PHOTO BY SETH KROLLIn 1973, after trying out two other groups, George Church finally found the perfect fit in Sung-Hou Kim’s X-ray crystallography lab, where he spent his second (and final) year as a Duke University undergraduate. “It was all of the things I was interested in—physics, math, biology, chemistry, and computers—in one. Kim had just come from MIT and was young and full of energy,” says Church, now a professor of genetics at Harvard University. “He saw a spark in me that few people had noticed before, and he treated me almost as an equal. I started to blossom in that lab.”
Church knew that he wanted to go to graduate school, and had already applied to the microbiology PhD program at Duke when he began to work...
“I then became a technician in Kim’s lab and planned to do that for the rest of my life until one day Sung-Hou said to me, ‘I don’t think you are cut out to be a technician, you have too many ideas. You should apply to graduate school.’” In 1977, Church says he did so halfheartedly, applying only to Harvard University’s molecular biology program, “which was crazy for someone who had flunked out of Duke.” To his surprise, he got in.
“In the 1990s, a big fraction of my lab was dedicated to computational biology, but we slowly started regressing back to my love of advancing technologies.”
Here, Church describes his first experiment, feeding prescription drugs to tadpoles; his reputation in graduate school; and an early epiphany that drove him to develop basic technologies to benefit all biologists.
Frogs as guinea pigs. Church was born in Florida in 1954 and adopted by his mother’s third husband at age 9, taking his surname. “He was a physician, and I was always fascinated by the things in his medical bag. He loved giving people thyroid hormone, so I got it into my head somehow that frogs were responsive to thyroid hormone. I probably didn’t read this in a book of experiments because I had mild dyslexia,” says Church. The 13-year-old Church added ground-up hormone pills to the water of a group of tadpoles and compared their growth to that of a group that developed in untreated water. “The hormones accelerated their growth considerably. I remember being excited about the result and presented this in my science class. This was a real experiment—with a control, even—in contrast with the usual emphasis then in science class on facts and not experiments.”
Nonstop science. In 1968, Church was sent to the Phillips Academy in Andover, Massachusetts, where he dove deeply into science. “It was an incredible place to learn,” he says. He kept a collection of carnivorous plants that he tried to make grow into giants by spiking their water with gibberellic acid, a hormone found in plant and fungi species. After completing the chemistry curriculum, Church was given independent access to the chemistry lab. “My favorite thing was to ask the chemistry professor to pull a random chemical off the shelf to see how fast I could identify it—it was like being a detective. Eventually he started giving me mixtures of chemicals.” Church, also drawn to computers and robots, discovered a lone computer in the basement of the math building during his freshman year. Using programming books and eavesdropping on conversations of fourth-year students taking computer labs, he taught himself how to program.
Fast track. Church entered Duke in 1972. “I wanted to go to the best warm-weather school I could get into,” he says. He placed out of one year’s worth of math and science courses and enrolled in summer courses and graduate-level courses his freshman year, sneaking the permission slip for a virology course in between other forms the professor had to sign. Church earned two undergraduate degrees in two years. “I was in a rush because I felt that all of the college courses were baby courses and because I had to pay for college. I saw this as my route to economic independence.”
Confessions of a wallflower. The summer before starting graduate school at Harvard University, Church lived in Boston, read molecular biology papers, and planned experiments, including techniques to improve DNA sequencing. He had already decided he wanted to join the lab of Walter Gilbert, one of the developers of DNA sequencing techniques, for which he would receive the Nobel Prize three years into Church’s time in his lab. In one course that fall, the professor spent several slides explaining a paper on which Church was the first author. “He had no clue that the author of that paper was in his class. That made me a bit more secure, but I was still full of insecurity. I was very, very shy that year. There was a Mad magazine type of student publication, and one of the issues had a matching game. In one column it said ‘3 words a day’ and that matched to ‘George’ in the second column. That was my reputation.”
Baby steps. Church’s first foray into technology development was to write sequencing software during a rotation in Don Wiley’s Harvard lab after helping to sequence the pBR322 plasmid with Greg Sutcliffe. “In a way, [this was] the first real synthetic biology,” Church says of the artificially constructed cloning vector. “I kept trying to develop technology rather than following protocols.” In Gilbert’s lab, Church worked on yeast mitochondrial introns, but “it was such an obscure field that even if I did it perfectly, I knew very few would really care. I thought that if I was going to put that much work into something, I wanted that to be basic enabling technology that all biologists could benefit from.” So Church set out to develop new sequencing methods. “I would come home to my girlfriend (now my wife) and say, ‘This is a really great day, I got a factor of 2 improvement in signal to noise. I have 10,000-fold more to go.’ I never felt frustrated or nervous. The factor of 2 would eventually turn into 10,000.”
Four years later, in 1984, Church published his direct genomic sequencing method, which extended Southern blotting to a new level of sensitivity and didn’t require cloning or amplification. Church also came up with the concept of increasing DNA sequencing throughput: mixing many DNA pieces in the same tube and reprobing and reimaging, concepts of barcoding and multiplexing that he published in 1988. These early tools later contributed to the automation and later generations of genomic sequencing.
Technology-focused. Church received a PhD in 1984, and, after a brief time at Biogen, joined Gail Martin’s lab at the University of California, San Francisco, to work on embryonic stem cells. He produced no publications, but “became comfortable with embryology and started the technology to read genomes and transcriptomes,” which he felt were missing, yet crucial, for stem cell studies. Church joined Harvard’s faculty in 1986, aiming to improve technologies for reading, writing, and testing genomes.
A year later, he received one of the first Human Genome Project grants. But while others on the project wanted to plow through the sequencing using existing, expensive technologies, Church thought the goal should be to streamline sequencing and lower its costs. “In the 1990s, a big fraction of my lab was dedicated to computational biology, but we slowly started regressing back to my love of advancing technologies,” says Church. Among the lab’s first contributions to sequencing was a method for clonally amplifying DNA in situ, developed in 1999 by an MIT engineering student. “This was something that no biologist in my lab wanted to touch,” says Church. In 2005, his lab optimized the technique into next-generation sequencing (NGS) and applied the approach to sequence a lab-evolved E. coli genome. By 2009, NGS brought down the cost of reading genomes a millionfold, according to Church.
Resurrecting species. In 2008, researchers used NGS to sequence a large portion of the mammoth genome from a recovered hair sample, discovering that the extinct creature is closely related to the Asian elephant. After commenting on the effort in The New York Times, Church says he began to think seriously about attempting to resurrect the species. “Like all kids, I was fascinated by large extinct creatures, and I tended to like the furry ones better,” he says. The project is small compared to others in his lab, but Church has plenty of volunteers wanting to tweak the elephant genome to resemble that of the mammoth. With private funding, the team has already replaced 15 genes in an elephant cell line with mammoth ones, including those that code for a cold-climate hemoglobin, long hair, small ears, and subdermal fat storage. “We have been developing the methods to engineer genomes of embryos in pigs. But for elephants, we lack key induced pluripotent elephant cells and reproductive technology,” says Church. The lab is also trying to develop organoid models to grow reproductive organs to eventually study mammalian—including elephant—development.
Due credit. The story of who turned CRISPR/Cas9 into a precise editing technology is complex, Church contends. After Emmanuelle Charpentier’s and Jennifer Doudna’s laboratories together demonstrated the system could be simplified and programmed to cut DNA in cell-free systems, Prashant Mali and Luhan Yang in Church’s lab demonstrated that Cas9 could be used instead for homologous recombination and in human induced pluripotent stem cells. The publication came out in the same issue of Science as a paper that reported similar findings by Church lab alumni, Le Cong and Feng Zhang, at MIT. For Church, CRISPR is just one of the latest genome engineering advances that include next-gen sequencing, synthesis, and large DNA assembly methods.
Creating new biology. In August, Church’s lab published a paper in Science that he describes as “the largest and most radical genome engineering project [to date].” The lab designed a 3.97-megabase-pair E. coli genome, replacing seven amino-acid codons with noncanonical synonymous alternative ones. When completely synthesized, the genome should have 62,214 codon replacements. The work is an improvement on a 2013 study in which Church and his colleagues swapped one codon for a synonymous one throughout the entire E coli genome. In the new work, the seven targeted codons were replaced in 63 percent of the synthetic genome. The researchers designed an E. coli genome with 87 50-kilobase-long segments containing the codon replacements and had commercial companies synthesize 3-kilobase-long fragments, which his lab then assembled into 50-kilobase fragments in yeast. So far, the team has replaced 55 of the wild-type segments with the novel ones and tested bacterial viability. Church’s lab is now working to test the rest of the genomic segments. The ultimate goal, according to Church, is to create a biocontained multivirus-resistant strain that is better suited for industry applications such as protein and chemical production. “This is also a good model. If you can get this to work in E. coli, you can get it to work in a lot of other organisms,” says Church.
Nature 2.0. The E. coli work lays the groundwork for synthesizing whole genomes of larger organisms, including that of humans, an endeavor Church and his colleagues are calling the Genome Project-Write. “The E. coli project is not quite complete, but what is clear is that we have done a lot of the hard part. . . . We think this is the flagship experiment for Genome Project-Write in that it shows how you can synthesize big pieces of DNA and then replace them in a chromosome by mainly using phage integrases. The point of the project is to improve technology and to bring down costs,” says Church whose group is already at work synthesizing long pieces of human DNA. “The synthesis is quite cheap, about $2,000, but assembly and testing of even a 4-million-base-pair genome is still quite challenging,” For Church, synthetic genome applications include virus-resistant agricultural species, as well as human cells for industrial production of human proteins, vaccines, and other therapies. “People who take time to study what the project actually is say, ‘Interesting.’ Some people imagined a parallel project aiming for parentless babies and that is clearly not what we are doing. We are constructing cell lines.”
Waiting for the right moment. “I think most of the failures in the lab are just ‘failures so far’—projects that haven’t worked yet or that have taken longer than what would be ideal, although I tend not to have particular time lines in mind for projects. I generally don’t give up a project, but may put it on the back burner at low priority. Some of these things, like nanopore sequencing, take 25 years before they work well.”
Grandfathered in. “My daughter and her family live right next door to me, which is delightful. My granddaughter is 21 months old, which is a really fun age (well, all of her ages are fun). Seeing her grow every day is like having another daughter with a 25-year difference in perspective. It really expands my horizons to see things through such wonderfully alien eyes.”
Perfect match. George Church met his wife, Ting Wu, in graduate school at Harvard. “She’s a way better scientist than I am. We’ve worked together on ultra conserved elements, Oligopaints technology, and the Personal Genetics Education Project, aimed at engaging the public on genetics.”
- Developed the first direct genome sequencing and DNA multiplexing methods that led to the first bacterial genome sequence in 1994 and, in 2003, to next-generation methods
- Spearheaded the Personal Genome Project, a way to engage the public in genomic and health data sharing
- As part of the BRAIN initiative, developed ways to encode data in DNA formats, including temporal records of events in living cells
- Intertwined genome reading and writing technologies that led to the largest (~4 million base pairs) synthetically engineered (recoded E. coli) genome to date
- Pioneered applications of CRISPR for organ transplants, aging reversal, and gene drives to eliminate malaria and Lyme disease