Team:Harvard/Results/Human Practices



Zinc fingers represent an exciting area of research, due in part to their potential clinical value in drug development and gene therapy. Yet with these applications comes responsibility: the genome--in particular, the human genome--cannot be altered lightly. If the public is to accept gene therapy as a viable treatment for genetic diseases, zinc fingers must be made as safe as possible, must be used in an ethical manner, and should be an open source, so both scientists and non-scientists can understand the technology behind their creation. Below we lay out several important issues and suggestions that we feel would benefit the human practice of zinc finger use and design.

In addition, we wanted to hear what leading researchers in the field of synthetic biology had to say about the future of zinc fingers and their impact on human practices. We interviewed Dr. George Church, Professor of Genetics at Harvard Medical School, and Dr. Keith Joung, Associate Chief of Pathology for Research at Massachusetts General Hospital and Associate Professor of Pathology at Harvard Medical School. Both scientists have performed groundbreaking work on zinc fingers and their related technologies, and their insight into the possibilities of zinc finger research and the challenges that lie ahead was invaluable to us.



While ZF nucleases are highly promising tool for gene therapy, significant concerns about the safety of its usage on humans remain. The concerns primarily stem from the off-target effects of ZF nucleases. Even a modest off-target effect could result in multiple double strand breaks on the genome leading to high genomic instability, potentially causing effects worse than the original defect.

Two approaches that would help mitigate this:

1. Explicitly design against non-specific sequences: As we design novel ZF proteins towards a sequence, it is essential to negatively design against all closely related sequences to reduce off-target effects. Further, it may be wise to compromise on binding affinity for higher specificity.

2. Design zero nuclease activity of the monomer – obligate dimer requirement: In order to induce a double strand break, two ZF arrays each attached to a FokI domain bind to a dsDNA. The FokI domains need to homodimerize to perform nuclease activity. However, low levels of FokI monomer nuclease activity could result in double strand breaks at unintended locations wherever a single ZF array-FokI chimera transiently binds. To reduce such background effect, a FokI domain (or other nuclease domains) have to be designed with zero monomer activity.


Important ethical issues surround therapeutic application of ZF nuclease based gene therapy. The efficacy of gene therapy could be offset by the poorly understood side effects of ZF nucleases. For instance, it is now known that deletion of CCR5 gene confers resistance to HIV. Therefore, efforts are underway to design ZF nucleases targeted to the CCR5 locus for deleting the gene. While this may certainly benefit fight against HIV, the side effects could cause a different, but just as fatal, illness. Enhanced oncogenicity could result from multiple off-target double strand breaks caused by ZF nucleases. Should we deny HIV therapy because of unknown side effects of the therapy? Or, is it okay to knowingly administer the therapy at the patient’s behest after he/she has been sufficiently educated? Dialogue on such ethical issues needs to progress in parallel with development of the therapy itself.

Ownership and Sharing

The objective of our project was to present an open-source technology for development of ZF proteins for any desired target sequence. Companies like Sangamo develop custom ZF proteins for your-favorite-gene at a price tag that is unaffordable to most academic laboratories. Here we showcase a nearly “reduced-to-practice” method with detailed protocols for any academic laboratory to repeat our method for their target sequence of choice. We wish to share our data and results with the community, highlighting our successes and failures to collectively advance of our knowledge of designing novel ZF proteins.


Our innovation lies in bringing together many technologies to create a general open-source method for designing novel ZF proteins. The plasmid-based one-hybrid selection system was previously known; we applied lambda-red recombineering to integrate the selection system on the genome in order to reduce background due to copy number variation of the plasmids. Chip-based DNA synthesis was previously developed a cheap source of DNA; we applied it towards generating nearly 55,000 designed ZF proteins. Our bioinformatics pipeline combined structure-based information and experimental binding data to develop frequency distribution tables for forward engineering. MAGE technology enabled facile genome modification without the need for a selection marker and allowed us to disable E.coli hisB and pyrF genes by inserting a stop codon.

With the increasing promise of gene therapy, we have uniquely combined various technologies to fulfill the unmet need for an open-source method to modify human genomes.