While investigating the possibility of using RNA interference to block herpes simplex virus 1 infection, the Beckman Research Institute’s John Rossi and colleagues stumbled across a discovery — detailed in the Feb. 8 issue of Nature Biotechnology — that may help researchers make siRNAs by in vitro transcription without having to worry that the oligos will trigger a cellular interferon response.
This, Rossi said, could result in the more widespread use of “T7 transcription kits, provided that you set up your transcripts such that you could easily get rid of the 5’ initiating nucleotides” — something his paper “clearly shows how to do.”
In evaluating the use of RNAi against HSV-1, Rossi told RNAi News that he had begun using siRNAs produced by bacteriophage T7 RNA polymerase “because it was less expensive and you could test a number of different sites, et cetera.” The siRNAs, he noted, were made using Ambion’s Silencer siRNA construction kit.
According to the Nature Biotechnology paper, prior to viral infection, the researchers transfected human embryonic kidney 293 cells with siRNAs targeting the early ICP4 gene transcript. Twenty hours later, recombinant HSV was added to the cultured cells. In the HSV virus, the researchers noted, the gene encoding VP20 was fused to the gene encoding the enhanced green fluorescent protein, in order to monitor viral infection.
“The observations were that when we made transcripts targeting sites in herpes, we would get potent inhibition of herpes infection, provided that we first transfected the siRNA into [HEK] 293 cells,” Rossi said. “But the control, which was an in vitro transcribed siRNA against [an irrelevant] target, also gave the same potent inhibition.”
The researchers’ hypothesis about what might be happening led them to try transfecting HEK-293 cells with chemically synthesized siRNAs — obtained from Dharmacon — targeting the same sites as the in vitro transcribed siRNAs. “They had no effect whatsoever,” Rossi noted.
“Since the first sets of experiments suggested there’s potent inhibition, we became suspicious that there might be interferon induction,” Rossi said. He added that inhibition of HSV infection also took place when medium from the siRNA-transfected cells was added to fresh cell culture at a ratio of one to 30.
According to the Nature Biotechnology article, after finding substantial amounts of both interferon alpha and interferon beta in the medium of transfected cells, the researchers tested HSV-1 infection using cells that had been previously treated with antibodies to both interferons (antibodies to just one of the interferons was not sufficient, they wrote) finding that HSV-1 infection would now occur.
“These results are consistent with the known mechanisms of interferon inhibition of HSV-1 and the anti-interferon mechanisms of the virus, which shuts down the host response during infection,” the researchers wrote in the paper. “In the present case, the interferons were induced before infection, presumably triggering the expression of antiviral genes.”
With it clear that the researchers were inducing interferon, “the question was: What was the component of the RNA that was inducing interferon?,” said Rossi. “We didn’t really have a clue — we felt that perhaps it was double-stranded RNA that was coming off these vectors.”
“So we made synthetic siRNAs that had three guanosines at the 5’ end — no [interferon induction] effect in [HEK-293 cells] whatsoever,” he said. “What was left was that the only difference between an in vitro transcribed RNA and one that’s chemically synthesized is the triphosphate that’s left at the 5’ end of the in vitro transcribed RNA.”
The triphosphate, Rossi said, was eliminated using “a process of cleavages. First, we used ribonuclease T1 to get rid of the Gs that are sticking off at the end,” he said. “Then, we used [calf intestine] phosphatase.”
The researchers wrote in the paper that “because ribonuclease T1 treatment … did not completely remove the 5’-pGGG, we reasoned that perhaps it or the adjacent Gs were involved in wobble base pairing with the 3’ terminal Us of the transcript, making this a poor substrate for single strand-specific ribonuclease and CIP.”
Rossi said that this was indeed the case, leading the researchers to postulate that “if we changed those Us to something that couldn’t base pair with G, like adenosine, now we could cleave off the Gs to a large extent.”
After doing so, and then adding ribonuclease T1 and CIP treatment, Rossi said that “we completely eliminated the interferon effect.”
Explaining the Results
“It became clear that we actually had this molecule that had something on the 5’ end — it looked like it was a triphosphate on the Gs,” Rossi said. “During the same period of time, just to eliminate the possibility of making double-stranded RNA, we also used single-stranded RNAs that we made with T7 — they also had the same potent sequence-non-specific induction of interferon.”
The ability of the ssRNAs to induce interferon, the authors of the Nature Biotechnology paper wrote, was lost when the RNAs were treated with CIP.
“It was the ribonuclease’s inability to digest the Gs that were wobble based-paired to the Us, and so the solution became quite straightforward: You put something at the 3’ end of your in vitro transcribed RNAs that can’t base pair with the Gs,” Rossi said.
Other polymerases, T3 and SP6, were tested. “Surprisingly, the stimulatory component of these transcripts is the 5’ initiating triphosphate,” wrote the paper’s authors. These findings have led to the development of an improved method for bacteriophage polymerase-mediated siRNA synthesis that incorporates two 3’ adenosines to prevent base-pairing with the initiating Gs, thereby allowing RNAse T1 and CIP to remove the initiating 5’ nucleotides and triphosphates of the transcripts.”
The data not only indicate that home-brew in vitro transcription RNAi is viable, but that “small RNAs with triphosphates can have very potent interferon effects, which is something that one could think about as a potential therapeutic application for treating viral infection,” Rossi added.