NEW YORK (GenomeWeb) – Researchers at the University of Washington have developed a method to enable researchers to perform site-directed, single amino acid mutagenesis.
By relying on microarray-based DNA synthesis and overlap-extension mutagenesis, the UW team led by Jay Shendure developed programmed allelic series, or PALS, an approach that introduces a single mutation into cDNA in a massively parallel fashion, as they reported in Nature Methods today.
Current approaches suffer from a limited ability to incorporate single base mutations into codons and are laborious, as only a few residues can be targeted at a time, and expensive.
"To overcome these limitations, we developed PALS, which combines low-cost, microarray-based DNA synthesis with overlap-extension mutagenesis to introduce one and only one mutation per cDNA template in a massively parallel fashion," Shendure and his colleagues said in their paper.
The PALS approach the researchers described starts with on-array synthesis of mutagenic primers for a target, each primer with a mutation nears its center.
The primers are flanked with adaptors so that certain ones can be removed using PCR. The primers are then annealed to and extended on a wild-type sense strand, and that substrate is then degraded and the mutants are amplified.
After the adaptors are snipped off, the megaprimers are extended against an antisense strand of the wild-type target gene. That substrate is then degraded, yielding full-length single mutant copies that are then amplified by PCR.
To develop full-length sequences, Shedure and his colleagues cloned the mutant libraries, added a random barcode to each clone, and performed hierarchical tag-directed sequencing to give a set of consensus haplotypes and linked tags.
To try out their method, the researchers constructed a PALS-based library for the DNA-binding domain of the yeast transcription factor Gal4. They targeted each codon of the Gal4 DNA-binding domain to be replaced with either a codon encoding one of the 19 other amino acids or a premature stop codon.
Here, the researchers reported that after cloning and subassembly, about 47 percent of full-length haplotypes had one programmed mutation in an otherwise wild-type background. Nearly all — 99.9 percent — of the programmed single-codon replacements were seen once, and 99.7 percent of them were seen at least five times.
By turning to p53, the researchers examined whether their method could easily be scaled up to target a full-length cDNA. Rather than using specified codons as they did for their Gal4 experiment, the researchers targeted p53 codons for replacement by degenerate triplets. This approach, they noted, reduced the number of microarray features required and allowed them to look at synonymous variants.
For this, they reported a lower rate of single-mutant haplotypes — 33 percent — which the researchers said was likely due to the increased probability of secondary errors on the longer templates. Still, the researchers said they observed 93.4 percent of the possible amino acid substitutions in p53 in clean, single-mutant clones.
Overall, Shendure and his colleagues noted that the mutations coverage by PALS was mostly uniform, though with a moderate N-terminus bias, which they saw in both the Gal4 and p53 studies.
Shedure and his colleagues also used PALS to conduct a deep mutational scan of Gal4. They introduced the Gal4 DNA-binding domain library into a two-hybrid reporter strain system. In it, the Gal4 gene was removed and replaced with the HIS3 gene, which fell under the control of the GAL1 promoter. If the Gal4 DNA-binding domain mutant could bind DNA to activate transcription and thus HIS3 expression, then the system could grow on a medium lacking histidine.
Based on this, they reported that, as suspected, premature stop codons were deleterious under selective though not permissive conditions, and about a third of the residues couldn't tolerate mutations.
This gave the researchers a picture of functional constraint. Gal4, they noted, binds DNA as a homodimer through a Zn2Cys6-class domain centered on a pair on Zn2+ ions, which help preserved the fold of the DNA-binding residues. Substations at any of the six chelating cysteines, they noted, upset this function.
By folding this data in with crystal structure information, the researchers further reported that that and other changes such as proline substitutions were deleterious.
"This trend is broadly observed in deep mutational scans of other proteins, likely reflecting disruption of protein secondary structure due to the proline residue kinking the backbone," the researchers said.
However, in the Gal4 DNA-binding domain linker region, the researchers suggested that proline substitutions might be beneficial as they decrease flexibility and could make DNA binding and transcriptional activation more favorable entropically.
"The combination of PALS mutagenesis, functional selection and deep sequencing provides a general framework to dissect the allelic heterogeneity of human genes and a path toward 'precomputed' functional annotation of the growing catalogs of variants of unknown significance," Shendure and his colleagues said.