NEW YORK (GenomeWeb) – A Stanford University group has come up with a high-throughput strategy for mapping RNA-protein interactions that involves repurposing the Illumina GAIIx instrument for use as a massively parallel RNA array.
Stanford applied physics researcher William Greenleaf gave an overview of the approach as part of his presentation on RNA-protein interaction profiling last week at the Biology of Genomes meeting at Cold Spring Harbor Laboratory.
He and his colleagues also outlined the method in a study published online in Nature Biotechnology last month, where they described findings from proof-of-principle experiments that looked at interactions between the Escherichia coli coat protein MS2 and RNA molecules with specific sequence changes or mutational combinations.
In particular, the team considered ways in which RNA mutations affecting the molecule's sequence and secondary structures altered its ability to associate and dissociate from the MS2 protein — information that's offering clues about the biophysics and evolutionary constraints of such interactions.
"Our results suggest that quantitative analysis of RNA on a massively parallel array (RNA-MaP) provides generalizable insight into the biophysical basis and evolutionary consequences of sequence-function relationships," Greenleaf and his co-authors wrote.
Greenleaf noted that experiments in the Nature Biotechnology study were done using an Illumina GAII instrument with minimal modifications, but said his team has since put together two instruments with more extensive optical modifications and is in the process of building a third modified instrument.
The group is not the first to take a crack at "pirating" or repurposing existing sequence equipment. For example, Greenleaf noted that his team's RNA array approach itself was inspired by a study appearing in Nature Biotechnology in 2011.
There, the Massachusetts Institute of Technology's Christopher Burge and co-authors from MIT and Illumina presented a strategy for adapting sequencing protocols as part of a high-throughput sequencing-fluorescent ligand interaction profiling, or HiTS-FLIP, scheme to quantify interactions between a given protein and DNA molecules.
Earlier this week, Illumina's Gary Schroth, an author on the HiTS-FLIP study, co-authored a Nature Methods study with Cornell University researchers that introduced an RNA-focused assay called "high-throughput sequencing-RNA affinity profiling" (HiTS-RAP) for quantifying RNA protein interactions.
In that protocol, researchers sequenced DNA clusters on the Illumina GAII before producing RNAs that were anchored onto the DNA and could be exposed to fluorescently labeled proteins — an approach they used for studying binding profiles of mutated RNA aptamers.
For their part, Greenleaf and colleagues used a slightly different strategy for forming the RNA array on the sequencing chip, though the result was similar: converting the GAII into a high-throughput imaging station for detecting fluorescently labeled proteins as they interact with RNA that's isolated on a chip.
For their Nature Biotechnology study, for example, the researchers designed RNA molecules with specific mutations, adding in sequences for RNA polymerase binding and initiation and for barcoding the molecules.
Each of the RNAs was reverse-transcribed to form single-stranded complementary DNAs, they explained, which was amplified and sequenced on another instrument.
After sequencing, the team put each chip onto a GAII instrument with modified optics and filters, converted single-stranded DNA back to double-stranded DNA, and added biotin sequence to the DNA so it could subsequently capture streptavidin proteins applied to the chip.
Finally, the researchers converted double-stranded DNA back to RNA using a RNA polymerase enzyme that stalled midway through the process, allowing them to wash out excess enzyme, and at the biotin-streptavidin blockade at the end of the DNA molecule.
Once the RNA array was formed, Greenleaf explained, the team could flow in fluorescently labeled proteins of interest, tallying up their binding and dissociation with lab-designed analysis software that correlates RNA sequence information to Illumina images — in this case, signals from tagged fluorescent proteins that interact with RNA.
Using that approach, the researchers were able to see how various point mutations in different regions of E. coli RNA molecules modified their interactions with the RNA-binding protein MS2 interactions.
"Because we could make this massive number of molecular variants and point mutations, we were able to partition the binding energy of the RNA-protein interaction into primary sequence determinant and secondary sequence determinants," Greenleaf said.
The primary sequence determinants describe binding patterns related to the order in which bases in an RNA molecule occur, he explained, while secondary determinants hinge on the resulting three-dimensional structure of the molecule.
For instance, he and his co-authors determined that the MS2 protein is prone to bind to specific secondary structures in some portions of the RNA molecule, including the stem-loop regions. In other RNA regions, though, protein binding was more closely linked to primary sequence determinants.
When they considered the protein's ability to associate with and dissociate from modified RNAs, meanwhile, the researchers observed differences in the effect of RNA alterations on these on- and off-rates. In particular, when mutations were introduced to the base of the RNA stem, the on-rate of the interacting protein typically diminished, though the dissociation rate remained the same.
Finally, the approach made it possible to ascertain evolutionary trajectories in the presence of mutation combinations and in the presence of proposed base-pairing intermediates in RNA evolution.
On the sample prep side, many of the steps needed to set up the high-throughput RNA-protein interaction experiments resemble those used to prepare Illumina sequencing samples, though additional steps are needed — for example to introduce sequences for binding RNA polymerase and initiating transcription and to produce tagged proteins.
So far, most experiments done in that lab have focused on RNAs with pre-specified mutation types, though the same approach is expected to be applicable to RNA transcripts isolated from a cell, provided RNA initiation sequences can be added in.
Greenleaf said it should be possible to look at interactions involving multiple proteins in the same experiment by using differently colored fluorescent probes and to track cooperative binding events involving different proteins.
Likewise, there is potential for increasing the number of RNA-interacting proteins that can be tested for interactions with the same set of RNAs. The ability to see several proteins simultaneously in the assay depends somewhat on the laser system used.
At the moment, the team is somewhat limited to looking at fluorescent tags that can be excited and detected by the laser and optical system in place, Greenleaf said. Even so, he noted that it should be fairly straightforward to add in other lasers or light sources to expand the repertoire of tags that can be tracked.
Because GAII chips are getting a bit harder to come by, the team is working on ways of performing similar experiments on the GAII using chips from other Illumina instruments. such as the MiSeq.
The same RNA array method should be possible with other surface-based sequencing platforms, Greenleaf said, though he and his colleagues have not tried their hand at that so far.
In addition to their work on RNA-protein interactions, the group is interested in using a similar strategy to study other RNA features — from targeted RNA-RNA interactions to whole transcriptome assessments and studies of RNA aptamers, lincRNAs, ribozymes, and the like.
"There's a lot to potentially do at high-throughput in RNA biochemistry with this platform," Greenleaf said, noting that the approach may be applied to "any sort of thing you can think of where you might want to understand the sequence determinance of affinity or even RNA thermodynamics."