalted by dissolution in methanol. The XAD-2 and IMAC purified supernatant was subjected to semipreparative HPLC to isolate individual peaks. Collected fractions were lyophilised using a Labconco FreeZone freezedryer. Three aliquots were subject to analysis by HRMS. uDFOA1 series (major peak): m/z ESI (positive ion). Found [M+H]+ 545.32934 (20), [C24H45N6O8]+ requires 545.32993. uDFOA1 series (minor peak): m/z ESI (positive ion). Found [M+H]+ 545.32919 (52), [C24H45N6O8]+ requires 545.32993. uDFOA2 (single peak): m/z ESI (positive ion). Found [M+H]+ 529.29807 (22), [C23H41N6O8]+ requires 529.29863. Sample Preparation: LC-MS-Q and LC-MS/MS-QQQ. Samples (0.1 mg) were dissolved in HPLC grade methanol (1 mL) for analysis by LC-MS-Q and LC-MS/MS-QQQ. Fe(III)loaded samples were generated by adding 100 µL of an iron(III) chloride solution (350 µM) to 100 µL of analyte and incubating for 1 h. Aliquots (10 µL) were analyzed in positive ion mode.

ASSOCIATED CONTENT Supporting Information Structures of DFOB analogues assembled from DB supplementation (Supplementary Chart 1), or of dimers from DBsupplementation (Supplementary Chart 2). LC-MS-Q (Supplementary Figure 1, LC), MS (Supplementary Figure 2, MS) and LC-MS/MS-QQQ data (Supplementary Figure 4, Supplementary Scheme 1, Supplementary Table 1) from DB-supplementation. LC-MS-QQQ (SIM: [M+H]+) and LC-MS-Q (Sum EIC: [M+H]+ + [M+2H]2+) from E-DBE-supplementation (Supplementary Figure 3). NMR spectra from uDFOA1 major peak (Supplementary Figure 5-6), uDFOA1 minor peak (Supplementary Figure 7-8), uDFOA2 (Supplementary Figure 9-10), uDFOA3 (Supplementary Figure 11-12). LC-MS/MS-QQQ data from dimers from E-DBE(Supplementary Figure 13-14, Supplementary Scheme 2) or DB(Supplementary Figure 16-17, Supplementary Scheme 3, Supplementary Table 2) supplementation. LC-MS-QQQ from standards (Supplementary Figure 15). This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author Phone: +61 2 9351 6738. E-mail: *[email protected].

ACKNOWLEDGMENTS The University of Sydney is acknowledged for funding support (bridging support, RC; full or partial postgraduate scholarships to TJT or CZS, respectively). The National Health and Medical Research Council (NHMRC) is acknowledged for funding the salary of MPG. W. Tieu is acknowledged for providing 4-[(5aminopentyl)(hydroxy)amino]-4-oxobutanoic acid.

REFERENCES (1) Bickel, H., Bosshardt, R., Gäumann, E., Reusser, P., Vischer, E., Voser, W., Wettstein, A., and Zähner, H. (1960) Metabolic products of Actinomycetaceae. XXVI. Isolation and properties of ferrioxamines A to F, representing new sideramine compounds, Helv. Chim. Acta 43, 2118–2128. (2) Bickel, H., Hall, G. E., Keller-Schierlein, W., Prelog, V., Vischer, E., and Wettstein, A. (1960) Metabolic products of Actinomycetaceae. XXVII. Constitutional formulas of ferrioxamine B, Helv. Chim. Acta 43, 2129–2138. (3) Neilands, J. B. (1981) Microbial iron compounds, Ann. Rev. Biochem. 50, 715–731. (4) Drechsel, H., and Winkelmann, G. (1997) Iron chelation and siderophores, In Transition Metals in Microbial Metabolism (Winkelmann, G., and Carrano, C. J., Eds.), pp 1–49, Harwood Academic, Amsterdam. (5) Winkelmann, G. (2002) Microbial siderophore-mediated transport, Biochem. Soc. Trans. 30, 691–696. (6) Budzikiewicz, H. (2010) Microbial siderophores, In Progress in the Chemistry of Organic Natural Products (Kinghorn, A. D., Falk, H., and Kobayashi, J., Eds.), pp 1–75, Springer-Verlag, New York. (7) Hider, R. C., and Kong, X. (2010) Chemistry and biology of siderophores, Nat. Prod. Rep. 27, 637–657. (8) Sandy, M., and Butler, A. (2009) Microbial iron acquisition: Marine and terrestrial siderophores, Chem. Rev. 109, 4580–4595. (9) Davis, B. A., and Porter, J. B. (2002) Results of long term iron chelation treatment with deferoxamine, In Iron Chelation Therapy (Hershko, C., Ed.), pp 91–125, Kluwer Academic / Plenum Publishers, New York. (10) Bergeron, R. J., and McManis, J. S. (1990) The total synthesis of desferrioxamines E and G, Tetrahedron 46, 5881–5888. (11) Bergeron, R. J., McManis, J. S., Phanstiel, O. I., and Vinson, J. R. T. (1995) A versatile synthesis of deferrioxamine B, J. Org. Chem. 60, 109–114. (12) Poreddy, A. R., Schall, O. F., Osiek, T. A., Wheatley, J. R., Beusen, D. D., Marshall, G. R., and Slomczynska, U. (2004) Hydroxamate-based iron chelators: Combinatorial syntheses of desferrioxamine B analogues and evaluation of binding affinities, J. Comb. Chem. 6, 239–254. (13) Peter, H. H. (1985) Industrial aspects of iron chelators: Pharmaceutical applications, In Proteins of Iron Storage and Transport (Spik, G., Montreuil, J., Crichton, R. R., and Mazurier, J., Eds.), pp 293–303, Elsevier, Amsterdam. (14) Schupp, T., Waldmeier, U., and Divers, M. (1987) Biosynthesis of desferrioxamine B in Streptomyces pilosus: Evidence for the involvement of lysine decarboxylase, FEMS Microbiol. Lett. 42, 135–139. (15) Schupp, T., Toupet, C., and Divers, M. (1988) Cloning and expression of two genes of Streptomyces pilosus involved in the biosynthesis of the siderophore desferrioxamine B, Gene 64, 179– 188. (16) Challis, G. L. (2005) A widely distributed bacterial pathway for siderophore biosynthesis independent of nonribosomal peptide synthetases, ChemBioChem 6, 601–611.



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

(17) Bentley, S. D., Chater, K. F., Cerdeño-Tárraga, A.-M., Challis, G. L., Thomson, N. R., James, K. D., Harris, D. E., Quail, M. A., Kieser, H., Harper, D., Bateman, A., Brown, S., Chandra, G., Chen, C. W., Collins, M., Cronin, A., Fraser, A., Goble, A., Hidalgo, J., Hornsby, T., Howarth, S., Huang, C.-H., Kieser, T., Larke, L., Murphy, L., Oliver, K., O’Neil, S., Rabbinowitsch, E., Rajandream, M.-A., Rutherford, K., Rutter, S., Seeger, K., Saunders, D., Sharp, S., Squares, R., Squares, S., Taylor, K., Warren, T., Wietzorrek, A., Woodward, J., Barrell, B. G., Parkhill, J., and Hopwood, D. A. (2002) Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2), Nature 417, 141–147. (18) Barona-Gomez, F., Wong, U., Giannakopulos, A. E., Derrick, P. J., and Challis, G. L. (2004) Identification of a cluster of genes that directs desferrioxamine biosynthesis in Streptomyces coelicolor M145, J. Am. Chem. Soc. 126, 16282–16283. (19) Kadi, N., Oves-Costales, D., Barona-Gomez, F., and Challis, G. L. (2007) A new family of ATP-dependent oligomerization-macrocyclization biocatalysts, Nat. Chem. Biol. 3, 652–656. (20) Meiwes, J., Fiedler, H.-P., Zähner, H., Konetschny-Rapp, S., and Jung, G. (1990) Production of desferrioxamine E and new analogues by directed fermentation and feeding fermentation, Appl. Microbiol. Biotechnol. 32, 505–510. (21) Konetschny-Rapp, S., Jung, G., Raymond, K. N., Meiwes, J., and Zaehner, H. (1992) Solution thermodynamics of the ferric complexes of new desferrioxamine siderophores obtained by directed fermentation, J. Am. Chem. Soc. 114, 2224–2230. (22) Soe, C. Z., and Codd, R. (2014) Unsaturated macrocyclic dihydroxamic acid siderophores produced by Shewanella putrefaciens using precursor-directed biosynthesis, ACS Chem. Biol. 9, 945–956. (23) Ledyard, K. M., and Butler, A. (1997) Structure of putrebactin, a new dihydroxamate siderophore produced by Shewanella putrefaciens, J. Biol. Inorg. Chem. 2, 93–97. (24) Hou, Z., Raymond, K. N., O'Sullivan, B., Esker, T. W., and Nishio, T. (1998) A preorganized siderophore: Thermodynamic and structural characterization of alcaligin and bisucaberin, microbial macrocyclic dihydroxamate chelating agents, Inorg. Chem. 37, 6630– 6637. (25) Brickman, T. J., Hansel, J.-G., Miller, M. J., and Armstrong, S. K. (1996) Purification, spectroscopic analysis and biological activity of the macrocyclic dihydroxamate siderophore alcaligin produced by Bordetella pertussis and Bordetella bronchiseptica, BioMetals 9, 191–203. (26) Boettcher, T., and Clardy, J. (2014) A chimeric siderophore halts swarming Vibrio, Angew. Chem. Int. Ed. 53, 3510– 3513. (27) Kadi, N., Arbache, S., Song, L., Oves-Costales, D., and Challis, G. L. (2008) Identification of a gene cluster that directs putrebactin biosynthesis in Shewanella species: PubC catalyzes cyclodimerization of N-hydroxy-N-succinylputrescine, J. Am. Chem. Soc. 130, 10458-10459. (28) Soe, C. Z., Pakchung, A. A. H., and Codd, R. (2012) Directing the biosynthesis of putrebactin or desferrioxamine B in Shewanella putrefaciens through the upstream inhibition of ornithine decarboxylase, Chem. Biodivers. 9, 1880–1890. (29) Soe, C. Z., Telfer, T. J., Levina, A., Lay, P. A., and Codd, R. (2015) Simultaneous biosynthesis of putrebactin, avaroferrin and bisucaberin by Shewanella putrefaciens and characterisation of complexes with iron(III), molybdenum(VI) or chromium(V), J. Inorg. Biochem. doi 10.1016/j.jinorgbio.2015.12.008, in press. (30) Chiani, M., Akbarzadeh, A., Farhangi, A., Mazinani, M., Saffari, Z., Emadzadeh, K., and Mehrabi, M. R. (2010) Optimization of culture medium to increase the production of desferrioxamine B (Desferal) in Streptomyces pilosus, Pak. J. Biol. Sci. 13, 546–550. (31) Keller-Schierlein, W., Mertens, P., Prelog, V., and Wasler, A. (1965) Metabolic products of microorganisms. XLIX. Ferrioxamines A1, A2, and D2, Helv. Chim. Acta 48, 710–723. (32) Ejje, N., Soe, C. Z., Gu, J., and Codd, R. (2013) The variable hydroxamic acid siderophore metabolome of the marine actinomycete Salinispora tropica CNB-440, Metallomics 5, 1519– 1528.

Page 22 of 23

(33) Simionato, A. V. C., de Souza, G. D., Rodrigues-Filho, E., Glick, J., Vouros, P., and Carrilho, E. (2006) Tandem mass spectrometry of coprogen and deferoxamine hydroxamic siderophores, Rapid Commun. Mass Spectrom. 20, 193–199. (34) Groenewold, G. S., Van Stipdonk, M. J., Gresham, G. L., Chien, W., Bulleigh, K., and Howard, A. (2004) Collision-induced dissociation tandem mass spectrometry of desferrioxamine siderophore complexes from electrospray ionization of UO22+, Fe3+ and Ca2+ solutions, J. Mass Spectrom. 39, 752–761. (35) Feistner, G. J., Stahl, D. C., and Gabrik, A. H. (1993) Proferrioxamine siderophores of Erwinia amylovora. A capillary liquid chromatographic/electrospray tandem mass spectrometric study, Org. Mass Spectrom. 28, 163–175. (36) Pluháček, T., Lemr, K., Ghosh, D., Milde, D., Novák, J., and Havlíček, V. (2016) Characterization of microbial siderophores by mass spectrometry, Mass Spectrom. Rev. 35, 35–47. (37) Lifa, T., Tieu, W., Hocking, R. K., and Codd, R. (2015) Forward and reverse (retro) iron(III)- or gallium(III)-desferrioxamine E and ring-expanded analogs prepared using metal-templated synthesis from endo-hydroxamic acid monomers, Inorg. Chem. 54, 3573–3583. (38) Keller-Schierlein, W., and Prelog, V. (1962) Metabolic products of Actinomycetaceae. XXXIV. Ferrioxamine G, Helv. Chim. Acta 45, 590–595. (39) Pakchung, A. A. H., Soe, C. Z., and Codd, R. (2008) Studies of iron-uptake mechanisms in two bacterial species of the Shewanella genus adapted to middle-range (Shewanella putrefaciens) or Antarctic (Shewanella gelidimarina) temperatures, Chem. Biodivers. 5, 2113–2123. (40) Schwyn, B., and Neilands, J. B. (1987) Universal chemical assay for the detection and determination of siderophores, Anal. Biochem. 160, 47–56. (41) Soe, C. Z., Pakchung, A. A. H., and Codd, R. (2014) Dinuclear [VVO(putrebactin))2(µ-OCH3)2] formed in solution as established from LC-MS measurements using 50V-enriched V2O5, Inorg. Chem. 53, 5852–5861.


Page 23 of 23

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


SYNOPSIS TOC (Word Style “SN_Synopsis_TOC”). If you are submitting your paper to a journal that requires a synopsis graphic and/or synopsis paragraph, see the Instructions for Authors on the journal’s homepage for a description of what needs to be provided and for the size requirements of the artwork.

Table of Contents artwork


Mixing Up the Pieces of the Desferrioxamine B ... - ACS Publications

Recommend Documents