A two-step sulfation in antibiotic biosynthesis requires a type III polyketide synthase

22
Accepted Manuscript Gust Lab 1 A two-step sulfation in antibiotic biosynthesis requires a type III polyketide synthase Xiaoyu Tang 1 , Kornelia Eitel 1 , Leonard Kaysser 1 , Andreas Kulik 2 , Stephanie Grond 3 & Bertolt Gust 1 * 1 Pharmaceutical Institute, University of Tübingen, Tübingen, Germany. 2 Institute of Microbiology and Infection Medicine, University of Tübingen, Tübingen, Germany. 3 Institute of Organic Chemistry, University of Tübingen, Tübingen, Germany. *e-mail: [email protected] Caprazamycins (CPZs) belong to a group of liponucleoside antibiotics inhibiting the bacterial MraY translocase, an essential enzyme involved in peptidoglycan biosynthesis. We have recently identified analogs that are decorated with a sulfate group at the 2-hydroxy of the aminoribosyl moiety, and we now report an unprecedented two-step sulfation mechanism during the biosynthesis of CPZs. A type III polyketide synthase (PKS) known as Cpz6 is employed in the biosynthesis of a group of new triketide pyrones that are subsequently sulfated by an unusual 3-phosphoadenosine-5-phosphosulfate (PAPS)-dependent sulfotransferase (Cpz8) to yield phenolic sulfate esters, which serve as sulfate donors for a PAPS-independent arylsulfate sulfotransferase (Cpz4) to generate sulfated CPZs. This finding is to our knowledge the first demonstration of genuine sulfate donors for an arylsulfate sulfotransferase and the first report of a type III PKS to generate a chemical reagent in bacterial sulfate metabolism. Sulfation of biomolecules is a vital process in all living organisms and is involved in detoxification, hormone regulation and drug metabolism. Sulfotransferases catalyze the transfer of a sulfate from the universal active donor molecule PAPS to proteins, sugars, antibiotics and a variety of other low-molecular-weight metabolites 1 . In addition to the extensively studied PAPS-dependent sulfotransferases, arylsulfate sulfotransferases (ASSTs) represent a group of PAPS-independent sulfotransferases that catalyze sulfotransfer from phenolic sulfate esters to another phenolic molecule 2 . Although ASSTs have been investigated for almost three decades, the genuine sulfate donor remains unknown 3 .

Transcript of A two-step sulfation in antibiotic biosynthesis requires a type III polyketide synthase

Accepted Manuscript Gust Lab

1

A two-step sulfation in antibiotic biosynthesis requires a type

III polyketide synthase

Xiaoyu Tang1, Kornelia Eitel1, Leonard Kaysser1, Andreas Kulik2, Stephanie Grond3 & Bertolt Gust1*

1Pharmaceutical Institute, University of Tübingen, Tübingen, Germany. 2Institute of Microbiology and Infection

Medicine, University of Tübingen, Tübingen, Germany. 3Institute of Organic Chemistry, University of Tübingen,

Tübingen, Germany.

*e-mail: [email protected]

Caprazamycins (CPZs) belong to a group of liponucleoside antibiotics inhibiting the bacterial

MraY translocase, an essential enzyme involved in peptidoglycan biosynthesis. We have recently

identified analogs that are decorated with a sulfate group at the 2-hydroxy of the aminoribosyl

moiety, and we now report an unprecedented two-step sulfation mechanism during the

biosynthesis of CPZs. A type III polyketide synthase (PKS) known as Cpz6 is employed in the

biosynthesis of a group of new triketide pyrones that are subsequently sulfated by an unusual

3-phosphoadenosine-5-phosphosulfate (PAPS)-dependent sulfotransferase (Cpz8) to yield

phenolic sulfate esters, which serve as sulfate donors for a PAPS-independent arylsulfate

sulfotransferase (Cpz4) to generate sulfated CPZs. This finding is to our knowledge the first

demonstration of genuine sulfate donors for an arylsulfate sulfotransferase and the first report of

a type III PKS to generate a chemical reagent in bacterial sulfate metabolism.

Sulfation of biomolecules is a vital process in all living organisms and is involved in detoxification,

hormone regulation and drug metabolism. Sulfotransferases catalyze the transfer of a sulfate from the

universal active donor molecule PAPS to proteins, sugars, antibiotics and a variety of other

low-molecular-weight metabolites1. In addition to the extensively studied PAPS-dependent

sulfotransferases, arylsulfate sulfotransferases (ASSTs) represent a group of PAPS-independent

sulfotransferases that catalyze sulfotransfer from phenolic sulfate esters to another phenolic molecule2.

Although ASSTs have been investigated for almost three decades, the genuine sulfate donor remains

unknown3.

Accepted Manuscript Gust Lab

2

Recently, we identified the biosynthetic gene cluster for the CPZs in Streptomyces sp.

MK730–62F2 (ref. 4) (cpz9-cpz31; Fig. 1). These compounds belong to a family of liponucleoside

antibiotics with potent activity against Mycobacterium tuberculosis5. Subsequently, we discovered a set of

2-O-sulfated CPZ analogs6 and showed that Cpz4, an ASST that is encoded ~4 kb upstream of the CPZ

cluster (Fig. 1), is responsible for the sulfation reaction7. In vitro characterization revealed that Cpz4

transfers a sulfate from a synthetic phenolic sulfate ester to the 2-hydroxy of the O-aminoribosyl moiety

of CPZs. This finding prompted us to investigate the sulfation mechanism in more detail with the prospect

to identify a genuine sulfate donor for an ASST. We concentrated on genes located between cpz4 and the

start (cpz9) of the CPZ cluster, which we previously determined not to be required for liponucleoside

formation4,6,7

. This region encodes a putative 3-hydroxy-3-methylglutaryl-CoA4 synthase (cpz5), a type III

PKS (cpz6), a 3 -phosphoadenosine-5 -phosphatase (cpz7) and a conserved hypothetical protein

(cpz8). Our curiosity was piqued especially by the presence of cpz6 because type III PKSs have been

shown to produce small phenolic compounds similar to what we expect of an ASST substrate. Moreover,

the cpz4-cpz8 region was found to be conserved in the biosynthetic cluster of liposidomycin, a sulfated

CPZ aglycone isolated from Streptomyces sp. SN-1061M6.

In this study, we used bioinformatic, genetic, biochemical and analytical methods to elucidate the

two-step sulfation mechanism during CPZ biosynthesis. A structurally new group of triketide 3,6-alkylated

pyrones are produced by the type III PKS Cpz6 and subsequently sulfated by the unusual

PAPS-dependent sulfotransferase Cpz8 to generate natural sulfate donors for the ASST Cpz4 for the

generation of sulfated CPZs.

RESULTS

Analysis of deletion mutants

To investigate the role of these genes in the production of sulfated CPZs, we generated a series of

mutants via RED/ET-mediated recombination8. We first individually deleted cpz5, cpz6, cpz7 and cpz8

in-frame on the cosmid cpzLK09 containing the CPZ gene cluster (Fig. 1). We then heterologously

expressed each of the mutated cosmids in Streptomyces coelicolor M512 and examined the extracts of

mutant cultures by HPLC and LC/ESI-MS analyses. The results showed that all of the tested mutants

have the ability to produce CPZ aglycones, indicating that they still contained the complete set of genes

required for CPZ aglycone biosynthesis (Fig. 2 and Supplementary Results, Supplementary Fig. 1).

Accepted Manuscript Gust Lab

3

We recently obtained identical results when we deleted the genes cpz1, cpz2 and cp3 as well as cpz1,

cpz2, cpz3 and cpz4 (ref. 7). However, in this case, the mutant that specifically lacked the ASST Cpz4 lost

the ability to produce the sulfated CPZ derivatives (Fig. 2a,b and Supplementary Fig. 1a,b). Notably,

deletion of cpz8 also resulted in a total abolishment of sulfated CPZ aglycone production (Fig. 2c and

Supplementary Fig. 1c). However, mutants lacking cpz7 or cpz6 still produced sulfated CPZ aglycones

(Fig. 2d,e and Supplementary Fig. 1d,e) but did so at reduced amounts ranging from 6.9% to 8.3% of

the total amount of CPZ aglycones, whereas strains containing the entire gene cluster (S. coelicolor

M512/cpzWP04( cpz1-3)) produced 20.5% of the total (Fig. 2). In addition, we observed production of

sulfated CPZ aglycones in all of the tested mutants of S. coelicolor M152/cpzWP08( cpz5) at the same

level (20.7%) as S. coelicolor M512/cpzWP04( cpz1-3) strains, suggesting that cpz5 is not involved in

the CPZ sulfation process (Fig. 2f and Supplementary Fig. 1f).

Cpz8 as a PAPS-dependent sulfotransferase

Analysis of the cpz8 mutant strain revealed that Cpz8 has an essential role in the sulfation mechanism. To

learn more about Cpz8, we subjected the enzyme to a secondary structure prediction-based homology

search (The Phyre Webserver)9 and found low similarity (19.6%) to a PAPS-dependent heparan sulfate

glucosamine 3-O-sulfotransferase 10

. To explore whether Cpz8 can indeed act as an sulfotransferase, we

cloned and expressed the gene for Cpz8 in Escherichia coli, which yielded soluble protein that was

purified in significantly high yields (120 mg l–1

) (Supplementary Fig. 2). We then added the purified

protein to a reaction mixture containing PAPS as sulfate donor and p-nitrophenol (pNP) as an acceptor.

Formation of p-nitrophenyl sulfate was detected by HPLC in comparison with synthetic standard and was

confirmed by LC-ESI-MS/MS (Fig. 3 and Supplementary Fig. 3a). Kinetic analysis of Cpz8 revealed

Michaelis-Menten kinetics yielding a Km value of 10.9 ± 1.9 M for PAPS and 959.2 ± 37.4 M for pNP, in

line with other PAPS-dependent sulfotransferases (Supplementary Fig. 3b). Subsequently, we

demonstrated that a variety of phenolic compounds were also accepted by Cpz8 as substrates such as

4-methylumbelliferone, 4-hydroxy-6-methyl-2-pyrone and 2-naphtol (Supplementary Fig. 3c–e) but not

3-hydroxy-2-methyl-4-pyrone, phloroglucinol or resorcinol (Supplementary Fig. 3f). Therefore, the

genetic and biochemical data strongly support that Cpz8 indeed acts as a PAPS-dependent

sulfotransferase in the biosynthesis of sulfated CPZs. Notably, we also found numerous Cpz8 homologs

in bacterial and fungal genomes, and almost all of them are annotated as hypothetical proteins

Accepted Manuscript Gust Lab

4

(Supplementary Fig. 4). Moreover, we also found several cpz8 homologs are located together with

genes probably involved in sulfate metabolism. For example, the genes encoding XOO3582 from

Xanthomonas oryzae, XALc_2793 from Xanthomonas albilineans and N47_P17040 from an uncultured

Desulfobacterium sp. are located next to the genes encoding for the sulfotransferase XOO3581, the

sulfotransferase XALc_2792 and the adenylyl-sulfate kinase N47_P17040, respectively. The fact that

these enzymes do not contain the conserved 5 -phosphosulfate-binding loop is unexpected as this is a

quintessential feature of all known PAPS-dependent sulfotransferases11

(Supplementary Fig. 5).

Cpz7 as a 3-phosphoadenosine-5-phosphatase

Analysis of the mutant lacking cpz7 revealed a decrease of a factor of three in the production of sulfated

CPZ aglycones in comparison to the total amount of CPZ aglycones produced by S. coelicolor

M512/CPZWP04( cpz1-3) (Fig. 2). cpz7 showed high similarity to rv2131c, which encodes a

3-phosphoadenosine-5-phosphatase from Mycobacterium tuberculosis that was demonstrated to

regulate the mycobacterial PAPS-level 12,13

. Our data is in agreement with data obtained from a rv2131c

gene inactivation experiment showing that production of sulfated glycolipids13

was decreased and

providing evidence of its role in the sulfate assimilation pathway12

. To further investigate the function of

Cpz7 in CPZ biosynthesis, we incubated the purified N-terminally His8-tagged Cpz7 (Supplementary Fig.

6a) with 3-phosphoadenosine-5-phosphate (PAP), yielding a new peak that corresponded to the AMP

standard (Supplementary Fig. 6b). Both the analysis of the knockout mutant and our in vitro assays

indicated that Cpz7 acts as a 3-phosphoadenosine-5-phosphatase, converting PAP to AMP and thereby

modulating the intracellular amount of PAP. PAP accumulation has been demonstrated to negatively

regulate activity of important bacterial enzymes such as oligoRNases, exonucleases,

phosphopanteheinyltransferases14

and PAPS-dependent sulfotransferases15

.

Cpz6 is essential for sulfation of CPZs

Cpz6 has sequence homology with type III PKSs, which are known for their biogenesis of a variety of

natural phenolic compounds. As ASSTs have been characterized to accept phenolic sulfate esters as

substrates, we speculated that the product of the type III PKS Cpz6 is converted by the PAPS-dependent

sulfotransferase Cpz8 to generate the sulfate donor for the ASST Cpz4. However, deletion of cpz6 only

resulted in a production of sulfated CPZ aglycones decreased by a factor of 2.5 when compared with S.

coelicolor M512/cpzWP04(cpz1-3) (Fig. 2). This result suggested that cpz6 might not be essential for CPZ

Accepted Manuscript Gust Lab

5

sulfation. A phylogenetic analysis showed that S. coelicolor encodes a close homolog of Cpz6, the type III

PKS Sco7221, that is responsible for germicidin biosynthesis16

. This led to the assumption that

germicidins might serve as alternative sulfate shuttles in the CPZ sulfation mechanism. To test our

hypothesis, we introduced the cosmid cpzLK09, containing the entire CPZ aglycon gene cluster as

positive control, cpzWP11(cpz8) as negative control and cpzWP09(cpz6), into the germicidin knockout

strain S. coelicolor M145/ sco7221 (ref. 16) and analyzed the extracts. HPLC and LC/ESI-MS data

showed that S. coelicolor M145/ sco7221/cpzLK09 readily produced CPZ aglycons as well as the

sulfated derivatives (Fig. 4a and Supplementary Fig. 7a). As expected, S. coelicolor

M145/ sco7221/cpzWP11(cpz8) lost the ability to produce the sulfated compounds (Fig. 4b and

Supplementary Fig. 7b). Both findings were consistent with the results obtained from the heterologous

producer S. coelicolor 512. Most notably, we also observed that the production of sulfated CPZ aglycones

was completely abolished in S. coelicolor M145/ sco7221/cpzWP09(cpz6) (Fig. 4c and Supplementary

Fig. 7c). This finding strongly supports that cpz6 is essential for the sulfation of CPZs. To further

corroborate our results, we incubated germicidin A in vitro with Cpz8 and PAPS, which resulted in a

product peak corresponding to sulfated germicidin A (Supplementary Fig. 8a). Thus, germicidin A can

indeed be used as alternative substrate by Cpz8. LC-ESI-MS/MS analysis also identified accumulation of

sulfated germicidins in S. coelicolor M512/cpzWP05(cpz1-4) mutants (Supplementary Fig. 8b). To the

best of our knowledge, this is the first time a type III PKS has been demonstrated to be involved in a

sulfation mechanism.

Presulficidins are products of the type III PKS Cpz6

To identify the products generated by Cpz6, we compared the metabolic profile of S. coelicolor

M512/cpzWP09(cpz6) with S. coelicolor M512/cpzWP11(cpz8) by HPLC. We expected that the products

of Cpz6 should be accumulated in a cpz8 mutant. Indeed, the UV chromatogram showed that at least

four peaks were missing in the type III PKS knockout strain (Fig. 5a). Furthermore, S. coelicolor

M512/cpzWP09(cpz6)/pXT19, a S. coelicolor M512/cpzWP09(cpz6) mutant containing an intact copy of

cpz6 under control of the strong constitutive promoter PermE*, not only restored but also increased

production of the four compounds detected (Fig. 5a). Next, these compounds were purified via silica gel

column and semi-preparative HPLC (Supplementary Note). High-resolution MS analysis resulted in

C13H20O3 (m/z 247.130434 [M + Na]+), C14H22O3 (m/z 261.145984 [M + Na]

+), C15H24O3 (m/z 275.162194

Accepted Manuscript Gust Lab

6

[M + Na]+) and C15H24O4 (m/z 291.156915 [M + Na]

+) as the molecular formulas of the compounds with

corresponding masses of 224 Da, 238 Da, 252 Da and 268 Da, respectively. Structural characterization

by extensive one- and two-dimensional NMR experiments (Supplementary Note) identified the

compounds as so-far-undescribed triketide pyrones that were further named presulficidins A–D (1–4) (Fig.

5b). The presulficidins are structurally related to the germicidins and type III PKS products from

Streptomyces griseus17

but form a new group of 3,6-alkylated 4-hydroxypyrones. To verify that the

presulficidins are indeed substrates for Cpz8, we used the compounds in an in vitro assay together with

PAPS. The subsequent LC/MS analysis demonstrated that 1–4 were readily sulfated by Cpz8

(Supplementary Fig. 9).

By inspecting structures 1–3, we could postulate that Cpz6 uses CoA- or ACP-activated iso-acyl

starter units from branched-chain fatty acid metabolism and uses one malonyl- and one methylmalonyl

unit to form the final triketide pyrones via Claisen condensation (Fig. 6a). Compound 4 represents a

derivative of 3 containing a terminal hydroxyl group at the saturated side chain. A metabolite with a mass

consistent with hydroxylated 1 (Supplementary Fig. 10a) could also be detected in extracts of S.

coelicolor M512/CPZWP11( cpz8). A similar hydroxylation in iromycin biosynthesis was found to be

mediated by a cytochrome P450 (CYP) monooxygenase18

. Because no homologous enzyme is encoded

in the CPZ gene cluster, we tested whether S. coelicolor M512 uses an endogenous CYP

monooxygenase for this reaction. The addition of 0.3 g l–1

of the CYP inhibitor ancymidol to S. coelicolor

M512/cpzWP011( cpz8) cultures increased production of 1 and 3 by 5- and 20-fold, respectively

(Supplementary Fig. 10b). In contrast, production of 4 was totally abolished (Supplementary Fig. 10b).

These findings strongly suggest that a CYP monooxygenase produced in S. coelicolor M512 is

responsible for the hydroxylation of the type III PKS products.

Two-step sulfation in CPZ biosynthesis

To prove our hypothesis for a two-step sulfation mechanism during CPZ biosynthesis (Fig. 6a), we

incubated purified Cpz4 and Cpz8 with PAPS, 1 and purified hydroxyacylcaprazol E (5) (a CPZ derivative

lacking the permethylated L-rhamnosyl- and 3-methylglutaryl-moiety4; Fig. 6b and Supplementary Fig.

11). Compound 5 has been demonstrated previously to be readily accepted by the arylsulfate

sulfotransferase Cpz4 (ref. 7). A new peak was detected by HPLC analysis with a retention time of 16.9

min (Fig. 6b–i). LC/MS analysis of the new compound (6) revealed a parent ion at m/z 880.5 [M-H]–,

Accepted Manuscript Gust Lab

7

consistent with the addition of a sulfate (+80 Da) to the liponucleoside substrate (m/z 800.5 [M-H]–)

(Supplementary Figs. 12a and 13a–i). MS/MS analysis of 6 confirmed that the fragmentation patterns

matched the sulfated hydroxyacylcaprazol E (Supplementary Fig. 12b). A trace amount of sulfated 1

was observed by LC/MS when 1, Cpz8 and PAPS were present in the reaction mixture (Supplementary

Fig. 13b), demonstrating that 1 is accepted by Cpz8. Exclusion of one of the components, that is, Cpz4,

Cpz8, PAPS or 1, failed to produce 6 (Fig. 6b and Supplementary Fig. 13a). The same results were

observed when 1 was replaced by 2–4 (Fig. 6b and Supplementary Fig. 13a,b). These findings strongly

support that a two-step sulfation mechanism is at work during the biosynthesis of sulfated CPZs (Fig. 6a).

We postulate that a similar sulfation pathway is also involved in the biosynthesis of other sulfated

liponucleoside antibiotics such as the liposidomycins6 and A-90289 (ref. 19). Moreover, we found that

several bacterial genomes contain a type III PKS, a PAPS-dependent sulfotransferase and an ASST

either located together or distributed within the genome (for example, Saccharopolyspora erythraea and

Conexibacter woesei; Supplementary Fig. 14).

To study the substrate specificity of Cpz8 toward phenolic compounds and presulficidins, we

measured 6 formation with different concentrations of sulfate donors in a two-enzyme assay including

both sulfotransferases. Assays containing 1, 2 or 3 resulted in the formation of 6 at concentrations of 10

M, 5 M and 2.5 M, with highest reaction velocities of 8.3 nM s–1

, 6.9 nM s–1

and 10.7 nM s–1

,

respectively (Supplementary Fig. 15a–c). However, further increased concentrations of the sulfate

donors reduced reaction velocities. The highest reaction velocity of 12.6 nM s–1

was observed with 4 at

100 M (Supplementary Fig. 15d). In contrast, the highest reaction velocities of 10.1 nM s–1

and 9.6 nM

s–1

were obtained for germicidin A and 4-hydroxy-6-methyl-2-pyrone but only at high substrate

concentrations of 1,000 M and 4,000 M, respectively (Supplementary Fig. 15e,f). Reaction velocities

with pNP and methylumbelliferone were lower than 1 nM s–1

, even at a high concentration of 4,000 M

(Supplementary Fig. 15g,h).

DISCUSSION

We have elucidated a two-step sulfation mechanism during CPZ biosynthesis involving the type III PKS

products presulficidins as sulfate shuttles. Type III PKSs are widely found in plants, bacteria and fungi20–24

,

but the physiological roles of their products remain unknown in most cases. Only a few bacterial type III

PKSs compounds have been assigned to specific biological functions, for example, antibiotic building

Accepted Manuscript Gust Lab

8

blocks25

, precursors of bacterial pigments26

, membrane and cell wall components27

and signaling

molecules28

. Our report on a type III PKS product as a chemical reagent in bacterial sulfate metabolism

substantially expands the known functional diversity of those enzymes.

Cpz8 was shown to accept a variety of phenolic compounds as sulfate acceptor substrates. To

prove that presulficidins are the bona fide sulfate shuttle in vitro, we initially tried to obtain kinetic data for

Cpz8. However, we failed to obtain these data for presulficidins. This might have been due to instability of

presulficidins or product inhibition by sulficidins in assays containing Cpz8. We therefore used a

two-enzyme assay for comparison of reaction velocities between presulficidins and germicidin A as well

as other commercially available phenolic compounds. Presulficidins had significantly higher reaction

velocities (greater than tenfold) than artificial substrates such as pNP or 4-MU. Although the 2-pyrone

derivatives germicidin A and 4-hydroxy-6-methyl-2-pyrone showed similar highest velocities as

presulficidins, the data were observed at over tenfold higher substrate concentrations. These data not

only suggest that sulfation of CPZs occurs with a preference for 2-pyrone derivatives as sulfate-delivering

molecules but also support that sulficidins (sulfated 1–3) are the genuine sulfate donors for Cpz4.

Furthermore, analysis of the gene deletion mutants also strongly supported that the Cpz6-produced

triketide pyrones (presulficidins) were the endogenous substrates for Cpz8. To the best of our knowledge,

this is the first report of a genuine sulfate donor for an ASST. Moreover, Cpz8 has sequence similarity to a

large group of hypothetical proteins found in sequence databases from bacterial and fungal genomes. We

thus speculate that Cpz8 represents a family of previously unrecognized sulfotransferases

(Supplementary Fig. 4). Hence, the Cpz8 family of sulfotransferase will make a fascinating subject for

future studies, in particular in regard to their use of their sulfate donor substrate.

In summary, we have depicted a two-step sulfation mechanism during CPZ biosynthesis (Fig.

6a). A type III PKS (Cpz6) is responsible for the formation of a structurally new group of triketide

3,6-alkylated pyrones 1–4 that are subsequently sulfated by an unusual PAPS-dependent

sulfotransferase (Cpz8) to generate the phenolic sulfate esters, sulficidins. Finally, the PAPS-independent

ASST Cpz4 transfers a sulfate from the sulficidins to generate sulfated CPZs. The intracellular amount of

PAP, a possible inhibitor of the PAPS-dependent sulfotransferase Cpz8, is reduced by the PAP

3-phosphatase Cpz7 by cleavage of the phosphate group at the 3 position of PAPS to generate AMP. As

the combination of a type III PKS, a PAPS-dependent sulfotransferase and an ASST exists in other

Accepted Manuscript Gust Lab

9

bacterial strains (Supplementary Fig. 14), it is plausible that this two-step mechanism is not limited to

secondary metabolism.

References

1. Chapman, E., Best, M.D., Hanson, S.R. & Wong, C.H. Sulfotransferases: structure, mechanism,

biological activity, inhibition, and synthetic utility. Angew Chem Int Ed Engl 43, 3526-48 (2004).

2. Malojcic, G. & Glockshuber, R. The PAPS-independent aryl sulfotransferase and the alternative

disulfide bond formation system in pathogenic bacteria. Antioxid Redox Signal 13, 1247-59

(2010).

3. Kobashi, K., Fukaya, Y., Kim, D.H., Akao, T. & Takebe, S. A novel type of aryl sulfotransferase

obtained from an anaerobic bacterium of human intestine. Arch Biochem Biophys 245, 537-9

(1986).

4. Kaysser, L. et al. Identification and manipulation of the caprazamycin gene cluster lead to new

simplified liponucleoside antibiotics and give insights into the biosynthetic pathway. J Biol Chem

284, 14987-96 (2009).

5. Igarashi, M. et al. Caprazamycin B, a novel anti-tuberculosis antibiotic, from Streptomyces sp. J

Antibiot (Tokyo) 56, 580-3 (2003).

6. Kaysser, L., Siebenberg, S., Kammerer, B. & Gust, B. Analysis of the liposidomycin gene cluster

leads to the identification of new caprazamycin derivatives. Chembiochem 11, 191-6 (2010).

7. Kaysser, L. et al. A new arylsulfate sulfotransferase involved in liponucleoside antibiotic

biosynthesis in streptomycetes. J Biol Chem 285, 12684-94 (2010).

8. Gust, B., Challis, G.L., Fowler, K., Kieser, T. & Chater, K.F. PCR-targeted Streptomyces gene

replacement identifies a protein domain needed for biosynthesis of the sesquiterpene soil odor

geosmin. Proc Natl Acad Sci U S A 100, 1541-6 (2003).

9. Kelley, L.A. & Sternberg, M.J. Protein structure prediction on the Web: a case study using the

Phyre server. Nat Protoc 4, 363-71 (2009).

10. Moon, A.F. et al. Structural analysis of the sulfotransferase (3-o-sulfotransferase isoform 3)

involved in the biosynthesis of an entry receptor for herpes simplex virus 1. J Biol Chem 279,

45185-93 (2004).

11. Kakuta, Y., Pedersen, L.G., Pedersen, L.C. & Negishi, M. Conserved structural motifs in the

sulfotransferase family. Trends Biochem Sci 23, 129-30 (1998).

12. Hatzios, S.K., Iavarone, A.T. & Bertozzi, C.R. Rv2131c from Mycobacterium tuberculosis is a

CysQ 3'-phosphoadenosine-5'-phosphatase. Biochemistry 47, 5823-31 (2008).

13. Hatzios, S.K. et al. The Mycobacterium tuberculosis CysQ phosphatase modulates the

biosynthesis of sulfated glycolipids and bacterial growth. Bioorg Med Chem Lett 21, 4956-9

(2011).

14. Mechold, U., Ogryzko, V., Ngo, S. & Danchin, A. Oligoribonuclease is a common downstream

target of lithium-induced pAp accumulation in Escherichia coli and human cells. Nucleic Acids

Res 34, 2364-73 (2006).

15. Pi, N. et al. Kinetic measurements and mechanism determination of Stf0 sulfotransferase using

mass spectrometry. Anal Biochem 341, 94-104 (2005).

Accepted Manuscript Gust Lab

10

16. Song, L. et al. Type III polyketide synthase beta-ketoacyl-ACP starter unit and ethylmalonyl-CoA

extender unit selectivity discovered by Streptomyces coelicolor genome mining. J Am Chem Soc

128, 14754-5 (2006).

17. Funabashi, M., Funa, N. & Horinouchi, S. Phenolic lipids synthesized by type III polyketide

synthase confer penicillin resistance on Streptomyces griseus. J Biol Chem 283, 13983-91

(2008).

18. Surup, F. et al. The iromycins, a new family of pyridone metabolites from Streptomyces sp. I.

Structure, NOS inhibitory activity, and biosynthesis. J Org Chem 72, 5085-90 (2007).

19. Funabashi, M. et al. The biosynthesis of liposidomycin-like A-90289 antibiotics featuring a new

type of sulfotransferase. Chembiochem 11, 184-90 (2010).

20. Moore, B.S. et al. Plant-like biosynthetic pathways in bacteria: from benzoic acid to chalcone. J

Nat Prod 65, 1956-62 (2002).

21. Austin, M.B. & Noel, J.P. The chalcone synthase superfamily of type III polyketide synthases.

Nat Prod Rep 20, 79-110 (2003).

22. Watanabe, K., Praseuth, A.P. & Wang, C.C. A comprehensive and engaging overview of the

type III family of polyketide synthases. Curr Opin Chem Biol 11, 279-86 (2007).

23. Abe, I. & Morita, H. Structure and function of the chalcone synthase superfamily of plant type III

polyketide synthases. Nat Prod Rep 27, 809-38 (2010).

24. Yu, D., Xu, F., Zeng, J. & Zhan, J. Type III polyketide synthases in natural product biosynthesis.

IUBMB Life 64, 285-95 (2012).

25. Katsuyama, Y. & Ohnishi, Y. Type III polyketide synthases in microorganisms. Methods Enzymol

515, 359-77 (2012).

26. Funa, N. et al. A new pathway for polyketide synthesis in microorganisms. Nature 400, 897-9

(1999).

27. Funa, N., Ozawa, H., Hirata, A. & Horinouchi, S. Phenolic lipid synthesis by type III polyketide

synthases is essential for cyst formation in Azotobacter vinelandii. Proc Natl Acad Sci U S A 103,

6356-61 (2006).

28. Aoki, Y., Matsumoto, D., Kawaide, H. & Natsume, M. Physiological role of germicidins in spore

germination and hyphal elongation in Streptomyces coelicolor A3(2). J Antibiot (Tokyo) 64,

607-11 (2011).

29. Flett, F., Mersinias, V. & Smith, C.P. High efficiency intergeneric conjugal transfer of plasmid

DNA from Escherichia coli to methyl DNA-restricting streptomycetes. FEMS Microbiol Lett 155,

223-9 (1997).

30. Doumith, M. et al. Analysis of genes involved in 6-deoxyhexose biosynthesis and transfer in

Saccharopolyspora erythraea. Mol Gen Genet 264, 477-85 (2000).

Acknowledgments

The authors thank G. Challis (University of Warwick) for providing S. coelicolor M145/ sco7221 and

germicidin A. We are grateful to R. Machinek and C. Zolke (Institute of Organic Chemistry, University of

Göttingen) for carrying out NMR measurements. We also thank A. Jones for reviewing the manuscript.

This work was supported by a grant from the Deutsche Forschungsgemeinschaft (SFB766) to K.E. and

Accepted Manuscript Gust Lab

11

X.T., a grant from the Graduate School ‘Promotionsverbund Antibakterielle Wirkstoffe’ of the University of

Tuebingen to X.T. and by the European Commission (IP005224, ActinoGen) to L.K.

Author contributions

X.T., L.K. and B.G. designed the research. X.T. and L.K. generated and analyzed the mutants. X.T.

performed the biochemical experiments and purified presulficidins. X.T. and K.E. purified

hydroxyacylcaprazol and purified the proteins. X.T., K.E., L.K. and B.G. analyzed the data. A.K. performed

MS analysis. S.G. elucidated the structure of presulficidins. X.T., S.G. and B.G. wrote the manuscript. B.G.

supervised the project.

Competing financial interests

The authors declare no competing financial interests.

Accepted Manuscript Gust Lab

12

Figure 1 | Structures of CPZs and sulfated CPZs and genetic organization of the CPZ biosynthetic

gene cluster (cpz9-cpz31). Genes supposed to be involved in the sulfation mechanism of capazamycins

are shown in red.

Accepted Manuscript Gust Lab

13

Figure 2 | HPLC profiles of S. coelicolor M512 mutant extracts. (a) S. coelicolor

M512/cpzWP04( cpz1-3) containing the entire CPZ gene cluster. (b) S. coelicolor

M512/cpzWP05( cpz1-4). (c) S. coelicolor M512/cpzWP11( cpz8). (d) S. coelicolor

M512/cpzWP10( cpz7). (e) S. coelicolor M512/cpzWP09( cpz6). (f) S. coelicolor

M512/cpzWP04( cpz5). I, CPZ E/F aglycones; II, CPZ C/D/G aglycones; III, CPZ A/B aglycones; IV,

sulfated CPZ E/F aglycones; V, sulfated CPZ C/D/G aglycones; VI, sulfated CPZ A/B aglycones. UV at

260 nm.

Accepted Manuscript Gust Lab

14

Figure 3 | Characterization of Cpz8 as PAPS-dependent sulfotransferase. HPLC analysis of Cpz8

assays with pNP as nongenuine sulfate acceptor after 1-h reaction (i), control reaction without Cpz8 (ii),

authentic pNP standard (iii), authentic p-nitrophenyl sulfate standard (iv).

Accepted Manuscript Gust Lab

15

Figure 4 | HPLC profiles of S. coelicolor M145/ sco7221 mutant extracts. (a) S. coelicolor

M145/ sco7221/cpzLK09 containing the entire CPZ gene cluster. (b) S. coelicolor

M145/ sco7221/cpzWP11( cpz8). (c) S. coelicolor M145/ sco7221/cpzWP09( cpz6). I, CPZ E/F

aglycones; II, CPZ C/D/G aglycones; III, CPZ A/B aglycones; IV, sulfated CPZ E/F aglycones; V, sulfated

CPZ C/D/G aglycones; VI, sulfated CPZ A/B aglycones. UV at 260 nm.

Accepted Manuscript Gust Lab

16

Figure 5 | Identification and structures of presulficidins. (a) HPLC profiles of extracts from S.

coelicolor M512/cpzWP09( cpz6) (i), S. coelicolor M512/cpzWP11(cpz8) (ii), S. coelicolor

M512/cpzWP09( cpz6)/pUWL201 (empty vector) (iii) and S. coelicolor

M512/cpzWP09( cpz6)/pXT19(pUWL201+cpz6) (iv). UV at 290nm. (b) Structures of presulficidin A–D

(1–4).

Accepted Manuscript Gust Lab

17

Figure 6 | In vitro analysis of the CPZ two step sulfation mechanism. (a) Proposed two-step sulfation

mechanism in CPZ biosynthesis. (b) HPLC analysis of two-enzyme assays demonstrating sulfation of

hydroxyacylcaprazol E (5) generating sulfated hydroxyacylcaprazol E (6) with Cpz4, Cpz8, PAPS and the

presence of 100 M of either presulficidin A (1) (i), 1 without Cpz4 (ii), 1 without Cpz8 (iii), 1 without PAPS

(iv), no presulficidins (v), presulficidin B (2) (vi), presulficidin C (3) (vii) or presulficidin D (4) (viii). UV at 290

nm.

Accepted Manuscript Gust Lab

18

ONLINE METHODS

Bacterial strains and general methods. S. coelicolor M512 (SCP1–, SCP2

–, actIIorf4, redD), S.

coelicolor M145/ sco7221 (a gift from G. Challis, Department of Chemistry, University of Warwick, UK)16

and their respective derivatives were maintained and grown on either MS agar (2% soy flour, 2% mannitol,

2% agar; components purchased from Carl Roth, Karlsruhe, Germany) or TSB medium (BD Biosciences).

E. coli strains were cultivated in LB medium (components purchased from Carl Roth) supplemented with

the appropriate antibiotics. DNA isolation and manipulations were carried out according to standard

methods for E. coli.

Production, extraction and detection of CPZs and their derivatives. Fifty milliliters of TSB medium

were inoculated with spore suspension of S. coelicolor M512 or a derivative thereof. The cultures were

incubated for 2 d at 30 °C at 200 r.p.m. For the production of CPZ aglycones, 1 ml of precultures were

inoculated into 100 ml of the production medium containing 1% soytone, 1% soluble starch and 2%

D-maltose adjusted to pH 6.7 (components purchased from BD Biosciences). The cultures were incubated

for 7 d at 30 °C at 200 r.p.m. The culture supernatant was adjusted to pH 4 and subsequently extracted

with an equal volume of n-butanol. The organic phase was evaporated and extracts were resolved in 500

l methanol. LC-ESI-MS/MS analysis of extracts of the mutants was performed on a Surveyor HPLC

system equipped with a Reprosil-Pur Basic C18 (5 m, 250 × 2 mm) column (Dr. Maisch, Ammerbuch,

Germany) coupled to a Thermo Finnigan TSQ Quantum triple quadrupole mass spectrometer (heated

capillary temperature, 320 °C; sheath gas, nitrogen).

Treatment with ancymidol. One hundred milligrams of ancymidol (Sigma-Aldrich) was dissolved in 1 ml

DMSO as a stock solution. One hundred and fifty microliters of DMSO, 15 l of the stock solution with 135

l DMSO or 150 l stock solution were added to 50-ml cultures of S. coelicolor M512/cpzWP11( cpz8) at

the beginning of the inoculation. The cultures were incubated for 7 d at 30 °C at 200 r.p.m. The culture

supernatant was adjusted to pH 5 and extracted with an equal volume of ethyl acetate. Ethyl acetate was

evaporated to dryness and dissolved in 500 l methanol for HPLC analysis using a Reprosil-Pur Basic

C18 (5 m, 250 × 2 mm) column (Dr. Maisch, Ammerbuch, Germany).

Generation of cpz5, cpz6, cpz7 and cpz8 mutants. An apramycin resistance (aac(3)IV)

cassette was amplified from plasmid pIJ773 (ref. 8) using primer pairs cpz05KO_F-cpz05KO_R,

cpz06KO_F-cpz06KO_R, cpz07KO_F-cpz07KO_R and cpz08KO_F-cpz08KO_R (Supplementary Note

Accepted Manuscript Gust Lab

19

and Supplementary Table 2). Targeted genes were replaced in E. coli BW25113/pIJ790 containing

cosmid cpzLK09 with intact CPZ gene cluster4 by using the PCR targeting system

8. Resulting cosmids

were confirmed by restriction analysis. Excision of the cassette was performed in E. coli BT340, taking

advantage of the FLP recognition sites adjacent to the resistance cassette. Positive cosmids were

screened for their apramycin sensitivity and verified by restriction analysis and PCR using the primer pairs

cpz05test_F-cpz05test_R, cpz06test_F-cpz06test_R, cpz07test_F-cpz07test_R and

cpz08test_F-cpz08test_R (Supplementary Table 2). Cosmids cpzWP08( cpz5), cpzWP09( cpz6),

cpzWP10( cpz7) and cpzWP11( cpz8) were transferred into E. coli ET12567 (ref. 29) and introduced

into S. coelicolor M512 or S. coelicolor M145/ sco7221 (ref. 16) by triparental intergeneric conjugation

with the help of E. coli ET12567/pUB307 (ref. 29). Kanamycin resistance clones were selected, confirmed

by PCR and designated as S. coelicolor M512/cpzWP08 ( cpz5), S. coelicolor M512/cpzWP09( cpz6),

S. coelicolor M512/cpzWP10( cpz7), S. coelicolor M512/cpzWP11( cpz8) and S. coelicolor

M145/ sco7221/cpzWP09( cpz6).

Genetic complementation. To generate expression plasmids for complementation of mutants, cpz6 was

amplified from cosmid cpzLK09 (ref. 4) using primer pair cpz6_fw_BamHI-cpz6_rv_HindIII

(Supplementary Table 2) and was cloned into pGEM-T (Promega). The BamHI/HindIII fragment

containing cpz6 was blunt-ended and subcloned into the EcoRV site of the expression vector pUWL201

(ref. 30) under the control of the ermE* promoter to give pXT19. DNA sequencing of the plasmid confirmed

the correct sequence. For protoplast transformation, the plasmids pXT19 and pUWL201 were transferred

into the nonmethylating E. coli strain ET12567, and DNA was isolated by standard procedures.

Transformation of the S. coelicolor mutant strains by polyethylene glycol–mediated protoplast

transformation finally generated S. coelicolor M512/cpzWP09( cpz6)/pXT19(cpz6) and S. coelicolor

M512/cpzWP09( cpz6)/pUWL201 (empty vector as control).

Construction of protein expression plasmids. The primer pair cpz7BamHI_F and cpz7XhoI_R

(Supplementary Table 2) were used for amplification of cpz7 from cosmid cpzLK09 containing the CPZ

gene cluster4. The PCR product was cloned into the BamHI and XhoI sites of pHis8 to obtain pXT20 (with

an N-terminal His tag). Amplification of cpz8 was accomplished with the primer pair cpz8BamHI_F and

cpz8HindIII_R (Supplementary Table 2). The resulting fragment was cloned into pGEM-T (Promega,

Mannheim, Germany) to give pXT2 and was verified by sequencing. cpz8 was subsequently cloned into

Accepted Manuscript Gust Lab

20

the BamHI and HindIII sites of pHis8 to give pXT5 and was confirmed by sequencing. pXT20 and pXT5

were transformed into E. coli Rosetta2TM (DE3)pLys (Novagen, Darmstadt, Germany).

Assay conditions for Cpz7. Cpz7 Assays were prepared on ice and carried out at 30 °C for 10 min. The

reaction mixture contained 50 mM Tris-HCl (pH 8.0), 0.5 mM MgCl2, 500 M PAP (purity 99%,

Sigma-Aldrich) and 0 nM, 75 nM, 150 nM, 300 nM or 600 nM of purified Cpz7. One hundred microliters of

ice-cold methanol was added to stop the reaction, and the tube was put on ice for 10 min. After

centrifugation of the assay at 15,000 r.p.m., the supernatant was injected to HPLC by using a Kinetex C18

100A column (Phenomenex, 100 × 4.6 mm, 2.6 ). Products were monitored at constant mobile phase

with 15 mM tetrabutylammonium hydrogen sulfate (Merck, Germany) and 28% acetonitrile in water at 261

nm.

Assay condition for Cpz8. The assay mixture for the reaction (50 l) consisted of 50 mM

2-(N-morpholino)ethanesulfonic acid (MES) (pH 6.5), 5 mM MgCl2, 20 M Cpz8, 200 M PAPS (purity

60%, Sigma-Aldrich) and 200 M of different substrates, including germicidin A (purity 95%, a gift from

G. Challis, Department of Chemistry, University of Warwick, UK)16

, pNP, methylumbelliferone (MU),

4-hydroxy-6-methyl-2-pyrone, 3-hydroxy-2-methyl-4-pyrone, 2-naphtol, phloroglucinol or resorcinol (all

purities 98%, Sigma-Aldrich). The reaction solutions were prepared on ice and incubated at 30 °C for 1

h. Reactions containing presulficidins were incubated for only 1 min. Reactions were terminated by adding

50 L ice-cold methanol and placing the tube on ice for 10 min. After centrifugation of the assay at 15,000

r.p.m. for 10 min, the supernatant was monitored by a Surveyor HPLC system equipped with a

Reprosil-Pur Basic C18 column (5 m, 250 × 2 mm; Dr. Maisch, Ammerbuch, Germany) coupled to a

Thermo Finnigan TSQ Quantum triple-quadrupole mass spectrometer (heated capillary temperature,

320 °C; sheath gas, nitrogen). Alternatively, the supernatant was monitored by HPLC using a Nucleosil

100-C18 column (3 m, 100 × 2 mm; Phenomenex, Germany) coupled to an ESI mass spectrometer

(LC/MSD Ultra Trap System XCT 6330; Agilent Technology). Analysis was carried out at a flow rate of 0.4

ml/min with a linear gradient from 10% to 100% of solvent B in 15 min (solvent A: water/formic acid (999:1);

solvent B: acetonitrile/formic acid (999.4:0.6)). Electrospray ionization (positive and negative ionization) in

Ultra Scan mode with capillary voltage of 3.5 kV and heated capillary temperature of 350 °C was used for

LC/MS analysis.

Accepted Manuscript Gust Lab

21

Kinetic analysis. For the determination of the Km values, assays consisted of 50 mM MES, pH 6.5, 5 mM

MgCl2, 20 M Cpz8 and near-saturating PAPS (2 mM) with variable pNP (25 M–4 mM) or near-saturating

pNP (4 mM) and variable PAPS (5 M–0.8 mM) in 50 l total volume. The reactions were performed at

30 °C for 10 min and terminated by the addition of 50 L ice-cold methanol. Product formation was

determined using HPLC. Each data point represents a minimum of three replicates. Kinetic constants

were obtained by nonlinear regression analysis using GraphPad Prism (GraphPad Software, La Jolla,

CA).

Conditions for the two-enzyme (Cpz4 and Cpz8) assay. The assay mixture for the reaction (50 l)

consisted of 50 mM MES (pH 6.5), 5 mM MgCl2, 0.25 M Cpz4, 1 M Cpz8, 100 M PAPS, 25 M

hydroxyacyl-caprazol E (purity 80%, determined by HPLC and LC/MS; Supplementary Fig. 11) and

100 M of presulficidins A–D, germicidin A, 4-hydroxy-6-Methyl-2-pyrone, pNP or MU. The reaction

solutions were prepared on ice and then incubated at 30 °C for 10 min. Reactions were terminated by the

addition of 50 L ice-cold methanol, and the tube was put on ice for 10 min. After centrifugation of the

assay at 15,000 r.p.m. for 10 min, the supernatant was monitored at 261 nm by HPLC using a Nucleosil

100-C18 column (3 m, 100 × 2 mm; Phenomenex, Germany) as follows: 0 min, 10% B; 20 min, 100% B;

25 min, 100% B; 26 min, 10% B; 32 min, 10% B. In addition, samples were analyzed by LC/MS and

MS/MS (as described above). For determination of reaction velocities, different amounts of Cpz8

substrates were used.

Purification of hydroxyacylcaprazol E (5). Compound 5 was purified from 1.8 L supernatant of S.

coelicolor M512/cpzLL06( cpz21)4. To direct production toward production of 5, 1.25g/L of palmitic acid

was added into to the production medium. After cultivation for 7 d at 30 °C at 200 r.p.m., the supernatant

was recovered by centrifugation at 8,000 r.p.m. for 15 min. The supernatant was adjusted to pH 4.0 with

12 M HCl and extracted with an equal volume of n-butanol, dried with Na2SO4 and evaporated to dryness.

The crude extract was washed twice by 200 ml ethyl acetate and then dissolved in methanol, mixed with

10 g silica gel and evaporated to dryness. The dried silica gel was applied on a silica column (50 g silica

gel, diameter = 4 cm and length = 30 cm) and washed sequentially with 500 ml of CH2Cl2/MeOH/H2O

(4:1:0.1), 500 ml of CH2Cl2/MeOH/H2O (2:1:0.2) and, finally, 500 ml of CH2Cl2/MeOH/H2O (1:1:0.2), and

washing fractions were collected. Fractions containing 5 were pooled together, evaporated and dissolved

in DMSO for reverse-phase semipreparative HPLC using a Multospher 120RP18 (Dr. Maisch, 250 × 8 mm,

Accepted Manuscript Gust Lab

22

5 ) connected to an Agilent 1100 HPLC instrument. Purified 5 was then verified by HPLC using a

Nucleosil 100-C18 column (3 m, 100 × 2 mm) coupled to an ESI mass spectrometer (LC/MSD Ultra Trap

System XCT 6330; Agilent Technology).

Isolation and purification of Cpz4, Cpz7 and Cpz8. Cpz4 was purified as described previously7. E. coli

Rosetta2TM (DE3)pLys containing pXT20 or pXT5 was cultivated in 1 L TB broth (components purchased

from Carl Roth, Germany) supplemented with 50 g/ml kanamycin and 50 g/ml chloramphenicol at 37 °C.

At a D600 nm of 0.7, the temperature was adjusted to 20 °C, and IPTG was added to a final concentration of

0.5 mM. After an additional 10-h cultivation at 20 °C, cultures were harvested, and 10 ml of buffer A (50

mM Tris-HCl, pH 8, 1 M NaCl, 10% glycerol, 10 mM -mercaptoethanol) supplemented with 0.5 mg/ml

lysozyme and 0.5 mM PMSF was added to the pellets. Cells were disrupted by sonication (Branson,

Danbury, CT) at 4 °C. The lysates were centrifuged at 18,000g for 45 min, and the supernatants were

applied to affinity chromatography using an Äktapurifier platform (GE Healthcare) equipped with a

His-TrapTM (34 m, 1.6 × 2.5 cm) HP column (GE Healthcare).The His-tagged proteins were eluted from

the column using a linear gradient from 0–100% imidazole (250mM) in buffer A over 10 min and collected

by a Frac-920TM system (GE Healthcare). Fractions were checked for the presence of the desired

proteins by SDS-PAGE. The purified protein was stored at –80 °C in aliquots.