October 1, 2011 (Vol. 31, No. 17)
Richard A. A. Stein M.D., Ph.D.
Vast Potential of Emerging Science Being Realized through Landmark Research
A fascinating observation made decades ago, and subsequently reinforced in the post-genomics era, is the lack of correlation between the genome size and the complexity of an organism, a concept that became known as the C-value paradox.
For example, the genome of the small protozoan Amoeba proteus, estimated to have 670 billion base pairs, is approximately 200 time larger than the human genome; the unicellular eukaryote Paramecium tetraurelia has approximately 40,000 genes; Drosophila melanogaster has approximately 13,000 genes; and the flowering plant Arabidopsis thaliana has around 28,000 genes, comparable to the 20,000–25,000 protein-coding genes from the human genome.
These findings sparked an interest in understanding how certain organisms can achieve much more complex tasks than others, despite using comparable numbers of genes. Alternative splicing and noncoding genomic regions are two of the features that explain the complexity of higher eukaryotes. An additional theme that emerged in recent years is the need to shift the focus from exploring isolated components toward visualizing the complexity of the biological networks, which often results from the modularity of the components and the novel combinations that they can establish.
Understanding how the different components of a circuit function together helps unveil evolutionary concepts, reveals basic principles behind their design, and facilitates the construction of novel synthetic circuits that can perform a desired function.
One of the most recent inter-disciplinary undertakings at the juncture between life sciences and engineering, synthetic biology emerges as a dynamic area of investigation that proposes to design and generate complex biological systems with novel properties. Recent progress in areas that include molecular biology, protein engineering, computational biology, and the development of the -omics sciences helped this vibrant discipline become reality.
One of the challenging topics in science has revolved around the fact that measurements of cellular components, such as mRNA or proteins, often reflect population averages but are less informative about individual cells. As a result, cell-to-cell differences are lost in the average. More recent approaches, powered to examine biological processes within individual cells, reveal a considerable cell-to-cell variability, or noise, even among genetically identical cells.
In 1957, Novick and Weiner reported that at intermediate inducer concentrations, an E. coli population consisted of cells fully induced for lac expression and cells that were not induced at all, with individual cells being able to switch stochastically between these two states. Intercellular variability is advantageous for cell populations that are exposed to fluctuating environments.
Cohen et al. revealed that cancer cells respond differently to the same therapeutic agent and, as a result, a small population may survive. The authors monitored the expression over time of almost 1,000 different proteins in human cancer cells, and found bimodal behavior in a subset of proteins, with increased cell-to-cell variability after addition of the anticancer drug camptothecin, an observation that explains the ability of some cells to escape anticancer therapy.
Previously, modeling studies proposed that negative feedback regulation could provide stability to biological systems, but for a long time, this was not tested experimentally. Becskei and Serrano designed a synthetic gene circuit in E. coli and demonstrated that negative feedback can considerably reduce variability in gene expression.
In this system, the tetracycline repressor-regulated promoter PLtetO1 drives the expression of TetR-EGFP, a fusion between the tetracycline repressor and the enhanced green fluorescence protein. Negative feedback is established because TetR represses transcription from PLtetO1. When the tet operator was replaced with the lac operator, or when TetR was replaced with the low binding affinity mutant TetRY42A, elimination of the feedback produced an unregulated system. This model revealed an over threefold decrease in gene-expression variability in the presence of the negative feedback, as compared to an unregulated system.
Around the same time, Elowitz and Leibler designed and constructed a synthetic oscillating network known as repressilator by using three transcriptional repressors that are not part of a naturally occurring biological system. In this system, E. coli LacI inhibits the transcription of the second repressor, tetR from the Tn10 tetracycline-resistance transposon, and the TetR protein inhibits the expression of a third gene, cI from bacteriophage λ.
The product of this gene, CI, inhibits lacI expression, completing the cycle. This system revealed that genetic components from multiple systems can be used and combined in a new genetic context to construct artificial networks with new properties.
Gardner et al. constructed a genetic toggle switch by using two repressors and two constitutive promoters, each of them inhibited by the repressor transcribed by the other promoter. This study brought a fundamental contribution to the field of synthetic biology, and represented a departure from the classical genetic engineering in the sense that DNA or protein engineering were replaced with an approach that involves manipulating the network architecture, a promising emerging concept in biotechnology and gene therapy.
Another study that pioneered the field of synthetic biology, conducted by Farmer and Liao, emerged from the observation that a common strategy used in life sciences—the induction of a particular enzyme or pathway—often caused cellular metabolic imbalances resulting in phenotypic changes.
Underscoring the importance of visualizing the dynamics of a metabolic pathway, as opposed to simply focusing on its genetic composition, the authors engineered, in E. coli, the synthesis pathway of lycopene, an antioxidant with applications in chemotherapy and neurodegenerative conditions, by coupling gene expression to the metabolic state of the cell.
Three of the five genes encoding enzymes in the lycopene synthesis pathway were placed under the lac promoter, and two genes, which control the rate limiting steps, were controlled by the glnAp2 promoter, whose transcription is initiated by the active, phosphorylated version of the response regulator NRI. A modified version of the E. coli Ntr regulon was generated, in which the sensor kinase NRII was deleted to make NRI more responsive to acetyl phosphate. This engineered strain produced 100 mg/L lycopene after growth in glucose-containing medium.
Promises for Clinical Medicine
The construction of synthetic genetic circuits was pioneered in bacteria, and over a decade elapsed until these concepts were applied to mammalian systems. Kemmer et al. reported the first synthetic closed-loop mammalian genetic circuit designed to correct a pathological state. The authors took advantage of a modified version of HucR, a bacterial transcriptional repressor from Deinococcus radiodurans, which binds its operator hucO in the absence of uric acid.
Upon sensing uric acid levels, this module triggers the synthesis of an engineered, secreted version of an Aspergillus flavus urate oxidase. In urate oxidase-deficient mice, an experimental model of hyperuricemia and gout, the authors reported that this system reduced blood urate concentration and uric acid deposition in the kidney.
In another tour de force that promises to reshape our perspectives on cancer therapy, Anderson et al. introduced the idea that engineered bacteria could integrate information from multiple sources to recognize, invade, and destroy mammalian cancer cells. The authors genetically engineered an E. coli strain expressing Yersinia pseudotuburculosis invasin and, by placing this gene under a Vibrio fisheri quorum sensing operon, an anaerobically induced promoter, or an arabinose-inducible promoter, they linked bacterial internalization to environmental cues such as hypoxia, cell density, or inducible stimuli. This strategy opens therapeutic promises to engineer bacterial cells that can recognize a tumor microenvironment.
Synthetic biology impacts several facets of the drug discovery process, and opens new perspectives toward addressing long-standing therapeutic challenges, such as infections caused by microbial biofilms, or antibiotic resistance, a crisis of global proportions. Illustrating the powerful applications in infectious diseases, Lu and Collins engineered the E. coli-specific T7 bacteriophage to express DspB, an Actinobacillus actinomycetemcomitans protein that hydrolyzes β-1,6-N-acetyl-D-glucosamine, a crucial adhesin from biofilms produced by many bacterial species.
The initial phage infection of a bacterial biofilm causes the intracellular production of progeny phage and DspB, which are both released by the lysing cells. DspB degrades the extracellular polymeric matrix of the biofilm, and the phages re-infect neighboring cells, reinitiating the cycle. This strategy, which removes the need to deliver enzymes to the sites of an infection, reduced the biofilm cell counts by 4.5 orders of magnitude.
In another study, the authors engineered a modified version of the M13 bacteriophage in which the lexA gene, a repressor of the bacterial SOS global response to DNA damage, was placed under the PLtetO inducible promoter. Combining antibiotic treatment with an antibiotic-enhancing bacteriophage increased the in vitro killing by several antibiotics and, in vivo, the survival of infected mice treated with ofloxacin.
Another landmark study reported the engineering of a yeast strain to produce artemisinin, a component of first-line combination therapies for uncomplicated malaria. The only commercial source of artemisinin is the Chinese plant Artemisia annua L. (wormwood), where the compound is present at low concentrations, making therapy expensive.
Chemically synthesizing this medication is also very costly. Ro et al. generated an artemisinic acid-producing Saccharomyces cerevisiae strain in three steps that involved manipulating its farnesyl pyrophosphate pathway to increase farnesyl pyrophosphate production and decrease its use for sterol synthesis; introducing the A. annua amorphadiene synthase gene to allow amorphadiene synthesis; and engineering an A. annua cytochrome P450 to facilitate amorphadiene oxidation into artemisinic acid. This engineered strain produced up to 100 mg/L artemisinic acid that was transported and retained on the outside of the yeast, facilitating subsequent purification.
Ajikumar et al. reported the metabolic engineering of E. coli to produce taxadiene, the first committed intermediate of the potent anticancer medication taxol, which is produced by the Pacific yew tree (Taxus brevifolia). In an approach called multivariate-modular pathway engineering, the authors combined the upstream isoprenoid pathway, native to E. coli, with the downstream, synthetic taxol pathway, and simultaneously varied gene expression within each module.
This allowed the expression of upstream and downstream pathways to be modulated by changing multiple parameters such as promoter strength or gene and plasmid copy numbers. This system yielded approximately 1 gram of taxadiene from a liter of culture, and represents a significant step toward producing taxol in a bacterial system.
Synthetic Genomes & New Fields
One of the promises of synthetic biology is the possibility to engineer user-friendly hosts that convert biomass into fuels. Steen et al. metabolically engineered E. coli by disrupting fadE, the gene that encodes the first enzyme in β-oxidation, overexpressing the genes encoding thioesterase and acyl-CoA ligase, and engineering the Zymomonas mobilis pdc and adhB genes that encode pyruvate decarboxylase and alcohol dehydrogenase.
This novel recombinant strain was able to produce biodiesel by the acyltransferase reaction between acyl-CoA and ethanol. Further engineering of this construct, by introducing an endoxylanase catalytic domain from Clostridium stercorarium and a Bacteroides ovatus, xylanase that were fused to the N-terminus of the E. coli OsmY protein, enabled growth on hemicellulose without the need for exogenously added enzymes.
Another recent landmark was the chemical synthesis and assembly of a bacterial genome, Mycoplasma mycoides JCVI-syn1.0, from computer-designed sequences, and its incorporation into a new cell that is controlled by the synthetic genome.
Pioneering the field of optogenetics, Yazawa et al. described a technology that allows protein-protein interactions to be controlled in response to light. This approach, called light-activated dimerization, relies on FKF1 and GIGANTEA (GI), two proteins that are involved in day-length measurement and control flowering in Arabidopsis thaliana.
Illumination with 450 nm light is detected by a domain within FKF1 and induces a covalent bond formation between a cysteine residue and flavin mononucleotide, forming an FKF1-GI heterodimer. In response to light, the G-protein Rac1 was recruited in this system to the plasma membrane and induced the local formation of lamellipodia.
With the birth of synthetic biology, science entered a new and exciting era. As practical applications for biotechnology and medicine already started to emerge, it is hard to envision an area that will not be impacted by this recent field. Amidst these transformations, it is noteworthy to remark that basic concepts, which in the early 1960s laid the foundations of gene regulation, are re-emerging within a more complex framework, to provide new perspectives about the dynamics of biological systems, and to unveil, as Jacques Monod so relevantly calls it, the music of the biosphere.
Richard A. Stein, M.D., Ph.D. ([email protected]), is carrying out research in molecular biology at Princeton University.