Fungal Biotransformation of Tetracycline Antibiotics - The Journal of

Jul 15, 2016 - ... Biosciences, The University of Queensland, St. Lucia, QLD 4072, Australia .... Microbiome-Mediated Biotransformation of Cane Toad B...
0 downloads 0 Views 740KB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Featured Article

Fungal Biotransformation of Tetracycline Antibiotics Zhuo Shang, Angela Aguslyarti Salim, Zeinab Khalil, Paul Vincent Bernhardt, and Robert John Capon J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.6b01272 • Publication Date (Web): 15 Jul 2016 Downloaded from http://pubs.acs.org on July 19, 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.

The Journal of Organic Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

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

The Journal of Organic Chemistry

Fungal Biotransformation of Tetracycline Antibiotics Zhuo Shang,a Angela A. Salim,a Zeinab Khalil,a Paul V. Bernhardtb and Robert J. Capon*a a

Institute for Molecular Bioscience, The University of Queensland, St. Lucia, QLD 4072,

Australia b

School of Chemistry and Molecular Biosciences, The University of Queensland, St. Lucia, QLD

4072, Australia

*[email protected] KEYWORDS (tetracycline antibiotics, biotransformation, marine-derived fungus, Paecilomyces sp.)

ACS Paragon Plus Environment

The Journal of Organic Chemistry

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

Page 2 of 29

Abstract

HO

OH O

5

NMe 2 OH

OH OH O

OH 5

OH O

antibiotics

CONH 2 OH

Seco-cyclines HO

OH

CONH 2

Oxytetracyclines

The commercial

O OH O

NMe 2 OH

OH OH O

O

CONH 2

Tetracyclines HO

OH

HO

tetracycline (3),

OH O

Hemi-cyclines

minocycline

(4),

chlortetracycline (5),

oxytetracycline (6) and doxycycline (7) were biotransformed by a marine–derived fungus Paecilomyces sp. to yield seco-cyclines A–H (9–14, 18 and 19) and hemi-cyclines A–E (20–24). Structures were assigned by detailed spectroscopic analysis, and in the case of 10 X-ray crystallography. Parallel mechanisms account for substrate-product specificity, where 3–5 yield seco-cyclines and 6 and 7 yield hemi-cyclines. The susceptibility of 3–7 to fungal biotransformation is indicative of an unexpected potential for tetracycline “degradation” (i.e. antibiotic resistance) in fungal genomes. Significantly, the fungal-derived tetracycline-like viridicatumtoxins are resistant to fungal biotransformation, providing chemical insights that could inform the development of new tetracycline antibiotics resistant to enzymatic degradation.

ACS Paragon Plus Environment

2

Page 3 of 29

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

The Journal of Organic Chemistry

Introduction We recently reported on a rare class of tetracycline-like polyketide (i.e. viridicatumtoxin A (1)) produced by an Australian marine mollusc-associated fungus, Paecilomyces sp. (CMB-MF010).1 In the course of those studies we noted that the minor 5-oxo co-metabolite, viridicatumtoxin B (2), was exceptionally potent against vancomycin-resistant Enterococci (VRE) (MIC 40 nM). Unfortunately, further in vivo antibiotic investigations stalled due to low fermentation yields. While an elegant total synthesis of (±)-2 had been reported,2,3 we elected to investigate the antiVRE properties of the viridicatumtoxin pharmacophore by transforming commercially available tetracycline (3), minocycline (4), chlortetracycline (5), oxytetracycline (6) and doxycycline (7) (Figure 1). More specifically, we hypothesized that 5-oxo-tetracyclines may exhibit enhanced anti-VRE properties, akin to those observed in the structure activity relationship between 1 and 2.

Figure 1. Examples of known viridicatumtoxins and tetracyclines

ACS Paragon Plus Environment

3

The Journal of Organic Chemistry

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

Page 4 of 29

While the chemical oxidation of 7 to the 5-oxo-doxycycline (8) had been noted (albeit not well documented),4 in our hands this transformation was not reproducible. Drawing inspiration from the ability of CMB-MF010 to biosynthesize the co-metabolites 1 and 2, we set out to investigate whether CMB-MF010 could mediate the oxidative biotransformation of 3–7 to yield 5-oxo-tetracyclines, as potential antibiotic surrogates of viridicatumtoxin B (2). Curiously, despite reports on the fungal biotransformation of fluoroquinolone antibiotics (e.g. flumequine,5 danofloxacin,6 norfloxacin7 and ciprofloxacin7), to the best of our knowledge there are no accounts of the biotransformation of tetracyclines by a defined filamentous fungus. This is a surprising knowledge gap, especially as tetracyclines are well known contaminants in urban, rural and industrial waste, where they can drive antibiotic resistance in pathogens, and potentially compromise microbiological processes that underpin waste management. These same environments are rich in fungal biodiversity.

Results and Discussion To develop methodology for the biotransformation of tetracyclines by Paecilomyces sp. (CMBMF010), we carried out preliminary analytical scale 30-day PYG agar and broth cultivations. As our commercial sample of tetracycline (3) was contaminated with ~10% 4-epi-tetracycline (3a), likely generated by autocatalytic acid mediated keto-enol tautomerism8 (Figure 2), biotransformation trials were performed on this mixture. Of note, conversion of 3 to 3a was enhanced 3–fold to ~30% when cultivated in media without inoculation (negative control), consistent with an autocatalytic interconversion. CMB-MF010 inoculated PYG agar cultivations of 3/3a revealed quantitative conversion to 9–11, while PYG broth cultivations exhibited modest conversion to 9 and 12. By contrast, PYG agar and broth cultivations on oxytetracycline (6)

ACS Paragon Plus Environment

4

Page 5 of 29

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

The Journal of Organic Chemistry

resulted in quantitative conversion to a single major product 20, and trace levels of 21. Based on these trials, multiple preparative-scale biotransformation studies were performed using 30-day PYG single agar plate cultivations of CMB-MF010, each loaded with the HCl salt of a commercial tetracycline antibiotic (40 mg). All biotransformation products were isolated by HPLC and the structures were assigned by detailed spectroscopic analysis.

Figure 2. Acid-mediated equilibration of 3 and 3a

A 30-day agar plate cultivation of CMB-MF010 treated with 3:3a (9:1) afforded >95% biotransformation to seco-cyclines A–C (9–11) (Figure S1). HRESI(+)MS analysis of 9 returned a quasi-molecular ion consistent with a molecular formula (C19H17NO8, ∆mmu –0.1) having lost the elements of CH and N(CH3)2. Comparison of 1D NMR (DMSO-d6) data (Tables 1–2) for 9 with 3 suggested conservation of rings C and D, loss of the ring A 4-N(CH3)2 with oxidation to a fully substituted p-quinone, and opening of ring B across C-11a to C-12a with associated loss of C-12. Diagnostic 2D NMR correlations (Figure 3) supported the assignment of the structure for seco-cycline A (9) as shown, with retention of the C-5a and C-6 configuration (evident in 3) validated by comparison to the co-biotransformation product 10.

ACS Paragon Plus Environment

5

The Journal of Organic Chemistry

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

Page 6 of 29

Table 1. 1H NMR (600 MHz, DMSO-d6) data for seco-cyclines A–D (9–12) Pos. 5 5a 7 8 9 11a 6-CH3 1-NH2 2-CONH2 10-OH a-c

seco-cycline A (9) α 2.76 (m)a β 2.17 (dd, 13.0, 12.1) 2.36 (m) 7.20 (dd, 7.8, 0.9) 7.57 (dd, 8.3, 7.8) 6.83 (dd, 8.3, 0.9) α 2.76 (m)a β 2.45 (dd, 18.8, 4.5) 1.54 (s) − − α 9.43 (br s) β 9.13 (br s) 12.6 (s)

seco-cycline B (10) α 2.75 (dd, 12.9, 3.6) β 2.14 (dd, 12.9, 11.7) 2.36 (m) 7.20 (d, 7.7) 7.56 (dd, 8.3, 7.7) 6.83 (d, 8.3) α 2.78 (dd, 18.5, 5.0) β 2.45 (dd, 18.5, 4.4) 1.54 (s) α 10.80 (d, 4.5) β 8.91 (d, 4.5) α 8.49 (d, 3.4) β 7.39 (d, 3.4) 12.6 (s)

seco-cycline C (11) α 2.64 (dd, 17.8, 6.0)b β 1.85 (dd, 17.8, 8.3)c 2.62 (m)b 6.99 (d, 6.8) 7.60 (dd, 8.2, 6.8) 6.94 (d, 8.2) α 3.27 (br d, 18.6) β 2.82 (dd, 18.6, 3.0) 1.81 (s)c

seco-cycline D (12) α 2.76 (dd, 13.2, 4.3) β 2.51 (dd, 13.2, 11.6) 3.17 (dd, 11.6, 4.3) 7.21 (d, 7.6) 7.66 (dd, 8.6, 7.6) 7.09 (d, 8.6) 3.50 (br s) − 1.93 (s)

− − α 9.46 (br s) β 9.24 (br s) 12.4 (s)

− − α 9.43 (br s) β 9.16 (br s) 11.4 (s)

Overlapping signals with assignments supported by HSQC correlations.

Table 2. 13C NMR (150 MHz, DMSO-d6) data for seco-cyclines A–D (9–12) Pos. 1 2 3 4 4a 5 5a 6 6a 7 8 9 10 10a 11 11a 12 12a 2-CONH2 6-CH3

9 nd 98.0 nd nd 118.3 22.0 43.3 71.0 150.8 116.4 137.0 115.6 161.3 114.6 203.9 39.0 − nd 172.1 30.2

10 155.6 95.8 177.2 158.9 114.8 21.9 43.3 71.1 150.8 116.4 137.0 115.6 161.3 114.6 203.9 39.1 − 180.0 169.3 30.3

11 175.0 98.7 178.7 153.9 117.3 22.0 34.1 79.3 145.8 115.5 138.0 117.5 161.8 114.0 202.3 40.3 − 179.3 172.4 27.9

12 nd 98.2 177.9 nd 115.1 21.5 52.5 86.7 144.4 115.5 137.8 119.7 162.0 112.6 194.7 62.1 170.6 nd 172.1 16.8

“nd” Undetected carbon signals due to tautomerization.

ACS Paragon Plus Environment

6

Page 7 of 29

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

The Journal of Organic Chemistry

Figure 3. Structures and diagnostic 2D NMR correlations for 9–11

HRESI(+)MS analysis of 10 revealed a quasi-molecular ion consistent with a molecular formula (C19H18N2O7, ∆mmu –0.2) for an NH2 substitution analogue of 9. Comparison of 1D NMR (DMSO-d6) data (Tables 1–2) for 10 with 9 revealed the only significant difference as the appearance of an amine moiety (δH 10.80 and 8.91), which diagnostic 2D NMR correlations positioned at C-1 (Figure 3). These considerations, together with X-ray crystallographic analysis (Figure 4), supported the structure for seco-cycline B (10) as shown. Significantly, as noted above for 9, the C-5a and C-6 configuration in 10 was retained from 3.

ACS Paragon Plus Environment

7

The Journal of Organic Chemistry

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

Page 8 of 29

Figure 4. ORTEP structure generated from the single crystal X-ray diffraction of 10 showing 30% probability ellipsoids (Note: different numbering systems are used for the structures in the text)

HRESI(+)MS analysis of 11 exhibited a quasi-molecular ion consistent with a molecular formula (C19H15NO7, ∆mmu +0.6) for a dehydrated analogue of 9. Comparison of the 1D NMR (DMSO-d6) data (Tables 1–2) for 11 with 9 revealed a downfield shift of C-6 (∆δC +8.3) and an upfield shift for C-5a (∆δC –9.2), C-6a (∆δC –5.0) and 6-CH3 (∆δC –2.3), consistent with the changes to the C-6 substitution. Diagnostic 2D NMR correlations (Figure 3), together with a facile acid-mediated conversion of 9 to 11, supported assignment of the structure for secocycline C (11) as shown, with the C-5a and C-6 absolute configuration retained from 3. Exposure of 3:3a (9:1) to a shorter duration 15-day agar plate cultivation of CMBMF010 yielded the additional minor biotransformation product seco-cycline D (12) (Figure S1). HRESI(+)MS analysis of 12 returned a quasi-molecular ion consistent with a molecular formula (C20H15NO9, ∆mmu +0.7) for a +CO2 homologue of 11. Comparison of the 1D NMR (DMSO-

ACS Paragon Plus Environment

8

Page 9 of 29

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

The Journal of Organic Chemistry

d6) data (Tables 1–2) for 12 with 11 revealed conversion of the C-11a methylene to a methine, and the appearance of a new ester/lactone carbonyl resonance. These considerations, together with diagnostic 2D NMR correlations (Figure 5), supported the structure assigned to secocycline D (12) as shown, with the C-5a and C-6 configuration retained from 3, and the C-11a configuration defined by a stereospecific β-facial lactonisation. Of note, 12 was cited (but not fully characterised) in two 1969 patents documenting the biotransformation of 3 by plant and fungi peroxidases, and by oxidation with m-chloroperbenzoic acid.9,10

Figure 5. Structure and diagnostic 2D NMR correlations for 12

Minocycline (4) underwent quantitative biotransformation when exposed to a 30-day agar plate cultivation of CMB-MF010, yielding seco-cyclines E–F (13 and 14) (Figure S2). HRESI(+)MS analysis of 13 and 14 revealed quasi-molecular ions corresponding to molecular formula C20H20N2O7 (∆mmu –0.4) and C20H21N3O6 (∆mmu –0.4) respectively. Analysis of 1D NMR (DMSO-d6) data (Table 3) suggested the biotransformation products akin to 9 and 10 derived from 3:3a. These considerations, together with diagnostic 2D NMR correlations (Figure 6), supported assignment of structures for seco-cyclines E (13) and F (14) as shown, with retention of the C-5a configuration from 4. Of note, a 2010 report described the Ag2CO3/EDTA (66% yield) and Hg(OAc)2 (88% yield) preparation of 4,11a-bridged derivatives of 4, which

ACS Paragon Plus Environment

9

The Journal of Organic Chemistry

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

Page 10 of 29

were readily converted to 13 (55% yield) via decarboxylative rearrangement in acidic aqueous phosphate buffer (pH 6.4) at 35 °C.11

Table 3. 1H and 13C NMR (600 MHz, DMSO-d6) data for seco-cyclines E and F (13 and 14) Pos. 1 2 3 4 4a 5 5a 6 6a 7 8 9 10 10a 11 11a 12a 2-CONH2 7-N(CH3)2 1-NH2 2-CONH2 10-OH a-f

seco-cycline E (13)

seco-cycline F (14)

δH (mult., J (Hz))

δC

δH (mult., J (Hz))

δC

− − − − − 2.50 (m)a,g 2.29 (m) α 3.18 (dd, 17.0, 4.0) β 2.63 (dd, 17.0, 5.1)b − − 7.57 (d, 8.8) 6.84 (d, 8.8) − − − α 2.64 (m)b β 2.53 (m)a,g − − 2.74 (br s)

nd 98.0 177.8 nd 117.3 28.1 33.2 30.5

155.5 95.9 177.2 158.7f 114.0 28.0 33.3 30.6

− − α 9.49 (br s) β 9.15 (br s) 12.36 (br s)

− − − − −

− − − − − 2.48 (dd, 7.0, 3.8)g 2.28 (m) α 3.17 (dd, 16.8, 3.7) β 2.61 (m)d − − 7.49 (br s) 6.81 (d, 8.9) − − − α 2.63 (m)d β 2.53 (m)g − − 2.67 (br s)d α 10.84 (d, 5.5) β 8.94 (d, 5.5) α 8.50 (d, 3.8) β 7.43 (d, 3.8) 12.30 (br s)

138.4c 138.4c 128.6 115.5 159.1 116.3 205.7 43.7 nd 172.1 45.0

Overlapping signals with assignments supported by HSQC and HMBC correlations; carbon signals due to tautomerization.

g

138.5e 138.5e 128.6 115.2 158.7f 116.4 205.8 43.9 180.2 169.3 44.8 − − − − −

Overlap with residual DMSO signal; “nd” Undetected

ACS Paragon Plus Environment

10

Page 11 of 29

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

The Journal of Organic Chemistry

Figure 6. Structures and diagnostic 2D NMR correlations for 13 and 14

Our commercial sample of 5 was contaminated with three isomers tentatively identified on the basis of HRESI(+)MS as 4-epi-chlortetracycline (15) (C22H23ClN2O8, ∆mmu +0.3), isochlortetracycline (16) (C22H23ClN2O8, ∆mmu –0.3) and 4-epi-iso-chlortetracycline (17) (C22H23ClN2O8, ∆mmu –0.4) (Figure 7).12,13 The acid-mediated epimerization of 5 to 15 likely proceeds as illustrated above for 3:3a (Figure 2), with the ring D Cl substituent further activating the C-11 carbonyl to intra-molecular lactonization and cleavage of C-11 to C-11a to yield 16 and 17 (Figure 8). Of note, the conversion of 5 to 15–17 was enhanced during cultivation without CMB-MF010 inoculation (negative control) (Figure S3). Notwithstanding chemical stability considerations, CMB-MF010 did biotransform 5 to seco-cyclines A (9) and B (10) (Figure S3), which were identified by HRESIMS and HPLC co-elution with authentic standards (prepared from 3). Furthermore, on the basis of HRESIMS and mechanistic considerations (Figure 11), we detected the presence of chloro analogues tentatively identified as seco-cyclines G (18) (C19H16ClNO8, ∆mmu +0.1) and H (19) (C19H17ClN2O7, ∆mmu +0.1) (Figure 7, Table S1).

ACS Paragon Plus Environment

11

The Journal of Organic Chemistry

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

Page 12 of 29

Figure 7. Degradation products 15–17, and CMB-MF010 biotransformation products 18 and 19, derived from 5

Figure 8. Mechanism for degradation of 5 to 15–17

Oxytetracycline (6) underwent quantitative biotransformation when exposed to 30-day agar plate cultivations of CMB-MF010, yielding hemi-cyclines A–B (20 and 21) (Figure S4). HRESI(−)MS analysis of 20 returned a quasi-molecular ion consistent with a molecular formula (C12H14O4, ∆mmu −0.5) suggestive of a substituted ring C/D tetralone. Analysis of the NMR

ACS Paragon Plus Environment

12

Page 13 of 29

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

The Journal of Organic Chemistry

(CD3OD) data (Table 4), and in particular diagnostic 2D NMR correlations (Figure 9), supported assignment of the structure for hemi-cycline A (20) as shown, with the C-6 configuration retained from 6. The C-5a configuration in 20 was assigned by comparison of J11aα,5a and J11aβ,5a measurements with 20a (Figure 9), a C-5a epimer produced by Streptomyces rimosus (R1059), a mutant strain used in the commercial production of 6.14 HRESI(+)MS analysis of 21 returned a quasi-molecular ion consistent with a molecular formula (C14H16O5, ∆mmu +0.1) for an acetate derivative of 20. Comparison of the 1D NMR (CD3OD) data (Table 4) for 21 with 20 revealed the only significant differences as deshielding of the C-5 methylene, and the appearance of resonances for a C-5 acetate moiety (δC 20.5, OCOCH3; δC 172.7, OCOCH3). These considerations, together with diagnostic 2D NMR correlations, and J11aα,5a and J11aβ,5a measurements, permitted assignment of the structure for hemi-cycline B (21) as shown (Figure 9).

Figure 9. Structures and diagnostic NMR data for 20, 21 and 20a

ACS Paragon Plus Environment

13

The Journal of Organic Chemistry

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

Page 14 of 29

Table 4. 1H and 13C NMR (600 MHz, CD3OD) data for hemi-cyclines A and B (20 and 21)

Pos. 5 5a 6 6a 7 8 9 10 10a 11 11a 6-CH3 5-OCOCH3 5-OCOCH3

hemi-cycline A (20)

hemi-cycline B (21)

δH (mult., J (Hz))

δC

δH (mult., J (Hz))

δC

α 3.89 (dd, 11.1, 5.8) β 3.45 (dd, 11.1, 8.0) 2.38 (m) − − 7.18 (dd, 7.7, 1.0) 7.52 (dd, 8.4, 7.7) 6.82 (dd, 8.4, 1.0) − − − α 2.96 (dd, 18.6, 5.4) β 2.89 (dd, 18.6, 4.3) 1.62 (s) − −

62.9

α 4.40 (dd, 11.4, 4.2) β 4.08 (dd, 11.4, 7.9) 2.54 (m) − − 7.17 (dd, 7.7, 1.0) 7.52 (dd, 8.4, 7.7) 6.83 (dd, 8.4, 1.0) − − − α 3.03 (dd, 18.4, 5.4) β 2.84 (dd, 18.4, 4.0) 1.64 (s) − 1.88 (s)

65.0

48.1 72.8 151.2 117.1 138.2 117.3 163.3 116.1 205.0 39.1 31.3 − −

46.3 71.6 151.1 117.0 138.2 117.3 163.2 116.3 204.5 39.4 31.4 172.7 20.5

Doxycycline (7) underwent ~90% biotransformation when exposed to 30-day agar plate cultivations of CMB-MF010, yielding hemi-cyclines C–E (22–24) (Figure S5). HRESI(+)MS analysis of 22 returned a quasi-molecular ion consistent with a molecular formula (C12H14O3, ∆mmu –0.5) for a deoxy analogue of 20. Analysis of NMR (CD3OD) data (Table 5), and in particular diagnostic 2D NMR correlations, and J11aα,5a and J11aβ,5a measurements, permitted assignment of the structure for hemi-cycline C (22) as shown (Figure 10), with the C-6 configuration retained from 7. HRESI(+)MS analysis of 23 returned a quasi-molecular ion consistent with a molecular formula (C12H12O4, ∆mmu –0.7) for an oxidized analogue of 22. Comparison of the 1D NMR (CD3OD) data (Table 5) for 23 with 22 revealed the absence of resonances for a C-5 methylene and appearance of resonances for a C-5 carboxylic acid (δC 177.0). These considerations, together with diagnostic HMBC correlations, and J11aα,5a and

ACS Paragon Plus Environment

14

Page 15 of 29

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

The Journal of Organic Chemistry

J11aβ,5a measurements, permitted assignment of the structure for hemi-cycline D (23) as shown (Figure 10), with the C-6 configuration retained from 7. HRESI(+)MS analysis of 24 returned a quasi-molecular ion consistent with a molecular formula (C12H12O5, ∆mmu +0.4) for an oxidized analogue of 23. Comparison of the 1D NMR (CD3OD) data (Table 5) for 24 with 23 revealed the absence of resonances for an aromatic methine, and the appearance of resonances for an oxygenated aromatic quaternary carbon (δC 145.3). These considerations, together with diagnostic 2D NMR correlations, and J11aα,5a and J11aβ,5a measurements, permitted assignment of the structure for hemi-cycline E (24) as shown (Figure 10), with the C-6 configuration retained from 7.

Figure 10. Structures and diagnostic NMR correlations for 22–24

ACS Paragon Plus Environment

15

The Journal of Organic Chemistry

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

Page 16 of 29

Table 5. 1H and 13C NMR (600 MHz, CD3OD) data for hemi-cyclines C–E (22–24) hemi-cycline C (22)

Pos. 5 5a 6 6a 7 8 9 10 10a 11 11a 6-CH3 a,b

hemi-cycline D (23)

δH (mult., J (Hz))

δC

3.50 (d, 6.8) 2.16 (m) 3.10 (qd, 7.2, 3.9) − 6.83 (dd, 7.5, 1.0) 7.44 (dd, 8.4, 7.5) 6.74 (dd, 8.4, 1.0) − − − α 2.99 (dd, 18.1, 5.3) β 2.65 (dd, 18.1, 4.6) 1.41 (d, 7.2)

64.8 43.6 35.8 150.3 120.3 138.1 116.3 163.7 116.8 205.6 37.9 21.9

hemi-cycline E (24)

δH (mult., J (Hz))

δC

δH (mult., J (Hz))

δC

− 3.04 (m)a 3.51 (qd, 7.2, 3.5) − 6.81 (dd, 7.5, 1.0) 7.43 (dd, 8.4, 7.5) 6.75 (dd, 8.4, 1.0) − − − α 3.01 (dd, 18.1, 4.9)a β 2.91 (dd, 18.1, 3.8) 1.44 (d, 7.2)

177.0 46.7 37.2 148.9 119.8 138.0 116.7 163.7 116.6 204.4 37.1

− 2.99 (m)b 3.42 (m) − 6.67 (dd, 8.1, 0.8) 6.99 (d, 8.1) − − − − α 2.97 (dd, 18.1, 4.8)b β 2.88 (dd, 18.1, 3.9) 1.41 (d, 7.2)

177.2 47.3 36.8 138.6 119.2 123.0 145.3 151.7 116.7 205.1 37.3

22.3

22.3

Overlapping signals with assignments supported by HSQC correlations.

The results from our studies demonstrate that 3–5 share a common mechanism of fungal biotransformation, leading to seco-cyclines A–H (9–14, 18 and 19), whereas 6 and 7 share an alternate mechanism of fungal biotransformation, leading to hemi-cyclines A–E (20–24) – drawing attention to a substrate-product specificity that correlates with C-5 hydroxylation. For example, as illustrated for 3:3a, we propose a mechanism initiated by enol-keto tautomerism to deliver a C-12 carbonyl, which undergoes enzymatic nucleophilic addition/oxidation and subsequent opening of ring B across the C-12 to C-12a bond (Figure 11). Importantly, deamination followed by oxidation assembles the quinone ring A common to seco-cyclines A (9) and D (12), with regioselective addition of NH3 via the doubly activated tetramic acid C-1, with concomitant loss of H2O, leading to the amino quinone ring A common to seco-cycline B (10). Two alternate routes for lactonization deliver seco-cyclines C (11) and D (12), with configurations dictated by pre-existing stereochemistry about C-6a in 11, and C-6 in 12. Following this same mechanism, minocycline (4) can yield seco-cyclines E (13) and F (14),

ACS Paragon Plus Environment

16

Page 17 of 29

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

The Journal of Organic Chemistry

while chlortetracycline (5) can yield seco-cyclines G (18) and H (19). That 5 can also yield 9 and 10 is attributed to partial de-chlorination to 3. By contrast, we propose that fungal biotransformation of the C-5 hydroxylated antibiotics oxytetracycline (6) and doxycycline (7) proceeds by a closely related but different mechanism. For example, as illustrated for 6 (Figure 12), we propose a mechanism initiated by the same enol-keto tautomerism, with enzymatic nucleophilic addition/oxidation and opening of ring B about the C-12 to C-12a bond. However, rather than de-amination (Figure 11), we hypothesize that ring A aromatization is achieved by an alternate elimination, with the C-5 hydroxy moiety influencing cleavage of the C-4a to C-5 bond to yield hemi-cyclines A (20) and B (21) (Figure 12). Following this same mechanism, doxycycline (7) can yield hemi-cycline C (22), with further oxidations leading to hemi-cyclines D (23) and E (24). These two highly conserved mechanisms (Figures 11 and 12) account for all the biotransformation products encountered during our investigations, and suggest a basis for predicting fungal biotransformation products for other members of the tetracycline family of antibiotics.

ACS Paragon Plus Environment

17

The Journal of Organic Chemistry

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

Page 18 of 29

Figure 11. Mechanism for the CMB-MF010 biotransformation of 3:3a to 9–12

Figure 12. Mechanism for the CMB-MF010 biotransformation of 6 to 20 and 21

ACS Paragon Plus Environment

18

Page 19 of 29

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

The Journal of Organic Chemistry

Although CMB-MF010 successfully biotransformed 3–7, none of these products matched our target 5-oxo scaffold. We therefore elected to investigate nine other filamentous fungi and three yeast isolates (see Experimental section). Intriguingly, all strains tested transformed 6 to 20, with high conversion rates (>70%, Figure S7). By contrast, 3 was far more resilient, with biotransformation limited to Fusarium sp. (CMB-MF017) with >95% conversion to 9–12, and Gliocladium roseum (ACM-4596) and Candida guillermondii (ACM-4553), with