A virus with harmful proteins, not a bottle of poison
When people ask whether a virus "produces toxins," they usually picture something like a bacterial toxin: a substance the microbe manufactures and releases to damage the host. SARS-CoV-2 is not described that way in this evidence. What the papers point to is different: parts of the virus's own protein machinery can be damaging. The spike protein, the piece on the virus surface that lets it grab cells, is discussed specifically for its toxicity [1]. That is a property of a viral building block, not a separate secreted poison.
The clearest direct demonstration comes from lab work on protein clumping. Researchers found that several SARS-CoV-2 proteins can assemble into amyloid aggregates, which are ordered, sticky protein clumps [2]. Among them, aggregates of the viral protein NSP11 were toxic to mammalian cells in culture [2]. That is a real, measured toxic effect, but it was shown in cells in a dish, not in patients, so it tells us the potential exists rather than how much it matters during an actual infection.
Why the spike protein draws so much attention
The spike protein is the virus's most exposed and most-studied component, and it is also the target of most vaccines. A review of the Delta and Lambda variants discusses the toxicity of the spike protein alongside the variants' increased infectivity and disease severity [1]. The same review notes that these variants carry more changes in their spike and other non-structural proteins than the original Wuhan strain, and that Lambda has substitutions plus a deletion in part of the spike that help it partly escape neutralizing antibodies in vaccinated people [1].
The important nuance is that this review is a summary of the field, not a new experiment, and it frames spike toxicity as part of a broader picture of variant behavior. So it supports the idea that the spike protein is a plausible source of harm, but it does not by itself prove that spike toxicity is what drives severe disease in patients. The link between variant mutations, greater toxicity, and worse outcomes is presented as an explanation under discussion, not a settled mechanism [1].
Don't confuse the virus's proteins with outside toxins
There is a separate line of research on toxins and COVID-19 that can easily be misread. One review examines the role of toxins, toxoids, and venoms in the pathophysiology of COVID-19 [3]. Read carelessly, that sounds like evidence that SARS-CoV-2 makes toxins. It isn't. That paper is about natural toxic substances from other sources and their possible pharmacological or disease-related effects, not about the virus manufacturing a toxin of its own [3].
Keeping these apart matters for the reader's actual question. The strongest direct evidence for a virus-linked toxic effect is the cell-culture finding that NSP11 aggregates harm mammalian cells [2]. The spike protein's toxicity is discussed as a contributing factor to disease severity [1]. Neither of these is the same as saying the virus secretes a toxin, and the toxins-and-venoms review does not change that [3].
What this means if you're trying to make sense of it
The practical takeaway is that "toxin" is the wrong frame but not a meaningless one. SARS-CoV-2 damages the body mainly through infection and the immune response it triggers, and some of its proteins appear to add direct injury on top of that. The spike protein is the best-supported candidate for direct toxicity [1], and NSP11 aggregates are the clearest lab demonstration of a viral protein harming cells [2].
What remains genuinely open is how much these protein-level effects contribute to illness in real people. The NSP11 result comes from cultured cells [2], and the spike-toxicity discussion comes from a review of variants rather than a controlled clinical study [1]. The papers themselves point toward more work on whether viral protein aggregation plays a role in protein-misfolding conditions and other human disease processes [2]. So the fair summary is: the virus carries proteins that can be toxic, but calling it a toxin producer overstates what has actually been shown.
Sources used in this answer
Severity, Pathogenicity and Transmissibility of Delta and Lambda Variants of SARS-CoV-2, Toxicity of Spike Protein and Possibilities for Future Prevention of COVID-19
A review of the Delta and Lambda variants discusses the toxicity of the SARS-CoV-2 spike protein and links variant mutations to increased infectivity, pathogenicity, and partial immune escape.
Amyloidogenic proteins in the SARS-CoV and SARS-CoV-2 proteomes
Several SARS-CoV and SARS-CoV-2 proteins form amyloid aggregates in vitro, and SARS-CoV-2 NSP11 aggregates were shown to be toxic to mammalian cell cultures.
Harnessing the Natural Toxic Metabolites in COVID-19
A review examines the role of toxins, toxoids, and venoms in the pathophysiology of COVID-19, focusing on externally derived toxic substances rather than a toxin made by the virus.