Karikó and Weissman’s Research, Paper by Paper

Katalin Karikó and Drew Weissman shared the 2023 Nobel Prize in Physiology or Medicine for a discovery about messenger RNA (mRNA), the cell’s working copy of a gene. The prize rests on a long run of laboratory papers — PubMed lists 34 that carry both their names, from 2000 to 2022. This page walks through the landmark ones in order. For each paper: the question they asked, what they actually did and in what (cells in a dish, mice, monkeys), what they found, and which obstacle it removed on the road to mRNA medicine. A one-picture map of the whole chain is in section 10.

A note on evidence before you start: none of the research papers described here tested a medicine or vaccine in people. Most were done in cells growing in a dish; later ones in mice, and a few in monkeys. The human trials of the COVID-19 vaccines are on the companion page From Pseudouridine to the COVID-19 Vaccines, and the vaccines’ safety record is on the main page, under The Safety Record. The three places where this record mentions snake venom phosphodiesterase, an enzyme long used as a laboratory tool, are noted paper by paper below and gathered together in section 10.

Table of Contents

  1. 1. Before They Met (1988–1998)
  2. 2. 2000–2004: An HIV Vaccine and the Inflammation Problem
  3. 3. 2005: The Discovery
  4. 4. 2007–2008: Pseudouridine mRNA Makes More Protein
  5. 5. 2010–2011: Why It Works, and How to Clean It
  6. 6. 2012: A Medicine Tested in Animals
  7. 7. 2013–2015: Delivery
  8. 8. 2017–2018: The First Vaccines in Animals
  9. 9. 2020: SARS-CoV-2 in Mice
  10. 10. What the Record Shows
  11. Key Research Papers
  12. Connections
  13. Featured Videos

1. Before They Met: Temple, 2-5A and the First mRNA Experiments (1988–1998)

Katalin Karikó earned her PhD in Szeged, Hungary, in 1982 and did postdoctoral research at the Hungarian Academy of Sciences in Szeged until 1985, then at Temple University in Philadelphia; in 1989 the University of Pennsylvania appointed her an assistant professor (the Nobel Foundation’s biography). Her first American papers were not about vaccines at all.

1988 · Temple University: the interferon “2-5A” pathway

The question. When a virus gets into a cell, the alarm hormone interferon switches on an enzyme that strings together a tiny molecule called 2-5A. 2-5A in turn switches on RNase L, an enzyme that shreds RNA so the virus cannot copy itself. Which proteins inside the cell actually grab hold of 2-5A?

What they did (test-tube biochemistry). R. J. Suhadolnik’s laboratory at Temple University — with Karikó as second author — made two light-activated versions of 2-5A (“photoaffinity probes”, which stick for good to whatever protein they are touching when a lamp is switched on). They built them with an enzyme from rabbit reticulocyte (young red blood cell) extracts and tested them on extracts of interferon-treated mouse cells.

What they found. The probes behaved like natural 2-5A — they switched RNase L on — and the two versions tagged quite different sets of 2-5A-binding proteins. To confirm the probes’ structure, the abstract names snake venom phosphodiesterase as one of three enzymes, with T2 RNase and a bacterial alkaline phosphatase, used to cut the new molecules apart so the pieces could be identified — the classic laboratory use of that enzyme. Its history, and every later mention, are on the Snake Venom Phosphodiesterase page.

Why it belongs here. This was not mRNA medicine, and it removed no obstacle to it. It belongs in the story because the same 2-5A and RNase L pathway returns in the pair’s 2011 paper (section 5).

Paper: Suhadolnik, Karikó, Sobol et al., Biochemistry, 1988 — PubMed PMID: 3242613

1989 and 1990 · Other laboratories: delivered mRNA can make protein

Two papers from other groups set the stage. In 1989 Malone, Felgner and Verma at the Salk Institute wrapped lab-made mRNA in a fatty carrier (a cationic liposome) and delivered it into cells growing in dishes — mouse, human, rat, frog and fruit-fly cells — where it was translated into protein. mRNAs given a protective “cap” and borrowed end-sequences from the beta-globin gene made at least 1,000-fold more of a test protein than mRNAs without them (cells in a dish; PubMed PMID: 2762315). In 1990 Wolff, Malone and colleagues injected RNA and DNA straight into mouse leg muscle and detected the encoded proteins in every case, with “no special delivery system” required (mouse; PubMed PMID: 1690918).

The obstacle they removed: the doubt that a living cell would read an mRNA delivered from outside. What remained — inflammation, low protein output and delivery — is laid out on the main page under Why mRNA Was Considered a Dead End.

1998 · Karikó at Penn: a better way into cells

In Penn’s neurosurgery department, Karikó worked on getting mRNA into cells. Her 1998 paper with Kuo, Barnathan and Langer found that conditioning the fatty carrier in phosphate buffer before mixing it with the nucleic acid raised the output of a test protein up to 56-fold from mRNA and up to 26-fold from DNA, in a human bone-tumour cell line (cells in a dish; PubMed PMID: 9518670). It removed a small obstacle — getting more mRNA into cells in a dish — but not the immune reaction, which was still unexplained.

Back to Table of Contents

2. 2000–2004: An HIV Vaccine and the Inflammation Problem

Drew Weissman, a physician and immunologist who had done postdoctoral research at the National Institutes of Health, set up his own research group at Penn in 1997. The Nobel committee describes what happened next: “A new colleague of Karikó at her university was the immunologist Drew Weissman. He was interested in dendritic cells, which have important functions in immune surveillance and the activation of vaccine-induced immune responses.” Dendritic cells are the immune system’s scouts: they pick up pieces of an invader and show them to T cells, which then learn to attack it. In the committee’s account, Weissman’s major goal was a vaccine against HIV-1, and together the two tested whether lab-made mRNA could be delivered to dendritic cells.

2000 · Journal of Immunology: HIV gag mRNA in dendritic cells (their first joint paper)

The question. Could mRNA load an HIV protein into dendritic cells so that they would train T cells against it — the core of a vaccine?

What they did (human cells in a dish). They delivered mRNA encoding HIV’s gag protein into human dendritic cells, then mixed those cells once with T cells.

What they found. Dendritic cells took up and translated the mRNA better than other antigen-presenting cells. One round of stimulation produced primary CD4+ and CD8+ T-cell responses in which 5–12.5% of the T cells were specific for the HIV protein. And the mRNA itself gave the dendritic cells a “maturation signal”. The authors wrote that the approach “has the potential to be a potent and effective anti-HIV T cell-activating vaccine.”

What it removed, and what it revealed. It showed that mRNA could carry a vaccine target into the right immune cells. The maturation signal looked like a bonus. The Nobel committee’s scientific background notes that it “initially was interpreted as a positive effect” and that “the negative consequences of innate immune activation by in vitro transcribed mRNA were not fully appreciated at this point.” That activation would become the next obstacle. Weissman’s NIAID grants of these years carried the title “RNA delivery for dendritic cell HIV antigen presentation” (see The NIH Grants).

Paper: Weissman, Ni, Scales et al., J Immunol, 2000 — PubMed PMID: 11035115

2002 · Journal of Biological Chemistry: mRNA outside a cell switches dendritic cells on

Following up, the team (Ni, Capodici, Cannon and colleagues, with Karikó and Weissman) characterised two signalling routes by which mRNA lying outside a cell activates dendritic cells, one of them the release of the inflammatory messenger TNF-alpha. They also identified the string of A’s on mRNA’s tail, poly(A), as a newly found trigger acting through what appeared to be a P2Y-type nucleotide receptor (cells in a dish). In plain terms: mRNA itself is something immune cells react to. Paper: PubMed PMID: 11821398.

2004 · Journal of Biological Chemistry: mRNA trips Toll-like receptor 3

The question. Which sensor is reacting? Toll-like receptors (TLRs) are the innate immune system’s tripwires. TLR3 was already known to respond to double-stranded RNA, which many viruses make while copying themselves.

What they did (cells in a dish). They built human kidney-derived cells carrying TLR3 plus a light-producing reporter, and exposed those cells — and human dendritic cells — to lab-made mRNA and to RNA from dying cells.

What they found. Lab-made mRNA switched TLR3 on, more strongly at higher doses. Dendritic cells matured in response, and an antibody that blocks TLR3 held that maturation back. RNA released from dying cells also made dendritic cells release interferon-alpha, and digesting it first with an RNA-cutting enzyme abolished the effect. Their conclusion: “RNA, likely through secondary structure, is a potent host-derived activator of TLR3.”

The obstacle it named. The inflammation problem now had an address — TLR3 — though not yet a fix.

Paper: Karikó, Ni, Capodici et al., J Biol Chem, 2004 — PubMed PMID: 14729660

Back to Table of Contents

3. 2005: The Discovery

This is the paper the Nobel Prize honours. The committee’s press release: “These seminal results were published in 2005, fifteen years before the COVID-19 pandemic.”

The question. Every cell is full of its own RNA, and the immune system leaves it alone. Why, then, do dendritic cells attack RNA made in a test tube? Natural RNA is not built only from the four plain letters A, U, G and C: cells decorate many of those letters chemically after making them (“nucleoside modifications”). Lab-made mRNA carries none. Could the missing decorations be what the sensors notice?

The clue. In the Nobel committee’s summary of the paper, “eukaryotic mRNA and tRNA, in which base modifications are abundant, did not stimulate a cytokine response while prokaryotic and in vitro-transcribed mRNA did.” Put simply: RNA from animal cells — including transfer RNA (tRNA), a kind rich in decorations — barely bothered the dendritic cells, while bacterial RNA and lab-made mRNA set them off.

What they did (cells in a dish). They made lab RNA in which one letter was swapped for a natural decorated version — 5-methylcytidine (m5C), N6-methyladenosine (m6A), 5-methyluridine (m5U), 2-thiouridine (s2U) or pseudouridine (Ψ: the same uridine letter, attached to the RNA backbone through a carbon atom instead of a nitrogen). They tested each one on human dendritic cells and on cells engineered to carry a single sensor: TLR3, TLR7 or TLR8.

What they found. In the abstract’s words: “RNA signals through human TLR3, TLR7, and TLR8, but incorporation of modified nucleosides m5C, m6A, m5U, s2U, or pseudouridine ablates activity.” Dendritic cells given modified RNA made “significantly less cytokines and activation markers” (cytokines are the inflammatory messengers). Bacterial and mitochondrial RNA set the cells off strongly; total RNA from mammals, rich in modifications, did not. The committee adds a detail from the figures: only the decorated uridines (m5U, s2U and Ψ) abolished dendritic-cell activation.

Their explanation — the “evolutionary origin” in the title. The innate immune system “may therefore detect RNA lacking nucleoside modification as a means of selectively responding to bacteria or necrotic tissue.” Undecorated RNA is a sign of germs or of damaged tissue; decorated RNA looks like self.

The obstacle it removed: inflammation — at least in cells in a dish. One swapped letter made lab RNA look like the body’s own. The sensors involved (Toll-like receptors) and the cells (dendritic cells) were themselves the subject of the 2011 Nobel Prize; see Beutler, Hoffmann & Steinman.

The patent dates from the same month: US patent 8,278,036, “RNA containing modified nucleosides and methods of use thereof”, inventors Karikó and Weissman, assigned to the Trustees of the University of Pennsylvania, priority date 23 August 2005, granted 2 October 2012. Its full text contains no “venom”, “snake” or “phosphodiesterase”; its worked examples cut RNA with a bacterial nuclease and test for bacterial toxin with the standard horseshoe-crab-blood (Limulus) test.

Paper: Karikó, Buckstein, Ni, Weissman, Immunity, August 2005 — PubMed PMID: 16111635

Back to Table of Contents

4. 2007–2008: Pseudouridine mRNA Makes More Protein

2007 · a review: what modified RNA could mean for medicine

In a review (no new experiments), Karikó and Weissman set their finding beside a known fact about DNA: the sensor TLR9 responds to DNA, and methylation of the CpG motifs in DNA blocks that signal, while TLR3, TLR7 and TLR8 respond to RNA. Their conclusion pointed at medicine: “RNAs containing modified nucleosides, and thus lacking immune-activating properties, have potential importance in clinical applications.” Paper: PubMed PMID: 17786850.

2008 · Molecular Therapy: more protein, in cells and in mice

The question. The paper states the problem in its opening lines: “in its present form, mRNA is unfeasible for clinical use because of its labile and immunogenic nature.” Would pseudouridine mRNA make more protein, last longer and stay quiet — not only in a dish but in a living animal?

What they did (cells, cell extracts and mice). They compared pseudouridine mRNA with ordinary mRNA in mammalian cells, in cell extracts, and injected into the veins of mice at 0.015–0.15 mg/kg.

What they found. Pseudouridine mRNA had “a higher translational capacity”. In the mice, both the injected mRNA and its protein turned up in the spleen at 1, 4 and 24 hours, at significantly higher levels when the mRNA carried pseudouridine. “Even at higher doses, only the unmodified mRNA was immunogenic, inducing high serum levels of interferon-alpha” — interferon-alpha being one of the body’s main antiviral alarm signals. The modified mRNA was also more stable.

The obstacle it removed: low protein output — now shown in an animal, not just in a dish. The authors called pseudouridine mRNA “a promising tool for both gene replacement and vaccination.” The Nobel committee names this paper among its three key publications.

The “snake venom” sentence. In the discussion, explaining why the modified mRNA lasts longer, the paper has its only mention of snake venom:

“A likely contributing factor to the enhanced translation observed with Ψ modification is an increase in biological stability of the mRNAs (Figure 4d). Indeed, higher resistance to hydrolysis by phosphodiesterases from snake venom and spleen has been reported when uridine was replaced with Ψ in dinucleotide substrates.[19]”

Reference 19 is a two-page 1965 chemistry paper by Naylor, Ho and Gilham (PubMed PMID: 4284810). The 2008 authors cite its observation — that two-letter RNA pieces containing pseudouridine resist being cut by those enzymes — as support for their own finding that pseudouridine mRNA is more stable. The enzyme is named nowhere else in the paper’s text, methods included; the full account is on the Snake Venom Phosphodiesterase page.

The funding note reads: “This work was supported by the National Institutes of Health (NIH) grants NIAID AI-050484, NHLBI HL87688, and NINDS NS-29331.” What each grant was for is on The NIH Grants page.

Paper: Karikó, Muramatsu, Welsh et al., Mol Ther, 2008 — PubMed PMID: 18797453

Back to Table of Contents

5. 2010–2011: Why It Works, and How to Clean It

2010 · Nucleic Acids Research: the PKR brake

The question. Why does pseudouridine mRNA make more protein? Until this paper, in the abstract’s words, “the reason for this enhancement has not been identified.”

What they did (test-tube and cell experiments, including cells lacking the enzyme). They looked at PKR, an antiviral enzyme that, once switched on by RNA, stops a cell making protein by tagging (phosphorylating) a starter protein called eIF-2α — a brake that keeps a virus from using the cell’s machinery.

What they found. Ordinary uridine mRNA switched PKR on, which tagged eIF-2α and blocked translation. Pseudouridine mRNA switched PKR on “to a lesser degree”, and its translation was not repressed. PKR grabbed uridine mRNA more efficiently in pull-down tests. And in cells with no PKR at all, the two kinds of mRNA were translated equally — showing that PKR was the difference.

The obstacle it removed: the unexplained protein gain became a mechanism that could be designed for. The Nobel committee names this paper too: “The effect was due to the reduced activation of an enzyme that regulates protein production.”

Paper: Anderson, Muramatsu, Nallagatla et al., Nucleic Acids Res, 2010 — PubMed PMID: 20457754

2011 · Nucleic Acids Research: the 2-5A pathway, again

The question. Does modification also quiet the interferon 2-5A system — OAS, the enzyme that makes 2-5A, and RNase L, which 2-5A switches on? This is the very pathway of Karikó’s 1988 Temple paper.

What they did (test-tube enzyme tests and cells, checked in cells and mice lacking RNase L). They exposed purified OAS and RNase L, and cells, to unmodified and modified lab RNA.

What they found. Unmodified RNA switched OAS on, made RNase L cut the cell’s ribosomal RNA, and was itself cut rapidly by RNase L. Modified RNA switched OAS on less efficiently, caused limited cutting, and resisted RNase L. Pseudouridine mRNA “is translated longer and has an extended half-life.”

The obstacle it removed: a second hidden brake, and with it the short working life of the mRNA.

The other two “snake venom” mentions are in this paper’s discussion, both in one sentence. It follows the remark that pseudouridine “has not previously been demonstrated to cause resistance to nucleases” (RNA-cutting enzymes):

“RNA containing Ψ was cleaved efficiently by RNase A, RNase H (36), RNase T1, RNase T2, nuclease P1 and snake venom phosphodiesterase, although there is some indication that pancreatic diesterase and snake venom phosphodiesterase may cleave Ψ-RNA with reduced efficiency (37).”

Reference 36 is a 2004 methods paper on detecting RNA modifications by Zhao and Yu (PubMed PMID: 15146083); reference 37 is the same 1965 study. As in 2008, the enzyme is named nowhere else in the text, methods included; see the Snake Venom Phosphodiesterase page.

Paper: Anderson, Muramatsu, Jha et al., Nucleic Acids Res, 2011 — PubMed PMID: 21813458

2011 · Nucleic Acids Research: HPLC cleaning removes double-stranded RNA

The question. Even pseudouridine mRNA left a “residual induction of type I interferons (IFNs) and proinflammatory cytokines”. Where was the leftover alarm coming from?

What they did (primary cells in a dish). They looked for contaminants made alongside the mRNA during test-tube production, and removed them with high-performance liquid chromatography (HPLC) — a fine separating column that sorts molecules by their chemical behaviour.

What they found. The culprits were contaminants “including double-stranded RNA” — the very shape that TLR3 and PKR evolved to detect. After HPLC, modified mRNA “does not induce IFNs and inflammatory cytokines and is translated at 10- to 1000-fold greater levels in primary cells.” Purified ordinary mRNA was translated better too, but “still induced high levels of cytokine secretion” — so both steps, the modified letter and the cleaning, were needed.

The obstacle it removed: double-stranded RNA contaminants.

Paper: Karikó, Muramatsu, Ludwig, Weissman, Nucleic Acids Res, 2011 — PubMed PMID: 21890902

Back to Table of Contents

6. 2012: A Medicine Tested in Animals

The question. Could a tiny dose of purified pseudouridine mRNA act as a medicine in a living animal? The test case was erythropoietin (EPO), the hormone that tells the bone marrow to make red blood cells (see the animation How Your Body Makes Red Blood Cells).

What they did (mice, plus a test in macaque monkeys). They injected mice with EPO-encoding mRNA packaged with a commercial transfection carrier, then measured EPO in the blood, counted young red cells (reticulocytes) and tracked the hematocrit — the share of the blood made up of red cells. In macaques, they injected rhesus-monkey EPO mRNA into the abdominal cavity.

What they found. A single injection of 100 ng (0.005 mg/kg) — one ten-millionth of a gram — raised blood EPO significantly within 6 hours, and the level held for 4 days. Ordinary uridine mRNA gave 10–100-fold less EPO, lasting only 1 day. The EPO worked: as little as 10 ng doubled reticulocyte numbers, and weekly 100 ng injections raised the hematocrit from 43 to 57%, which held with continued treatment. “Even when a large amount of pseudouridine-mRNA was injected, no inflammatory cytokines were detectable in plasma.” In the macaques, blood EPO rose significantly.

The obstacle it removed: “does it work in an animal?” — yes, at sub-microgram doses in mice, with a rise in blood EPO in monkeys too. The paper thanks NIH grants R01NS029331 and R42HL87688. The second is a small-business (STTR) grant from the NHLBI to RNARx, the small company the two had formed, titled “Erythropoietin-encoding mRNA for treatment of anemia” (2007–2013); see The NIH Grants and Anemia.

Paper: Karikó, Muramatsu, Keller, Weissman, Mol Ther, 2012 — PubMed PMID: 22334017

Back to Table of Contents

7. 2013–2015: Delivery

In 2013 Karikó moved to industry. The Nobel Foundation’s biography: “In 1989, she was appointed Assistant Professor at the University of Pennsylvania, where she remained until 2013. After that, she became vice president and later senior vice president at BioNTech RNA Pharmaceuticals.” She kept publishing with Weissman. The problem now was the last of the three: getting fragile mRNA into the body’s cells.

2014 · a review with Şahin and Türeci

Writing with Uğur Şahin and Özlem Türeci in Nature Reviews Drug Discovery (a review, no new experiments), Karikó — listed at both BioNTech and Penn — surveyed lab-made mRNA as “a potential new drug class”. The review reported that “mRNA-based cancer immunotherapies and infectious disease vaccines have entered clinical development”, alongside newer uses: replacing missing proteins, making stem cells, and gene editing with mRNA-encoded “designer nucleases”. Paper: PubMed PMID: 25233993.

2015 · Journal of Controlled Release: lipid nanoparticles, six routes into mice

The question. Lipid nanoparticles (LNPs) — microscopic fat bubbles — were already carrying another kind of RNA, short-interfering RNA, in clinical trials, but “little is known about the potential of LNPs to deliver mRNA.” Where does LNP-packed modified mRNA go, and how long does it keep working?

What they did (mice). Norbert Pardi, Weissman and colleagues, including Karikó and lipid-nanoparticle scientists from Vancouver, Canada, packed HPLC-purified mRNA containing 1-methylpseudouridine and encoding firefly luciferase (a light-producing marker protein) into LNPs. They injected mice at 0.005–0.250 mg/kg by 6 different routes and watched the glow by imaging.

What they found. Injected under the skin, into muscle or into the skin itself, the mRNA was translated at the injection site for up to 10 days. Delivered into the windpipe, it made protein in the lung for several days. Injected into a vein or the abdominal cavity (and, less so, into muscle or windpipe), the particles travelled through the body and the mRNA was translated in the liver for 1–4 days.

The obstacle it removed: delivery — in the authors’ words, “LNPs are appropriate carriers for mRNA in vivo.” Note the letter used: this paper already used 1-methylpseudouridine, a close relative of pseudouridine.

Paper: Pardi, Tuyishime, Muramatsu et al., J Control Release, 2015 — PubMed PMID: 26264835

2015 · a commentary: “mRNA: Fulfilling the Promise of Gene Therapy”

In a two-page commentary in Molecular Therapy (no new experiments; Karikó now listed at BioNTech RNA Pharmaceuticals), Weissman and Karikó discussed another group’s report that mRNA engineered to be rich in the letters G and C, made without modified nucleosides and purified by HPLC, also avoided immune activation. They recalled that the immune reaction to lab-made (“in vitro transcribed”, IVT) mRNA had blocked its use for protein-replacement therapy and that “this problem was recently solved by the introduction of modified nucleosides into the IVT mRNA”. They set out the experiments they thought were still needed, and summed up the field: “to generate an optimal IVT mRNA for protein therapy, one needs to reduce or completely eliminate its U content without interfering with its translatability.” Paper: PubMed PMID: 26321183.

2015 · another group: N1-methylpseudouridine

The same year, a team at Ghent University and MIT (Andries, Mc Cafferty, De Smedt, Weiss, Sanders and Kitada) tested that close relative of pseudouridine, N1-methylpseudouridine (m1Ψ). Their abstract calls the modified-letter approach “the Kariko paradigm”. In cell lines and mice, m1Ψ mRNA, alone or combined with m5C, gave up to roughly 13- to 44-fold more of a test protein than the pseudouridine versions, depending on the comparison; in cells it also set off less innate immune activity and left the cells healthier (cells and mice; PubMed PMID: 26342664).

m1Ψ is the letter in the COVID-19 vaccines: the Nobel committee’s scientific background states that both COVID-19 mRNA vaccines “had complete substitutions of uridine with N1-methylpseudouridine (m1Ψ)”. See also the 2021 review by Nance and Meier (PubMed PMID: 34075344) and a 2017 study by Svitkin and colleagues reporting that m1Ψ raises the number of ribosomes reading each mRNA (PubMed PMID: 28334758).

Back to Table of Contents

8. 2017–2018: The First Vaccines in Animals

2017 · Nature: one shot against Zika, in mice and monkeys

The question. Zika virus had “recently emerged as a pandemic associated with severe neuropathology in newborns and adults”, with no vaccine. Could a single low dose of a modified-mRNA vaccine in lipid nanoparticles protect against it?

What they did (mice and non-human primates). Pardi, Weissman and a large team — including Karikó, now at BioNTech, and NIAID scientists — injected into the skin an mRNA vaccine encoding two Zika surface proteins (pre-membrane and envelope) from a strain from the 2013 outbreak, then exposed the animals to the virus.

What they found. The vaccine “elicited potent and durable neutralizing antibody responses in mice and non-human primates” (neutralizing antibodies block a virus from entering cells). 30 µg protected mice against Zika given 2 weeks or 5 months after vaccination, and a single 50 µg dose protected non-human primates against a challenge at 5 weeks.

The obstacle it removed: “can one shot protect a monkey?” — for this vaccine, in these animals, yes. The paper’s support line: “Supported by the NIH, NIAID Duke Center for HIV/AIDS Vaccine Immunology AI100645.” More on the disease: Zika Virus.

Paper: Pardi, Hogan, Pelc et al., Nature, 2017 — PubMed PMID: 28151488

2018 · Journal of Experimental Medicine: why the antibodies are strong

Antibodies get better in germinal centres — small training camps in the lymph nodes where B cells, coached by T follicular helper (Tfh) cells, refine their antibodies and become long-lived memory and plasma cells. In this paper (mice and non-human primates), skin injections of modified-mRNA vaccines encoding various viral surface proteins drove strong Tfh and germinal-centre B-cell responses, which went with “long-lived and high-affinity neutralizing antibodies and durable protection”. In the authors’ comparisons the mRNA vaccines “outperformed adjuvanted protein and inactivated virus vaccines and pathogen infection”. And, they wrote, “The incorporation of noninflammatory, modified nucleosides in the mRNA is required for the production of large amounts of antigen and for robust immune responses.”

Paper: Pardi, Hogan, Naradikian et al., J Exp Med, 2018 — PubMed PMID: 29739835

2018 · a review: “a new era in vaccinology”

Pardi, Hogan, Porter and Weissman reviewed the field (no new experiments). mRNA vaccines, they wrote, had been held back by “the instability and inefficient in vivo delivery of mRNA”. “Recent technological advances have now largely overcome these issues, and multiple mRNA vaccine platforms against infectious diseases and several types of cancer have demonstrated encouraging results in both animal models and humans.” Paper: PubMed PMID: 29326426. The HIV line that started the collaboration continued too: a 2019 paper characterised HIV-1 modified-mRNA vaccines in rabbits and rhesus macaques (PubMed PMID: 30974332); see HIV/AIDS.

Back to Table of Contents

9. 2020: SARS-CoV-2 in Mice

SARS-CoV-2, the virus that causes COVID-19, was first identified in late December 2019 and January 2020, and its genetic sequence was posted publicly on 10–11 January 2020. Because an mRNA vaccine changes only the protein it encodes, the platform built in the papers above could be pointed at a new virus quickly.

The question. How strong is the immune response to a single dose of modified-mRNA vaccines encoding the SARS-CoV-2 spike protein — either the whole spike, or just its receptor-binding domain, the part that latches onto human cells?

What they did (mice). Laczkó, Pardi and colleagues, with Weissman and Karikó (listed at Penn and BioNTech), immunized mice once with lipid-nanoparticle vaccines and measured their T cells, B cells and antibodies.

What they found. One dose induced “strong type 1 CD4+ and CD8+ T cell responses, as well as long-lived plasma and memory B cell responses”, with “robust and sustained neutralizing antibody responses”. In laboratory tests the antibodies “do not show antibody-dependent enhancement of infection in vitro” — antibody-dependent enhancement being a feared effect in which antibodies help a virus into cells instead of blocking it.

The obstacle it removed: a new virus — the platform gave strong responses against SARS-CoV-2 in mice after one dose. The paper’s interest statement notes that Weissman and Karikó are named on patents describing nucleoside-modified mRNA, and that Karikó was then a BioNTech employee. The human trials of the two COVID-19 mRNA vaccines, which began in 2020, are on the companion page From Pseudouridine to the COVID-19 Vaccines.

Paper: Laczkó, Hogan, Toulmin et al., Immunity, 2020 — PubMed PMID: 32783919

Back to Table of Contents

10. What the Record Shows

Read in order, the papers form a chain. Each asked a narrow question, answered it in cells or animals, and removed one obstacle: the 2005 paper the inflammation; the 2008 and 2010 papers the low protein output, and the reason for it; the 2011 papers a second brake and the double-stranded RNA contaminants; the 2012 paper the doubt that it would work as a medicine in an animal; the 2015 paper the delivery; the 2017 paper a vaccine that protected monkeys; and the 2020 paper the new virus, in mice. The Nobel committee puts it in one line: “Karikó and Weissman had eliminated critical obstacles on the way to clinical applications of mRNA.”

A left-to-right chain of Karikó and Weissman’s research from 2000 to 2020. Each column puts an obstacle above the paper that removed it: immune alarms found in 2000–2004; inflammation removed in 2005; low protein output in 2008 and 2010; leftover double-stranded RNA in 2011; a working medicine in mice and macaques in 2012; delivery by lipid nanoparticles in 2015; a one-shot Zika vaccine in mice and monkeys in 2017; and a SARS-CoV-2 vaccine in mice in 2020. Three lanes underneath show what each step was tested in — cells, then animals — and that none of these papers tested anything in people. EACH OBSTACLE, AND THE PAPER THAT REMOVED IT Karikó and Weissman, 2000 → 2020 · top row: what stood in the way · bottom row: the paper that removed it WHAT STOOD IN THE WAY Can mRNA teach immune cells? Inflammation: sensors attack lab-made RNA Too little protein made Leftover alarm from double- stranded RNA Does it work in an animal? Getting it into the body Can one shot protect a monkey? A new virus: SARS-CoV-2 2000–2004 2005 2008 · 2010 2011 2012 2015 2017 2020 mRNA works, but trips immune alarms modified bases silence TLR3, TLR7 and TLR8 pseudouridine mRNA: more protein (PKR) HPLC cleaning removes it EPO mRNA makes more red cells lipid nano- particles carry it one Zika shot protects mice and monkeys spike mRNA vaccine works in mice TESTED IN CELLS TESTED IN ANIMALS TESTED IN PEOPLE mice mice, macaques mice mice, monkeys mice none of these papers · COVID-19 vaccines were first tested in people in 2020 Tiers as each paper’s abstract states them · cells = test tube or cell culture

Read the picture left to right: each column is one obstacle above the paper that removed it, and the lanes underneath show the evidence climbing from cells to mice and monkeys — with no step in this chain tested in people.

Where snake venom phosphodiesterase is mentioned

Dr. Bryan Ardis said on a May 2025 podcast: “And in every one of those research studies, they state snake venom phosphodiesterase, snake venom…” A clip posted by his own show names the 2008 and 2011 papers. In the record laid out above, the enzyme is mentioned in three places:

PubMed lists 34 papers with both Karikó and Weissman as authors, from 2000 to 2022; none is dated 2009. Of the 22 published from 2008 on, 20 have open full text, and in those 20 the word “venom” occurs in two papers, three times in all — the 2008 and 2011 sentences quoted in sections 4 and 5. In both of those papers, those discussion sentences are the only places the enzyme is named; the methods sections do not mention it. The 12 joint papers from 2000 to 2007 are not in the open full-text archive. The 2005 patent’s text contains no “venom”, “snake” or “phosphodiesterase”. The enzyme’s history as a laboratory tool over more than seventy years, and Dr. Ardis’s statements set beside the papers’ exact sentences, are on the Snake Venom Phosphodiesterase page; every grant these papers thank is listed on The NIH Grants page.

Back to Table of Contents

Key Research Papers

  1. Suhadolnik RJ, Karikó K, Sobol RW Jr, et al. (1988). 2- and 8-azido photoaffinity probes. 1. Enzymatic synthesis, characterization, and biological properties of 2- and 8-azido photoprobes of 2-5A and photolabeling of 2-5A binding proteins. Biochemistry. — PubMed PMID: 3242613
  2. Weissman D, Ni H, Scales D, et al. (2000). HIV gag mRNA transfection of dendritic cells (DC) delivers encoded antigen to MHC class I and II molecules, causes DC maturation, and induces a potent human in vitro primary immune response. J Immunol. — PubMed PMID: 11035115
  3. Karikó K, Ni H, Capodici J, et al. (2004). mRNA is an endogenous ligand for Toll-like receptor 3. J Biol Chem. — PubMed PMID: 14729660
  4. Karikó K, Buckstein M, Ni H, Weissman D (2005). Suppression of RNA recognition by Toll-like receptors: the impact of nucleoside modification and the evolutionary origin of RNA. Immunity. — PubMed PMID: 16111635
  5. Karikó K, Muramatsu H, Welsh FA, et al. (2008). Incorporation of pseudouridine into mRNA yields superior nonimmunogenic vector with increased translational capacity and biological stability. Mol Ther. — PubMed PMID: 18797453
  6. Anderson BR, Muramatsu H, Nallagatla SR, et al. (2010). Incorporation of pseudouridine into mRNA enhances translation by diminishing PKR activation. Nucleic Acids Res. — PubMed PMID: 20457754
  7. Anderson BR, Muramatsu H, Jha BK, et al. (2011). Nucleoside modifications in RNA limit activation of 2'-5'-oligoadenylate synthetase and increase resistance to cleavage by RNase L. Nucleic Acids Res. — PubMed PMID: 21813458
  8. Karikó K, Muramatsu H, Ludwig J, Weissman D (2011). Generating the optimal mRNA for therapy: HPLC purification eliminates immune activation and improves translation of nucleoside-modified, protein-encoding mRNA. Nucleic Acids Res. — PubMed PMID: 21890902
  9. Karikó K, Muramatsu H, Keller JM, Weissman D (2012). Increased erythropoiesis in mice injected with submicrogram quantities of pseudouridine-containing mRNA encoding erythropoietin. Mol Ther. — PubMed PMID: 22334017
  10. Pardi N, Tuyishime S, Muramatsu H, et al. (2015). Expression kinetics of nucleoside-modified mRNA delivered in lipid nanoparticles to mice by various routes. J Control Release. — PubMed PMID: 26264835
  11. Andries O, Mc Cafferty S, De Smedt SC, et al. (2015). N(1)-methylpseudouridine-incorporated mRNA outperforms pseudouridine-incorporated mRNA by providing enhanced protein expression and reduced immunogenicity in mammalian cell lines and mice. J Control Release. — PubMed PMID: 26342664
  12. Pardi N, Hogan MJ, Pelc RS, et al. (2017). Zika virus protection by a single low-dose nucleoside-modified mRNA vaccination. Nature. — PubMed PMID: 28151488
  13. Pardi N, Hogan MJ, Naradikian MS, et al. (2018). Nucleoside-modified mRNA vaccines induce potent T follicular helper and germinal center B cell responses. J Exp Med. — PubMed PMID: 29739835
  14. Laczkó D, Hogan MJ, Toulmin SA, et al. (2020). A Single Immunization with Nucleoside-Modified mRNA Vaccines Elicits Strong Cellular and Humoral Immune Responses against SARS-CoV-2 in Mice. Immunity. — PubMed PMID: 32783919

PubMed Topic Searches

  1. PubMed: every paper by Karikó and Weissman together
  2. PubMed: pseudouridine mRNA and immune activation
  3. PubMed: nucleoside-modified mRNA lipid-nanoparticle vaccines

External Authoritative Resources

Back to Table of Contents

Connections

Back to Table of Contents