Published online 7 June 2010 | Nature | doi:10.1038/news.2010.283


Mosquito saliva may signal infection outbreaks

A mosquito's sweet tooth could help researchers to detect deadly viruses.

MosquitoMosquitoes left traces of their saliva, containing viral RNA, on cards soaked in honey.Paul Zborowski

Baiting mosquito traps with cards soaked in honey, and then analysing viral RNA in saliva left by mosquitoes that feed on them, may be a way of tracking the spread of some diseases.

To assess whether mosquito populations are harbouring dangerous viruses, researchers often use traps baited with carbon dioxide or light to attract the insects, which are then ground up and subjected to genetic analyses to identify any viruses. But this procedure does not distinguish between viruses that are safely confined to the mosquitoes' gut and those that have migrated to their salivary glands to be released in saliva when the insects bite a host. An alternative approach is to analyse blood samples from animals, such as chickens and pigs, for antibodies that signal the presence of pathogens. Both methods put people at risk of exposure to the viruses.

Andrew van den Hurk of the Queensland Health Forensic and Scientific Services in Coopers Plains, Australia, and his colleagues have developed a method for collecting mosquito saliva by allowing the insects to feed on honey-drenched cards placed in a trap filled with carbon dioxide. The cards are infused with chemicals that preserve nucleic acids but inactivate viruses, enabling researchers to collect them safely. The team report their new approach in Proceedings of the National Academy of Sciences1.

Honey trap

The researchers infected mosquitoes in the lab with West Nile, Ross River or chikungunya viruses. About 10 days later, they captured more than 90 mosquitoes, placed each one in a separate vial and allowed them to feed on the honey-soaked cards for 2 days. The honey contained blue dye, so that the tint in their gut would later indicate whether the insects had ingested it. The team then used a genetic test to analyse viral RNA on the cards. They found that many mosquitoes had consumed the honey, and that more than 70% of cards tested positive for the three viruses. Almost all cards that mosquitoes had fed on tested positive for the viruses they carried.

The team next tested their approach in the field. Their results showed that traps containing honey-soaked cards attracted more mosquitoes than those without cards, with more than 75% of mosquitoes consuming honey while in the traps. Each week, the team collected cards and trapped mosquitoes and shipped them to a lab, where they were tested for Ross River and Barmah Forest viruses. Viral RNA was found on the cards and in the mosquitoes that fed on the cards.


The approach is promising because it detects viruses only when mosquitoes are capable of transmitting them. Viruses in mosquito saliva can be transmitted, but those in the gut cannot infect a new host when a mosquito bites. But the usefulness of the cards may vary according to the mosquito species and the geographical region, van den Hurk says. For instance, Aedes aegypti, which spreads dengue, chikungunya and yellow fever viruses, prefers blood meals over honey. "The kinds of mosquitoes they trapped with this method are not necessarily the most important vectors for some viruses," says Scott Weaver, who studies virus–mosquito interactions at the University of Texas Medical Branch in Galveston.

The method does not indicate which species, or how many mosquitoes, deposited viruses on the cards. As a result, it would be nearly impossible to quantify the risk of infection on the basis of the amount of viral RNA on the cards, says Phil Lounibos, a medical entomologist at the University of Florida in Vero Beach. "It would be more valuable for the quick and dirty detection of viruses," he says. Once a virus is spotted, he adds, scientists could then use more comprehensive analyses to determine whether the mosquito preys on humans, and to calculate the infection rate in mosquitoes and assess the potential threat.

Next, van den Hurk will compare the sensitivity of the approach with those of other standard methods, such as the use of animals. "At this stage, we really don't want to say that one system is better than another," he says. 

Commenting is now closed.