Abstract
Synthetic biology has advanced to the point where some pathogens can be manufactured from scratch. This technical leap has beneficent implications for medical research and vaccine design, but it also raises concerns that the technology could be used to produce a deadly pathogen for nefarious use. Addressing these concerns, the Department of Health and Human Services (HHS) released their Screening Framework Guidance for Providers of Synthetic Double-Stranded DNA on October 13, 2010. They took the right approach: The oversight framework for gene synthesis companies included in this guidance is adaptable to new technical developments and changing risks, it can be implemented immediately, it can be readily adopted by other countries, and it will cost little. Though there have been some calls to increase the regulatory controls on synthetic biology, these should be resisted. For now, at least, the oversight is appropriate to the risks.
Acquiring a pathogen by this route isn't easy, however. There are many more straightforward ways to obtain a pathogen, including getting a sample from another person or laboratory, or isolating it from nature or from a patient or animal sample. There are technical limits to creating new, never-before-seen pathogens, or even mash-ups of existing pathogens, and it is a tedious, lengthy process. It is also not simple to “boot up” a genome and create an infectious organism. Acquiring the genetic material encoding Ebola virus is not equivalent to possessing an infectious Ebola virus sample.
Although the process is difficult, there is one step that can be outsourced: Key parts of a genome could be ordered from a gene synthesis company. For a price, a gene synthesis company could produce a string of genetic material, from a few dozen DNA base pairs (bp) up to more than 50,000bp. This would save someone a great deal of time if they were intent on manufacturing a pathogen from scratch. It is this step that the Department of Health and Human Services (HHS) has recently addressed in their Screening Framework Guidance for Providers of Synthetic Double-Stranded DNA, released on October 13, 2010. 1
Key Elements of the Guidance
The guidance asks providers of double-stranded DNA to screen the DNA sequence of the order as well as screens for the customer, combining a technological approach with common sense.
Sequence screening: All double-stranded (ds) DNA orders are screened against GenBank, the National Institutes of Health (NIH) genetic sequence database, which is an annotated collection of all publicly available DNA sequences. If the DNA order is a “best match” to a regulated pathogen, it would raise a red flag. Customer screening: A red flag would also be raised if something about a customer were suspicious or out of the ordinary—for example, if the customer tried to pay in cash, or asked that the order be sent to a private address. Customer screening is also expected to reflect U.S. trade restrictions and export controls, and companies need to check customers against “denied” lists maintained by the Departments of Commerce, State, and Treasury. If sequence screening determines that the genetic material is, in fact, only allowable to those with select agent clearance, the customer must be in compliance with the Select Agent Regulations. Follow-up screening: If red flags are raised, the company should take additional steps to verify customer legitimacy. If that does not satisfy concerns, or if assistance is needed, the company is encouraged to contact the nearest FBI Field Office Weapons of Mass Destruction (WMD) Coordinator or, for international orders, the Department of Commerce's Office of Export Enforcement.
Existing Oversight of Synthetic Biology
The HHS guidance is just one part of the oversight of synthetic biology in the United States, which includes the PATRIOT Act, the Select Agent Regulations, and self-regulation through harmonized industry standards.
PATRIOT Act: Regardless of whether a biological sample is synthetic, occurs naturally, is infectious, or is a select agent, the PATRIOT Act carries criminal and civil penalties for those who possess biological agents that cannot be justified for prophylactic, protective, or peaceful purposes.
2
The bottom line is that misuse of biology is illegal. Select Agent Regulations: Some pathogens and toxins are also covered under the Select Agent Regulations (SAR), which regulate possession and transfer of more than 80 pathogens in the U.S. For international shipping, there are regulations in the Export Administration Regulations (EAR). The Select Agent Regulations primarily regulate functional, infectious pathogens, not pathogen genomes or genome fragments. However, the complete genetic material for select agent toxins is forbidden without select agent clearance, as are several complete viral genomes that are relatively straightforward to boot up, including foot-and-mouth disease virus and some tickborne encephalitis viruses.
3
Self-Regulation: There is also self-regulation in synthetic biology that extends beyond national borders. Many gene synthesis companies belong to 1 of 2 industry groups that have developed harmonized screening standards that they apply to their customer orders. This outreach is greater than the U.S. could do alone. The International Gene Synthesis Consortium members include GeneArt (Regensburg/Bavaria), DNA 2.0 (Menlo Park, CA), Blue Heron (Bothell, WA, a division of OriGene), GenScript (Piscataway, NJ), and Integrated DNA Technologies (Skokie, IL, but with facilities in Belgium, the UK, and Germany).
4
The International Association Synthetic Biology includes ATG:Biosynthetics GmbH (Germany), Biomax Informatics AG (United States), Entelechon GmbH (Germany), febit Holding GmbH (Germany), and Sloning BioTechnology GmbH (Germany).
5
The HHS Guidance
The HHS guidance is a good addition to the mix of federal regulation and self-regulation that govern synthetic biology. It is easily updatable, as it builds on an existing database, GenBank. This should allow oversight to reflect current knowledge of pathogens, and it should reduce costs for implementation. There is no need for the government to build a separate database specifically for the purposes of screening DNA synthesis orders. GenBank is already heavily used, freely available, and constantly being updated. Researchers are, in fact, required by many journals to submit their sequences to GenBank, and the database itself is part of an international collaboration for data exchange with other European and Japanese genetic databases. Because the screening mechanism is tied to an already existing database, it is a relatively simple matter to include more sequences as more biological agents are put on the select agent list or become known as potentially dangerous.
The principles of the guidance are adaptable to international use. The customer screening steps make sense regardless of country, and, instead of the U.S. Select Agent Regulations, an international company may focus its screening on the Australia Group pathogens, for example, or its own nation's prohibited organisms. The fact that some steps of the sequence screening process can be automated is also an inducement for international companies to sign on, because the time it takes an employee at a gene synthesis company to fulfill an order will not increase significantly.
This guidance is unlikely to cause great delays in getting orders filled, which will benefit legitimate users and enhance the use of gene synthesis screening as a security tool. Right now, it is generally cheaper and faster to get both short and long stretches of single-stranded and double-stranded DNA from specialized companies. However, if legitimate scientists experience higher costs or delays in outsourcing this step, more might turn to in-house facilities or would make the products they need themselves. This is obviously bad for the companies, but it might also diminish the effectiveness of order screening as a screening tool if few people outsourced this step.
It makes sense that the guidance is limited to those pathogens that are regulated by the Select Agent Regulations or by export controls, as these pathogens may be more difficult to acquire through other means. Synthesis companies are encouraged to develop more extensive screening, and, if they screen against GenBank, they might be concerned if an order is focused on another pathogen. The HHS guidance acknowledges that this is an area that is difficult to regulate right now: “… due to the complexity of determining pathogenicity and because research in this area is ongoing and many such agents are not currently encompassed by regulations in the U.S., generating a comprehensive list of such agents to screen against is not currently feasible and hence is not provided in this Guidance.” 1 (p9)
Resist Regulations Based on DNA Sequences
One of the scarier aspects of synthetic biology is the prospect of dangerous sequences being ordered, patched together, and released as a pathogen that has never been naturally encountered. The National Academies Committee was asked to examine whether it would be possible to replace the current select agent list with an oversight system that predicts whether a DNA sequence would result in a dangerous organism. If pathogenicity could be predicted, gene synthesis companies could then be asked to screen orders more broadly for potentially dangerous sequences, even if those sequences aren't currently part of a select agent.
The NAS panel concluded in their report Sequence-Based Classification of Select Agents: A Brighter Line that it is currently impossible to predict function/pathogenicity from sequence, and that “[f ]or the foreseeable future, the only reliable predictor of the hazard posed by a biological agent is actual experience with that agent.” 6 (p3)
The select agent regulations, in general, take this functional approach to regulating known hazards, or biological organisms known to be functionally infectious. For example, a select agent clearance is required to possess Rift Valley fever (RVF) virus in a laboratory. The complete genetic material that could produce a functional Rift Valley fever virus would also be covered under the regulations. However, fragments of the RVF genome are not covered. This is pragmatic, as there are only a few genetic variations that will produce a functional, infectious pathogen, but an infinite number of possible fragments of the genetic material similar to a pathogen that will not, in fact, be infectious. A regulatory approach that is focused on function is less likely to impose regulatory burdens on legitimate, nonpathogenic work.
Relying on the function of a biological agent versus the DNA sequence might be confusing to legitimate users of gene synthesis products, however. How is a scientist to know if the DNA sequence he or she orders is enough like a select agent that it would qualify as one? How many mutations in the genetic sequence are needed before it is no longer a select agent, without actual experience with the agent? In response to this type of ambiguity, the NAS committee described (but did not necessarily endorse) a sequence-based classification system that “would better define, in terms of DNA sequence, what is meant by each name on the existing Select Agent list.” 6 Instead of being defined by its infectious function, a pathogen would be defined by its genetic sequence.
Building such a sequence-based classification system would require a significant up-front investment—a sequence database with a focus on select agents would be needed to define the “minimal parts list” of each select agent. Once the proposed database is active, gene synthesis providers would use a yellow-flag system in which orders that are in sequence close to the select agent (through a predetermined algorithm) would be flagged as “sequences of concern.”
As the NAS Committee itself noted, construction of such a classification system might be very resource intensive. Creating a centralized sequence database, annotating those sequences, and creating a review process would be prohibitively costly, and it would be limited to select agents. It would invariably be too slow to keep pace with new advances in biology, and so it would be frustratingly slow for legitimate researchers.
The yellow-flag system would not be a contribution to biosafety either. If prediction of pathogenicity from sequence is not possible, it is also not possible to predict safety requirements from a DNA sequence. Instead, researchers should have the training and resources to deal with any potential infectious surprise, regardless of the source.
Sensible Steps, But No Silver Bullet
The measures taken by the U.S. government and practitioners to prevent the misuse of synthetic biology have been comprehensive and appropriate to the risks, but it is not possible to totally prevent potential misuse. Synthetic biology is an internationally diffuse activity, and technical barriers will continue to fall over time. For example, mix-and-match sequences from pathogens are unlikely to be functionally infectious, and so they are not likely to be public health threats. However, in the future, predicting function from sequence might be technically easier, making designer pathogens a real threat. The danger of misuse of synthetic biology cannot be entirely prevented, so our goal should be to prevent dangers where possible and be ready to respond quickly to misuse.