Synthetic protein scaffolds provide modular control over metabolic flux

From Wiki FKKT

Jump to: navigation, search

(Ana Dolinar)

Dueber, J. E. et al. 2009. Synthetic protein scaffolds provide modular control over metabolic flux. Nature Biotechnology 27: 753-759.


The article was written by John E. Dueber and his colleagues and published in Nature Biotechnology in the year 2009. In the research, they used synthetic protein scaffolds to organize metabolic enzymes and thus improve the effectiveness of metabolite synthesis. Following writing will cover the basics of metabolic engineering, protein scaffolds, substrate channeling and mevalonate synthesis as well as experimental results from Dueber’s paper and a brief comparison of other scaffolding molecules (i.e. DNA, RNA).

Contents

Introduction

Metabolic engineering

Metabolic engineering is a term, defined in 1991 by James Edward Bailey as ‘’Improvement of cellular activities by manipulation of enzymatic transport and regulatory functions of the cell with the use of recombinant DNA technology’’ [1]. These days it is used as a term for describing modifications of metabolic pathways leading to improved or rewired synthesis of desired products [2]. Traditional metabolic engineering is focused mainly on improving and rewiring biosynthesis (Figure 1A: Schematic representation of traditional metabolic engineering.). Contrary, synthetic biology is usually based on establishing completely new biosynthetic pathways in cells (Figure 1C: Schematic approach of synthetic biology.). Many research papers these days are using the techniques from both research fields to interchange metabolic enzymes and production hosts to maximize the production yield (Figure 1B: Representation of metabolic engineering which is combined with synthetic biology.) [3].

Protein scaffolds

Protein scaffolds exist in laboratory as well as in nature. Natural protein scaffolds are observed in yeast (e.g. Ste5 protein coordinates mitogen activated protein kinase (MAPK) pathways) and in mammals (e.g. PSD-95 binds neuronal receptors, other scaffold proteins and actin; this complex coordinates neuronal response to stimuli), as well as in plants (e.g. protein Rack1A (Receptor for Activating C Kinase1) originates from Arabidopsis thaliana and is responsible for interactions with proteins in photosynthetic pathway, stress-response pathways, and in ribosomal pathway) [4]. Scaffold proteins are also found in bacteria and viruses where they regulate host’s immune response [5]

Organized metabolic enzymes have many advantages compared to unorganized enzymes, i.e. metabolic intermediates are not lost in cell during the reaction and they are harder to accumulate (which is great if they are toxic). Moreover, scaffolds allow us to determine the number of enzymes participating in the reaction, so we can better overcome the bottleneck reactions. The ability to spatially control enzymes represents a promising tool for posttranslational regulation of metabolic pathways. [6] Engineered protein scaffolds are used to alter existing signalling and metabolic pathways. In yeast, researchers put together two signaling pathways to form a synthetic signaling pathway with combined characteristics. First pathway is mating pathway, induced by alpha-factor and has its own scaffold protein Ste5. Second pathway regulates the response to osmotic changes and is induced by high osmolarity, that pathway, too, has its natural scaffold protein, Pbs2. They synthesized artificial scaffold which consisted of N-terminal part of Ste5 and C-terminal part of Pbs2. This scaffold is responsive to an alpha-factor and mediates the osmo-response. [7] Engineered metabolic pathway is further discussed on the example of mevalonate synthesis in the following chapter.

Substrate channeling

Substrate channeling is a feature, which can be observed in different natural enzymatic systems such are pyruvate dehydrogenase complex, tryptophan synthase, and various polyketide synthases. In the process of transferring the intermediate toward next enzyme, channeling provides a secure environment for unstable intermediates, prevents diffusion of intermediates and enhances turn-over rate either by sequestering intermediates from cytosol or covalently binding them to enzymatic complexes. If the metabolite could divergently enter more metabolic pathways, substrate channeling also provides the necessary metabolite pool for the pathway. [6][8]

Mevalonate synthesis

Mevalonate (anionic form of mevalonic acid) is produced from acetyl-coenzyme A (acetyl-CoA) in three enzymatic reactions. First are two acetyl-CoA condensed into acetoacetyl-CoA via acetoacetyl-CoA thiolase. Acetoacetyl-CoA is then condensed with another acetyl-CoA via hydroxyl-methylglutaryl-CoA synthase (HMGS) to form 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA). HMG-CoA is then reduced to mevalonate via hydroxyl-methylglutaryl-CoA reductase (HMGR). This step is rate limiting reaction in the pathway. Mevalonate is fourth metabolite in the mevalonate pathway (Scheme of mevalonate pathway). Final product of the mevalonate pathway is farnesyl-pyrophosphate which can be further converted into many different molecules (squalene, heme A, ubiquinones, sterols …). [9]

Design of scaffold and enzymes

Synthetic protein scaffold consisted of three different binding domains and three enzymes necessary for the synthesis of mevalonate. Dueber and colleagues used acetoacetyl-CoA thiolase AtoB from Escherichia coli and HMGS and HMGR from Saccharomyces cerevisiae. For production organism, they selected the bacterium Escherichia coli, therefore HMGS and HMGR were codon-optimized for expression in Escherichia coli (Figure 1A: Schematic representation of engineered mevalonate pathway.). [6]

  • Their scaffolding strategy is based on substrate channeling. They assayed different scaffold architectures. Firstly, they wanted to confirm improvement of mevalonate production using synthetic complex of HMGS and HMGR. Formation of a synthetic complex is based on interaction between SH3 ligand and respective binding domain SH3 which were previously added to C-terminus of HMGS and to N-terminus of HMGR, respectively. They evaluated different ratios between HMGS and HMGR through modulating the number of SH3 ligands, fused to HMGS (Figure 2A: Schematic representation of various HMGS-HMGR complexes.). [6]
  • Secondly, they separated scaffolding from enzymes. Design of the scaffold is based on three different interaction domains separated by flexible linkers. Enzymes AtoB, HMGS and HMGR were fused with a ligand for GBD domain, SH3 domain, and PDZ domain, respectively. Scaffold consisted from one GBD domain and varying number of SH3 (x) and PDZ (y) domains (abbreviated G1SxPy). GBD domain is a 78 amino acids long fragment from rat protein N-WASP (Neural Wiskott–Aldrich syndrome protein). GBD ligand is a fragment from protein Cdc 42 (cell division control protein 42) which interacts with N-WASP [10]. SH3 (Src homology 3) domain is N-terminal domain of mouse protein c-Crk and binds guanine nucleotide exchange factor C3G with high affinity [11]. SH3 ligand is 11-mer of mostly proline and arginine [12]. PDZ domain originates from mouse protein α-syntrophin, 53,5 kDa big protein which functions as a modular adapter that can bind to transmembrane receptors and ion channels and recruit other proteins [13]. Amino acid sequences for both ligands and binding domains are listed in supplementary table 1. Enzymes with fused ligands were put under transcriptional control of PTET and scaffold was under transcriptional control of PBAD (Figure 3A: Representation of a synthetic scaffold with enzymes and respective binding domains.). Both promoters are inducible, PTET with tetracycline or its analog and PBAD with arabinose. [6]
  • Furthermore, they wanted to determine the optimal scaffold design. As observed in supplementary figure 4A, they took the scaffold with optimal numbers of binding domains (G1S2P2) and rearranged sequence of domains. For the control experiment, they replaced scaffold with green fluorescent protein. [6]
  • Lastly, to demonstrate the generality of modular scaffolding they improved the biosynthesis of glucaric acid. Scaffold G1S1P1 was reused from previous experiments. Enzymes myo-inositol-1-phosphate synthase (Ino1) from Saccharomyces cerevisiae and myo-inositol oxygenase (MIOX) from mice were fused with SH3 ligand or PDZ ligand, respectively, and heterologously expressed in Escherichia coli. Another enzyme, urinate dehydrogenase (Udh) from Pseudomonas syringae, was introduced into production organism for successful conversion of glucose into glucaric acid (Figure 6, left panel: Reused scaffold for production of glucaric acid.). [6][14]

Results

  • In their first experiment, they confirmed an assumption that co-recruitment of HMGS and HMGR, two enzymes which catalyze bottleneck reactions in the synthesis of mevalonate, will improve its production. Result is shown in figure 2C. An increase in mevalonate titers was observed in all experiments with co-recruited HMGS and HMGR. HMGS was fused with a variable number of SH3 ligands which interacted with SH3 domain, fused to HMGR. Approximately tenfold increase in mevalonate production was observed with six SH3 domains and when number of SH3 domains was higher, mevalonate production decreased to around sevenfold compared to unscaffolded production. Dueber and colleagues reasoned that (a) HMGS might be misfolded due to too many SH3 ligands, (b) enzymes might spatially obstruct each other or/and (c) HMGR could be unevenly arranged around HMGS thus unable to perform maximum catalytic activity. [6]
  • When they introduced specific polypeptide for scaffold, ascertainment of scaffold architecture was necessary. Tested scaffolds consisted of one GBD domain and SH3 and PDZ domains (one to four each). As can be seen in figure 3B, scaffold G1S2P2 outdid all other combinations resulting in approximately 77-fold increase in mevalonate production. As the scaffold was under control of PBAD, they also evaluated the optimal concentration of arabinose for scaffold induction. As we can see in figure 3C, supplementary table 4, and supplementary figure 8, optimal concentration for induction varies between different scaffolds with lowest inducer concentration needed for G1S4P4 scaffold and highest inducer concentration necessary for G1S2P2 scaffold. [6]
  • Comparison of three enzyme-scaffold (AtoB, HMGS and HMGR; G1S2P2) and two enzyme-scaffold (HMGS and HMGR; G0S2P2) was also made. Results (Supplementary figure 5, lower panel) show an increase in mevalonate production, compared to experiment with no scaffold, in both cases. Higher increase was observed with G1S2P2 scaffold (77-fold) versus G0S2P2 scaffold (8,5-fold). [6]
  • Furthermore, they needed to determine which organization of domains would be optimal for highest mevalonate production. They took the optimal scaffold G1S2P2 and rearranged each domain to form two new scaffolds G1S1P2S1 and G1S1P1S1P1. Both rearranged scaffolds show decrease in mevalonate production compared to the optimal scaffold, however their production is approximately 10-fold (G1S1P2S1) or 20-fold (G1S1P1S1P1) higher than without scaffold (Supplementary figure 4). Results, therefore, confirmed the hypothesis that scaffold architecture could drastically affect the synthesis of desired product. [6]
  • The implementation and performance of scaffolds were further validated in separate control experiment. They used five different constructs, one with green fluorescent protein instead of scaffold and four with G1S2P2 scaffold. From these, in one enzymes had their respective binding ligand and in other they were without it. In other two, they introduced glutathione-S-transferase with PDZ ligand as a competitor under control of PTET or PBAD. Result is shown in figure 4A, right panel and confirms the effectiveness of scaffold. Competitor ligands occupy binding sites instead of enzymes and therefore decrease production of mevalonate. Decrease in production is greater if the competitor is expressed together with scaffold (9-fold improvement in production versus 27-fold improvement in production if competitor is expressed together with enzymes; improvements are compared to production without scaffold). [6]
  • Scaffolds also allow lower enzyme expression with equivalent mevalonate production as without scaffold. This result is shown in figure 4B, where scaffolded pathway produced more mevalonate at low inducer concentrations than unscaffolded pathway. At high inducer concentration, production of mevalonate was comparable between scaffolded and unscaffolded pathway. They interpreted these results as inability of scaffold to sustain its function in the environment with high enzyme concentrations. [6]
  • Both optical density of bacterial cultures (Supplementary figure 6) and growth curves (Figure 5) indicate that scaffolded pathway needs lower induction level of enzymes to achieve equal or higher titers of mevalonate than unscaffolded pathway. As it can be seen in figure 5 (upper panel) high induction of either scaffolded or unscaffolded pathway is related to decreasing in growth rate of bacteria. [6]
  • Lastly, Dueber and colleagues demonstrated that scaffold could be used for enhanced production of other metabolites. Design of experiment (as shown in figure 6) is described in details in the previous chapter. Although initial titers of glucaric acid are already high, they were able to further improve the production by 200 % with scaffolded enzymes. [6]

Discussion

  • Dueber and colleagues successfully implemented a scaffold protein in Escherichia coli and used it to improve the production of mevalonate by 77-fold. Advantages of their approach are (a) generalization of scaffolding as it can be applied to different enzymatic pathways and (b) modularity of the design which enables easy addition of binding domains to the scaffold for more complex pathways with greater number of enzymes. [6] The production of glucaric acid was further developed in the year 2010 when Moon and colleagues published the article Use of modular, synthetic scaffolds for improved production of glucaric acid in engineered Escherichia coli. They upgraded the before mentioned scaffold for enzymes Ino1 and MIOX with additional binding site for enzyme Udh and further improved the titer of glucaric acid by 50 % (compared to previous scaffold[6]; it is approximately 5-fold improvement compared to unscaffolded pathway) resulting in maximum titer around 2,5 g/L. [15]
  • Another experiment by Wang and Yu in the year 2012 verified the generality of scaffolding approach. Using different scaffold architectures (from G1S1P1 and G0S0P0 for control to G1S4P4), they applied the scaffolding in yeasts for the production of resveratrol, type of natural stilbenoid. SH3 ligand was fused to 4-coumarate:CoA ligase 1 (4CL1) and stilbene synthase (STS) was added a PDZ ligand. Results show the optimal scaffold to be G1S2P4. Despite the increase in resveratrol production in scaffolded pathway over unscaffolded pathway, the amount of produced resveratrol is nevertheless low compared to several other reports about engineering the metabolic pathway of resveratrol. [16]
  • However, when talking about scaffolds, we cannot let DNA-scaffolds slip away. In the year 2009, Wilner and his colleagues published an article about topological DNA-scaffolds. They were able to construct hexagon shaped DNA nanostructures – two- and four-hexagon DNA strips – which had binding sites for functionalized enzymes (i. e. enzymes or cofactors with added oligonucleotides complementary to the binding sites). Schematic representation is shown in figure 2. Cofactor NAD+ was combined with NAD+-dependent glucose dehydrogenase and glucose oxidase was associated with horseradish peroxidase. Both functionalized pairs exhibited greater activity in the presence of scaffolds. All these experiments were performed without the presence of cells. [17]
  • Nonetheless, research with DNA-scaffolds was also performed in bacteria. Conrado and his colleagues designed a modular scaffold for production of resveratrol in Escherichia coli. To confirm the modularity of the design, they employed the scaffold for production of two other metabolic products, 1,2-propanediol and mevalonate. Scaffold is based on interaction between zinc-fingers (ZFs) and DNA. Depending on the enzymatic complex they wanted to implement, they first constructed the plasmids with desirable sequence of binding sites for adequate ZFs which were fused to the enzymes. For the production of resveratrol, they chose 4-coumarate:CoA ligase (4CL) and stilbene synthase (STS) and scaffold of one binding site for each enzyme (as seen in figure 1A). Production of 1,2-propanediol and mevalonate is more complex and requires three enzymes, however this allows more flexibility with scaffold design (Figure 1B). Selected enzymes for the production of 1,2-propanediol originate from Escherichia coli: methylglyoxal synthase (MgsA), 2,5-diketo-D-gluconic acid reductase (DkgA) and glycerol dehydrogenase (GldA). Escherichia coli acetoacetyl-CoA thiolase (AtoB), Saccharomyces cerevisiae hydroxy-methylglutaryl-CoA synthase (HMGS) and Saccharomyces cerevisiae hydroxymethylglutaryl-CoA reductase (HMGR) were selected for producing mevalonate. Various increases of metabolite production were observed among all scaffolds confirming the role of scaffold architecture in the effectiveness of enzymatic conversion, although DNA scaffolds exhibited relatively low improvement in metabolite production – approximately 5-fold increase is maximal improvement.[18]
  • Yet another shape of scaffolds is represented by RNA assemblies. Delebecque and her colleagues performed a research on improving bioproduction of hydrogen using RNA aptamers and suitable binding proteins as scaffolds. Using RNA assembly as scaffold for [FeFe]-hydrogenase and ferredoxin resulted in 48-fold increase of hydrogen production. Additionally, RNA assemblies are capable of forming divergent shapes such are nanotubes, 2D layers, and ribbons. As a main advantage of RNA assemblies compared to protein scaffolds, authors emphasised the precision of complex formations even on nanometer scale.[19]

Concluding remarks

Scaffolding is proven to be applicable in bacteria and yeasts, however I think that engineered mammalian scaffolds will follow shortly. Researchers implemented various types of scaffolds in metabolic and signalling engineering, with metabolic engineering being slightly more represented. Among all presented scaffolds and assemblies for metabolic engineering, Dueber and his colleagues were able to achieve the greatest increase in metabolite production.

Reasonably, all types of scaffolds have their advantages and disadvantages. Protein scaffolds could be constructed from a wide variety of protein binding domains, yet they are prone to faster degradation than plasmid DNA scaffolds. RNA assemblies are as well prone to fast degradation in the cellular environment. All scaffolds are modular and thus enable simple modifications for different applications in metabolic and other types of engineering. DNA-binding domains which interact with DNA scaffolds are virtually sequence-independent. By sequence-independent is meant the fact that we can design DNA- binding domains for almost all DNA sequences, not that DNA-binding proteins are interacting with any DNA, regardless of its sequence. Two of such DNA-binding domains are previously mentioned zinc-fingers, and the other is TAL (transcription activator-like) effector. [6][18][19] Production of molecules that are now days synthesized principally through chemical synthesis could benefit a great deal with eased enzymatic biosynthesis. Additionally to cheaper production, biosynthesis facilitates the chiral selectivity of reactions hence eliminating undesired by-products, reduces the chemical waste, and in some instances it offers a more economic way of production without the need of high pressure, expensive catalysts or extreme temperatures. [6][15] [20]

References

  1. Bailey, J. E. 1991. Toward a science of metabolic engineering. Science 252: 1668–1674.
  2. Stephanopoulos, G. 2012. Synthetic biology and metabolic engineering. ACS Synthetic biology 1: 514−525.
  3. Nielsen, J. and Keasling, J. D. 2011. Synergies between synthetic biology and metabolic engineering. Nature Biotechnology 29: 693−695.
  4. Kundu, N., Dozier, U., Deslandes, L., Somssich, I. E. and Ullah, H. 2013. Arabidopsis scaffold protein RACK1A interacts with diverse environmental stress. Plant Signaling & Behavior 8: e24012.
  5. Good, M. C., Zalatan, J. G. and Lim, W. A. 2011. Scaffold proteins: hubs for controlling the flow of cellular information. Science 332: 680-686.
  6. 6.00 6.01 6.02 6.03 6.04 6.05 6.06 6.07 6.08 6.09 6.10 6.11 6.12 6.13 6.14 6.15 6.16 6.17 6.18 6.19 6.20 Dueber, J. E., Wu, G. C., Malmirchegini, G. R., Moon, T. S., Petzold, C. J., Ullal, A. V., Prather, K. L. J. and Keasling, J. D. 2009. Synthetic protein scaffolds provide modular control over metabolic flux. Nature Biotechnology 27: 753-759.
  7. Park, S. H., Zarrinpar, A. and Lim, W. A. 2003. Rewiring MAP kinase pathways using alternative scaffold assembly mechanisms. Science 299: 1061−1064.
  8. Zhang, Y. H. P. 2011. Substrate channeling and enzyme complexes for biotechnological applications. Biotechnology Advances 29: 715−725.
  9. Nelson, D. L. and Cox, M. M. 2012. Lehninger Principles of biochemistry. W.H. Freeman; Sixth Edition. ISBN-10: 1429234148
  10. Kim, A.S., Kakalis, L.T., Abdul-Manan, N., Liu, G.A. and Rosen, M.K. 2000. Autoinhibition and activation mechanisms of the Wiskott-Aldrich syndrome protein. Nature 404: 151-158.
  11. Wu, X., Knudsen, B., Feller, S. M., Zheng, J., Sali, A., Cowburn, D., Hanafusa, H. and Kuriyan, J. 1995. Structural basis for the specific interaction of lysine-containing proline-rich peptides with the N-terminal SH3 domain of c-Crk. Structure 3: 215-226.
  12. Nguyen, J.T., Turck, C.W., Cohen, F.E., Zuckermann, R.N. and Lim, W.A. 1998. Exploiting the basis of proline recognition by SH3 and WW domains: design of N-substituted inhibitors. Science 282: 2088-2092.
  13. Schultz, J., Hoffmuüller, U., Krause, G., Ashurst, J., Macias, M. J., Schmieder, P. Schneider-Mergener, J. and Oschkinat, H. 1998. Specific interactions between the syntrophin PDZ domain and voltage-gated sodium channels. Nature Structural biology 5: 19-24.
  14. Moon, T. S., Yoon, S. H., Lanza, A. M., Roy-Mayhew J. D. and Prather, K. L. 2009. Production of glucaric acid from a synthetic pathway in recombinant Escherichia coli. Applied and environmental microbiology 75: 589−595.
  15. 15.0 15.1 Moon, T. S., Dueber, J. E., Shiue, E. and Prather, K. L. J. 2010. Use of modular, synthetic scaffolds for improved production of glucaric acid in engineered E. coli. Metabolic engineering 12: 298-305.
  16. Wang, Y. and Yu, O. 2012. Synthetic scaffolds increased resveratrol biosynthesis in engineered yeast cells. Journal of Biotechnology 157: 258-260.
  17. Wilner, O. I., Weizmann, Y., Gill, R., Lioubashevski, O., Freeman, R. and Willner, I. 2009. Enzyme cascades activated on topologically programmed DNA scaffolds. Nature Nanotechnology 4: 249-254.
  18. 18.0 18.1 Conrado, R. J., Wu, G. C., Boock, J. T., et al. 2012. DNA-guided assembly of biosynthetic pathways promotes improved catalytic efficiency. Nucleic acids research 40: 1879-1889.
  19. 19.0 19.1 Delebecque, C., J., Linder, A., B., Silver, P. A. and Aldaye, F. A. 2011. Organization of intracellular reactions with rationally designed RNA assemblies. Science 333: 470-474.
  20. Keasling, J. D. 2012. Building with biology: a constructive debate. Nature 492: 188.


SB students resources

Personal tools