Singlet Oxygen as the Virus-Killing Agent: Plasma and Plasma-Activated Water Against Three Bacteriophages (Guo et al., 2018)

Plasma treatment fills water with many Reactive Oxygen and Nitrogen Species (RONS), and this study set out to find which of them inactivates viruses in Plasma-Activated Water (PAW). The 2018 paper in Applied and Environmental Microbiology by Li Guo, Ruobing Xu, Lu Gou, Zhichao Liu, Yiming Zhao, Dingxin Liu, Lei Zhang, Hailan Chen and Michael G. Kong (received 3 April 2018, accepted 12 June 2018) tries to name the culprit for viruses. Using bacteriophages (viruses that infect bacteria) as stand-ins for pathogenic viruses, the authors report that both direct plasma and plasma-treated water inactivated all three phages tested, and that the effect was almost entirely blocked by two chemicals that mop up singlet oxygen, an energised form of ordinary oxygen. Their conclusion is that singlet oxygen, rather than hydrogen peroxide, nitrite or nitrate, played "a primary role".

What was the set-up?

The plasma came from a surface discharge: a flat high-voltage electrode and a grounded hexagonal mesh separated by a polytetrafluoroethylene layer, run at a power density of 0.2 W/cm². It sat 8 mm above about 8 ml of liquid, roughly 1 mm deep, in a sealed box flushed at 4 litres/min with argon containing 1% artificial air. The authors say the argon raised the output of reactive species compared with discharge in air alone. This is one variety of the Non-Thermal Plasma (Cold Atmospheric Plasma) devices covered in this dossier.

Three phages were chosen because their genetic material differs: T4 (double-stranded DNA), Φ174 (single-stranded DNA) and MS2 (RNA). They were either exposed to the plasma directly or mixed in equal volume with water that had been treated beforehand, then left for 1 h at 22°C before surviving infectious particles were counted as plaque-forming units (PFU).

What chemistry was found in the water?

After 2 min of treatment, standard assay kits gave hydrogen peroxide at 221 μM, nitrite at 8 μM and nitrate at 216 μM. Short-lived species were probed by electron spin resonance with four "spin traps", each meant to capture particular species. The trap for singlet oxygen gave an adduct reading of 145 μM, the trap for hydroxyl radicals only 0.5 μM. The paper is careful that these adduct figures show only relative amounts, not true concentrations; on that basis it infers that hydroxyl radicals were scarce and singlet oxygen comparatively plentiful.

How well did it inactivate the phages?

The reported numbers are large:

• Direct plasma, T4. 40 s cut infectivity from 8.7 × 10^10 to 3.7 × 10^5 PFU/ml; 80 s left about 400 PFU/ml; 100 s abolished it.
• Direct plasma, Φ174 and MS2. 30 s brought reductions of roughly 4.2 and 4.6 orders of magnitude, with near-complete loss at 60 s. The authors read this as these two phages being more sensitive than T4.
• Treated water, T4. Water given 60 s or 120 s of plasma, then incubated with phage for 1 h, lowered T4 from 5.8 × 10^11 to 6.0 × 10^6 and to about 20 PFU/ml respectively.
• Longer contact. Using water given half the plasma time needed for near-complete inactivation (60 s for T4, 40 s for Φ174 and MS2), mixed 1:1 with phage suspension, T4 fell about 7.2 orders after 4 h and 8.8 after 8 h; Φ174 fell to 124 PFU/ml after 6 h and MS2 to 21 PFU/ml after 4 h. Formaldehyde, used as a comparison virus-inactivating agent, achieved under 2 orders in 8 h. Its strength is given as 1‰ in the results text but 1% in the figure legend and methods.

On storage, water kept for 10 days at 22°C was reported to be only slightly less active than fresh water, with no clear difference between glass and plastic tubes or between light and dark storage (the latter results are in supplementary figures).

Why singlet oxygen?

The argument proceeds in two steps. First, the authors made up a solution of the three long-lived species alone, at 500 μM hydrogen peroxide, 75 μM nitrite and 500 μM nitrate. That is more of each than the plasma-treated water was measured to contain, yet it reduced phage counts by only 1.6 to 3 orders of magnitude. They take this to mean the long-lived species are not the main agents.

Second, they added scavengers aimed at individual short-lived species: mannitol for hydroxyl radicals, tiron for superoxide, ebselen for peroxynitrite, and sodium azide and L-histidine for singlet oxygen. On their own the scavengers did not obviously harm the phages. With direct plasma, azide and histidine almost entirely removed the killing effect while the others made no significant difference; the same pattern held for treated water whether the scavenger was added before or after plasma exposure. A separate probe the paper calls highly selective for singlet oxygen, trans-1-(2′-methoxyvinyl)pyrene (described in a supplementary figure), detected it in both kinds of treated water.

What does the damage look like?

For T4, the authors report that much of the extracted genomic DNA stuck in the loading wells of an agarose gel instead of forming a single band, but ran normally once cut with the enzyme DraI. They interpret this as DNA cross-linked to itself or to coat proteins. Protein gels showed some bands weakening, and electron microscopy showed treated phages clumped into large aggregates. Treated water did less damage than direct plasma, which the authors attribute to its lacking some short-lived species and UV. Their proposed chemistry, drawn from cited literature rather than measured here, is that singlet oxygen links cysteine residues into disulfide bridges (encouraging clumping) and joins guanine in DNA to lysine in proteins.

What the paper does not establish

• No human or animal viruses were tested. Phages were used explicitly as surrogates; the paper mentions poliovirus and foot-and-mouth disease virus only as motivation.
• Clean water only. The authors themselves note that singlet oxygen also reacts with organic compounds, so organic matter would weaken the effect, and they restrict the suggested uses to space and water disinfection.
• The mechanism rests on scavengers. The singlet-oxygen identification depends on two scavengers plus a probe reported only in supplementary material; the spin-trap figures are relative.
• No quenching step is described. The methods move from incubation to serial dilution and plating without naming a chemical neutralizer, the gap that The Neutralization Problem: Why Plasma-Activated Water Kill Rates May Be Overstated argues can inflate kill figures.
• One device, one gas mix. All results come from a single argon-rich surface discharge; the paper does not show the same species balance in air-only or other reactors.

How it fits the dossier

The paper argues that hydrogen peroxide, nitrite and nitrate are not the main agents here, but its evidence on that point is a single comparison: one mixture of the three, tested together, proved weaker than the plasma-treated water. The species were not tested separately. A related finding appears in Adding Iodide to Plasma-Activated Water: Faster Kills, a Proposed Hypoiodous Acid Mechanism, and a Salmonella Exception (McClenaghan et al., 2025), where a singlet-oxygen scavenger also abolished killing of E. coli, although there removing peroxide did too. Its report of activity lasting 10 days also sits awkwardly beside the neutralization paper's modest effect of day-old water on fresh bacteria. The two used different organisms, water and devices, so the results are not directly comparable.