Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 May 25.
Published in final edited form as: Chem Biol. 2012 May 25;19(5):647–655. doi: 10.1016/j.chembiol.2012.04.010

Tailoring Enzymes Involved in the Biosynthesis of Angucyclines Contain Latent Context-Dependent Catalytic Activities

Pekka Patrikainen 1, Pauli Kallio 1, Keqiang Fan 2, Karel D Klika 3, Khaled A Shaaban 4, Pekka Mäntsälä 1, Jürgen Rohr 4, Keqian Yang 2, Jarmo Niemi 1, Mikko Metsä-Ketelä 1,*
PMCID: PMC3361699  NIHMSID: NIHMS372073  PMID: 22633416

Summary

Comparison of homologous angucycline modification enzymes from five closely related Streptomyces pathways (pga, cab, jad, urd, lan) allowed us to deduce the biosynthetic steps responsible for the three alternative outcomes, gaudimycin C, dehydrorabelomycin and 11-deoxylandomycinone. The C-12b-hydroxylated urdamycin and gaudimycin metabolites appear to be the ancestral representatives from which landomycins and jadomysins have evolved as a result of functional divergence of the ketoreductase LanV and hydroxylase JadH, respectively. Specifically, LanV has acquired affinity for an earlier biosynthetic intermediate resulting in a switch in biosynthetic order and lack of hydroxyls at C-4a and C-12b, whereas in JadH, C-4a/C-12b dehydration has evolved into an independent secondary function replacing C-12b hydroxylation. Importantly, the study reveals that many of the modification enzymes carry several alternative, hidden or ancestral catalytic functions, which are strictly dependent on the biosynthetic context.

Introduction

Enzymes have been classically renowned for their precision and ability to act as highly stereo- and regiospecific catalysts. However, in recent years the promiscuity of enzymes has raised considerable attention (Copley, 2003; Khersonsky et al., 2006; Hult and Berglund, 2007). This holds especially true for proteins involved in the biosynthesis of natural products, which have been valued for their ability to utilize “unnatural” substrates (Niemi et al., 2008). In effect, the phenomenon appears to be an inherent property of secondary metabolic pathways, where selective pressure for diversification of function may supersede efficiency and accuracy. Substrate promiscuity may therefore be considered as a highly beneficial trait, which facilitates the rapid evolution of the biosynthetic pathways that ultimately has led to the exceptional numbers of metabolites isolated from natural sources. The diversification process requires adaptation of the enzymes to new biosynthetic contexts and may involve changes in substrate recognition, reaction stereo- and regiochemistry, or even appearance of a completely new function. Such re-functionalization would be more challenging to accomplish with more efficient and fine-tuned enzymes such as those found in primary metabolism. Importantly, the broad substrate specificity of enzymes in secondary metabolism has enabled the use of pathway engineering, e.g. combinatorial biosynthesis, for generation of novel natural products (Rix et al., 2002; Niemi et al., 2008). Therefore detailed understanding of changes in enzyme function would also be important for elucidating the boundaries and possibilities of current pathway engineering technologies.

Angucyclines are a specific group of aromatic polyketides, which have been associated with various biological activities (Antal et al., 2005; Jakeman et al., 2009; Korynevska et al., 2007; Lombo et al., 2009). These compounds are found widely in nature and are mainly produced by soil dwelling Streptomyces bacteria. The biosynthesis of angucyclines diverges from other type II aromatic polyketides by the action of the specific cyclase that closes the fourth ring of the polyketide into an angular orientation producing UWM6 (1) (Figure 1) (Kulowski et al., 1999; Metsä-Ketelä et al., 2003). The biosynthesis continues through a cascade of oxidation/reduction reactions that result in the formation of the gaudimycin (cab/pga) (Palmu et al., 2007), landomycin (lan) (Weber et al., 1994), urdamycin (urd) (Rohr and Zeeck, 1987) and jadomycin (jad) (Ayer et al., 1991) metabolites (Figure S1). The enzymes responsible for the modification of the polyketide aglycone belong to two well-characterized families, NADPH-dependent flavoprotein mono-oxygenases and short-chain alcohol dehydrogenases/reductases (SDRs). The biosynthetic genes are distributed in the clusters in a conserved order (Figure 2) and the gene products share substantial sequence identity, which has made functional predictions challenging based on primary amino acid sequence alone. Each gene cluster encodes for a NADPH-dependent flavoprotein hydroxylase (PgaE, UrdE, LanE, CabE and JadH), and an additional flavoprotein oxygenase that may be found in several different combinations with an SDR component (Figure 2).

Figure 1.

Figure 1

Open in a new tab

The proposed angucycline redox modification reactions and the structures of the metabolites in the native pga, cab, lan, urd, and jad pathways as determined in vitro. See also Figure S1.

Figure 2.

Figure 2

Open in a new tab

Graphical Representation of the Gene Clusters pga, cab, urd, lan and jad under Study. The genes encoding for the enzymes involved in angucyclione modification are flavin-dependent oxygenases (black) and short-chain alcohol reductases/oxygenases (dark grey). The clusters also show genes involved with the aglycone biosynthesis (white), tailoring (jadG, light grey) and regulation (pgaY, light grey).

Recent studies have shown that on the gaudimycin pathway, the flavoenzyme PgaE utilized 2,3-dehydro-UWM6 (2) as a substrate and catalyzed two consecutive steps, C-12 (3) and C-12b (4) hydroxylations, which were followed by the action of the SDR enzyme CabV to produce gaudimycin C (5) (Figure 1) (Kallio et al., 2011). The oxidation of the hydroquinone to the C-7,12-quinone structure is generally thought to occur nonenzymatically. The two reactions catalyzed by PgaE were dependent on the presence of NADPH and molecular oxygen and were shown to occur at different phases due to competitive inhibition between 2 and 3. The second reaction catalyzed by PgaE initiated only after most of the original substrate was used, which allowed the isolation and use of the intermediate 3 as an alternative substrate. Efficient coupling of the second PgaE reaction to that of CabV was a strict requirement for the formation of 5, since in the absence of the SDR the biosynthesis diverted onto a degradation pathway (Kallio et al., 2011). The ketoreduction step could be performed in an identical manner by either CabV or the two-domain PgaM, which implicated indirectly that the N-terminal oxygenase domain of PgaM was not required during conversion of 2 to 5 (Kallio et al., 2008b; Kallio et al., 2011). The flavoenzyme JadH from the jadomycin pathway has also been investigated in detail (Chen et al., 2010), which has revealed that somewhat surprisingly, JadH behaves differently to PgaE despite the considerable amino acid sequence identity (58%). While both enzymes utilize 2 as a substrate and catalyze C-12 hydroxylation, JadH additionally is able to perform a C-4a/C-12b dehydration reaction as seen in the structure of dehydrorabelomycin (7) (Figure 1) instead of the secondary hydroxylation at C-12b. Furthermore, conversion of 2 into jadomycin A has been established using cell-free extracts containing the flavoprotein mono-oxygenases JadH and JadF together with JadG, which is homologous to anthrone oxygenases (Chen et al., 2005). The functional assignments of the gene products from the landomycin and urdamycin pathways are mainly based on molecular genetic evidence. On the urdamycin pathway, UrdE was originally postulated to be responsible for the C-12b hydroxylation (Decker and Haag, 1995), but later results suggested that the hydroxylation would be catalyzed by the oxygenase domain of UrdM (Faust et al., 2000), while the SDR domain was associated with the ketoreduction at C-6 (Mayer et al., 2005). On the landomycin pathway, LanV has been proposed to be responsible for the aromatization of ring A and 6-ketoreduction in landomycin A biosynthesis (Mayer et al., 2005).

In the present study, we have characterized enzymes from five different angucycline pathways (pga, cab, urd, lan and jad) in order to determine the functional differences which explain the natural end-product diversity. We have used different combinations of enzymes and reaction conditions to understand the order of the biosynthetic steps, and the co-substrate and oxygen dependence of the individual enzymes. The first part of the study concentrates on the functions of the enzymes from the lan, urd and jad pathways in their native contexts. In the latter part we examine the flexibility and specificity of the target enzymes in different combinations (combinatorial enzymatic synthesis) and conditions in order to deduce their catalytic behavior under non-native environments.

Results and Discussion

The Flavoenzyme UrdE and the Reductase Domain of UrdM are Sufficient for Production of Gaudimycin C

Purified UrdE was shown to convert 2 in two separate NADPH/O2-dependent steps (Figure S2) as previously described for PgaE (Kallio et al., 2011), and when coupled to the independently expressed SDR domain UrdMred, the corresponding reaction efficiently produced 5 (Figure 3A and 3B). This was equivalent to the biosynthetic reaction order UrdE-UrdE-UrdMred (Figure 4A) (Kallio et al., 2011), and attributed UrdE responsible for the consecutive hydroxylations at positions C-12 and C-12b (Figure 1). The significance of the finding was evident as UrdE had been previously associated with only one of the two hydroxylation steps in urdamycin biosynthesis (Decker and Haag, 1995; Faust et al., 2000). The C-6 ketoreduction step in the biosynthesis of 5 was catalyzed by UrdMred proposedly in the same way as by CabV (Kallio et al., 2011). In the native context UrdMred is translated as part of the two-domain flavoprotein oxygenase fusion and as a separate C-terminal SDR in analogy to PgaM (Kallio et al., 2008a). In the case of PgaM, the fusion form appears to be essential for the stability of the SDR (Kallio et al., 2008a), as the C-terminal half could not be recovered by itself. In contrast, both the full-length UrdM and the oxygenase domain UrdMox alone were highly unstable, whereas the SDR UrdMred domain could be isolated in soluble form at quantities of over 200mg per liter culture. The results show that the N-terminal flavoprotein oxygenase domain of UrdM is not required for C-12b hydroxylation, even though the enzyme has been linked with the reaction in the past (Faust et al., 2000). Instead, one possibility may be that the oxygenase domain has a similar role to JadF (Pahari et al., 2012) and LanM2 (Kharel et al., 2012), which have been very recently shown to be responsible for the release of the thioester-tethered polyketide from the acyl carrier protein.

Figure 3.

Figure 3

Open in a new tab

Representative HPLC Traces at 428 nm of the Native In Vitro Reactions with 2 as a Substrate. (A) The substrate 2, (B) UrdE/UrdMred product 5 (C) LanE/LanV product 8, (D) JadH product 7 (E) JadH production of 7 was not affected by the presence of SDRs. Spontaneous formation of 7 (Kallio et al., 2011) was avoided by ensuring quantitative conversion of 3 before metabolite extraction. See also Figure S2.

Figure 4.

Figure 4

Open in a new tab

Consumption of 2 (■) and Formation of Products (○) in Reactions Carried Out to Completion at Different NADPH Concentrations. (A) A typical profile of a reaction producing the 12b-hydroxylated product 5. Characteristically the NADPH-dependent steps do not overlap, and the product is not formed before 2 has been depleted as in the CabE/CabV, PgaE/PgaM (Kallio et al., 2011) and UrdE/UrdMred in vitro reactions (B) A typical profile of a reaction producing 8. The two NADPH-dependent steps occur simultaneously and the product is formed already during the conversion of 2 as in the LanE/LanV in vitro reaction. See also Figure S3.

Enzymatic Synthesis of 11-Deoxylandomycinone by LanE and LanV

Also LanE was shown to utilize 2 and to catalyze two successive oxygen-dependent steps (Figure S2) in the same manner to UrdE, PgaE (Kallio et al., 2011) and CabE (Koskiniemi et al., 2007). At first this was a surprising finding as natural landomycins do not contain a hydroxyl group at C-12b. However, incubation of LanE and LanV together with 2 resulted in the appearance of a previously undetected metabolite in HPLC (Figure 3C), while 5 was not observed. LC-MS and UV-Vis properties of this new compound were in agreement with the landomycin pathway intermediate 11-deoxylandomycinone (8) (Figure 1) (Shaaban et al., 2011), and further supported by the molecular formula C19H14O5 ([M-H]-observed 321.0780, calculated 321.0763) inferred by ESI-HR-MS. The structure was subsequently confirmed as 8 by comparison to an authentic standard by co-elution in HPLC, UV/Vis-spectra and CD-spectra (Figure S2). The structure of the product supported the functional assignment of LanE as the C-12 hydroxylase (Zhu et al., 2005), LanV as the C-6 ketoreductase (Mayer et al., 2005) and further clarified that one of the enzymes was responsible for an additional dehydration between C-4a/C-12b.

The Order of the Tailoring Steps has Changed in Landomycinone Biosynthesis

Time-course analysis of the coupled LanE/LanV reaction showed that 8 was accumulating already during the conversion of 2 (Figure 4B). This was a direct proof that the LanV reaction took place immediately after the C-12 hydroxylation step, but prior the C-12b hydroxylation which was still inhibited at this phase (Kallio et al., 2011). The chronology indicated that LanV possessed catalytic affinity towards the C-12 hydroxylated intermediate 3, which was in clear contrast to the homologous enzymes involved in the biosynthesis of 5 (Figure 4A) (Kallio et al., 2011). To confirm that the steps after the initial C-12 hydroxylation were independent of oxygen as suggested by the structure of 8, the reaction was conducted under anaerobic conditions using the isolated 3 as the substrate. Consequently, 8 was produced (Figure 5A), but only in the simultaneous presence of both LanE and LanV. Despite the incomplete conversion, this had important implications in relation to the function of LanE in the native context: The experiment verified the catalytic order in the biosynthesis of 8 to be LanE-LanV-LanE, and attributed LanE indispensable for two separate catalytic steps. However, the second step could not be a hydroxylation because it was accomplished in the absence of oxygen. Instead, the chemical changes associated with the formation of 8 proposed that it would be the C-4a/C-12b-dehydration (Figure 1) equivalent to the secondary catalytic function of JadH (Chen et al., 2010). The NADPH titration profile (Figure 4B) confirmed that LanV used one equivalent of the cosubstrate in the production of 8 which was consistent with the expected reaction mechanism of SDR enzymes catalyzing similar ketoreduction reactions (Kavanagh et al., 2008).

Figure 5.

Figure 5

Open in a new tab

Representative HPLC Traces of Reaction Products Obtained with Isolated 3 as Substrate at 449nm. (A) LanE/LanV under anaerobic conditions yields 8, (B) LanE/LanV in the presence of O2 yields 5, (C) JadH/CabV in the presence of O2 yields 5 (D) JadH/UrdMred in the presence of O2 yields 5 (E) JadH/LanV in the presence of O2 yields 5. See also Figure S5.

Confirmation of the Opposite Stereoselectivities of LanV and CabV/UrdMred

The absolute stereochemistry of 8 has been previously confirmed as 6R via total synthesis, (Roush and Neitz, 2004; Zhu et al., 2005) while to date only the relative stereochemistry of 5 has been known. To ascertain the absolute configuration for 5, experimental and theoretically calculated ECD spectra were compared (Figure 6A, 6B and S4), which was in good agreement for the 4aR,6S,12bS enantiomer shown in Figure 1. Similar experiments conducted for 2 (Figure 6C, 6D and S4) also indicated an R configuration at C-4a, which thus supported the assignment for 5. The theoretically calculated ECD spectrum of 8 was also in good agreement for an R configuration at C-6 (Figures 6E, 6F and S4) for this compound. The rigid polyaromatic aglycones limited the conformational freedom to rings A and B, and as a result only two (ca. 1:1 ratio), one and two (ca. 2:1 ratio) conformers of significant population were found for 2, 5 and 8, respectively, by molecular modeling. Importantly, the results verified that CabV and UrdMred were responsible for the 6S configuration in 5, whereas LanV catalyzed the formation of the opposite 6R configuration in 8 in the native pathways.

Figure 6.

Figure 6

Open in a new tab

Measured and Theoretical ECD Spectra of 5, 2 and 8. (A) Experimental and (B) predicted ECD spectrum for the 4aR,6S,12bS enantiomer of 5, (C) measured and (D) calculated ECD spectrum for the 4aR,12bR enantiomer of 2. The calculated spectrum was based on two molecules of 2 with one in a C-4-in conformation and the other in a C-4-out conformation, (E) experimental and (F) predicted spectrum for the 6R enantiomer of 8. Three molecules of 8 with two molecules having an equatorially disposed HO-6 group and the third containing an axially disposed HO-6 group were used in the calculations. See also Figure S4 and Table S1.

JadH Catalyzes Dehydrorabelomycin Formation in a Single Step

Despite high sequence identity with PgaE (58%), CabE (58%), UrdE (61%) and LanE (57%), JadH differed in function from the other hydroxylases and catalyzed the conversion of 2 into 7 (Figure 3D) via C-12 oxygenation and C-4a/C-12b dehydration (Figure 1) (Chen et al., 2010; Chen et al., 2005). One equivalent of NADPH was consumed in the process (Figure S3) and product degradation fragments were not observed. The C12-hydroxylated product 3 could not be detected or isolated from the reaction, and in parallel, co-incubation with the purified SDRs from the cab, urd or lan pathways did not have any effect on the reaction profile and only 7 was produced (Figure 3E). In addition, JadH failed to convert isolated 3 into 7 under equivalent conditions (data not shown), further implying that i) the native JadH reactions may proceed in a closely-coupled sequence which cannot dissociated from one another, and possibly that ii) the dehydration at C-4a/C-12b needs to take place first before the C-7/C-12 dihydroquinone formation can occur.

Timing of 6-ketoreduction Determines Hydroxylase Activity

Next the enzymes from the parallel pathways were assayed in different combinations to deduce their functions in non-native contexts. Incubation of the hydroxylases PgaE, CabE, UrdE or LanE in any combination with either CabV or UrdMred and 2 resulted in the formation of 5. In contrast, the corresponding coupled reactions with LanV produced exclusively 8. This was an unambiguous proof that outcome of the reaction was solely determined by the SDR and not the hydroxylase per se. Against the deduced native functions, this could now be explained by the substrate specificities of the different SDRs towards either 3 or 4, which ultimately defined the sequential reactions to proceed towards landomycins or the 12b-hydroxylated metabolites, respectively. Thus it appears that a simple decrease in KM for 3 is the underlying cause for the functional diversification of LanV from the other SDRs, and the key step responsible for the evolution of the landomycin antibiotics. This change in catalytic specificity, however, is not apparent at the level of primary structure, as the sequence identity of LanV with its presumed ancestral counterparts CabV, UrdMred, and PgaMred still remains exceptionally high (over 65%) for members of the SDR family (Jornvall et al., 1995; Persson et al., 2003).

The results also demonstrated that the hydroxylases, including LanE, all functioned in an identical context-dependent manner. In the presence of CabV or UrdMred the catalysis proceeded via the hydroxylase-catalyzed C-12b oxygenation as in gaudimycins and urdamycins, whereas with LanV the reaction continued towards the dehydration at C-4a/C-12b as observed in landomycins. This showed that all the hydroxylases carried two alternative catalytic modes (hydroxylation and dehydration), the choice between which was determined by the timing of the SDR reaction. Thus, the lack of the C-12b hydroxyl group in landomycins is not due to a lack of an appropriate hydroxylase in the pathway but, intriguingly, the surrounding biosynthetic context in which the oxygenation reaction is overhauled by the previous step.

The C-6 Stereoselectivity of LanV is Context Dependent

We then further corroborated our hypothesis on the reaction order by controlling the availability of oxygen with isolated intermediate 3 as a substrate. As shown above, incubation of LanV with LanE and 3 under strict anaerobic conditions resulted in the formation of 8 (Figure 5A). In striking contrast, when the experiment was repeated in the presence of oxygen, conversion to 5 was observed (Figure 5B). Having now deduced the steps in the alternative pathways, this drastic change could be explained by the ability of both LanV and LanE to utilize and compete for the substrates 3 and 4. The hydroxylases apparently have higher affinity for 3 in comparison to LanV, and when O2 is available this resulted in C-12b hydroxylation followed by C-6 ketoreduction as in the native pga, urd and cab pathways (Figure 1). In the absence of oxygen the C-12b hydroxylation is blocked and, consequently, the reaction order is reversed: LanV catalyzes C-6 ketoreduction, which then allows the hydroxylase to perform C-4a/C-12b dehydration as described for the formation of 8 (Figure 1). This scenario is equivalent to the native reaction sequence in landomycin biosynthesis, with the difference that here the C-12b hydroxylation was inhibited by the anaerobic conditions and not the presence of 2. In support of this proposed sequence, the same outcome could also be accomplished by co-addition of a small amount of 2 even when oxygen was present. The corresponding reactions with the SDRs CabV and UrdMred only produced trace quantities of the lan product 8 (Figure S5), indicating that the SDRs had a poor affinity towards 3 when the native pathway was restrained.

In addition to the switch in the reaction order itself, the experiment verified that LanV was in effect able to catalyze the formation of both 6R and 6S configurations in 8 and 5, respectively. The main determinant of the stereospecificity appears to be the C-12b substituent missing in 3 and present in 4, possibly by affecting the substrate orientation in the active site.

JadH Carries the Ancestral C-12b Hydroxylase Activity

The JadH reaction with 3 failed to produce 7, but instead, resulted in gradual disappearance of the substrate as previously associated with enzyme-catalyzed degradation by PgaE (Figure S2) (Kallio et al., 2008b; Kallio et al., 2011). When the reaction was repeated in the presence of either CabV, UrdMred or LanV, notably, 5 was obtained (Figure 5C, 5D and 5E). Although the conversion was not quantitative, the finding demonstrated that JadH had retained the ability of the presumed ancestral enzymes to catalyze hydroxylation at position C-12b, a function masked by the C-4a/C-12b dehydratase activity in the native pathway. The experiment also encapsulated the utmost context-dependence of the enzymes, as in the coupled JadH/LanV reaction neither of the enzymes functioned according to their native functions, but catalyzed reactions characteristic for the corresponding enzymes in the gaudimycin and urdamycin pathways (Figure 1).

Concluding Remarks

The study clarifies the redox modification steps of the urd, lan and jad angucyclinone pathways, and identifies the native enzyme-catalyzed steps responsible for the structural diversity of the alternative end-products 5, 8, and 7, respectively (Figure 1). By comparing the catalytic functions of homologous proteins in vitro, we demonstrated how in the course of enzyme evolution the SDR LanV and the flavoprotein hydroxylase JadH have differentiated into novel catalytic species. Specifically, LanV has acquired catalytic affinity towards the earlier biosynthetic intermediate 3 resulting in the switch of biosynthetic reaction order, and promoting LanE to catalyze C-4a/C-12b dehydration in place of the C-12b hydroxylation. In JadH the corresponding C-4a/C-12b dehydratase activity appears to have evolved into an SDR-independent secondary function, which effectively outcompetes the parallel C-12b oxygenase activity. Importantly, LanV and JadH represent the first branching-points of the landomycin and jadomycin pathways, respectively, and thus appear to be ultimately responsible for the evolution of the corresponding angucycline families.

As one of the most important findings in this work, many of the enzymes were characterized by innate catalytic context-dependence, and could be coerced to perform several alternative functions by altering the reaction set-ups. For example, the hydroxylases PgaE, CabE, UrdE and LanE could all be harnessed to produce either 5 or 8 depending on the SDR involved, irrespective of their original functions. In a similar manner, under specified conditions also JadH could participate in the biosynthesis 5 in addition to the native reaction producing 7. As for the SDRs, LanV could be applied to produce both 5 and 8, depending on the reaction conditions. From an evolutionary viewpoint this was fascinating as it demonstrated that both JadH and LanV have retained the ability to perform their ancestral catalytic activities, and at the same time underlined the immense flexibility of the enzymes for acquisition of novel functions, which has enabled the biosynthesis of the vast number of secondary metabolites isolated from natural sources.

Significance

In this study we demonstrate the critical role of functional diversification of biosynthetic enzymes in the evolution of angucycline secondary metabolites in Streptomyces bacteria. We show that homologous flavoenzyme oxygenases and short chain alcohol dehydrogenases/reductases (SDRs) from related pathways (pga, cab, urd, lan, jad) are responsible for three alternative redox modification patterns, yielding either gaudimycin C, dehydrorabelomycin or 11-deoxylandomycinone when provided with the common substrate 2,3-dehydro-UWM6. Detailed functional comparison of the enzymes allowed us to identify specific catalytic divergences which were not apparent based on sequence homologies. Functional characterization also revealed latent context-dependent catalytic activities which were suppressed in the native pathways; we show that the enzymes harbor explicit functional flexibility and ability to catalyze several alternative reactions depending on the substrate, the reaction conditions, and the other enzymes present. The results support the view that catalytic flexibility may be a common trait for biosynthetic enzymes involved in secondary metabolism pathways as a way to adapt to rapidly changing biosynthetic environments, and to increase chemical diversity at low metabolic cost.

Experimental Procedures

Production and Purification of Primary Substrate

The substrate 2 was produced in Streptomyces lividans TK24/pMC6BD (Metsä-Ketelä et al., 2003) and purified as described earlier (Kallio et al., 2011).

Cloning, Expression and Purification of LanE and LanV

The lanE and lanV genes were amplified with Phusion DNA polymerase (Finnzymes) using cosmid H2-26 (Westrich et al., 1999) as a template. The following primers were used: 5′-TTA GAT CTG ACG CTG CGG TGA TCA TCG-3′ and 5′-TTT AAG CTT ACG GCT TCA GAA GGC GCG-3′ for lanE and 5′-TTA GAT CTG GCA ATC TCA CCG GAA AA-3′ and 5′-TTT AAG CTT TTT CTC TCT CTT CGG AAT TCG GCT GCA-3′ for lanV. The amplified genes were subcloned into pBHBΔ (Kallio et al., 2006) as BglII/HindIII fragments. Constructs pBHBΔ/LanE and pBHBΔ/LanV were verified by sequencing (Eurofins MWG Operon), which revealed some inconsistencies in the lanE sequence in comparison to the deposited data after two independent cloning experiments. The revised sequence of lanE was deposited to GenBank under accession number JQ782417. LanE (53.7 kDa) and LanV (27.3 kDa) were over-expressed as N-terminal His-tagged fusion proteins in Escherichia coli strain TOP10 (400 mL LB medium in 2 L Erlenmeyer, 0.02% L-arabinose induction, incubation at 24 °C with 180 rpm shaking for 21 h). The proteins were purified to homogeneity in a single step using TALON Co2+-affinity resin (Clontech) in a disposable chromatography column (Bio-Rad) and eluted with 250 mM imidazole. Purity of the proteins was confirmed with sodium dodecyl sulfate-polyacrylamide gel electrophoresis and the concentration of the proteins was estimated using Bradford dye binding method. LanE and LanV were stored at −20 °C in 10 mM Na3PO4 * 12 H2O, 50 mM NaCl, 125 mM imidazole and 50% glycerol (pH 7.6).

Cloning, Expression and Purification of UrdE,UrdM, UrdMox, and UrdMred

The urdE, urdM, urdMox and urdMred genes were amplified with Phusion DNA polymerase (Finnzymes) using Streptomyces fradiae AO (Trefzer et al., 2001) chromosomal DNA as a template. The following primers were used: 5′-TTA GAT CTG ATG CTT CAG TCA TCG TCG-3′ and 5′-TTT AAG CTT GTC AGG ACT GCT GGA TCC-3′ for urdE, 5′-TTA GAT CTG TCG CGC CCT CTC TGG AC-3′ and 5′-TTT AAG CTT TCA TCC GCG TGA CTC AGC-3′ for urdM, 5′-TTA GAT CTG TCG CGC CCT CTC TGG AC-3′ and 5′-TTT AAG CTT TCA GCC GTC GAG CCC G-3′ for urdMox and 5′-TTA GAT CTG GCA AGC TCA CCG GAA AG-3′ and 5′-TTT AAG CTT TCA TCC GCG TGA CTC AGC-3′ for urdMred. The amplified genes were subcloned into pBHBΔ as BglII/HindIII fragments and the constructs pBHBΔ/UrdE, pBHBΔ/UrdM, pBHBΔ/UrdMox and pBHBΔ/UrdMred were verified by sequencing (Eurofins MWG Operon). In a similar manner to lanE, the sequence of the gene urdM was revised and deposited to GenBank under the accession number JQ782418. UrdE (54.7 kDa), UrdM (71.1 kDa), UrdMox (41.5 kDa) and UrdMred (27.2 kDa) were over-expressed as N-terminal His-tagged fusion proteins in E. coli strain TOP10 (500 mL 2×YT medium in 2 L Erlenmeyer, 0.02% L-arabinose induction, incubation at 25 °C with 250 rpm shaking for 21 h). The proteins were purified to homogeneity in a single step using a 5 mL HisTrap Ni2+-affinity column (Amersham Biosciences) and eluted with 0.5 M imidazole. Purity of the proteins was confirmed with sodium dodecyl sulfate-polyacrylamide gel electrophoresis and the concentration of the proteins was estimated using Bradford dye binding method. The proteins were stored at −20 °C in 10 mM Na3PO4 * 12 H2O, 75 mM NaCl, 125-250 mM imidazole and 50% glycerol (pH 7.4).

Expression and Purification of PgaE, CabE, CabV and JadH

The proteins PgaE (Koskiniemi et al., 2007), CabE (Koskiniemi et al., 2007), CabV (Kallio et al., 2011) and JadH (Chen et al., 2010) were heterologously produced and purified as described earlier.

Enzyme Reactions In Vitro

In vitro enzyme reactions were carried out in conditions described earlier (Kallio et al., 2011). The enzyme concentrations used in the reactions were 160 nM JadH, 150 nM PgaE, 600 nM CabE, 150 nM LanE, 350 nM UrdE and 150-6000 nM CabV or LanV. Isolation of substrate and products from in vitro reactions was performed with repeated chloroform extractions as described earlier (Kallio et al., 2008b). Solid phase extraction (SPE) columns (Discovery DSC-18/SUPELCO) were used to isolate 3 (Kallio et al., 2011). Reactions where 3 was used as a substrate were extracted with SPE columns, eluted with MeOH and incubated with PgaE and NADPH in order to degrade all unused substrate which would be seen as 7 in HPLC (Kallio et al., 2011). This analysis confirms that 7 seen in HPLC is 7 and not 3. Anaerobic reaction conditions were achieved using a Thunberg cuvette or glucose oxidase/catalase treatment as described earlier (Kallio et al., 2011).

HPLC Analysis of the Substrate and Products

The compounds extracted from reactions were analyzed by RP-HPLC (Shimadzu VP series chromatography system with a diode array detector,SPD-M10Avp, Merck LiChrospher 100 RP-18/5 μM column, gradient from 15% acetonitrile with 0.1% HCOOH to 100% acetonitrile, flow rate 0.5 mL/min, Ttot = 60 min). The substrate 2 and the products 5, 7 and 8 were identified based on retention times and UV-Vis spectra and compared on the basis of the peak areas at 406, 428, 459 or 449 nm, respectively.

LC-MS Analysis

The exact mass of the 8 from in vitro reaction was determined with a high resolution mass spectrometer (micrOTOF-Q, Bruker Daltonics) using negative ESI connected to RP-HPLC (Agilent Technologies 1200 series, Waters SunFire C18/3.5μm column, gradient from 15% acetonitrile with 0.1% HCOOH to 100% acetonitrile, flow rate 0.2 mL/min, Ttot = 45 min).

Molecular Modeling, Experimental and Calculated Optical Rotation (OR) and Electronic Circular Dichroism (ECD) Spectra

DFT quantum chemical calculations were performed using Gaussian09 (Frisch et al., 2009) (version A.01) and analyzed using GUIs GaussView (version 3.07) and GaussSum (O’Boyle et al., 2008) (version 2.2). Structures were first geometry-optimized using the M06-2X hybrid meta density functional (Zhao and Truhlar, 2008a; Zhao and Truhlar, 2008b) with the 6-31G(d) basis set. IR frequencies, ECD (Autschbach et al., 2002; Hansen and Bak, 1999) and OR (Ruud and Helgaker, 2002; Stephens et al., 2001; Stephens et al., 2002; Stephens et al., 2005; Stephens et al., 2008) properties were calculated for these structures at the same level of theory.

Optical rotations were measured at 22 °C using a Jasco DIP-360 polarimeter equipped with a 1 dm pathlength cell and the sodium D lines (589.3 nm) at the concentrations indicated in methanol. [α]22d for 5 is +20.6° (c 0.0669, methanol), measured; +309.3° (4aR,6S,12bS), calculated, for 2 +34.1° (c 0.0180, methanol), measured; +14.0° (4aR,12bR), calculated for a 1:1 ratio and for 8 - 71.2° (c 0.00503, methanol), measured; −62.7° (6R), calculated for 1.5:1 ratio.

ECD spectra were measured over the range of 200-650 nm at 25 °C using a Chirascan™ circular dichroism spectrometer equipped with a 1 cm pathlength cell at the concentrations indicated in methanol. The calculated ECD spectra were found to be sufficiently modeled for bands located at wavelengths longer than 200 nm (the experimental cutoff for methanol) using 24 excited states for single molecules since only minimal changes occurring for these bands when more than 24 excited states were used; in the case of the multiple-molecule calculations, the molecules were spaced ca. 200 Å apart and the number of excited states was increased proportionately. All calculations were performed in the gas phase and optimized structures are depicted in Figure S4 and their atomic coordinates in Table S1. Frequency calculations were conducted to confirm that the optimized structures were true minima on the potential energy surface by not providing imaginary frequencies and to obtain the thermodynamic contributions at 298.15 K (frequencies were unscaled). The experimental and calculated ECD spectra are presented in Figures 6 and S4.

Supplementary Material

01
02

Highlights.

  • Comparison of tailoring enzymes from five pathways led to synthesis of three products

  • The enzymes responsible for divergence of landomycin and jadomycin were identified

  • Many of the tailoring enzymes carried alternative, hidden or ancestral functions

  • The study clarified the evolutionary relationships of various angucycline metabolites

Acknowledgments

We thank Andriy Luzhetskyy and Andreas Bechthold for cosmid H2-26 and Streptomyces fradiae AO strain. This study was supported by the Academy of Finland (grant no.s 121688 to J.N., 127844 to P.M. and 136060 to M.M.-K.), the National Natural Science Foundation of China (grant no 31130001 to K.Y.), Turun Yliopistosäätiö (to K.D.K) and US National Institutes of Health (grants no CA091901 and CA102102 to J.R.). The Center for Scientific Computing is acknowledged for computational resources.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Antal N, Fiedler HP, Stackebrandt E, Beil W, Stroch K, Zeeck A. Retymicin, galtamycin B, saquayamycin Z and ribofuranosyllumichrome, novel secondary metabolites from Micromonospora sp. Tu 6368. I. Taxonomy, fermentation, isolation and biological activities. J. Antibiot. (Tokyo) 2005;58:95–102. doi: 10.1038/ja.2005.12. [DOI] [PubMed] [Google Scholar]
  2. Autschbach J, Ziegler T, van Gisbergen SJA, Baerends EJ. Chiroptical properties from time-dependent density functional theory. I. Circular dichroism spectra of organic molecule. J. Chem. Phys. 2002;116:6930–6940. [Google Scholar]
  3. Ayer SW, McInnes AG, Thibault P, Walter JA, Doull JL, Parnell T, Vining LC. Jadomycin, a novel 8H-benz[b]oxazolo[3,2-f]phenanthridine antibiotic from from Streptomyces venezuelae ISP5230. Tetrahedron Lett. 1991;32:6301–6304. [Google Scholar]
  4. Chen Y, Fan K, He Y, Xu X, Peng Y, Yu T, Jia C, Yang K. Characterization of JadH as an FAD- and NAD(P)H-dependent bifunctional hydroxylase/dehydrase in jadomycin biosynthesis. ChemBioChem. 2010;11:1055–1060. doi: 10.1002/cbic.201000178. [DOI] [PubMed] [Google Scholar]
  5. Chen YH, Wang CC, Greenwell L, Rix U, Hoffmeister D, Vining LC, Rohr J, Yang KQ. Functional analyses of oxygenases in jadomycin biosynthesis and identification of JadH as a bifunctional oxygenase/dehydrase. J. Biol. Chem. 2005;280:22508–22514. doi: 10.1074/jbc.M414229200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Copley SD. Enzymes with extra talents: moonlighting functions and catalytic promiscuity. Curr. Opin. Chem. Biol. 2003;7:265–272. doi: 10.1016/s1367-5931(03)00032-2. [DOI] [PubMed] [Google Scholar]
  7. Decker H, Haag S. Cloning and characterization of a polyketide synthase gene from Streptomyces fradiae Tu2717, which carries the genes for biosynthesis of the angucycline antibiotic urdamycin A and a gene probably involved in its oxygenation. J. Bacteriol. 1995;177:6126–6136. doi: 10.1128/jb.177.21.6126-6136.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Faust B, Hoffmeister D, Weitnauer G, Westrich L, Haag S, Schneider P, Decker H, Kunzel E, Rohr J, Bechthold A. Two new tailoring enzymes, a glycosyltransferase and an oxygenase, involved in biosynthesis of the angucycline antibiotic urdamycin A in Streptomyces fradiae Tu2717. Microbiology. 2000;146:147–154. doi: 10.1099/00221287-146-1-147. [DOI] [PubMed] [Google Scholar]
  9. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Scalmani G, Barone V, Mennucci B, Petersson GA, et al. Gaussian09, Revision A.01. Gaussian, Inc; 2009. [Google Scholar]
  10. Hansen AE, Bak KL. Ab-initio calculations of electronic circular dichroism. Enantiomer. 1999;4:455–476. [Google Scholar]
  11. Hult K, Berglund P. Enzyme promiscuity: mechanism and applications. Trends Biotechnol. 2007;25:231–238. doi: 10.1016/j.tibtech.2007.03.002. [DOI] [PubMed] [Google Scholar]
  12. Jakeman DL, Bandi S, Graham CL, Reid TR, Wentzell JR, Douglas SE. Antimicrobial activities of jadomycin B and structurally related analogues. Antimicrob. Agents Chemother. 2009;53:1245–1247. doi: 10.1128/AAC.00801-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jornvall H, Persson B, Krook M, Atrian S, Gonzalez-Duarte R, Jeffery J, Ghosh D. Short-chain dehydrogenases/reductases (SDR) Biochemistry. 1995;34:6003–6013. doi: 10.1021/bi00018a001. [DOI] [PubMed] [Google Scholar]
  14. Kallio P, Sultana A, Niemi J, Mäntsälä P, Schneider G. Crystal structure of the polyketide cyclase AknH with bound substrate and product analogue: implications for catalytic mechanism and product stereoselectivity. J. Mol. Biol. 2006;357:210–220. doi: 10.1016/j.jmb.2005.12.064. [DOI] [PubMed] [Google Scholar]
  15. Kallio P, Liu Z, Mäntsälä P, Niemi J, Metsä-Ketelä M. A nested gene in Streptomyces bacteria encodes a protein involved in quaternary complex formation. J. Mol. Biol. 2008a;375:1212–1221. doi: 10.1016/j.jmb.2007.11.044. [DOI] [PubMed] [Google Scholar]
  16. Kallio P, Liu Z, Mäntsälä P, Niemi J, Metsä-Ketelä M. Sequential action of two flavoenzymes, PgaE and PgaM, in angucycline biosynthesis: chemoenzymatic synthesis of gaudimycin C. Chem. Biol. 2008b;15:157–166. doi: 10.1016/j.chembiol.2007.12.011. [DOI] [PubMed] [Google Scholar]
  17. Kallio P, Patrikainen P, Suomela JP, Mäntsälä P, Metsä-Ketelä M, Niemi J. Flavoprotein hydroxylase PgaE catalyzes two consecutive oxygen-dependent tailoring reactions in angucycline biosynthesis. Biochemistry. 2011;50:5535–5543. doi: 10.1021/bi200600k. [DOI] [PubMed] [Google Scholar]
  18. Kavanagh KL, Jornvall H, Persson B, Oppermann U. Medium- and short-chain dehydrogenase/reductase gene and protein families: the SDR superfamily: functional and structural diversity within a family of metabolic and regulatory enzymes. Cell Mol. Life Sci. 2008;65:3895–3906. doi: 10.1007/s00018-008-8588-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kharel MK, Pahari P, Shaaban KA, Wang G, Morris C, Rohr J. Elucidation of post-PKS tailoring steps involved in landomycin biosynthesis. Org. Biomol. Chem. 2012 doi: 10.1039/c2ob07171a. DOI: 10.1039/C2OB07171A. [DOI] [PubMed] [Google Scholar]
  20. Khersonsky O, Roodveldt C, Tawfik DS. Enzyme promiscuity: evolutionary and mechanistic aspects. Curr. Opin. Chem. Biol. 2006;10:498–508. doi: 10.1016/j.cbpa.2006.08.011. [DOI] [PubMed] [Google Scholar]
  21. Korynevska A, Heffeter P, Matselyukh B, Elbling L, Micksche M, Stoika R, Berger W. Mechanisms underlying the anticancer activities of the angucycline landomycin E. Biochem. Pharmacol. 2007;74:1713–1726. doi: 10.1016/j.bcp.2007.08.026. [DOI] [PubMed] [Google Scholar]
  22. Koskiniemi H, Metsä-Ketelä M, Dobritzsch D, Kallio P, Korhonen H, Mäntsälä P, Schneider G, Niemi J. Crystal structures of two aromatic hydroxylases involved in the early tailoring steps of angucycline biosynthesis. J. Mol. Biol. 2007;372:633–648. doi: 10.1016/j.jmb.2007.06.087. [DOI] [PubMed] [Google Scholar]
  23. Kulowski K, Wendt-Pienkowski E, Han L, Yang K, Vining LC, Hutchinson CR. Functional characterization of the jadI gene as a cyclase forming angucyclinones. J. Am. Chem. Soc. 1999;121:1786–1794. [Google Scholar]
  24. Lombo F, Abdelfattah MS, Brana AF, Salas JA, Rohr J, Mendez C. Elucidation of oxygenation steps during oviedomycin biosynthesis and generation of derivatives with increased antitumor activity. Chembiochem. 2009;10:296–303. doi: 10.1002/cbic.200800425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mayer A, Taguchi T, Linnenbrink A, Hofmann C, Luzhetskyy A, Bechthold A. LanV, a bifunctional enzyme: aromatase and ketoreductase during landomycin A biosynthesis. ChemBioChem. 2005;6:2312–2315. doi: 10.1002/cbic.200500205. [DOI] [PubMed] [Google Scholar]
  26. Metsä-Ketelä M, Palmu K, Kunnari T, Ylihonko K, Mäntsälä P. Engineering anthracycline biosynthesis toward angucyclines. Antimicrob. Agents Chemother. 2003;47:1291–1296. doi: 10.1128/AAC.47.4.1291-1296.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Niemi J, Metsä-Ketelä M, Schneider G, Mäntsälä P. Biosynthetic Anthracycline Variants. Top. Curr. Chem. 2008;282:75–99. [Google Scholar]
  28. O’Boyle NM, Tenderholt AL, Langner KM. Cclib: a Library for Package-Independent Computational Chemistry Algorithms. J. Comput. Chem. 2008;29:839–845. doi: 10.1002/jcc.20823. [DOI] [PubMed] [Google Scholar]
  29. Pahari P, Kharel MK, Shepherd MD, van Lanen SG, Rohr J. Enzymatic total synthesis of defucogilvocarcin M and its implications for gilvocarcin biosynthesis. Angew. Chem. Int. Ed. Engl. 2012;51:1216–1220. doi: 10.1002/anie.201105882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Palmu K, Ishida K, Mäntsälä P, Hertweck C, Metsä-Ketelä M. Artificial reconstruction of two cryptic angucycline antibiotic biosynthetic pathways. ChemBioChem. 2007;8:1577–1584. doi: 10.1002/cbic.200700140. [DOI] [PubMed] [Google Scholar]
  31. Persson B, Kallberg Y, Oppermann U, Jornvall H. Coenzyme-based functional assignments of short-chain dehydrogenases/reductases (SDRs) Chem. Biol. Interact. 2003;143-144:271–278. doi: 10.1016/s0009-2797(02)00223-5. [DOI] [PubMed] [Google Scholar]
  32. Rix U, Fischer C, Remsing LL, Rohr J. Modification of post-PKS tailoring steps through combinatorial biosynthesis. Nat. Prod. Rep. 2002;19:542–580. doi: 10.1039/b103920m. [DOI] [PubMed] [Google Scholar]
  33. Rohr J, Zeeck A. Metabolic products of microorganisms. 240. Urdamycins, new angucycline antibiotics from Streptomyces fradiae. II. Structural studies of urdamycins B to F. J. Antibiot. 1987;40:459–467. doi: 10.7164/antibiotics.40.459. [DOI] [PubMed] [Google Scholar]
  34. Roush WR, Neitz RJ. Studies on the synthesis of landomycin A. Synthesis of the originally assigned structure of the aglycone, landomycinone, and revision of structure. J. Org. Chem. 2004;69:4906–4912. doi: 10.1021/jo049426c. [DOI] [PubMed] [Google Scholar]
  35. Ruud K, Helgaker T. Optical rotation studied by density-functional and coupled-cluster methods. Chem. Phys. Lett. 2002;352:533–539. [Google Scholar]
  36. Shaaban KA, Stamatkin C, Damodaran C, Rohr J. 11-Deoxylandomycinone and landomycins X-Z, new cytotoxic angucyclin(on)es from a Streptomyces cyanogenus K62 mutant strain. J. Antibiot. 2011;64:141–150. doi: 10.1038/ja.2010.121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Stephens PJ, Devlin FJ, Cheeseman JR, Frisch MJ. Calculation of Optical Rotation Using Density Functional Theory. J. Phys. Chem. A. 2001;105:5356–5371. [Google Scholar]
  38. Stephens PJ, Devlin FJ, Cheeseman JR, Frisch MJ, Rosini C. Determination of absolute configuration using optical rotation calculated using density functional theory. Org. Lett. 2002;4:4595–4598. doi: 10.1021/ol0201714. [DOI] [PubMed] [Google Scholar]
  39. Stephens PJ, McCann DM, Cheeseman JR, Frisch MJ. Determination of absolute configurations of chiral molecules using ab initio time-dependent Density Functional Theory calculations of optical rotation: how reliable are absolute configurations obtained for molecules with small rotations? Chirality. 2005;17:S52–64. doi: 10.1002/chir.20109. [DOI] [PubMed] [Google Scholar]
  40. Stephens PJ, Pan JJ, Devlin FJ, Cheeseman JR. Determination of the absolute configurations of natural products using TDDFT optical rotation calculations: the iridoid oruwacin. J. Nat. Prod. 2008;71:285–288. doi: 10.1021/np070502r. [DOI] [PubMed] [Google Scholar]
  41. Trefzer A, Fischer C, Stockert S, Westrich L, Kunzel E, Girreser U, Rohr J, Bechthold A. Elucidation of the function of two glycosyltransferase genes (lanGT1 and lanGT4) involved in landomycin biosynthesis and generation of new oligosaccharide antibiotics. Chem. Biol. 2001;8:1239–1252. doi: 10.1016/s1074-5521(01)00091-6. [DOI] [PubMed] [Google Scholar]
  42. Weber S, Zolke C, Rohr J, Beale JM. Investigations of the biosynthesis and structural revision of landomycin A. J. Org. Chem. 1994;59:4211–4214. [Google Scholar]
  43. Westrich L, Domann S, Faust B, Bedford D, Hopwood DA, Bechthold A. Cloning and characterization of a gene cluster from Streptomyces cyanogenus S136 probably involved in landomycin biosynthesis. FEMS Microbiol. Lett. 1999;170:381–387. doi: 10.1111/j.1574-6968.1999.tb13398.x. [DOI] [PubMed] [Google Scholar]
  44. Zhao Y, Truhlar DG. Density functionals with broad applicability in chemistry. Acc. Chem. Res. 2008a;41:157–167. doi: 10.1021/ar700111a. [DOI] [PubMed] [Google Scholar]
  45. Zhao Y, Truhlar D. The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor. Chem. Acc. 2008b;120:215–241. [Google Scholar]
  46. Zhu L, Ostash B, Rix U, Nur-E-Alam M, Mayers A, Luzhetskyy A, Mendez C, Salas JA, Bechthold A, Fedorenko V, Rohr J. Identification of the function of gene lndM2 encoding a bifunctional oxygenase-reductase involved in the biosynthesis of the antitumor antibiotic landomycin E by Streptomyces globisporus 1912 supports the originally assigned structure for landomycinone. J. Org. Chem. 2005;70:631–638. doi: 10.1021/jo0483623. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS372073-supplement-01.doc (620.5KB, doc)
02
NIHMS372073-supplement-02.xlsx (19.3KB, xlsx)

RESOURCES