Engineering Viruses is Hard. Let’s Make it Easier.

The virus φX174 has locked inside itself a puzzling conundrum. While the virus utilizes 11 different proteins, there is not enough viral DNA to encode all of them. The combined length of the 11 proteins is 2327 amino acids, but the genome is only 5386 nucleotides in length. Ignoring regulatory sequences, the most that a 5386-base genome can encode is a mere 1795 amino acids (= 5386/3). That's 532 amino acids short, leaving nearly a quarter of the proteins of φX174 mysteriously unaccounted for.

The answer to this puzzle shows how flexible biology can be. It turns out that the φX174 genome does encode all 11 proteins, but that some of the protein-coding genes overlap. This is possible because the genetic code—the set of rules for converting a DNA or RNA sequence to a protein sequence—operates such that three DNA or RNA bases (a "codon") code for one amino acid. Two protein-coding sequences can be overlaid in the same DNA sequence by utilizing overlapping sets of codons (see the figure; alternative ATG start codons underlined).

This overlapping genetic system means that φX174 can efficiently package its genome into its viral coat. And, since a smaller genome requires fewer resources, more viruses can be built from the finite resources of each E. coli cell before the cell bursts. However, the φX174 genetic system is by virtue of its extreme compactness and overlapping genes hard to engineer, and so a team at Stanford University decided to "decompress" the φX174 genome¹.

As expected, a "full-length" φX174 genome with all overlaps removed but all coding sequences preserved would be much longer, and therefore the resulting viruses would be less fit. Since the question of interest is the effect of the overlaps, rather than the effect of the genome length, the team used a synthetic φX174 genome with most of the F gene deleted to accommodate the increased length created by extending out overlaps. Thus a fair comparison can be made between the decompressed φX174.1f genome and the wild type φX174 genome. The F gene was knocked into the genome of their E. coli strain, allowing the virus to function normally.

With this new "φX174.1f" genome designed, the team used synthetic DNA and PCR amplification from the wild type φX174 genome for assembly. They used yeast to help assemble and maintain the synthetic genome, in a similar fashion as Craig Venter's team did when assembling the synthetic Mycoplasma genome². Obviously E. coli cannot maintain the φX174 genome as it can routine plasmids, since the genome instructs the cell's destruction by viral infection. However, yeast (eukaryotes, like us) are perfectly happy to serve as an archival system for maintaining the genomes of viruses that infect bacteria.

The authors found that E. coli transformed with the yeast-archived viral genomes succumbed to φX174 infection. Cleverly, the F gene knocked into the E. coli genome (to compensate for the truncated version in their decompressed virus) was regulated by an inducible promoter. By changing the concentration of the sugar rhamnose in the growth media, the researchers could control the severity of the viral infection. Not surprisingly, the wild-type φX174 outperformed their decompressed version. When agar plates covered with E. coli are infected with φX174, clear circles called plaques form, a visible mark of the dismemberment of billions of E. coli cells. Infection with wild-type φX174 produced plaques on average 4.9 mm across; infection with their decompressed strain under full rhamnose concentration produced plaques on average 4.1 mm across.

This new virus could be an important platform for viral synthetic biology. The genome is small enough to be made cheaply, so that the design, tinker, redesign process is feasible. Viruses like φX174 are far simpler than the bacteria they infect, so why not start here? Famous and well-funded biologists like Craig Venter and George Church³ are trying to design new bacteria from scratch. That's quite a tall order. I really want them to succeed, but I have my doubts.

New designs rarely work the first time, even when we are talking about only a couple genes sitting on a plasmid in E. coli. Designing a whole genome will surely require extensive tinkering and revision. Since viral genomes are far simpler, it seems to make sense to start there. I think a completely new virus could with relative ease be put together from parts strewn across the catalog of thousands of known genes (viral or otherwise). That's the sort of design work that we are talking about with synthetic cells; nobody expects us to make a genome full of new genes, let alone a completely new biochemistry. Designer cells will be made of known genes, assembled together by human design. What we learn from making designer viruses could help us in the pursuit of new living cells of human design.

Image Credits: both made by me; use freely with attribution

References:

1. Jaschke, P. R. et al. A Fully Decompressed Synthetic Bacteriophage øX174 Genome Assembled and Archived in Yeast. Virology (2012). Published online.

2. Gibson, D. G. et al. Creation of a Bacterial Cell Controlled by a Chemically Synthesized Genome. Science 329, 52–56 (2010).

3. Forster, A. C. & Church, G. M. Toward Synthesis of a Minimal Cell. Molecular Systems Biology 2, 45 (2006).