The heterologous biosynthesis of erythromycin A through E. coli includes the following experimental steps: 1) genetic transfer; 2) heterologous reconstitution; and 3) product analysis. Each step will be explained in the context of the motivation, potential, and challenges in producing therapeutic natural products using E. coli as a surrogate host.
The heterologous production of complex natural products is an approach designed to address current limitations and future possibilities. It is particularly useful for those compounds which possess therapeutic value but cannot be sufficiently produced or would benefit from an improved form of production. The experimental procedures involved can be subdivided into three components: 1) genetic transfer; 2) heterologous reconstitution; and 3) product analysis. Each experimental component is under continual optimization to meet the challenges and anticipate the opportunities associated with this emerging approach.
Heterologous biosynthesis begins with the identification of a genetic sequence responsible for a valuable natural product. Transferring this sequence to a heterologous host is complicated by the biosynthetic pathway complexity responsible for product formation. The antibiotic erythromycin A is a good example. Twenty genes (totaling >50 kb) are required for eventual biosynthesis. In addition, three of these genes encode megasynthases, multi-domain enzymes each ~300 kDa in size. This genetic material must be designed and transferred to E. coli for reconstituted biosynthesis. The use of PCR isolation, operon construction, multi-cystronic plasmids, and electro-transformation will be described in transferring the erythromycin A genetic cluster to E. coli.
Once transferred, the E. coli cell must support eventual biosynthesis. This process is also challenging given the substantial differences between E. coli and most original hosts responsible for complex natural product formation. The cell must provide necessary substrates to support biosynthesis and coordinately express the transferred genetic cluster to produce active enzymes. In the case of erythromycin A, the E. coli cell had to be engineered to provide the two precursors (propionyl-CoA and (2S)-methylmalonyl-CoA) required for biosynthesis. In addition, gene sequence modifications, plasmid copy number, chaperonin co-expression, post-translational enzymatic modification, and process temperature were also required to allow final erythromycin A formation.
Finally, successful production must be assessed. For the erythromycin A case, we will present two methods. The first is liquid chromatography-mass spectrometry (LC-MS) to confirm and quantify production. The bioactivity of erythromycin A will also be confirmed through use of a bioassay in which the antibiotic activity is tested against Bacillus subtilis. The assessment assays establish erythromycin A biosynthesis from E. coli and set the stage for future engineering efforts to improve or diversify production and for the production of new complex natural compounds using this approach.
Erythromycin A is a polyketide antibiotic produced by the Gram-positive soil bacterium Saccharopolyspora erythraea, and current production has been incrementally improved to ~10 g/L through decades of traditional mutagenesis and screening protocols and more recently through process optimization schemes 1-6. Mutagenesis and screening strategies are common in antibiotic natural product development as a result of difficulties in culturing and/or genetically manipulating native production hosts and because of the readily available antibiotic activity or improved growth phenotypes to aid selection. In the case of erythromycin A, S. erythraea is limited by a slow growth profile and the lack of more direct genetic manipulation techniques (relative to organisms like E. coli), thus, hampering rapid improvements in production and the biosynthesis of new derivatives. Having recognized the production issues and unlocked diversification possibilities with compounds like erythromycin A, the research community began to pursue the idea of heterologous biosynthesis (Figure 1) 7. These efforts coincided with available sequence information for the erythromycin A gene cluster 8-11. It should be emphasized that the number of sequenced complex natural product gene clusters has greatly expanded 12-16, providing the impetus for continued efforts in heterologous biosynthesis to access encoded medicinal potential. To do so, heterologous reconstitution requires that the new host meet the needs of the specific biosynthetic pathway. E. coli provides technical convenience, a wide-spanning set of molecular biology techniques, and metabolic and process engineering strategies for product development. Yet, when compared to native production hosts, E. coli does not exhibit the same level of complex natural product production. It was therefore unknown whether E. coli could serve as a viable heterologous option for complex natural product biosynthesis. However, it was assumed that E. coli would be an ideal host organism if heterologous biosynthesis could be accomplished.
With this goal in mind, initial efforts began to produce the polyketide aglycone 6-deoxyerythronolide B (6dEB) through E. coli. However, native E. coli metabolism could not provide appreciable levels of the propionyl-CoA and (2S)-methylmalonyl-CoA precursors needed to support 6dEB biosynthesis nor could the new host post-translationally modify the deoxyerythronolide B synthase (DEBS) enzymes. To remedy these issues, a metabolic pathway composed of native and heterologous enzymes was built into E. coli such that exogenously fed propionate was converted intracellularly to propionyl-CoA and then (2S)-methylmalonyl-CoA; during the engineering to complete this pathway, an sfp gene was placed into the chromosome of E. coli BL21(DE3) to produce a new strain termed BAP1. The Sfp enzyme is a phosphopantetheinyl transferase capable of attaching the 4′-phosphopantetheine cofactor to the DEBS enzymes 17,18. The three DEBS genes (each ~10 kb) were then placed on two separately selectable expression vectors containing inducible T7 promoters. After a key adjustment of post-induction temperature (to 22 °C), the DEBS genes were coordinately expressed within BAP1 in an active state capable of generating 6dEB 19.
The pursuit of full erythromycin A biosynthesis then began using an analogous gene cluster from Micromonospora megalomicea or a hybrid pathway composed of genes from S. erythraea, S. fradiae, and S. venezuelae which produced the intermediates erythromycin C and 6-deoxyerythromycin D, respectively 20-22. Recently, our group has extended these efforts by producing erythromycin A (the most clinically-relevant form of erythromycin) through E. coli. In contrast to previous work, our strategy coordinately expressed the 20 original S. erythraea genes needed for polyketide biosynthesis, deoxysugar biosynthesis and attachment, additional tailoring, and self-resistance (Figure 2). In total, 26 (native and heterologous) genes were engineered to allow E. coli to produce erythromycin A at 4 mg/L 23,24. This result established complete production of a complex polyketide natural product using E. coli and serves as a basis to leverage this new production option or pursue new ones.
The text below is specific for the erythromycin A antibiotic, but the steps are designed to be generally applicable to other natural products as candidates for heterologous biosynthesis.
1. Erythromycin A Genetic Cluster Transfer
2. E. coli Biosynthetic Reconstitution
3. Product Analysis
The desired outcome of this approach is the production of a fully bioactive natural product from the E. coli heterologous host. This is best represented by the LC-MS results used to confirm and quantify production (Figure 6) and the antibacterial bioassay used to confirm final activity (Figure 7). In the overall scheme of heterologous biosynthesis, this result defines success. Once accomplished, research efforts then turn to optimization (both at the cellular and process scales) and molecular engineering. The final objectives include economical production processes for the original compound and control over the heterologous system such that rational modifications to the process can be made to produce variants of the original compound with the potential for a heightened or broadened activity spectrum.
Failure to produce the desired compound then triggers numerous contingency plans to assess which portion of the approach is prohibiting successful biosynthesis. Many of these trouble-shooting mechanisms are in-built within the procedure outlined above. In our experience, it is important, at a minimum, to confirm successful gene expression (SDS-PAGE analysis of soluble biosynthetic enzymes) of the newly introduced gene cluster prior to attempting to assess biosynthesis of the formed natural product. Post expression, as indicated above, the use of protein folding chaperones, enhanced gene copy number, and process temperature have been effective means of allowing eventual biosynthesis.
Figure 1. An overall depiction of the heterologous natural product biosynthetic scheme featuring the erythromycin A compound. The experimental steps required include 1) design and transfer of the erythromycin gene cluster from S. erythraea to E. coli; 2) reconstitution of biosynthesis within E. coli; and 3) product analysis of erythromycin A.
Figure 2. A. The erythromycin gene cluster as organized in the native S. erythraea chromosome. B. Erythromycin A biosynthesis. Within E. coli, the Sfp enzyme is required for posttranslational modification of the DEBS1, 2, and 3 polyketide synthase enzymes (each >300 kDa); this is denoted by the connecting arm associated with each ACP domain. The DEBS enzymes form an α2β2γ2 complex subdivided into module units that catalyze the consecutive condensations of one starter propionyl-CoA unit and six (2S)-methylmalonyl-CoA extender units to form the polyketide 6-deoxyerythronolide B (6dEB). KS-ketosynthase; AT-acyl transferase; ACP-acyl carrier protein; KR-ketoreductase (with the inactive KR noted in Module 3); DH-dehydratase; ER-enoyl reductase; TE-thioesterase. A PrpE (propionyl-CoA synthetase) and PCC (propionyl-CoA carboxylase, composed of two protein subunits) are required to convert exogenous propionate to the starting substrates propionyl-CoA and (2S)-methylmalonyl-CoA. The genes for Sfp and PrpE were previously engineered within the BAP1 E. coli chromosome 19. Conversion of the 6dEB molecule to erythromycin is accomplished by the deoxysugar biosynthesis and attachment, additional tailoring, and self-resistance enzymes. Click here to view larger figure.
Figure 3. A. Isolation of individual genes required for erythromycin A biosynthesis (excluding the three DEBS genes). Individual plasmid design including the PCR oligonucleotide primers used for isolating eryCI; primer restriction sites have been indicated in bold and complementary sequence has been underlined. Also included is the restriction analysis of the resulting pET21 and pET28 vectors containing individual genes and operons. B. i. Consolidation of the biosynthetic genes needed for polyketide tailoring as operons for introduction to E. coli. ii. Restriction site and other expression elements involved with operon construction and design; X/S represents the compatible-cohesive ligation of XbaI and SpeI sites which results in a new sequence unrecognized by either restriction enzyme. Click here to view figure 3.
Figure 4. A. SDS-PAGE analysis of the individual genes presented in Figure 3. The asterisks denote those genes expressed with 5′ leader sequences resulting from the pET28 expression vector. B. Phenotypic assessment of ErmE (erythromycin A resistance). Erythromycin A within the solid medium used to test E. coli strains harboring plasmids with or without the ermE erythromycin A resistance gene. Cell growth is rescued with with ermE expression.
Figure 5. A. Schematic of the full erythromycin A pathway introduced to E. coli, including a plasmid for the GroEL/ES chaperonin (pGro7) and a plasmid containing an extra copy of eryK. B. Transformation comparison between chemical and electro methods using both combination (i) and sequential (ii) introduction of plasmids. C. Plasmid stability of transformed strain. Click here to view larger figure.
Figure 6. A standard LC-MS trace of erythromycin A produced from E. coli.
Figure 7. A. Solid phase bioassay of filter disks containing i. extract from an uninduced BAP1 E. coli negative control; ii. commercially available erythromycin A (positive control); and iii. erythromycin A extracted from an induced BAP1 E. coli system. The disks were plated on a lawn of B. subtilis bacteria and resulting zones of inhibition indicate antibiotic activity. B. Assay demonstrated in liquid phase using 96-well plate time course samples. Click here to view larger figure.
Critical steps in heterologous biosynthesis are encountered at each of the three procedural points in the process: 1) genetic transfer; 2) biosynthetic reconstitution; and 3) product analysis. A problem at any stage will derail the ultimate objective of establishing heterologous biosynthesis. Perhaps the most challenging aspect of the process is establishing reconstituted biosynthesis, since this is absolutely required to allow successful analysis. However, reconstitution is dependent upon careful design and transfer of the genetic material responsible for eventual biosynthesis. Likewise, established, sensitive analytical methods (like LC-MS) can detect small titer levels that otherwise may be interpreted as lack of production. It is the culmination of these situations that makes heterologous biosynthesis challenging.
With regards to the erythromycin A example, confirmed gene sequences isolated by PCR were individually tested for expression prior to consolidating the genes for coordinated expression and natural product formation. During initial attempts at product analysis, intermediate compounds were first identified only when a chaperonin (GroEL/ES) was co-expressed with the required gene cluster. The final product was only produced when an additional copy of the gene responsible for the last step in biosynthesis (eryK) was included. Both of these measures benefitted from the sensitive LC-MS analytical method used to initially assess production.
Now that production has been confirmed, the E. coli cellular background offers numerous engineering opportunities. The first of which is to optimize production of the native erythromycin A compound. This would offer an alternative production route with potential savings in process time and cost. Furthermore, the erythromycin A pathway may be manipulated for the purpose of diversifying product formation. New analogs offer the potential of new antibiotic (or other therapeutic) activity. These same opportunities exist for general applications of heterologous biosynthesis and provide the primary motivation for the approach, further supported by the challenges encountered with many original producing organisms. Continual applications and improvements of the approach described in this article will increase the frequency of future heterologous biosynthetic success.
The authors have nothing to disclose.
The authors thank the NIH (AI074224 and GM085323) and NSF (0712019 and 0924699) for funding to support projects dedicated to heterologous biosynthesis.
Name of the reagent | Company | Catalogue number | Comments (optional) |
PCR machine | Eppendorf | Mastercycler personal | |
Dimethyl sulfoxide (DMSO) | Fisher | BP231 | |
Electroporator | BioRad | Micropulser | |
IPTG | Fisher | BP1620 | |
Sodium propionate | Sigma | P1880 | |
L-arabinose | Sigma | A3256 | |
Refrigerated Shaker | Thermo Scientific | MaxQ 4000 | |
Microfuge | Eppendorf | Centrifuge 5415D | |
pGro7 | Takara | Chaperone Plasmid Set (3340) | |
pET21, pET28, pCDF-Duet-1 | EMD Chemicals | 69742-3, 69864, 71340 | |
LC-MS | Applied Biosystems | 3200 Q-Trap | |
Ethyl acetate | Sigma | 270989 | |
Methanol | Sigma | 322415 | |
Vacuum centrifuge | Eppendorf | Concentrator 5301 | |
Rotary Evaporator | Buchi | R-200 |