Our Molecular Design is a proposal of a process. It is centered on the in silico generation of DNA nanocages by the Polygen Software, like the experimental work as well, were evidence of functionality of this software is explored. In the previous section previous section our project and its motivations were justified, now here we enter in further details this proposed process, expliciting how find a biochemical pathway of interest, how 3D model it on a cage for evaluate its spatial characteristics, how the construction is supposed to be done at lab and how to test and measure the scaffolded system built. All this is summarized in the following infographic:




On the first step of the design process is used the KEGG Database (Kanehisa et al, 2000) for "manually" finding possible natural pathways for biocatalysis. Another interesting option are the recent computational tools for discovery of new synthetic metabolic pathways, like the Metabolic Tinker tool, where the inputs are a desired substract, the pathway’s lenght and a target product. Our criteria was searching for biosynthetic pathways with simple, well characterized and relevant bioproducts. The penicillin pathway fitted like a glove. The disruptive action of this antibiotic in the bacterial cell wall formation (Yocum et al, 1979) is responsible for a medical revolution .The historical importance of the first antibiotic "blockbuster" (Newman et al, 2000) lead to a great number of studies and information about it is readily available making it a reasonable first option to develop our model.

Also, instead of focusing only on a single drug, for an expanded proof-of-concept design we prefered a common biochemical pathway which is used for the production of many natural drugs: the mevalonate pathway. This route is used to generate isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) which are precursors for compounds such as cholesterol, vitamin K, heme and more (Buhaescu et al, 2007). The designs of those pathways are discussed forward.



After selecting the suitable enzymes, we searched for the protein structures in Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB-PDB) [Berman et al, 2000] and the PDB files were then uploaded to UCSF Chimera [Pettersen et al, 2004] in order to model molecularly the assembly of the proteins to the cage


The penicillin biosynthetic path found in KEGG displayed the following reactions:

1) 3 ATP + L-2-aminohexanedioate + L-cysteine + L-valine + H2O = 3 AMP + 3 diphosphate + N-[L-5-amino-5-carboxypentanoyl]-L-cysteinyl-D-valine

Enzyme: 4MB0

2) N-[(5S)-5-amino-5-carboxypentanoyl]-L-cysteinyl-D-valine + O2 = isopenicillin N + 2 H2O

Enzyme: 2IVJ

3) phenylacetyl-CoA + isopenicillin N + H2O = CoA + penicillin G + L-2-aminohexanedioate

Enzyme: 2X1C

Which can be simplified in the global reaction:

3 ATP + L-cysteine + L-valine + O2 + phenylacetyl-CoA = 3 AMP + 3 diphosphate + CoA + penicillin G

In order to avoid CoA production as a byproduct, we added a third reaction . This was also found in the KEGG database after screening the reactions involving phenyl-acetyl CoA . This search yield this reaction :

4) ATP + Phenylacetic acid + CoA <=> AMP + Diphosphate + Phenylacetyl-CoA

Enzyme: 2Y4N

These reactions use commercially available reagents both as catalytic and consumable reagents:





Phenylacetic acid




Mevalonate - Sesquiterpenes - Artemisinin and Paclitaxel

In this example, we expanded the pathway until formation farnesyl pyrophosphate, which is a compound crucial to the synthesis of sesquiterpenoids. These compounds can be further modified chemically, generating a variety of odorants, bioactive compounds, flavouring agents (De Carvalho et al, 2006) and even major pharmaceutical blockbusters such as Paclitaxel (Guo et al, 2006), a major anti-cancer drug, and artemisinin, a Nobel Prize winning antimalarial drug (Wen et al, 2011)

Firstly we used the Metabolic Tinker (McClymont et al, 2013)to find suitable reactions starting from mevalonate and generating farnesyl pyrophosphate. This powerful bioinformatic tool takes as input the maximum number of compounds in each path, a similarity threshold, the substrate of choice and the product of choice and generates as output a variety of possible pathways, calculating the dG0 and indicating the Rhea (Alcantara et al, 2012) reactions.

Metabolic Tinker - Selected pathway (mevalonate to farnesyl pyrophosphate, maximum of 5 intermediates and similarity threshold of 0,5)


[RESULT: 5 ] ⇒ (R)-mevalonate {17066, 17065, 17068} (R)-5-phosphomevalonate {16342,16344, 16341} (R)-5-diphosphomevalonate {23733, 23732, 23735} isopentenyl diphosphate{19364, 19362, 19361} farnesyl pyrophosphate(3-) [dG0 = -121.4 ] [represents 81 unique paths]

Even though this pathway was not the one with the best dG0 value, the others required a variety of intermediates which would increase the complexity and the cost of our model.

For the sake of reducing the use of other reagents, thus making the pathway more elegant, looked for possible enzymes in KEGG.


KEGG Map for the same pathway

The analysis of the metabolic map indicated us 2 more enzymes which could prove helpful . These enzymes catalyse the intermediate reactions number 4 and 5 shown below.

We repeated the steps described previously to model the pathway with PDB files [Jinglin Fu et al, 2012] and UCSF Chimera (Pettersen, E. F, 2004)We selected the desired molecules both from the KEGG enzyme files or from the enzyme names obtained in the Metabolic Tinker. We handpicked proteins from pathogenic bacteria or E.coli, making it possible to infer that the optimal activity of the proteins would be similar and around 37ºC. The links for the chosen enzymes are available below each reaction.



1) (R)-mevalonate + ATP => (R)-5-phosphomevalonate + ADP + H+

Enzyme 2OI2:

(R)-5-phosphomevalonate + ATP < ? > (R)-5-diphosphomevalonate + ADP

Enzyme 1K47

(R)-5-diphosphomevalonate + ATP <=> ADP + CO2 + isopentenyl diphosphate + phosphate

Enzyme 2HK2

Isopentenyl diphosphate <=> Dimethylallyl diphosphate

Enzyme 1PVF

dimethylallyl diphosphate + isopentenyl diphosphate < ? > diphosphate + geranyl diphosphate

Enzyme 4LFG

geranyl diphosphate + isopentenyl diphosphate < ? > (2E,6E)-farnesyl diphosphate + diphosphate

Enzyme 1RQJ


Global reaction:

3 (R)-mevalonate + 6 ATP => (2E,6E)-farnesyl diphosphate+ 6 ADP + CO2

Although this cage does not produce a commercially important product itself, the possibilities of adding further modifications is of great importance. To assess the dimension of the possibilities of product formation from farnesyl pyrophosphate we used KEGG again:

Sesquiterpenoid synthesis pathways :

The map above shows a collection of products that can be readily made with the output of our cage . We sifted through them and selected 4 in order to provide a glance at the big picture:


Anti-malarial (Artemisinin) precursor:

(2E,6E)-farnesyl diphosphate <=> (+)-amorpha-4,11-diene + diphosphate -

The amorpha-4,11-diene is a small molecule produced as an intermediate in the generation of artemisinin by Artemisia annua [Wallaart, T. E. et al, 2001] . The further oxidation by CYP71AV1 produces dihydroartemisinic acid which is then converted to artemisinin by non enzymatic photo oxidation [Acton et al, 2002]. Although the mechanism of action is not fully understood, it might fight malaria through reactive oxygen species (ROS) production or/and inhibiting plasmodial sarcoplasmic reticulum calcium ATPase (SERCA), unleashing lethal consequences to Plasmodium falciparum and other Plasmodium species [Golenser et al, 2006] .Its semi-synthetic derivative artemether is one of WHO’s essential drugs [WHO, 2013]


Anti-tumoral (Taxol) precursor:

trans,trans-Farnesyl diphosphate + Isopentenyl diphosphate <=> Diphosphate + Geranylgeranyl diphosphate

Geranylgeranyl diphosphate <=> Taxa-4(5),11(12)-diene + Diphosphate

This first reaction would generate Geranylgeranyl diphosphate, which is the first compound in the diterpenoid biosynthesis, many of which are depicted in this KEGG map:

Among the many valuable substances inside this category it is worth highlighting Taxa-4(5),11(12)-diene, a precursor of the drug taxol (paclitaxel). This drug, originally extracted from Taxus brevifolia [Wani et al, 1971] is capable of controlling tumor acting on the cytoskeleton microtubules, blocking the depolymerization steps essential for proper cell cycle completion [Horwitz, S. B., 1993] . Taxol (or paclitaxel) is also a WHO essential medicine (Who, 2013) .

The following reagents are consumed in the mevalonate pathway reactions:




The next step is the actual ligation between DNA and the enzymes of choice. There are many available methods to perform this step:

The vast majority of approaches to couple proteins to scaffolds rely on linkers attached to nucleophilic amino acid side chains of residues such as cysteine and lysine or on the addition of various modifications on the N-terminal end of the proteins.

We propose two methods for protein anchoring to the DNA:



These crosslinkers must be heterofunctional so that one and binds to the DNA and another binds a protein functional group. The spacer arms of the crosslinkers should be small to keep the enzyme close to the cage’s edges. One of the DNA strands must have a 5’ end modification that could be a primary amine (-NH2) (figure 1) or a thiol (-SH) (figure 2) that, after the assembly of the cage, would form a nick gazing the cage’s interior.(Erben et al, 2006)

Using our software we can program predict where in the arris the nick will take place and where it’s facing:

Figure 2. DNA nicked with the addition of a primary amine group at the 5’ end. Carbons (grey), oxygens (red), phosphorus (orange), nitrogens (blue) and hydrogens (white). Only the backbone of the DNA is been shown. Image generated in the software UCFS Chimera.


Figure 3. DNA nicked with the addition of a thiol group at the 5’ end.Carbons (grey), oxygens (red), phosphorus (orange), sulfurs (yellow) and hydrogens (white). Only the backbone of the DNA is been shown. Image generated in the software UCFS Chimera.

Aiming to chose the best crosslinker, each enzyme must be analysed individually about the number and position of the possible functional groups for the linking:

  • Primary amines (-NH2), that exists at the N-terminus and in residues of lysine.
  • Thiols (-SH), that exists in residues of cysteine.
  • Carboxylic acids (-COOH),that exists in residues of glutamic acid and aspartic acid.

Since usually the proteins have more than one site with those functional groups the ligation may be random and not directed. The choice of the target functional group at the enzyme must be based in the number of existing sites (bigger the number, bigger the chance of successful linkage) and if one of these functional groups is part of the catalytic site (linking to this place would block the catalytic site).

Figures 4, 5, 6 and 7: Structures with substrates and ions for the four enzymes on the Penicilin Pathway - respectively: Phosphopantothenate synthetase, Isopenicillin N Synthase, Isopenicillin N acyltransferase and Phenylacetate-CoA ligase.


The four enzymes that compose the pathway of synthesis of penicillin are represented above. Some functional groups (possible linking sites) are highlighted: primary amine (red), carboxylic acid (blue) and thiol (yellow). These images were obtained from the software Chimera.

Table 1. Functional groups linked by the heterofunctional crosslinkers in the DNA and in the protein, the name and spacer arm size of each possible crosslinker is indicated. The possible crosslinkers were investigated and selected according to Thermofisher’s Crosslinking Reagents Handbook.


Aldehyde Tag

One recently developed approach is the generation of modified recombinant proteins that bear a site-specific sequence (Carrico et al, 2007). This amino acid sequence (CxPxR) is the consensus sequence recognized by Formylglycine Generating Enzyme (FGE) homologous proteins, which catalyze the conversion of that specific cysteine into a formyl-glycine residue, which bears a reasonably reactive and unique aldehyde in the protein.

The selection of a suitable region for the linkage site can then be performed in silico and the DNA sequence corresponding to (CxPxR, where x is any amino acid) added either in inside loops or in the ends of the coding sequence. The generation of aldehyde-tagged proteins is straightforward for molecular biology laboratories and a comprehensive protocol by Rabuka and colleagues is available (Rabuka et al, 2012).

After producing the enzyme in the model of choice (Chinese Hamster Ovary cells or E. coli, for example (Rabuka et al, 2012) it is necessary to couple the enzymes to the DNA. This step has already been described using a variety of DNA modifications, such as dimethoxytrityl(DMT)-protected aminooxy-modified DNA or Hydrazino-iso-Pictet−Spengler (HIPS) modified DNA in relatively mild conditions (Liang et al, 2014).)

Regarding the methods, the aminooxy - aldehyde reaction has relatively large background use (Broyer et al, 2011.) with applications not only for basic research, but for applied, such as vaccine development (Lees A, 2005.). However, the hydrazone/oxime bond formed is less stable than the one achieved with the novel HIPS ligation (Agarwal, 2013 ) which provide a C-C bond that would assure that the DNA and the protein remain linked for longer periods.

Following, the ligation should be validated and the DNA-protein complex separated from the mixture, in order to avoid formation of empty cages.



The easiest way to infer the protein-polynucleotide binding is through SDS-PAGE and visualization using protein staining methods (Coomassie Blue, for example (Steinberg, 2009)) or ssDNA staining methods (SYBR GREEN, SYBR GOLD, GEL STAR, Diamond Nucleic Acid Dye and others) *.The polynucleotide-protein complex should appear as a slower migration band in relation to the unconjugated enzymes and polynucleotides.

If you want to be really sure that the DNA is responsible for the shift, a further step of hybridization with antisense fluorescent probes could determine whether the shifting agent is truly the polynucleotide of choice (Liang et al, 2014).

Even though this migration shift assay is useful for linkage validation, it does not provide the enzyme-DNA complex for the nanocage assembly. In order to obtain pure DNA-enzyme complex, one approach to perform HPLC with anion exchange columns verifying the constitution of the eluates by measuring 260 and 280 nm absorbances, as described by Liang and colleagues (Liang et al, 2014) This step is also valuable for the determination of yield and optimization of the method. After performing HPLC, repeating the SDS-PAGE step with the purified and the crude samples can be an useful tool for assessing the presence of impurities.

DNA intercalating agents such as ethidium bromide may not be optimal, due to the fact that single stranded DNA polynucleotides are being used (Vardevanyan, 2003)



Now it is important to assess whether the binding with DNA caused any loss or increase of efficiency of the enzyme of choice. One of the ways of measuring the integrity of the protein is using circular dichroism, a technique which measures the effects of protein structures on polarized light to provide insights regarding the secondary structures (alpha helices and beta sheets, for example) and, consequently, about the tertiary structure (Kelly et al, 2005) .

This technique can also be used for DNA-protein conjugates, as exemplified by the work of Flory and colleagues. (Flory et al, 2014) They used Peptide Nucleic Acid (PNA)-protein conjugates as templates which hybridize with DNA polyhedral scaffolds. Even though we do not use PNA in our approach, the use of circular dichroism for assessing enzyme integrity is nevertheless a good option If the proper observations are made regarding, for example, the avoidance of measurements in the UV region, due to strong absorption by nucleotides, it is possible to get a good idea of proper conjugation.

The measurement of enzymatic activity is essential for ensuring the proper function of the conjugated complexes. These enzyme assays can assume many configurations that will depend on the pathway chosen and many full books on the subject are available (Cornish-Bowden, A. (2014).) (Wingard, L. B., Katchalski-Katzir, E., & Goldstein, L. (Eds.). (2014).). It is possible, for example, to do spectrophotometric/colorimetric [Jinglin Fu, 2012] and radiochemical[Sterri, S. H., 1985] assays.

Even though each desired reaction has different gold-standards, there are some "universal" methods for kinetic assays based on, for example, cross correlation of substrates labeled with two different fluorophores that are both split or joined together [Kettling, U., Koltermann, A., Schwille, P., & Eigen, M. (1998)] or in time-resolved fluorescence energy transfer (TR-FRET) based approaches with conjugation of the enzymes with a streptavidin–TB and fluorescent label of the substrate [Schiele, F., Ayaz, P., & Fernández-Montalván, A. (2015)]

An interesting approach that doesn’t require that many additional steps is the use of isothermal titration microcalorimetry (Todd, M. J., & Gomez, J. (2001).) , which uses pseudo-first order conditions (obtained after titration) and calculates the reaction rate based on thermic differences. The use of modern titration calorimeters allows assumption of a direct relation between this change in temperature and reaction speed and is suitable for enzymes in all E C. classification [Webb, E. C. (1992).]. groups .

The application of the enzyme activity assay of choice to the enzymes before and after the conjugation with DNA is, thus, the final step before assembling the catalytic nanocages. If the enzyme does not retain the desired activity values, a change of the reactions conditions ( temperature, pH, salt concentration and substrate saturation) or even the type of linker used could be helpful.



The software presented at the dry lab allows us to predict the nucleotide sequences to build a broad variety of nanocage geometries and different edges sizes. Some parameters may help to define the ideal cage geometry and size to host an specific enzymatic pathway.

The velocity of a global reaction of an enzymatic pathway depends on the diffusion of product of one enzyme to the next enzyme in the pathway. Thus, the distance between the enzymes helps to define the efficiency. Although in the nanocage the enzymes are compartmentalized, depending on the cages edge size the efficiency can also be manipulated. Reducing the cage size may bring the cages together and improve the global efficiency but at some point we suppose that it may starts to spoil the cage assembling. The geometry and the cage size can be designed to keep an optimal distance to avoid the assembling problem but also keep the high efficiency.

It is important to notice too that some enzymatic pathways have a limiting enzyme, which has the slowest enzyme activity and thus, defines the global reaction velocity. Using a cage geometry and size different to the ideal for a a pathway with a defined number of enzymes enables the possibility to anchor the slow enzyme more times and improve this step of the pathway.

The number of possible sites for enzyme anchoring is equal to the number of possible nicks (free 5’ DNA in one strand), which is the same as the number of different polynucleotides used. As each polynucleotide is directly responsible for the formation of all edges of one face, the total number of sites available for enzyme coupling is equal to the number of faces in the structure. This is only holds true for the faces with double stranded edges, as the faces originating from the truncation of the structure (with ssDNA edges) do not present suitable sites for anchoring.

The enzymes must be anchored at the middle of the double stranded edges for 3 main reasons. The first is that it is not possible to anchor them in the single strand sites because the presence of a "nick" in the region would disrupt the formation of the cage. The second reason for increasing the distances from the edges, because of steric hindrance: the proximity to the vertex would make it difficult to assemble the cage. The third one is that the region close to the single strand is more unstable since it has a bigger linear momentum (figure 8 | temporary supressed) [Icovelli et al, 2014], and, as a consequence would also give more freedom of movement to the enzymes. Therefore, the anchoring in the middle of the edges makes the enzymes more stable. With that in mind, we could also hypothesize that cage geometries with more truncations and the same length of edges would have a bigger freedom of movement.

The type of substrate that needs to enter the cage and the type of project that needs to get out must also be considered when choosing the cage’s geometry and size. Some molecules have a size that prevents it from entering or getting out of a cage, while others has a negative charge and are repulsed by the natural negative charge of the DNA. These two problems can be worked out by changing the sizes and geometry of the "holes". Bigger "holes" can be designed to allow bigger molecules to passage and also the negatively charged molecules, since the charge per area of the cage reduces.

Figure 8: Dynamics simulation of the kinectic behaviour of the nucleotides on a truncated octahedron generated by Polygen - a kind courtesy of Iacovelli et al. Mind the lower momentum of the nucleotides in the middle of the edge compared with the ones nearer the vertexes, where the DNA simple strands are used to make the curves.

We choose the truncated octahedron geometry for the cage to be build in the wet lab because it’s described in the literature as a stable cage (Falconi et al, 2009).

Figure 9. 3D scheme of a cage with a truncated octahedron geometry hosting the enzymes that compose the enzymatic pathway of synthesis of penicillin. Crosslinkers are not been shown. Image generated by the software Chimera.


Figure 10 and 11. 3D Scheme of two possible cages that could host the enzymatic pathway of artemisinin: a truncated dodecaedro and an hexagonal prism. Crosslinkers are not been shown. Image generated by the software Chimera.



In the end of these steps, the reagents for the DNA catalytic nanocage synthesis will be prepared.

Many protocols of nanometric DNA polyhedron assembly use high temperatures and conditions not so amenable to the enzymes [He et al, 2008]. However, in order to avoid denaturation of your enzymes, a milder process has to be performed.

The DNA nanocage assembly can, nonetheless, be performed at lower temperatures or even at room temperature, even though the yield might be compromised (Erben et al, 2006). The determination of an assembly protocol depends on the DNA structure chosen and on the number, positioning, size and format of the enzymes in the pathway of choice

One thing to have in mind is that the chemical conjugation process will make it impossible to join all the polynucleotide ends with ligases. Even though that this indeed decreases the stability of the final product, the vast amount of hydrogen bonds and the ligations in the other sites should make it possible to perform it without major complications.




The assembly of the nanocages can be performed in a variety of ways. Using agarose or acrylamide gels it is possible to check the electrophoretic mobility of the complex. Although the enzyme linked nanocage is heavier than the pure DNA polyhedron, a proper folding process would be seen only as a small shift. This is due to the fact that the final format is about the same in either cases. (Erben et al, 2006)

A confirmation of the assembly can be performed with many imaging methods. The use of techniques such as cryo-negative staining electron microscopy [De Carlo, S., 2011] and Atomic Force Microscopy [Andersen, E. S. et al, 2009] would be valuable approaches, being possible to see/infer the actual presence of proteins inside our nanocages. Using methodologies such as dynamic light scattering (DLS) it is possible as well to determine the size of the obtained particles, due to the differences in light scattering depending on particle dimensions (Berne, B. J., & Pecora, R. (2000).). This is a complementary methodology to confirm proper assembly of the cages.


The kinetic assays for the cages should be chosen in a similar way as described for the protein-ssDNA conjugates.


One of the challenges faced by the industry nowadays is to meet productivity with a need for environmental friendly approaches. This need gave rise to the expanding field of green chemistry (Sheldon, R. A, 2012) . Our initiative also aim facing those challenges. So we had to come up with a method to capture our cages back while maintaining their functionality.

Magnetic nanoparticles (MNPs) These nanoparticles can be made out of many metals [Lu et al, 2007], but the iron oxides magnetite (Fe3O4) and maghemite (γ-Fe2O3) are the most used, with wide scope of applications [ Laurent et al 2008]. Functionalization of iron oxide MNP compounds has been used for improving processes in areas ranging from nanohydrometallurgy - to recover copper trapping nanoparticles after electrodeposition [Toma, 2015] - to a plethora of enzyme immobilization approaches [Netto et al, 2013].

These nanoparticles can be functionalized in a variety of ways, such as streptavidin coating, which can be produced in the lab [Gong et al, 2013] or acquired from commercial sources . Streptavidin is a protein that has the ability to strongly bind to a small molecule called biotin (Kd~10^-13, [Weber et al, 1992]). We propose the insertion of biotinylated polynucleotides in the construction of the nanocages. The TEG-biotin molecules can be linked to the backbone of the nucleotide chains via spacers (Aldrich, S. (2015)) avoiding hindrance issues, specially if the modification is added to the single strand region of the cage . This is a reasonable strategy to facilitate the their retrieval through the attachment onto a streptavidin coated MNP. The MNP-nanocage association is theoretically simple and can be applied to any DNA cage, re-enforcing the modularity of the project.

We are not alone in this idea. Many groups are involved with this kind of functionalization, creating a strong scientific and technical support for this proposal. For example, the coupling of 3’-end biotinylated DNA aptamers to streptavidin-coated Cross-linked Iron Oxide (CLIO) magnetic nanoparticles has successfully been done regarding the retention of aptamer functional/structural properties (Yang, J. et al (2016)). Moreover, the coupling of streptavidin to nanocages has already been accomplished with 5’ biotinylation of strands, connecting the proteins to tetrahedra (Zhang, C et al. (2012)). Even though the cages used by Zhang et al were based on slightly different approaches than the ones projected by Polygen, they set the way for the efficient biotin-streptavidin functionalization of our "nanoboxes".

The diameter of the circumscribed sphere of our truncated octahedron (produced as a proof-of-concept for the Polygen software) is about 12 nm. The use of 20-100 nm streptavidin coated MNPs is, thus, one suitable option, enabling coupling of several cages to one structure. In order to achieve maximum efficacy and optimize the process, the systematic testing of MNPs with different sizes and different matrix materials should be performed before the final product is developed.

Figure 12: Illustration of how the nanocges could be connected with superparamagnectic nanobeads.


Taking all this by consideration we ratify, as a final step towards the Expanded Nanocage manufacturing, the functionalization of the cages with magnetic nanoparticles through streptavidin/biotin binding, thus creating a recoverable scaffold. That enables the execution of multiple batches and facilitates the retrieval of the products of interest, greatly enhancing the efficiency of the bioproductive process.



Fisher, A. K., Freedman, B. G., Bevan, D. R., & Senger, R. S. (2014). A review of metabolic and enzymatic engineering strategies for designing and optimizing performance of microbial cell factories. Computational and structural biotechnology journal, 11(18), 91-99.

McClymont, K., & Soyer, O. S. (2013). Metabolic tinker: an online tool for guiding the design of synthetic metabolic pathways. Nucleic acids research, 41(11), e113-e113.

Kanehisa, M., & Goto, S. (2000). KEGG: kyoto encyclopedia of genes and genomes. Nucleic acids research, 28(1), 27-30.

Yocum, R. R., Waxman, D. J., Rasmussen, J. R., & Strominger, J. L. (1979). Mechanism of penicillin action: penicillin and substrate bind covalently to the same active site serine in two bacterial D-alanine carboxypeptidases. Proceedings of the National Academy of Sciences, 76(6), 2730-2734.

Newman, D. J., Cragg, G. M., & Snader, K. M. (2000). The influence of natural products upon drug discovery. Natural product reports, 17(3), 215-234.

Buhaescu, I., & Izzedine, H. (2007). Mevalonate pathway: a review of clinical and therapeutical implications. Clinical biochemistry, 40(9), 575-584

Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., & Ferrin, T. E. (2004). UCSF Chimera—a visualization system for exploratory research and analysis. Journal of computational chemistry, 25(13), 1605-1612.

Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., … Bourne, P. E. (2000). The Protein Data Bank. Nucleic Acids Research,28(1), 235–242.

De Carvalho, C. C., & da Fonseca, M. M. R. (2006). Biotransformation of terpenes. Biotechnology advances, 24(2), 134-142.

Wen, W., & Yu, R. (2011). Artemisinin biosynthesis and its regulatory enzymes: Progress and perspective. Pharmacognosy reviews, 5(10), 189.

Corsello, M. A., & Garg, N. K. (2015). Synthetic chemistry fuels interdisciplinary approaches to the production of artemisinin. Natural product reports, 32(3), 359-366

Guo, B. H., Kai, G. Y., Jin, H. B., & Tang, K. X. (2006). Review-Taxol synthesis

McClymont, K., & Soyer, O. S. (2013). Metabolic tinker: an online tool for guiding the design of synthetic metabolic pathways. Nucleic acids research,41(11), e113-e113.

Alcantara, R., Axelsen, K. B., Morgat, A., Belda, E., Coudert, E., Bridge, A., ... & Steinbeck, C. (2012). Rhea—a manually curated resource of biochemical reactions. Nucleic acids research, 40(D1), D754-D760.

Wallaart, T. E., Bouwmeester, H. J., Hille, J., Poppinga, L., & Maijers, N. C. (2001). Amorpha-4, 11-diene synthase: cloning and functional expression of a key enzyme in the biosynthetic pathway of the novel antimalarial drug artemisinin. Planta, 212(3), 460-465.

Teoh, K. H., Polichuk, D. R., Reed, D. W., Nowak, G., & Covello, P. S. (2006). Artemisia annua L.(Asteraceae) trichome-specific cDNAs reveal CYP71AV1, a cytochrome P450 with a key role in the biosynthesis of the antimalarial sesquiterpene lactone artemisinin. Febs Letters, 580(5), 1411-1416.

Acton, N., & Roth, R. J. (1992). On the conversion of dihydroartemisinic acid into artemisinin. The Journal of Organic Chemistry, 57(13), 3610-3614.

Golenser, J., Waknine, J. H., Krugliak, M., Hunt, N. H., & Grau, G. E. (2006). Current perspectives on the mechanism of action of artemisinins.International journal for parasitology, 36(14), 1427-1441.

World Health Organization. (2013). WHO model list of essential medicines: 18th list, April 2013.

Wildung, M. R., & Croteau, R. (1996). A cDNA clone for taxadiene synthase, the diterpene cyclase that catalyzes the committed step of taxol biosynthesis. Journal of Biological Chemistry, 271(16), 9201-9204.

Wani, M. C., Taylor, H. L., Wall, M. E., Coggon, P., & McPhail, A. T. (1971). Plant antitumor agents. VI. Isolation and structure of taxol, a novel antileukemic and antitumor agent from Taxus brevifolia. Journal of the American Chemical Society, 93(9), 2325-2327.

Horwitz, S. B. (1993). Taxol (paclitaxel): mechanisms of action. Annals of oncology: official journal of the European Society for Medical Oncology/ESMO, 5, S3-6

Erben, C. M., Goodman, R. P., & Turberfield, A. J. (2006). Single‐molecule protein encapsulation in a rigid DNA cage. Angewandte Chemie, 118(44), 7574-7577

Liang, S. I., McFarland, J. M., Rabuka, D., & Gartner, Z. J. (2014). A Modular Approach for Assembling Aldehyde-Tagged Proteins on DNA Scaffolds. Journal of the American Chemical Society, 136(31), 10850-10853.

Carrico, I. S., Carlson, B. L., & Bertozzi, C. R. (2007). Introducing genetically encoded aldehydes into wproteins. Nature chemical biology, 3(6), 321-322

Rabuka, D., Rush, J. S., Wu, P., & Bertozzi, C. R. (2012). Site-specific chemical protein conjugation using genetically encoded aldehyde tags.Nature protocols, 7(6), 1052-1067.)

Steinberg, T. H. (2009). Protein gel staining methods: an introduction and overview. Methods in enzymology, 463, 541-563

Vardevanyan, P. O., Antonyan, A. P., Parsadanyan, M. A., Davtyan, H. G., & Karapetyan, A. T. (2003). The binding of ethidium bromide with DNA: interaction with single-and double-stranded structures. Experimental and molecular medicine, 35(6), 527-533.].

Kelly, S. M., Jess, T. J., & Price, N. C. (2005). How to study proteins by circular dichroism. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics, 1751(2), 119-139.

Flory, J. D., Simmons, C. R., Lin, S., Johnson, T., Andreoni, A., Zook, J., ... & Fromme, P. (2014). Low temperature assembly of functional 3D DNA-PNA-protein complexes. Journal of the American Chemical Society, 136(23), 8283-8295.

Broyer, R. M., Grover, G. N., & Maynard, H. D. (2011). Emerging synthetic approaches for protein–polymer conjugations. Chemical Communications,47(8), 2212-2226

Lees, A. (2005). U.S. Patent Application 11/044,866

Agarwal, P., Kudirka, R., Albers, A. E., Barfield, R. M., de Hart, G. W., Drake, P. M., ... & Rabuka, D. (2013). Hydrazino-Pictet-Spengler ligation as a biocompatible method for the generation of stable protein conjugates.Bioconjugate chemistry, 24(6), 846-851.

Cornish-Bowden, A. (2014). Principles of enzyme kinetics. Elsevier

Wingard, L. B., Katchalski-Katzir, E., & Goldstein, L. (Eds.). (2014).Immobilized enzyme principles: applied biochemistry and bioengineering (Vol. 1). Elsevier

Kettling, U., Koltermann, A., Schwille, P., & Eigen, M. (1998). Real-time enzyme kinetics monitored by dual-color fluorescence cross-correlation spectroscopy. Proceedings of the National Academy of Sciences, 95(4), 1416-1420

Schiele, F., Ayaz, P., & Fernández-Montalván, A. (2015). A universal homogeneous assay for high-throughput determination of binding kinetics.Analytical biochemistry, 468, 42-49

Webb, E. C. (1992). Enzyme nomenclature 1992. Recommendations of the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology on the Nomenclature and Classification of Enzymes (No. Ed. 6). Academic Press

Todd, M. J., & Gomez, J. (2001). Enzyme kinetics determined using calorimetry: a general assay for enzyme activity?. Analytical biochemistry,296(2), 179-187.)

He, Y., Ye, T., Su, M., Zhang, C., Ribbe, A. E., Jiang, W., & Mao, C. (2008). Hierarchical self-assembly of DNA into symmetric supramolecular polyhedra.Nature, 452(7184), 198-201.

Berne, B. J., & Pecora, R. (2000). Dynamic light scattering: with applications to chemistry, biology, and physics. Courier Corporation.

Jinglin Fu, Minghui Liu, Yan Liu, Neal W. Woodbury, and Hao Yan; "Interenzyme Substrate Diffusion for an Enzyme Cascade Organized on Spatially Addressable DNA Nanostructures"; 2012, Journal os the American chemical society

Sterri, S. H., Johnsen, B. A., & Fonnum, F. (1985). A radiochemical assay method for carboxylesterase, and comparison of enzyme activity towards the substrates methyl [1-14 C] butyrate and 4-nitrophenyl butyrate. Biochemical pharmacology, 34(15), 2779-2785.

Jinglin Fu, Minghui Liu, Yan Liu, Neal W. Woodbury, and Hao Yan; "Interenzyme Substrate Diffusion for an Enzyme Cascade Organized on Spatially Addressable DNA Nanostructures"; 2012, Journal os the American chemical society

De Carlo, S., & Harris, J. R. (2011). Negative staining and cryo-negative staining of macromolecules and viruses for TEM. Micron, 42(2), 117-131.

Andersen, E. S., Dong, M., Nielsen, M. M., Jahn, K., Subramani, R., Mamdouh, W., ... & Kjems, J. (2009). Self-assembly of a nanoscale DNA box with a controllable lid. Nature, 459(7243), 73-76.

Zhang, C., Tian, C., Guo, F., Liu, Z., Jiang, W., & Mao, C. (2012). DNA‐Directed Three‐Dimensional Protein Organization. Angewandte Chemie International Edition, 51(14), 3382-3385.

Iacovelli, F., Alves, C., Falconi, M., Oteri, F., Oliveira, C. L., & Desideri, A. (2014). Influence of the single‐strand linker composition on the structural/dynamical properties of a truncated octahedral DNA nano‐cage family. Biopolymers, 101(10), 992-999.

Sheldon, R. A. (2012). Fundamentals of green chemistry: efficiency in reaction design. Chemical Society Reviews, 41(4), 1437-1451.

Lu, A. H., Salabas, E. E., & Schüth, F. (2007). Magnetic nanoparticles: synthesis, protection, functionalization, and application. Angewandte Chemie International Edition, 46(8), 1222-1244.

Laurent, S., Forge, D., Port, M., Roch, A., Robic, C., Vander Elst, L., & Muller, R. N. (2008). Magnetic iron oxide nanoparticles: synthesis, stabilization, vectorization, physicochemical characterizations, and biological applications. Chemical reviews, 108(6), 2064-2110

Toma, H. E. (2015). Magnetic nanohydrometallurgy: a nanotechnological approach to elemental sustainability. Green Chemistry, 17(4), 2027-2041.

Netto, C. G., Toma, H. E., & Andrade, L. H. (2013). Superparamagnetic nanoparticles as versatile carriers and supporting materials for enzymes.Journal of Molecular Catalysis B: Enzymatic, 85, 71-92.

Gong, P., Peng, Z., Wang, Y., Qiao, R., Mao, W., Qian, H., ... & Shi, S. (2013). Synthesis of streptavidin-conjugated magnetic nanoparticles for DNA detection. Journal of nanoparticle research, 15(4), 1-11.

Weber, P. C., Wendoloski, J. J., Pantoliano, M. W., & Salemme, F. R. (1992). Crystallographic and thermodynamic comparison of natural and synthetic ligands bound to streptavidin. Journal of the American Chemical Society, 114(9), 3197-320

Aldrich, S. (2015), Custom Oligonucleotide Modifications Guide, Sigma Aldrich, St. Louis, MO, USA

Yang, J., Donolato, M., Pinto, A., Bosco, F. G., Hwu, E. T., Chen, C. H., ... & Hansen, M. F. (2016). Blu-ray based optomagnetic aptasensor for detection of small molecules. Biosensors and Bioelectronics, 75, 396-403

Zhang, C., Tian, C., Guo, F., Liu, Z., Jiang, W., & Mao, C. (2012). DNA‐Directed Three‐Dimensional Protein Organization. Angewandte Chemie International Edition, 51(14), 3382-3385.

Falconi, M., Oteri, F., Chillemi, G., Andersen, F. F., Tordrup, D., Oliveira, C. L., ... & Desideri, A. (2009). Deciphering the structural properties that confer stability to a DNA nanocage. ACS nano, 3(7), 1813-1822.