PROCESS FOR THE PREPARATION OF SIMVASTATIN
Field of the invention
The field of the invention relates to a process for the preparation of simvastatin and analogues thereof through expression of customised hybrid modular polyketide synthases in a host, preferably one containing the lovastatin-producing polyketide synthase gene cluster. Background of the invention
The compounds lovastatin, pravastatin, mevastatin and simvastatin are highly potent anticholesterolaemic agents that act as inhibitors of HMG-CoA reductase, an enzyme involved in the rate-limiting step in cholesterol biosynthesis in humans. With the exception of simvastatin that is a synthetically made drug, the above-mentioned compounds are naturally occurring and are obtained through fermentation. Lovastatin, a product of Aspergillus terreus, possesses a 2-methylbutyrate side chain linked at C-8 position of the hexahydronaphthalene ring system. Simvastatin on the other hand possesses a 2,2-dimethylbutyrate side chain at the C-8 position. This subtle sutructural difference has been shown to render simvastatin much more potent than lovastatin.
The increased potency of simvastatin over its analogues has necessitated a need for a chemical synthesis that is efficient, eco-friendly, economically viable and high yielding. Numerous chemical syntheses of simvastatin have been reported worldwide since its discovery in 1984.
One route (Hoffman et al, 1984) involves deesterification of the butyrate side- chain of lovastation, followed by several distinct chemical steps that involve lactonization, hydroxy group protection/deprotection and reesterification with the appropriate 2,2- dimethylbutyrate side chain. This process results in low overall yield.
Another method (Sleiteinger et al, 1986) involves direct methylation of the lovastatin side chain using a metal alkylamide and methylhalide. This method suffers from problems concerning the purity of the final product as it leads to several by-products that have to be separated from the target compound. Another method (Verhoeven et al, 1989), while addressing the problems of low overall yield and purity, involves as many as six chemical steps that utilize reagents that are unsafe to handle on industrial scale.
The recent method of Kumar et al (1998) involves conversion of lovastatin to simvastatin using fewer chemical steps. However, this method employs expensive chemical reagents and also results in low overall yield.
The present state of the art for the production of simvastatin, therefore, leaves a lot to be desired and there is an urgent need to obtain the compound by means other than chemical, thereby avoiding the above-mentioned problems that are associated with chemical syntheses. Objects of the invention
It is therefore, the object of the present invention to provide a process for the preparation of simvastatin or analogues or derivatives thereof which avoids the drawbacks of prior art methods.
It is another important object of the present invention to provide a process for the preparation of simvastatin by a non chemical route.
It is another also the aim of the invention to provide a process resulting in high yield and purity.
Summary of the invention
The above and other objects of the present invention are achieved by the method of the present inventoin by biological expression, wherein said method comprises: a), providing a host having a customised gene encoding for a polyketide synthase. b). fermenting the said host to obtain simvastatin or analogs or derivatives thereof.
Another aspect of the invention relates to a polyketide synthase gene that is customised to obtain the 2,2-dimethylbutyrate side chain of simvastatin.
Another aspect of the invention relates to the host being a eukaryote or a prokaryote. Another aspect of the invention relates to the host being a lovastain-producer.
Another aspect of the invention relates to the host being selected from Aspergillus spp..
Another aspect of the invention relates to the Aspergillus spp. being the lovastain- producing Aspergillus terreus. Another aspect of the invention relates to the host being the lovastain-producing
Aspergillus terreus mutant that is unable to produce the 2-methylbutyrate side chain.
Another aspect of the invention relates to the host being Aspergillus terreus/ARl.
Another aspect of the invention relates to a method wherein said domains for the customised polyketide synthzase are obtained from fatty acid synthase.
Another aspect of the invention relates to a method wherein the selected domain nucleic acids are selected from the group comprising type I modular polyketide synthase, type II PKS and type I iterative PKS.
Another aspect of the invention relates to a method wherein the selected domain nucleic acids are type I modular PKS.
Another aspect of the invention relates to a method wherein the polyketide synthase gene is one selected from the group provided in Figures 6.
Another aspect of the invention relates to a method wherein the polyketide synthase gene is one selected from the group provided in Figures 8. The invention also provides for simvastatin or analogs or derivatives thereof whenever prepared by the novel process.
The invention also provides for a host as modified according to the objectives of the invention.
The invention also provides for a construct encoding for a polyketide synthase according to the objectives of the invention. Description of the drawings
Figure 1 The anticholesterolaemic compounds that act as HMG-CoA reductase inhibitors.
Figure 2 The primary organization of genes and their corresponding proteins of the erythromycin-producing type I modular polyketide synthase.
Figure 3 The primary organization of genes and their corresponding proteins of the lovastatin-producing polyketide synthase. The hypothetical iterative functioning of the
LNKS is illustrated by analogy to the ery PKS. The diketide is synthesised by LDKS.
LNKS - lovastatin nonaketide synthase; LDKS - lovastatin diketide synthase; MeT - methyl transferase.
Figure 4 Domain organization in lovastatin diketide synthase (LDKS).
Figure 5 The composition of hybrid modular polyketide synthase A template (HMPKSA) that is customised to synthesise the 2,2-dimethylbutyrate side chain (22DMB). Figure 6 The hybrid modular polyketide synthase Al-4 (HMPKSA1-4) constructed to synthesise 22DMB; the engineered sequence that is at the junctions of the assembled fragments is shown in capitals.
Figure 7 The composition of the hybrid modular polyketide synthase B template (HMPKSB) that is customised to synthesise analogues of 22DMB.
Figure 8 The hybrid modular polyketide synthase Bl-2 (HMPKSB1-2) constructed to synthesise analogue of 22DMB.
Figure 9 Integration of plasmid pSIM10H4S through homologous double recombination into the wild-type A. terreus strain resulting in the construction of a lovF mutant.
Detailed description of the invention
The invention describes a process for the synthesis of simvastatin. The process comprises of the following steps: i). construction of a customised gene coding for a polyketide synthase (PKS) that is geared to synthesise the 2,2-dimethylbutyrate side chain of simvastatin. ii). transforming a host, preferably one containing the lovastatin-producing PKS gene cluster with a plasmid bearing the above said PKS gene and fermentation thereof to produce simvastatin.
By describing the process herein with reference to simvastatin, the process and the invention also envisages analogs and derivatives thereof and references to simvastatin for the purposes of the specification also include references to its analogs and derivatives.
The host selected may be any prokaryote or a eukaryote that expresses the lovastatin PKS. Therefore, any host in addition to those know to express lovastatin PKS may also be employed for the transformation thereof with a plasmid bearing the 2,2- dimethylbutyrate synthesizing PKS gene.
The process described herein dispenses with the number of steps normally required for the synthesis of simvastatin as compared with the chemical syntheses of the target that have been reported so far in the scientific literature. The primary factor responsible for this is an aspect of the invention that involves the synthesis of the 2,2-dimethylbutyrate side chain of simvastatin through the use of customised PKS enzymes in a host, especially a lovastatin-producing one. The use of such enzymes for synthesis of the side chain followed by its in vivo priming of the hexahydronaphthalene ring of simvastatin circumvents the production thereof by the lengthy chemical synthesis route. Moreover, as the customised PKS enzymes are expressed by a host organism, the cost of synthesising the drug is drastically reduced when compared with chemical synthesis. Also, the tight stereocontrol that is normally the feature of a PKS-catalysed reaction provides an advantage over chemical synthesis.
Polyketides such as erythromycin, tetracycline and lovastatin are all synthesised by a class of enzymes called the Polyketide synthases (PKS). PKSs have been classified into
essentially three groups - type I, type II and type I iterative. PKSs of the modular variety (called as type I PKS) are giant multienzymes (with a molecular weight ranging from 100- lOOOKDa) that synthesise medicinally valuable drugs like the antibacterial erythromycin, immunosuppressant rapamycin and the anticancer epothilone B. Type I PKSs are composed of enzymes like the ketoacyl synthase (KS), acyl carrier protein (ACP), acyl transferase (AT), dehydratase (DH), enoyl reductase (ER), keto reductase (KR) and thioesterase (TE; Donadio et al, 1991; Schwecke et al, 1995). These enzymes (also known as domains) are covalently linked with each other to form multienzymes. The multienzymes, through composite domains, catalyse the condensation and subsequent processing of small carbon units, ultimately giving rise to complex polyketides like erythromycin and rapamycin. By processing, it is meant that chemical groups belonging to the carbon units are altered or removed. The deciphering of the assembly-line like nature of PKSs (Figure 2) has resulted in numerous genetic engineering efforts in order to make drug analogues and new bio-active molecules. Investigations have shown that the inherent so-called 'programming' of PKSs can be altered by adding, removing or inactivating the PKS domains or sets thereof, called modules (McDaniel et al, 1999; Ranganathan et al, 1999, Gokhale et al, 1999). The resulting hybrid PKSs then function as directed to synthesise the desired target polyketide. There are now more than four hundred PKS domains whose DNA sequence information is publicly available. Type II PKSs on the other hand are composed of a set of discrete domains that interact with each other by forming a freely dissociable complex. Type II systems synthesise aromatic polyketides like tetracyclines.
The third class of PKSs are the type I iterative. These are systems where the polyketide part of the molecule is synthesised by just one module (that has covalently linked domains) that functions iteratively. Lovastatin and methylsalicylic acid are synthesised by this class of polyketide synthases. The type I iterative module shows significant homology to animal fatty acid synthases, and like the latter, it possesses all the reduction-domains (DH, ER and KR).
Studies on the biosynthesis of lovastatin using separate incorporation of 13C-, 2H- and 18O-labelled acetate, 13C-methionine and I8O2 have shown that it is a polyketide having a backbone of nine intact acetate units linked in a "head-to-tail" manner (Yoshizawa et al, 1994; Moore et al, 1985).
The lovastatin biosynthetic gene cluster has recently been isolated from A. terreus and sequenced (Figure 3; Kennedy et al., 1999). Two large open reading frames lovB and
lovF have been identified as those coding for the nonaketide-producing 'lovastatin nonaketide synthase' (LNKS) and 'lovastatin diketide synthase' (LDKS) respectively. LNKS has been shown to be a type I iterative module while LDKS is predicted to be non- iterative, much like the type I DEBS modules that synthesise erythromycin (Figure 2). Kennedy and coworkers have proposed that the 500-odd amino acid region at the C- terminus of LNKS is important for LNKS-LDKS intermodular interaction and necessary for the 2-methylbutyrate diketide chain to be transferred onto the nonaketide part of lovastatin. Another gene, lovD has been proposed to code for a transesterase that could help in bringing the two polyketide chains together. An A. terreus mutant with the inactivated LDKS gene (lovF) was shown to accumulate monacolin J, the immediate precursor to lovastatin that is devoid of the side chain, thus directly implicating LDKS in 2-methylbutyrate synthesis (Kenedy et al, 1999).
The LDKS-catalysed polyketide biosynthesis starts through the uptake of a small carbon unit called starter acid that primes the KS (defined as KS in LDKS, Figure 4) of the PKS and is one of the units involved in the first condensation reaction between the said starter acid (in this case acetyl-CoA) and extension acid (malonyl-CoA) that is selected by the AT domain immediately downstream of the KS. The sequential order of domains present in LDKS is as shown in Figure 4. The presence of a methyltransferase domain (MeT) in LDKS argues well for the addition of a methyl group at the C-2 position being catalysed by this enzyme. Recently, PKS modules containing such "embedded" MeT domains have been reported for the yerseniabactin and epothilone PKSs (Suo et al, 2001; Molnar et al, 2000).
In the case of a type I modular PKS like the one shown in Figure 2, polyketide biosynthesis starts through the uptake of the starter acid that primes the first AT (defined as ATO in HMPKSB, Figure 7) of the PKS and, like in LDKS, it is one of the units involved in the first condensation reaction between the said starter acid and extension acid that is selected by the AT domain immediately downstream of the first AT. The choice of the starter acid is dependent upon the degree of stringency that the first AT domain employs in the uptake. For example, the first AT domain of the erythromycin-producing PKS specifically selects propionyl coenzyme A (propionylCoA) as the starter acid whereas the first AT domain of the avermectin-producing PKS has been shown to accept a wide variety of starter acids (Dutton et al, 1991). The above said avermectin PKS domain is thus most suited for generating polyketide molecules that require to be synthesised from starter acids that are not used naturally by any of the PKS enzymes known in the art.
Therefore, preferred first AT domains in some of the customised type I PKSs envisaged by the invention is from the avermectin-producing PKS.
The 2,2-dimethylbutyrate side chain of simvastatin is synthesised through the use of customised PKSs. The customisation is based on identification of the stereochemical features of the said side chain. By customisation, it is meant that the domain selection is carried out in view of the identified stereochemical features.
For the construction of customised PKSs, the DNA fragments that correspond to the desired domains are Ugated with each other to give a gene that is expressed in a suitable host in order to obtain the above said side chain. The blueprint for the customised PKS therefore is the customised gene that is incorporated in a host.
The desired PKS DNA fragments are assembled by any one of the methods known in the art. Particularly preferred is the method taught in Ranganathan (2000) as it is a time and cost-effective alternative to the conventional methods, especially suited for multi- component DNA assembly. The incorporation of the customised PKS genes into host organisms is accomplished through the use of molecular biology techniques that are known to a person of skill (see Methods). Hosts that are envisaged for the synthesis of the simvastatin side chain are of the prokaryotic and eukaryotic variety, such as actinomycetes, Aspregillus spp., s/9 line of insect cells, P. pastoris, and especially Aspregillus spp. In particular, the invention discloses the transformation of a mutant strain of the lovastatin-producing Aspregillus terreus that is unable to synthesize the LDKS protein with the plasmid bearing the customised PKS gene. This procedure results in the replacement of the 2-methylbutyrate side chain with the 2,2-dimethylbutyrate side chain in the above said lovastatin-producing strain, ultimately resulting in the production of only simvastatin from the same strain.
Without prejudice to the spirit or scope of the invention, the invention will be described with reference to the following illustrative but non-limiting Examples: Example 1
In accordance with the preceding disclosure, the synthesis of simvastatin was carried out by keeping to the following sequence: i). Synthesis of the 2,2-dimethylbutyrate side chain (22DMB) of simvastatin. ii). Transformation of a mutant of A. terreus with a plasmid bearing the 22DMB- producing PKS gene.
Example 1.1 Synthesis of 22DMB
22DMB was synthesised in a host through the expression of customised PKSs HMPKSAl -4 (Figure 6). Example 1.1.1
Composition of the hybrid modular polyketide synthase A (HMPKSA) template
The HMPKS A template (Figure 5) comprises of the following enzymatic domains in the given sequence: keto synthase - KS1, acyltransf erase - ATI, dehydratase - DH1, methyltransferase - MeTl, enoylreductase - ER1, ketoreductase - KR1 and acyl carrier protein - ACPI. The domains are covalently linked with each other, to give a multi- enzyme HMPKS A template. Example 1.1.2
Construction of the hybrid modular polyketide synthase Al (HMPKSAl) i). Selection of PKS domain DNA sequence For the construction of HMPKSAl, the selection of domains from publicly available DNA sequences was as follows: KS 1 - from LDKS of the lovastatin PKS ATI - from module 7 of the epothilone PKS DH1 - from LDKS of the lovastatin PKS MeTl - from LDKS of the lovastatin PKS ER1 - from LDKS of the lovastatin PKS KR1 - from LDKS of the lovastatin PKS ACPI - from LDKS of the lovastatin PKS ii). Amplification of the selected PKS domain DNA The above said PKS domain DNA sequences were amplified from the genomic
DNA, using the Polymerase Chain Reaction (PCR) and by following a modification of the protocol of Saiki et al (1985). For ease of DNA fragment assembly, some of the above said domains were co-amplified as single DNA fragments (Figure 6); the total number of fragments that were amplified was four, out of which two fragments (3 and 4) contained DNA sequences corresponding to multiple domains. The source of DNA that acted as template for PCR was as indicated below. The nucleotide positions in the template are shown according to the same disclosed in prior art (Kennedy et al, 1999; Molnar et al, 2000). The oligonucleotide primers (shown as 5' to 3') used for each PCR are:
iii). Assembly of the above said PKS domain DNA fragments in order to construct the final expression plasmid pARHMPKSAl
As the construction of pARHMPKSAl requires a multi-component DNA assembly, the method of Ranganthan (2000) was employed for the purpose. The DNA sequences corresponding to the constituent domains of HMPKSAl were amplified by
PCR to incorporate the two recognition sequences for Xbal (5'TCTAGA3' and 5'TCTAGATC3') at the 5' and 3' ends of the DNA fragment respectively. The DNA sequence 5OATC3' is recognised by the Dam methylase gene of E. coli (Geier and Modrich, 1979). All PCR products were phosphorylated and ligated to Smal-cut, dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers. No mistakes in the amplified products were detected. All four plasmids were then used to transform the E.coli ET12567 dam" strain (MacNeil et al, 1992). Isolated DNA was then digested with Xbal and the desired fragments isolated and purified. Fragment 4 was then ligated to Xbal-cut pCJR24HS vector that is derived from the S. erythraea expression vector pCJR24 (Rowe et al, 1998). The pCJR24HS vector has a unique Xbal site and contains hygromycin as well as thiostrepton-resistance gene as markers for identifying successful integrands. Construction of pCJR24HS is described in the Methods section. The ligation products were used to transform E. coli DH10B electrocompetent cells and the desired plasmid was isolated using ampicillin as a resistance marker. This plasmid (pSIMAl) can only be singly cleaved with Xbal despite possessing two Xbal recognition sequences, as one of the sites (situated at the 3 ' end of 4) has been methylated by the E. coli Dam methylase. Plasmid pSIMAl was digested with Xbal, ligated to fragment 3, and the ligation products treated as mentioned above to yield pSIMA2. DNA fragment 2 was then added to finally yield plasmid pSIMA3. This plasmid was then digested with Ndel and Xbal and ligated with the KS1 fragment (1) previously digested with the same two enzymes. The ligated products were used to transform E. coli DH10B electrocompetent cells and the final expression plasmid pARHMPKSAl, containing the customised HMPKSAl gene, was isolated. The junctions where the domains were joined are shown in Figure 6. iv). Expression of plasmid pARHMPKSAl in a host
Plasmid pARHMPKSAl was used to transform the mutant strain A. terreus/ARl (for the construction of this strain see Methods 1.3). Particularly preferred was the transformation procedure of Punt et al. (1992). Thiostrepton-resistant colonies were selected upon integration of the vector into the A. terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin and thiostrepton. v). Growth of host
The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of simvastatin The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 900 mg of crude product. A sample of this crude product was analysed by GC-MS. The peak corresponding to the molecular mass 418 was observed, indicating the presence of simvastatin. The latter was subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR. The preferred method for isolation of simvastatin was based on the findings of Kennedy et al. (1999). 1 litre of fermentation broth was found to produce 51 mg of simvastatin.
Example 1.1.3
Construction of the hybrid modular polyketide synthase A2 (HMPKSA2) i). Selection of PKS domain DNA sequence
For the construction of HMPKSA2, the selection of domains from publicly available DNA sequences was as follows:
KS 1 - from LDKS of the lovastatin PKS
ATI - from mas of the M tuberculosis H37Rv DH1 - from LDKS of the lovastatin PKS
MeTl - from LDKS of the lovastatin PKS
ER1 - from LDKS of the lovastatin PKS
KR1 - from LDKS of the lovastatin PKS
ACPI - from LDKS of the lovastatin PKS ii). Amplification of the selected PKS domain DNA
The above said PKS domain DNA sequences were amplified from genomic DNA using PCR as described in Example 1.1.2. Only one DNA fragment needed to be amplified (fragment 5). The rest of the DNA fragments required for the construction of
HMPKSA2, namely fragments 1, 3 and 4 had been prepared earlier, as described in Example 1.1.2. The nucleotide positions in the template are shown according to the same disclosed in prior art (Cole et al., 1998). The source of DNA, nucleotide positions of the template and the oligonucleotide primers (shown as 5' to 3') used for each PCR were as follows:
iii). Assembly of the above said PKS domain DNA fragments in order to construct the final expression plasmid pARHMPKSA2 For the construction of pARHMPKSA2, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. The PCR products were phosphorylated and ligated to Smal-cut, dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers. No mistakes in the amplified products were detected. The plasmids were then used to transform the E.coli ET12567 dam" strain. Isolated DNA was digested with Xbal and the desired fragments isolated and purified. Fragment 4 was then cloned in Xbal-c t pCJR24HS vector using the procedure described in Example 1.1.2. The resulting plasmid pSIMAl was digested with Xbal and the other DNA fragments, namely, 3 and 5 were sequentially added to yield plasmid pSIMA23. This plasmid was then digested with Ndel and Xbal and ligated with fragment 1 previously digested with the same two enzymes. The ligated products were used to transform E. coli DH10B electrocompetent cells and the final expression plasmid pARHMPKSA2, containing the customised HMPKSA2 gene, was isolated. The junctions where the domains were joined are shown in Figure 6. iv). Expression of plasmid pARHMPKSA2 in a host
Plasmid pARHMPKSA2 was used to transform the mutant strain A. terreus/ARl. Particularly preferred was the transformation procedure of Punt et al. (1992). Thiostrepton-resistant colonies were selected upon integration of the vector into the A.
terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin and thiostrepton. v). Growth of host
The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of simvastatin
The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 1.1 g of crude product. A sample of this crude product was analysed by GC-MS. The peak corresponding to the molecular mass 418 was observed, indicating the presence of simvastatin. The latter was subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR. The preferred method for isolation of simvastatin was based on the findings of Kennedy et al. (1999). 1 litre of fermentation broth was found to produce 44 mg of simvastatin.
Example 1.1.4
Construction of the hybrid modular polyketide synthase A3 (HMPKSA3) i). Selection of PKS domain DNA sequence
For the construction of HMPKSA3, the selection of domains from publicly available DNA sequences was as follows:
KS 1 - from LDKS of the lovastatin PKS
ATI - from module 7 of the epothilone PKS
DH1 - from module 7 of the epothilone PKS
MeTl - from module 7 of the epothilone PKS ER1 - from LDKS of the lovastatin PKS
KR1 - from LDKS of the lovastatin PKS
ACPI - from LDKS of the lovastatin PKS ii). Amplification of the selected PKS domain DNA
The above said PKS domain DNA sequences were amplified from genomic DNA using PCR as described in Example 1.1.2. The DNA fragment that needed to be amplified
(fragment 6) contained DNA sequences corresponding to two domains (Figure 6). The rest of the DNA fragments required for the construction of HMPKS A2, namely fragments 1, 2 and 4 had been prepared earlier, as described in Example 1.1.2.
The source of DNA, nucleotide positions of the template and the oligonucleotide primers (shown as 5' to 3') used for each PCR were as follows:
iii). Assembly of the above said PKS domain DNA fragments in order to construct the final expression plasmid pARHMPKSA3
For the construction of pARHMPKSA3, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. The PCR products were phosphorylated and ligated to Smal-c t, dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers. No mistakes in the amplified products were detected. The plasmids were then used to transform the E.coli ET12567 dam" strain. Isolated DNA was digested with Xbal and the desired fragments isolated and purified. Fragment 4 was then cloned in Xbal-cut pCJR24HS vector using the procedure described in Example 1.1.2. The resulting plasmid pSIMAl was digested with Xbal and the other DNA fragments, namely 6 and 2 were sequentially added to finally yield plasmid pSEVIA33. This plasmid was then digested with Ndel and Xbal and ligated with fragment 1 previously digested with the same two enzymes. The ligated products were used to transform E. coli DH10B electrocompetent cells and the final expression plasmid pARHMPKSA3, containing the customised HMPKS A3 gene, was isolated. The junctions where the domains were joined are shown in Figure 6. iv). Expression of plasmid pARHMPKSA3 in a host
Plasmid pARHMPKSA3 was used to transform the mutant strain terreus/ARl. Particularly preferred was the transformation procedure of Punt et al. (1992).
Thiostrepton-resistant colonies were selected upon integration of the vector into the A. terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin and thiostrepton. v). Growth of host The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of simvastatin
The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 1.0 g of crude product. A sample of this crude product was analysed by GC-MS. The peak corresponding to the molecular mass 418 was observed, indicating the presence of simvastatin. The latter was subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR. The preferred method for isolation of simvastatin was based on the findings of Kennedy et al. (1999). 1 litre of fermentation broth was found to produce 70 mg of simvastatin.
Example 1.1.5
Construction of the hybrid modular polyketide synthase A4 (HMPKSA4) i). Selection of PKS domain DNA sequence For the construction of HMPKS A4, the selection of domains from publicly available DNA sequences was as follows:
KS 1 - from LDKS of the lovastatin PKS
ATI - from module 4 of the erythromycin PKS
DH1 - from module 4 of the erythromycin PKS MeT 1 - from module HMWP 1 of the yersiniabactin PKS
ER1 - from LDKS of the lovastatin PKS
KR1 - from LDKS of the lovastatin PKS
ACPI - from LDKS of the lovastatin PKS ii). Amplification of the selected PKS domain DNA The above said PKS domain DNA sequences were amplified from genomic DNA using PCR as described in Example 1.1.2. For ease of DNA fragment assembly, some of the above said domains were co-amplified as single DNA fragments (Figure 6); the total number of fragments that were amplified were two (7 and 8), out of which one fragment
(7) contained DNA sequence corresponding to two domains. The rest of the DNA
fragments required for the construction of HMPKS A3, namely fragments 4 and 1 had been prepared earlier, as described in Examples 1.1.2. The source of DNA (Glasser et al., 1998; Donadio et al., 1991), nucleotide positions of the template and the oligonucleotide primers (shown as 5' to 3') used for each PCR were as follows:
iii). Assembly of the above said PKS domain DNA fragments in order to construct the final expression plasmid pARHMPKSA4
For the construction of pARHMPKSA4, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. The PCR products were phosphorylated and ligated to Swαl-cut, dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers.
No mistakes in the amplified products were detected. The plasmids were then used to transform the E.coli ET12567 dam" strain. Isolated DNA was digested with Xbal and the desired fragments isolated and purified. Fragment 4 was then cloned in Xbal-cvA. pCJR24HS vector using the procedure described in Example 1.1.2. The resulting plasmid pSIMAl was digested with Xbal and the other DNA fragments, namely, 8 and 7 were
sequentially added to finally yield plasmid pSIMA43. This plasmid was then digested with Ndel and Xbal and ligated with fragment 1 previously digested with the same two enzymes. The ligated products were used to transform E. coli DH10B electrocompetent cells and the final expression plasmid pARHMPKSA4, containing the customised HMPKS A4 gene, was isolated. The junctions where the domains were joined are shown in
Figure 6. iv). Expression of plasmid pARHMPKSA4 in a host
Plasmid pARHMPKSA4 was used to transform the mutant strain terreus/ARl.
Particularly preferred was the transformation procedure of Punt et al. (1992). Thiostrepton-resistant colonies were selected upon integration of the vector into the A. terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin and thiostrepton. v). Growth of host
The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of simvastatin
The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 1.5 g of crude product. A sample of this crude product was analysed by GC-MS. The peak corresponding to the molecular mass 418 was observed, indicating the presence of simvastatin. The latter was subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR. The preferred method for isolation of simvastatin was based on the findings of Kennedy et al. (1999). 1 litre of fermentation broth was found to produce 55 mg of simvastatin.
Example 1.2
Synthesis of 22DMB analogues
22DMB analogues were synthesised in a host through the expression of customised PKSs HMPKSB 1-2 (Figure 8). Example 1.2.1
Composition of the hybrid modular polyketide synthase B (HMPKSB) template
The HMPKSB template (Figure 7) comprises of the following enzymatic domains in the given sequence: acyltransferase - ATO, acyl carrier protein - ACP0, ketosynthase -
KS2, acyltransferase - AT2, dehydratase - DH2, methyltransferase - MeT2,
enoylreductase - ER2, ketoreductase - KR2 and acyl carrier protein - ACP2. The domains are covalently linked with each other, to give a multi-enzyme HMPKSB template.
Example 1.2.2 Construction of the hybrid modular polyketide synthase Bl (HMPKSB 1) i). Selection of PKS domain DNA sequence
For the construction of HMPKSB 1, the selection of domains from publicly available DNA sequences was as follows:
ATO - from loading module of the avermectin PKS ACPO - from module 7 of the epothilone PKS
KS2 - from module 7 of the epothilone PKS
AT2 - from module 7 of the epothilone PKS
DH2 - from module 7 of the epothilone PKS
MeT2 - from module 7 of the epothilone PKS ER2 - from LDKS of the lovastatin PKS
KR2 - from LDKS of the lovastatin PKS
ACP2 - from LDKS of the lovastatin PKS ii). Amplification of the selected PKS domain DNA
The above said PKS domain DNA sequences were amplified from genomic DNA using PCR as described in Example 1.1.2. For ease of DNA fragment assembly, some of the above said domains were co-amplified as single DNA fragments (Figure 8); the total number of fragments that were amplified were two (9 and 10), out of which one fragment
(9) contained DNA sequence corresponding to two domains. The rest of the DNA fragments required for the construction of HMPKSB 1, namely fragments 2, 6 and 4 had been prepared earlier, as described in Examples 1.1.2. and 1.1.4. The nucleotide positions in the template are shown according to the same disclosed in prior art (Ikeda et al., 1999).
The source of DNA, nucleotide positions of the template and the oligonucleotide primers
(shown as 5' to 3') used for each PCR were as follows:
iii). Assembly of the above said PKS domain DNA fragments in order to construct the final expression plasmid pARHMPKSBl
For the construction of pARHMPKSBl, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. The PCR products were phosphorylated and ligated to Smαl-cut, dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers. No mistakes in the amplified products were detected. The plasmids were then used to transform the E.coli ET12567 dam" strain. Isolated DNA was then digested with Xbal and the desired fragments isolated and purified. Fragment 4 was then cloned in Xbal-c t pCJR24HS vector using the procedure described in Example 1.1.2. The resulting plasmid pSIMAl was digested with Xbal and the other DNA fragments, namely, 6, 2, and 9 were sequentially added to finally yield plasmid pSIMB14. This plasmid was then digested with Ndel and Xbal and ligated with fragment 10 previously digested with the same two enzymes. The ligated products were used to transform E. coli DH10B electrocompetent cells and the final expression plasmid pARHMPKSBl, containing the customised HMPKSB 1 gene, was isolated. The junctions where the domains were joined are shown in Figure 8. iv). Expression of plasmid pARHMPKSBl in a host
Plasmid pARHMPKSBl was used to transform the mutant strain A. terreus/ARl. Particularly preferred was the transformation procedure of Punt et al. (1992). Thiostrepton-resistant colonies were selected upon integration of the vector into the A.
terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin and thiostrepton. v). Growth of host
The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of simvastatin
The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 1.2 g of crude product. A sample of this crude product was analysed by GC-MS. Two peaks corresponding to the molecular masses 418 and 432 were observed, indicating the presence of simvastatin and an analogue of simvastatin where the side chain is derived from propionylCoA as the starter acid (3, Figure 1). The two compounds were subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR. The preferred method for isolation of simvastatin was based on the findings of Kennedy et al. (1999). 1 litre of fermentation broth was found to produce 100 mg of simvastatin and 82 mg of the analogue.
Example 1.2.3
Construction of the hybrid modular polyketide synthase B2 (HMPKSB2) i). Selection of PKS domain DNA sequence
For the construction of HMPKSB2, the selection of domains from publicly available DNA sequences was as follows:
ATO - from loading module of the avermectin PKS
ACP0 - from loading module of the avermectin PKS KS2 - from module 4 of the eryhjtromycin PKS
AT2 - from module 4 of the eryhjtromycin PKS
DH2 - from module 4 of the eryhjtromycin PKS
MeT2 - from module 7 of the epothilone PKS
ER2 - from module 4 of the eryhjtromycin PKS KR2 - from module 4 of the eryhjtromycin PKS
ACP2 - from LDKS of the lovastatin PKS ii). Amplification of the selected PKS domain DNA
The above said PKS domain DNA sequences were amplified from genomic DNA using PCR as described in Example 1.1.2. The total number of fragments that were
amplified were five (11, 12, 13, 14 and 15; Figure 8). DNA fragment 7, required for the construction of HMPKSB2 had been prepared earlier as described in Example 1.1.5. The nucleotide positions in the template are shown according to the same disclosed in prior art (Donadio et al., 1991). The source of DNA, nucleotide positions of the template and the oligonucleotide primers (shown as 5' to 3') used for each PCR were as follows:
iii). Assembly of the above said PKS domain DNA fragments in order to construct the final expression plasmid pARHMPKSB2
For the construction of pARHMPKSB2, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. The PCR products were phosphorylated and ligated to Smal-cut, dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers. No mistakes in the amplified products were detected. The plasmids were then used to transform the E.coli ET12567 dam" strain. Isolated DNA was digested with Xbal and the desired fragments isolated and purified. Fragment 15 was then cloned in .X7j>αl-cut pCJR24HS vector using the procedure described in Example 1.1.2. The resulting plasmid pSIMB21 was digested with Xbal and the other DNA fragments, namely, 14, 13, 7 and 12 were sequentially added to finally yield plasmid pSIMB25. This plasmid was then digested with Ndel and Xbal and ligated with fragment 11 previously digested with the same two enzymes. The ligated products were used to transform E. coli DH10B electrocompetent cells and the final expression plasmid pARHMPKSB2, containing the customised HMPKSB2 gene, was isolated. The junctions where the domains were joined are shown in Figure 8. iv). Expression of plasmid pARHMPKSB2 in a host
Plasmid pARHMPKSB2 was used to transform the mutant strain A. terreus/ AR1. Particularly preferred was the transformation procedure of Punt et al. (1992). Thiostrepton-resistant colonies were selected upon integration of the vector into the A. terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin and thiostrepton. v). Growth of host
The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of simvastatin The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 1.6 g of crude product. A sample of this crude product was analysed by GC-MS. Two peaks corresponding to the molecular masses 418 and 446 were observed, indicating the presence of simvastatin and an analogue of simvastatin where the side chain is derived from 2-methylbutyrylCoA as the starter acid (4, Figure 1). The two compounds were subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR. The preferred method for isolation of simvastatin was based on the findings of Kennedy et al. (1999). 1 litre of fermentation broth was found to produce 109 mg of simvastatin and 75 mg of the analogue. Materials
Chemicals were of analytical grade or the best commercially available. Water was purified by the Millipore Milli-Q water purification system (Millipore S.A., France). All antibiotics were bought from Sigma Chemical Company, MO, USA. Pfu DNA polymerase was purchased from Boeringer, Germany. PCR reactions were preformed on a programmable Robo Cycler Gradient 40 (Stratagene, USA). Automated DNA sequencing was carried out on double-stranded DNA templates using an automated ABI 373 A sequencer (Applied Biosystems). 1H NMR spectra were recorded at 500 MHz on a Bruker DRX-500. 13C NMR spectra were recorded at 100 MHz on a Bruker DRX-400. Gas chromatography mass spectometry (GC-MS) was performed on a Finnigan MAT GCQ instrument. Analytical reverse phase high performance liquid chromatography mass spectrometry (HPLC-MS) analysis was carried out using Phenomenex Prodigy 5μ ODS3 100 A column with the following dimensions: 250 X 4.6 mm.
Methods Method 1.1
Bacterial Methods
Routine cloning and transformation procedures for E. coli and Aspergillus were carried out using the teachings of Sambrook et al (1989), Ballance and Turner (1985) and
Punt and Hondel (1992). Electrocompetent E. coli DH10B cells were made using the method of Dower et al (1988).
Method 1.2
Bacterial strains strain genotype use Reference
E. coli TGlrecO reeO::Tn5 routine cloning Kolodner et al., 1985
E. coli mcrA, U(mrr- routine cloning Gibco BRL,
DH10B™ hsdRMS- USA mcrBC), deoR, rec l, end Al, φ80d/αcZDM15
E. coli ET12567 dam', dcm", alleviates MacNeil et al., hsdM" restriction 1992 barrier to transformation
A. terreus Lovastatin gene
ATCC20542 producer overexpression
A. terreus/ARl LovF mutant, gene over monacolin J expression producer
Plasmids
Vector Reporter Siz Application Reference genes e pUC18 AmpR, 2.7 cloning, Yanisch-Perron et lacVOZ i' kb sequencing al, 1985
pCJR24 AmpR, ' TsrR 4.7 Streptomyces Rowe et al, 1998 kb expression system
Method 1.3
Construction of A. terreus/ARl strain
The strain A. terreus/ARl was constructed by integrating vector pSEVI10H4S into the A. terreus wild-type chromosome.
Method 1.3.1
Construction of expression plasmid pSEV110H4S i). Selection of DNA fragments
For the construction of pSEVI10H4S, the selection of domains from publicly available DNA sequences was as follows: H - hygromycin-resistance gene S - levansucrase gene
010 - hypothetical protein ORF 10 from the lovastatin-producing A. terreus 4 - from LDKS of the lovastatin PKS
ii). Amplification of the desired vector DNA components
The desired DNA sequences were amplified from plasmid DNA, using PCR as described in Example 1.1.2. The nucleotide positions in the template are shown according to the same disclosed in prior art (Quandt and Hynes, 1993; Mahenthiralingam et al, 1998). The source of DNA, nucleotide positions of the template and the oligonucleotide primers (shown as 5' to 3') used for each PCR were as follows:
iii). Assembly of the above said DNA fragments in order to construct plasmid pSEVI10H4S
For the construction of pSIM10H4S, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. The PCR products were phosphorylated and ligated to Smal-cυ , dephosphorylated pUC18 vector and then used to transform E. coli DH10B electrocompetent cells. The desired plasmids containing the amplified DNA fragments were isolated and sequenced using standard pUC forward and reverse primers. No mistakes in the amplified products were detected. The plasmids were then used to transform the E.coli ET12567 dam" strain. Isolated DNA was then digested with Xbal and the desired fragments isolated and purified. DNA fragment S was then cloned in .ATj l-cut pUC18 vector using the procedure described in Example 1.1.2. The resulting plasmid pSEVISl was digested with Xbal and the other DNA fragments, namely, 4, H and O10 were inserted as Xbal fragments to finally yield plasmid pSIM10H4S. The integrity of the plasmid was confirmed by restriction enzyme mapping and selection on hygromycin.
iv). Integration of plasmid pSIM10H4S in a host
Plasmid pSIM10H4S was used to transform the wild-type lovastatin-producing strain A terreus ATCC20542. Particularly preferred was the transformation procedure of Punt et al. (1992). Hygromycin-resistant colonies were selected upon integration of the vector into the A. terreus chromosome. Single transformants were picked and grown on plates supplemented with hygromycin, following which the colonies were selected on plates containing 10% sucrose. v). Growth of host
The above said single transformants were grown in one litre of liquid media for seven days. Media composition was based on the findings of Szakacs (1998). Cells were removed by centrifugation and the supernatant collected. vi). Isolation of monacolin J The above said supernatant was saturated with NaCl and extracted three times with equal volumes of ethyl acetate at pH 4.0. The solvent was evaporated to yield 880 mg of crude product. A sample of this crude product was analysed by GC-MS. A peak corresponding to the molecular mass 320 was observed, indicating the presence of monacolin J. The peak corresponding to the presence of lovastatin could not be seen, thus confirming the integrity of the lovF mutant strain. Monacolin J was subsequently isolated and fully characterised by high-pressure liquid chromatography (HPLC), 1H ID, 2D NMR and 13C NMR.
Method 1.4 Construction of expression plasmid pCJR24HS
For the construction of pCJR24HS, the method of Ranganthan (2000) was employed, as described earlier in Example 1.1.2. DNA fragment S (Method 1.3.1) was cloned in Xbal-cut pCJR24 vector using the procedure described in Example 1.1.2. The resulting plasmid pCJR24S was digested with Xbal and the other DNA fragment, namely the hygromycin-resistance gene H (Method 1.3.1) was inserted as an Xbal fragment to finally yield plasmid pCJR24HS. The integrity of the plasmid was confirmed by restriction enzyme mapping and selection on hygromycin. References Ballance, D.J, Turner, G. Gene. 1985, 36, 321. Cole, S.T, Brosch, R, Parkhill, J, Gamier, T, Churcher, C, Harris, D, Gordon, S.V, Eiglmeier, K, Gas, S, Barry III, C.E, Tekaia,F, Badcock, K, Basham, D, Brown, D, Chillingworth, T, Connor, R, Davies, R, Devlin, K, Feltwell, T, Gentles, S, Hamlin, N, Holroyd, S, Hornsby, T,. Jagels, K, Krogh, A, McLean, J, Moule, S, Murphy, L, Oliver, S, Osborne, J, Quail, M.A, Rajandream, M.A, Rogers, J, Rutter, S, Seeger, K, Skelton, S, Squares, S, Sqares, R, Sulston, J.E, Taylor, K, Whitehead, S, Barrell, B.G. Nature. 1998, 393, 537. Donadio, S, Staver, M.J, McAlpine, IB., Swanson, S.I, Katz, L. Science. 1991, 252,
675. Dower, W.J, Miller, IF. Ragsdale, C.W. Nuc. Acids Res. 1988, 16, 6127.
Dutton, C.I, Gibson, S.P, Goudie, AC, Holdom, K.S, Pacey, M.S., Ruddock, IC. J
Antibiot. 1991, 44, 357.
Gokhale, R.S, Tsuji, S.Y, Cane, D.E, Khosla, C. Science. 1999, 284, 482.
Glaser, P, Rusniok, C, Buchrieser, C, Couve, E, Frangeul, L, Billault, A, Kunst, F, Carniel, E. Pubmed accession number AL031866, 1998.
Hoffman, W. F, Smith, R. L, Willard, A. K. U. S. Patent No. 4444784, 1984.
Ikeda, H, Nonomiya, T, Usami, M, Ohta, T, Omura, S. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 9509.
Kennedy, I, Auclair, K, Kendrew S.G, Park C, Vederas IC, Hutchinson C.R Science. 1999, 284, 1368.
Kolodner, R, Fishel, R.A,Howard, M. J. Bacteriol. 1985, 163, 1060.
Kumar, Y, Thaper, R. K, Misra, S, Kumar, D. S. M, Khanna, I M. U. S. Patent No.
5763646, 1998.
MacNeil, D.I, Gewain, KM, Ruby, C.L, Dezeny, G, Gibbons P.H, MacNeil, T. Gene. 1992, 111, 61.
Mahenthiralingam, E, Marklund, B.I, Brooks, L.A, Smith, D.A, Bancroft, G.I,
Stokes, R.W. Infect Immun. 1998, 66, 3626.
McDaniel, R, Thamchaipenet, A, Gustafsson, C, Fu, H, Betlach, M, Ashley, G. Proc.
Natl. Acad. Sci. USA. 1999, 96, 1846. Molnar, I, Schupp, T, Ono, M, Zirkle, R, Milnamow, M, Nowak-Thompson, B, Engel,
N, Toupet, C, Stratmann, A, Cyr, D.D, Gorlach, I, Mayo, IM, Hu, A, Goff, S,
Schmid, I, Ligon, IM. Chem. Biol. 2000, 7, 97.
Moore, R.N, Bigam, G, Chan, IK, Hogg, A.M., Nakashima, T.T, Vederas, IC, J. Am.
Chem. Soc., l9$5, 107, 3694. Punt, I, Hondel, V.D. Methods Enzymol. 1992, 216, 447.
Quandt, I, Hynes, M.F. Gene. 1993, 127, 15.
Ranganathan, A, Timoney, M, Bycroft, M, Cortes, I, Thomas, IP, Wilkinson, B,
Kellenberger, L, Hanefeld, U, Galloway, I.S, Staunton, I, Leadlay, P.F. Chem Biol.
1999, 10, 731. Ranganathan A. PCT - International Patent No. WO 00/77181 A2. 2000.
Rowe, C.I, Cortes, I, Gaisser, S, Staunton, I, Leadlay, P.F. 1998, Gene, 216,215.
Saiki, R.K, Gelfand, D.H., Stoffel, S, Scharf, S.I, Higuchi, R, Horn, G.T, Mullis, K B,
Erlich, H.A. Science. 1988, 239, 487.
Sambrook, I, Fritsch, E.F, Maniatis, T. A laboratory Manual (2nd edn), Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 1989.
Schwecke, T, Aparicio, IF, Molnar, I, Konig, A, Khaw, L.E, Haydock, S.F, Oliynyk, M, Caffrey, P, Cortes, I, Lester, IB, Staunton, I, Leadlay, P.F. Proc. Natl. Acad. Sci. USA. 1995, 92, 7839.
Sleteinger, M, Verhoeven, T. R, Volante, R. P. U. S. Patent No. 4582915, 1986. Verhoeven, T. R, Askin, D. U. S. Patent No. 4820850, 1989. Suo, Z, Tseng, C.C, Walsh, C.T. Proc. Natl. Acad. Sci. USA. 2001, 98, 99. Szakacs, G, Morovjan, G, Tengerdy, R.P. Biotechnol. Lett. 1998, 20, 411. Yanisch-Perron, C, Vieira, I, Messing, I Gene 1985, 33, 103.
Yoshizawa, Y, Witter, D.I, Liu, Y.Q, Vederas, IC, J. Am. Chem. Soc, 1994, 116, 2693.