Synthetic Biology: Tools and Applications

I

Synthesis and Engineering Tools in Synthetic Biology

Chapter 1 New Tools for Cost-Effective DNA Synthesis

Chapter 2 Protein Engineering as an Enabling Tool for Synthetic Biology

Chapter 3 Pathway Engineering as an Enabling Synthetic Biology Tool

Chapter 4 From Biological Parts to Circuit Design

Chapter 1

New Tools for Cost-Effective DNA Synthesis

Nicholas Tang¹, Siying Ma¹ and Jingdong Tian¹, ², ¹Duke University, Durham, NC, USA, ²Chinese Academy of Sciences, Tianjin, China

Introduction

DNA synthesis is a powerful enabling technology and yet a limiting step in synthetic biology. Decreasing costs in DNA synthesis will open new frontiers and project concepts that would not be feasible to most scientists at current levels. The cost of column synthesized oligonucleotides has dropped 10-fold over the past 15 years,¹ and is currently around USD0.08–0.2 per nucleotide.² Gene synthesis is a more expensive process, and involves generating longer DNA constructs from overlapping oligonucleotides. In 2000, the market cost of gene synthesis was approximately USD10 per base.³ Since then, prices have dropped nearly 50-fold in 10 years, to as low as USD0.2 per base pair, with average error rates of about 1 in 300–600 bases.¹,² Prices continue to drop by a factor of 1.5 per year, in a manner akin to Moore’s law.¹ By 2005, there existed 39 gene synthesis vendors worldwide, and the number has increased since then. In the past couple of years, more companies have begun to emerge to take on the challenge of adapting new synthesis technologies for the market. It would be interesting to see dramatic changes to the gene synthesis market in the near future.

Oligonucleotide Synthesis

Column Oligonucleotide Synthesis

Chemical DNA synthesis can be used for applications as common as primer, linker, or probe synthesis. Here we discuss oligonucleotide synthesis for the application of gene and genome synthesis. Chemical assembly of DNA using programmable synthesizers is now a routine procedure. The most common and reliable system for chemical synthesis involves synthesizing individual oligonucleotides in small columns. A series of valves and pumps introduce the correct nucleotide monomers and reagents required for growing oligomers of DNA in a stepwise manner. Chemical oligodeoxynucleotide synthesis is different from enzymatic DNA synthesis in living cells in that it is a cyclical process that elongates nucleotides from the 3′ to the 5′-end. The starting complex for chemical synthesis of DNA consists of an initial acid-activated nucleoside phosphoramidite tethered with a spacer to a solid-support controlled pore glass (CPG) or polystyrene (PS) bead. The advent of solid-phase DNA synthesis made automation possible by eliminating purification steps to remove intermediates or unreacted reagents. The column is simply rinsed with anhydrous acetonitrile to remove these reagents, and then purged with argon to remove the remaining acetonitrile.

The cyclic addition of additional monomers to the existing oligonucleotide chain occurs in four steps: deprotection, activation/coupling, capping, and oxidation (Fig. 1.1). Each additional nucleoside which is added to the growing chain has a 5′DMT protection group. This assembly is called a phosphoramidite. The four-step phosphoramidite chemistry is the method of choice for most commercial DNA synthesizers because the yields are more accurate and homogeneous than other methods.⁴ First, a strong acid is used to de-block the 5′-O-4,4′-dimethoxytrityl (DMT) group, removing the protecting group from the nucleotide chain and exposing a reactive OH group. In the next step, 1H-tetrazole and the dissolved phosphoramidite are simultaneously added to the column. Tetrazole, a weak acid, protonates the trivalent phosphorus on the 3′-end of the monomer. This results in a slow displacement of the secondary amine and formation of a highly reactive tetrazolide that then immediately couples with the OH group. At this point, the added phosphoramidite is coupled to the existing chain. Uncoupled 5′-OH groups are blocked by an acylating capping reagent, usually acetic anhydride, to minimize deletion products. Finally, the unstable phosphite triester internucleotide linkage between nucleotides is oxidized to a more stable pentavalent phosphotriester. The end results of this process are oligonucleotide strands that are bound to beads. Each phosphate bond contains a methyl group, which can be removed by chemical treatment in the reaction column. The 5′ terminus of the last nucleotide can be deprotected through detritylation of the DMT group, and phosphorylated by T4 kinase. DNA strands can also then be cleaved from the spacer linker off the solid support.

Figure 1.1 The solid-phase, four-step oligodeoxynucleotide synthesis cycle.

The described four-step synthesis procedure has been the basis of fully automated DNA synthesizers with up to 1536 sequence throughputs.⁵ Throughput evolved from 2–4 individual sequences in initial synthesizers manufactured by Applied Biosystems, to 96 well plates in 1995.⁶ Lashkari et al. used computer-controlled solenoid valves to deliver bulk reagents through Teflon tubes into a microwell plate. Since then, parallel synthesis using multiplexed reagent delivery lines has allowed for synthesis in other microwell plate formats.

Optimizations in reaction chemistry include a two-step cycle synthesis, which reduces costs by eliminating several reagents.⁷ A peroxy anion is used as a nucleophile to remove a 5′-carbonate and oxidize the internucleotide phosphite triester. Deprotection with peroxy anion under mildly basic conditions can eliminate depurination, a side reaction that leads to mutations in synthetic DNA. If further developed, the two-step synthesis process can make oligonucleotide synthesis simpler, and consequently more robust.

Microarray Oligonucleotide Pool Synthesis

The major costs for gene synthesis are attributed to oligonucleotide synthesis, sequence verification, and labor for processing steps. Microarray-enabled oligonucleotide pool synthesis effectively tackles oligonucleotide costs. Since their inception in 1995, microarrays have dramatically revolutionized genomics with massive parallelism and automation. Microarrays are 2D solid-phase arrays used to assay or screen biological materials like nucleic acids, proteins, or cells. Oligonucleotide arrays can be used for a variety of designs, including gene expression screens, SNP genotyping, comparative genomic hybridization (CGH), tiling, ChIP-on-chip, microRNA, resequencing, and aptamer screening.⁸ More relevantly, oligonucleotide microarrays can offer significant reductions in DNA synthesis cost due to their dense and massively parallel feature designs. For example, reducing costs by scaling down reagent volumes in resin-based synthesis is restricted to decreasing the diameter of capillaries. Simply eliminating the rinsing of lines in instrumentation makes at least a 10-fold reduction in cost.⁶ This is because acetonitrile is often used in high volumes for rinsing and dissolving reagents, making it one of the most expensive bulk reagents.⁹ Cost for reagents in a custom inkjet microarray slide is less than USD50 per slide, due to the low volumes of phosphoramidite and tetrazole necessary for the miniaturized platform.¹⁰ Microarray synthesis not only lowers costs, but also makes synthesis more environmentally friendly. High-density arrays can offer 10⁴–10⁶ unique oligonucleotides, and can reduce costs by at least an order of magnitude.¹¹,¹² The price for 3912 90-mers from LC Sciences is about USD 1000; the Agilent 55 k chip, ~USD 7000, which translates to USD 0.0025 per base.¹³ These costs do not include downstream costs for gene assembly.

The variety in technologies and techniques for microarray-based oligonucleotide synthesis is expected to offer reductions in cost and improvements in throughput in the coming years. In contrast, the cost of column-synthesized oligonucleotides has remained constant over the past six years and is unlikely to decrease significantly.¹⁴ The microarray technologies that exist in the DNA synthesis market include ink-jet printing (Agilent, Protogene), photosensitive 5′ deprotection (Nimblegen, Affymetrix, Flexgen), photo-generated acid deprotection (Atactic/Xeotron/Invitrogen, LC Sciences), and electrolytic acid/base arrays (Oxamer, Combimatrix/Customarray).

Optimizations that have been made to microarray synthesis include synthesis on PDMS (poly(dimethylsiloxilane))¹⁵ and COC (cyclic olefin copolymer)¹⁶ substrates as low-cost and flexible alternatives to glass. Ma and coworkers showed that oxCOC, a hybrid substrate composed of COC and RF sputtered SiO2, takes benefits from both constituents.¹⁷ COC offers low density, resistance to organic solvents, high stiffness, and UV transparency, while SiO2 offers useful surface linkers for phosphoramidite chemistry. Furthermore, oxCOC can be manipulated for large-scale production. With soft lithography, a PDMS stamp can be made to imprint channels or wells on COC before thin-film deposition of SiO2.¹⁸ The stamps are disposable and can be created in a non-cleanroom setting with silicon molds, although the silicon molds must be fabricated via photolithography. A study involving inkjet microchip synthesis demonstrated that using a COC chip with patterned silica features reduced the error rate of synthesized oligonucleotides from one in 200 bases to one in 600 bases,¹⁶ which is equivalent to high-quality column synthesis.

The dominant costs are now enzymatic processing, cloning, and sequencing.¹² A recent study by Tian’s group addresses the limitations of processing steps using an integrated combination of isothermal nicking, strand displacement amplification (nSDA), and polymerase cycling assembly (PCA), reducing the cost even further to USD 0.005 per base with an error frequency of <0.2 errors/kb.¹⁹ Such reduced cost could make gene and gene library synthesis more widely accessible.

Microfluidic and Fluidic Systems

Microfluidics allows for the control and manipulation of small volumes of liquids. These features make it particularly useful for a variety of applications like PCR and cell screening. More specifically, microfluidics can decrease the space and cost requirements of DNA synthesis. A recent study reports a programmable microfluidic synthesis platform that can synthesize ~100 pmol of each unique oligonucleotide, which is substantial enough not to require PCR amplification steps before gene assembly.²⁰ The output levels from this study better match the amounts necessary for gene assembly. Using microfluidics for DNA synthesis reduces reagent consumption by 100-fold compared to conventional column synthesis which generates far more oligonucleotides than necessary for gene assembly. On the other hand, although microarrays can synthesize a large number of oligonucleotides, they produce only 10⁶–10⁸ molecules per spot and yield less than 2 fmol. This yield is six orders of magnitude lower than those of column-synthesized oligonucleotides,²⁰,²¹ which are too low for gene assembly without amplification. Although amplification methods have been refined, PCR amplification can introduce errors and increase overall labor and processing costs. Additionally, there is a 100-fold reduction of reagent consumption compared to other solid-phase synthesis technologies that have less efficient deblocking steps.²⁰ DNA synthesis with microfluidics as performed in this study can further advance lab-on-a-chip (LOC) or Micro Total Analysis Systems (µTAS). However, the 17–24 mer length and 1 in 153 bp error rates could still be refined to be suitable for longer gene constructs.

Although microfluidics allows for massively parallel microscale DNA synthesis, there are a few key technical challenges. One challenge is organization of addressable synthesis units. Xiao et al.²² used soft lithography to synthesize oligonucleotide arrays on glass surfaces. During the coupling step of synthesis, pre-cast PDMS microstamps transferred a mixture of nucleoside monomer and tetrazole onto the glass slides surface. Like photomasks in photolithography, different stamps dictate predefined areas for the coupling of transferred monomer. 20 mers were synthesized with high coupling efficiency, with a stepwise yield of 97%.

Blair et al.²³ used 365 nm wavelength ultraviolet-light-emitting diodes (UV-LEDs) as a cost-effective alternative light source for addressing and directing oligonucleotide synthesis inside glass capillaries. A string of UV-LEDs were positioned along the length of a glass capillary. The inside walls were functionalized for oligonucleotide synthesis using photolabile 2-nitrophenyl propoxycarbonyl (NPPOC) chemistry.²⁴ The glass capillaries were designated capillary synthesis cells (CSC) in their fluidic system. Because the spectrum of the UV-LED contained no emission below 360 nm, no filters were needed to prevent DNA damage from short wavelength radiation. 70 mers were successfully synthesized and used for gene assembly.

Another technical challenge is the chemical resistance of microfluidic substrates. For example, elastomers are a popular material for microfluidic devices because of their fabrication cost and labor benefits over silicon. However, elastomers must be chosen or modified to be chemically resistant to the organic solvents used in oligonucleotide synthesis. Popular elastomers like poly(dimethylsiloxane) (PDMS) degrade, swell and clog in contact with DNA synthesis reagents, which results in as high as a 90% flow rate drop.²⁵ Moorcroft et al.¹⁵ circumvented this issue by substituting both the oxidation and deprotection solvents in conventional DNA synthesis. PDMS microchannels were first molded by a soft lithography process. Then, the surface was functionalized by silanizing with 3-glycidoxypropyltrimethoxysilane (GPTMS) and adding a PEG spacer. 21 mers were synthesized on the PEG-Silane-PDMS chip. The Quake group used a chemically resistant photocurable perfluoropolyether (PFPE) for their microfluidic synthesis chambers.²⁶ To avoid functionalization, the synthesis chambers were made of PFPE, while oligonucleotide synthesis was performed on porous silica beads. The device was used to synthesize 60 pmol of 20 mers oligonucleotides with 60-fold less reagent consumption than conventional synthesis.

Another chemically resistant substrate is carbon. Carbon offers superior heat and chemical stabilities compared to glass, allowing for robust linkage of oligonucleotide probes.²⁷ Phillips et al. functionalized carbon surfaces for light-directed oligonucleotide array synthesis using RF plasma treatment and 9-decene-1-ol. The authors used two types of carbon as substrates for DNA synthesis: glassy carbon and CVD diamond. Oligonucleotide synthesis was then performed on these surfaces with photolabile NPPOC chemistry. Despite robust oligonucleotide linkage, carbon-based DNA probe arrays²⁸,²⁹ are not yet commonly used, due to their low-density probe coverage.

Hua and Gulari²⁵ reported a platform that integrated a pneumatic microvalve array in a microfluidic system to precisely control the flow of fluidics for parallel oligonucleotide synthesis. They coated the PDMS valve membranes with parylene to make them chemically resistant and compatible with aggressive chemical reagents. These valve membranes were sandwiched between a PDMS and a silicon layer. The PDMS layer provided air channels, while the fluid reactions occurred on the silicon layer. The pressurized air channels closed the valves and restricted fluids from leaving. Sixteen multiplexed air channels were sufficient to control 12 870 reactors. 30 mers were successfully synthesized with a stepwise synthesis yield of ~99.5%.

Photolithography

Some of the earliest efforts to synthesize oligonucleotides on microarrays involved using physical photolithographic masks to direct a light pattern on nucleoside monomers with photolabile protecting groups.³⁰,³¹ Light patterns are directed to areas on the array to remove photolabile protecting groups from the growing oligomers. After deprotection, a selected phosphoramidite monomer is spread over the entire surface and only the exposed areas are activated for coupling of that particular phosphoramidite. Masks must be prefabricated for each cycle so that the next bases on the synthesized oligonucleotides are mapped to the areas of exposure. Affymetrix Inc. applied this technology to large-scale fabrication of high-density GeneChip probe arrays.

The costs for optics, light sources, and pre-fabricated photolithographic masks can be prohibitive. This is partly because, in principle, the number of necessary masks is four times the number of bases in the oligonucleotides. Texas Instrument’s DMD (Digital Micromirror Device) eliminates the necessity for physical masks, greatly reducing processing time and costs for light-directed oligonucleotide synthesis. A DMD device consists of an electromechanically controlled array of micromirrors.³² It is typically used for DLP (digital light processing) projection display systems, with resolutions as high as 2073 600 pixels (SVGA). With flexible programmable light patterns, the generation of arrays can be automated.

(R,S)-1-(3,4-(methylenedioxy)-6-nitrophenyl)ethyl chloroformate (MeNPOC) or 2-(2-nitrophenyl)propoxycarbonyl (NPPOC) can be used in photo-deprotection to block function groups on the linker or monomers.³²,³³ In one study, a 600×800 DMD array of 16 μm wide micromirrors was used to synthesize 41 mer oligonucleotides.³²,³³ In this work, sub-populations of oligonucleotides were eluted, amplified, and assembled. The authors extrapolated that a chip with 786 432 unique 40 mer oligonucleotides could potentially be used to assemble a >15 Mb construct. FlexGen’s in-house synthesizer, the Flexarrayer, generates oligonucleotides in a similar manner. Nucleotide deprotection is activated by a laser before the nucleotides are washed and bound to activated spots to produce 60 mers.

Another study used a photoacid generator (PAG), triarylsulfonium hexafluoroantimonate, to perform the deblocking step.³⁴ As with conventional acid-deblocking nucleotide phosphoramidite chemistry, acids are generated in the ditritylation step, but with a UV photolytic process. The resulting PGA solution is at a high enough mM concentration to effectively remove the DMT group on a nucleotide, and is more efficient than photo-cleavage methods. The main advantage of this method is that it relies on conventional phosphoramidite chemistry, without the cost or availability limitations of photolabile protection groups. Photoacid generator (PAG) may liberate free radicals to produce errors in synthesis. Serafinowski and Garland³⁵ developed and used two photosensitive esters, R-phenyl-4,5-dimethoxy-2-nitrobenzyltrichloroacetate and R-phenyl-4,5-dimethoxy-2,6-dinitrobenzyltrichloroacetate, as PAG agents for synthesis to avoid these problems.

Zhou et al.¹¹,³⁶ coupled the PAG technique with a parallelized microfluidic platform. An array chip was installed in a flow-through cartridge connected to a commercial oligonucleotide synthesizer (Expedite 8909). The chips used differential surface tension to isolate reaction sites, and allowed for the miniaturization of synthesis. Cleaved oligonucleotides were of high enough quantities and qualities to be successfully assembled in 10 kb constructs with PCR and ligation.

Electrochemical Arrays

Another method of DNA synthesis involves using localized electrochemical reactions for the deblocking step. An array of platinum microelectrodes is covered with an electrolyte solution. Current is simultaneously applied to individually addressable microelectrodes with a semiconductor circuit, so that the electrolytes near activated anodes are oxidized and release acid for deblocking. In one study, the electrolyte used was 25 mM hydroquinone and 25 mM benzoquinone with 25 mM tetrabutylammoniumhexafluorophosphate in anhydrous acetonitrile.³⁷ By pairing each anode with adjacent cathodes which can reduce acid, the produced acid is confined. The acid then diffuses to the layer of substrate which contains the oligonucleotides being synthesized. The electrodes were made by thin-film photolithography of 50-nm thick iridium metal and were durable enough to use over 500 cycles without deteriorations. The authors have demonstrated the synthesis of short 17 bp oligonucleotides with complete electrochemical deblocking in as little as 9 seconds.

Customarray’s oligonucleotides arrays are also synthesized using a semiconductor-based electrochemical synthesis process. Rather than using cathodes to localize generated acids, synthesis is performed on a polymer surface over the surface of the semiconductor. The porous polymer slows down acid diffusion to prevent cross-contamination between local electrodes. The surface also increases the oligonucleotide density during synthesis. Customarray’s 90 K microarrays contain 94 000 unique oligonucleotides and 25 µm features, and are capable of synthesizing oligonucleotides of up to 50 bp lengths.

Besides feature density, another advantage of electrochemical synthesis is flexibility. Because photolithographic masks are not involved, it is less expensive and requires less labor to change the array design. Additionally, electrochemical detritylation is a relatively efficient chemistry; the deblocking step can be complete in seconds and any side reactions from the electrolyte chemicals are minimal. Efficient chemistry and a lack of moving parts contribute to a total synthesis time of less than 24 hours.

Inkjet Printing

In 1981, Hood and Caruther modified liquid phase phosphite-triester chemistry for solid-phase DNA synthesis on polymer supports.³⁸ Solid-phase synthesis not only became the method of choice for conventional DNA synthesis, but also opened DNA synthesis to automated and miniaturized strategies on microarrays by eliminating the need to purify synthetic intermediates or unreacted reagents. The first inkjet DNA synthesizers used in-house microfabricated piezoelectric actuators.⁸ Piezoelectric inkjet DNA synthesizers operate similarly to commercial inkjet printers. Instead of ink, the inkjet head contains four phosphoramidite fluid channels, one activator channel, and one optional modified base channel. In this work, the discrete features on the substrate could hold and segregate 100 pL droplets using pre-patterned surface tension (Fig. 1.2).

Figure 1.2 DNA microchip synthesis with piezoelectric inkjet technology. The microarray synthesis platform consists of a controller PC, DAQ-electronics, and a servo controller, which together control the movement and the printing of the inkjet print head onto slide surfaces. Droplets are dispensed onto as many as 10 000 spot features that are each less than 100 µm in