Science Wiki

Dengue virus (DENV; species Orthoflavivirus denguei) is an enveloped, positive-sense, single-stranded RNA virus of the genus Orthoflavivirus, family Flaviviridae, and the cause of dengue fever and severe dengue. It is transmitted between vertebrates by female Aedes mosquitoes, principally Aedes aegypti and Aedes albopictus. Four antigenically distinct types, DENV-1 to DENV-4, circulate in human populations, alongside ancestral sylvatic lineages in forest cycles. A fifth type was proposed in 2013 on the basis of a single isolate from Sarawak, Malaysia, but this has not been accepted and the variant is best regarded as part of a sylvatic lineage.

Dengue virus is the most widespread and fastest-spreading arbovirus in the world. Modelling and serological surveys estimate around 390 million infections each year, of which about a quarter cause illness, across more than 130 countries, with roughly half the world's population living where transmission is possible.

At the molecular level the virus is a well-studied model for class II membrane fusion, temperature-dependent changes in particle structure, long-range folding of the RNA genome, and antibody-dependent enhancement. Because replication in the mosquito depends directly on temperature, with transmission peaking near 29 °C in Aedes aegypti, the virus's range and seasonality are closely tied to climate, to El Niño, to the warming of cities, and to climate change.

Taxonomic Classification and Sylvatic Origins

The International Committee on Taxonomy of Viruses places dengue virus in the realm Riboviria, order Amarillovirales, family Flaviviridae, genus Orthoflavivirus, species Orthoflavivirus denguei. Within the genus, the four dengue types form a group distinct from those containing Japanese encephalitis virus and West Nile virus, Zika virus, and yellow fever virus.

Dengue virus did not originate in humans or in urban mosquitoes. Molecular dating and field isolations, beginning with Albert Rudnick's work in Malaysia in the 1960s and extended by Nikos Vasilakis, Scott Weaver, Edward Holmes, and others, indicate that the ancestor of all dengue viruses circulated in forest canopies in Southeast Asia. Estimates of the age of this ancestor vary and are sensitive to the assumptions used.

Two sylvatic cycles are known:

• In Asia, the virus circulates in Malaysia, Borneo, Indonesia, and Vietnam between wild macaques and leaf monkeys and canopy-dwelling mosquitoes of the Aedes niveus group. Sylvatic lineages of all four types have been found or inferred in this cycle.
• In West Africa, a separate cycle involving only DENV-2 occurs in gallery forests from Senegal to Nigeria, between African green monkeys, patas monkeys, and baboons and canopy mosquitoes including Aedes furcifer, Ae. luteocephalus, and Ae. taylori. Whether this represents an introduction from Asia or an older divergence is unsettled.

The four human types did not arise from one another after entering people. Each descends from a separate transfer out of a forest cycle into rural human populations in Southeast Asia. The dates of these transfers differ between types and remain uncertain, though all are more recent than the divergence of the types themselves. Transmission was initially maintained by Aedes albopictus in villages. Only after Aedes aegypti spread through Asian ports in the nineteenth century did the modern urban cycle become established.

Purifying Selection and Genotype Competition

The viral polymerase lacks proofreading, so replication is error-prone, producing a population of closely related variants within each infected host rather than a single sequence.

Despite this, dengue virus changes more slowly over time than viruses with a single host, such as influenza. The constraint comes from having to replicate in two very different hosts: a warm-blooded mammal with antibodies, T cells, and interferons, and a cold-blooded insect with RNA interference and different membrane lipids. A mutation that helps in human cells often harms replication in the mosquito, so most amino acid changes are removed by selection. Estimates of the strength of this constraint vary between genes.

The four types share roughly 65 to 70% of their amino acid sequence. Within each type, strains differ by up to a few per cent and are grouped into genotypes, which differ in how readily they spread and in the severity of disease they cause:

Genotypes of the four dengue virus types

The displacement of the American DENV-2 genotype by an Asian one across the Americas after 1981 illustrates how genotypes compete. Work by Rebeca Rico-Hesse, Eva Harris, Mariano Garcia-Blanco, and colleagues identified three advantages of the incoming lineage, acting in both hosts:

1. A change in the envelope protein. A single amino acid difference at position 390 of the E protein is associated with substantially better replication in human immune cells, raising the level of virus in the blood and so the chance of infecting a feeding mosquito.
2. Better establishment in the mosquito. In feeding experiments, the Asian genotype infected a higher proportion of mosquito guts and reached the salivary glands two to four days sooner. Because a mosquito must survive this interval to transmit, a shorter delay has a disproportionate effect.
3. Changes in the untranslated region. Mutations in the structured region at the end of the genome increase production of a non-coding RNA fragment that suppresses both human interferon responses and mosquito RNA interference. A similar change accompanied a later replacement event in Puerto Rico.

Pseudo-Icosahedral Envelope Architecture

A mature particle is about 50 nm across: a core of RNA and capsid protein, a lipid bilayer taken from the host cell, and an outer protein shell.

The shell contains 180 copies each of the envelope (E) and membrane (M) proteins. Rather than forming the projecting pentamers and hexamers of a conventional icosahedral shell, the E proteins lie flat against the membrane as 90 pairs, arranged in groups of three to form flat rafts, thirty of which tile the surface. The arrangement is usually described as pseudo-icosahedral for this reason.

Each E protein consists of three domains joined to a membrane anchor:

• Domain I is the central organizing region, carrying a sugar group that affects particle stability and the pH at which fusion occurs.
• Domain II is elongated and holds the pair together. At its tip lies the fusion loop, the hydrophobic segment that inserts into the host membrane during entry; in the mature pair it is tucked into a pocket on the partner protein. This domain also carries a sugar group recognized by the human receptor DC-SIGN.
• Domain III has an immunoglobulin-like fold, binds host receptors, and carries the targets of the most potent type-specific neutralizing antibodies.

Furin Cleavage and Immature Particle Maturation

Newly formed particles bud into the endoplasmic reticulum as larger, spiky structures about 60 nm across, in which the E proteins are paired with a precursor protein, prM, and project outward as sixty trimeric spikes. A domain of prM sits over the fusion loop, preventing the particle from fusing with the cell's own membranes on its way out.

As the particle passes through the trans-Golgi network, the compartment becomes acidic. The spikes collapse and the proteins fold flat against the membrane, exposing a cleavage site on prM to the host enzyme furin. Furin cuts prM into a soluble fragment and the membrane-anchored M protein. While the particle remains in the acidic compartment, the cut fragment stays attached, still covering the fusion loop. Only on release into the neutral environment outside the cell does it fall away, leaving an infectious particle.

Cleavage is inefficient in dengue virus, because the sequence around the cut site is a poor match for furin's preference. Infected cells therefore release a mixture of fully mature, partly mature, and fully immature particles, in proportions that vary with cell type. On partly mature particles, the cleaved regions allow entry while the uncleaved regions present targets for antibodies against prM, which contribute to antibody-dependent enhancement.

Temperature-Dependent Virion "Breathing"

Work by Shee-Mei Lok, Michael Rossmann, and colleagues showed that the mature particle changes shape with temperature as it passes between mosquito and human:

• At mosquito temperatures, around 25 to 28 °C, the particle remains compact and smooth, with the membrane fully covered and internal regions inaccessible to antibodies.
• At human body temperature and above, the surface proteins of some strains, best documented for DENV-2, rotate and move outward, expanding the particle to roughly 55 nm and opening gaps that expose the underlying membrane. The extent varies between strains and types.
• The immunological consequence is that regions normally hidden, including the fusion loop, become briefly accessible. Antibodies that recognize the intact pair of E proteins lock the particle closed and neutralize it. Antibodies binding the transiently exposed fusion loop generally do not neutralize and can instead promote uptake into immune cells.

Untranslated Regions and Genome Cyclization

The genome serves both as a messenger RNA, read by ribosomes from one end, and as a template for the viral polymerase, which copies it from the other. A ribosome and a polymerase moving toward each other on the same molecule would collide. The virus avoids this by folding the genome into a circular form, a mechanism worked out largely by Andrea Gamarnik and colleagues.

Structured elements in the untranslated regions

Endoplasmic Reticulum Remodeling and Lipophagy

After entry and release into the cytoplasm, the genome is translated on the endoplasmic reticulum. Within hours, the viral membrane proteins, together with host proteins that shape membranes, remodel the ER into three connected structures, visualized by Ralf Bartenschlager and colleagues:

1. Convoluted membranes, dense networks containing the viral protease, apparently the main site where the viral protein is cut up.
2. Vesicle packets, spherical invaginations of the ER membrane about 80 to 90 nm across, each containing the replication machinery and the double-stranded RNA intermediate. Each connects to the cytoplasm through a narrow pore roughly 10 nm wide. This allows the building blocks of RNA to enter and finished genomes to leave, while keeping the double-stranded intermediate hidden from the host sensors that would otherwise detect it.
3. Budding sites directly opposite these pores, where emerging genomes are packaged by capsid protein and bud into the ER as immature particles.

Building thousands of such structures requires large amounts of membrane lipid and energy, which the virus obtains by redirecting host metabolism. The viral NS3 protein recruits the host enzyme that makes fatty acids to the replication sites. Separately, viral proteins trigger the breakdown of the cell's stored fat droplets, releasing fatty acids that are burned in mitochondria to supply the energy the viral helicase requires.

Comparative Biology Across Primate and Mosquito Hosts

The virus in its two hosts

Suppression of Human Type I Interferon and STAT2 Degradation

Human cells detect viral RNA through cytoplasmic sensors that signal through mitochondria, and detect leaked mitochondrial DNA through a separate pathway. Both converge on the production of type I interferon, which acts on neighbouring cells to switch on hundreds of antiviral genes.

Dengue virus interferes at several points:

1. Disguising its RNA. The viral NS5 protein adds methyl groups to the end of the genome in the same pattern as host messenger RNA, so that a host restriction factor does not recognize it as foreign.
2. Blocking RNA sensing. The non-coding RNA fragment produced from the end of the genome binds a host enzyme needed to activate the principal RNA sensor, while viral proteins interfere with the sensor's relocation to mitochondria and with its partner protein there.
3. Cutting host signalling proteins. The viral protease cleaves a protein central to DNA sensing, and also cleaves proteins involved in mitochondrial fusion.
4. Blocking interferon signalling. Several viral proteins prevent the phosphorylation steps required to switch on antiviral genes in response to interferon.
5. Destroying a human signalling protein. Work by Ashour, Morrison, Adolfo García-Sastre, and colleagues showed that NS5 binds human STAT2 and recruits a host enzyme that marks it for destruction, leaving the cell unable to respond to interferon.

This has a notable consequence. Dengue proteins act on human and other primate versions of these targets but not on the mouse versions. Ordinary mice are therefore resistant to the virus, and mouse models require either animals lacking interferon receptors or animals carrying the human genes.

Counteracting Mosquito RNA Interference via sfRNA

Mosquitoes have no antibodies or interferons. Their main defence is RNA interference, in which viral double-stranded RNA is chopped into short fragments that direct the destruction of matching viral sequences, supplemented by a second small-RNA pathway and by signalling pathways producing antimicrobial proteins.

The virus's principal countermeasure is the same non-coding RNA fragment produced from the end of its genome, which binds the enzymes of the RNA interference pathway and acts as a decoy. Infection also reduces the activity of the mosquito's antimicrobial signalling pathways, particularly in the salivary glands.

Trait-Based Thermal Performance Curves

Because the mosquito's body temperature follows the air, transmission depends on temperature through eight separate processes, combined by Mordecai and colleagues into a single expression:

where a is the biting rate, b and c the transmission probabilities in each direction, μ the mosquito death rate, PDR the rate of viral development in the mosquito, EFD the eggs laid per female per day, p_EA the survival of larvae, MDR the rate of larval development, and γ the rate at which people stop being infectious.

Temperature dependence of transmission by Aedes aegypti

Diurnal Temperature Fluctuations and Climate Change Drivers

Temperature is not constant through the day, and because these relationships are curved, the effect of a fluctuating day differs from that of a steady one at the same average. Experiments by Louis Lambrechts, Krijn Paaijmans, and Thomas Scott showed two opposite effects:

• In warm conditions, large daily swings carry the mosquito above its upper limit each afternoon, reducing infection and survival compared with the small swings typical of humid coastal cities and indoor rooms.
• In cool conditions, near or below the threshold for transmission, large swings warm the mosquito into the permissive range for several hours each afternoon, allowing transmission where the daily average alone would suggest it is impossible.

Because transmission by Aedes aegypti peaks at a higher temperature than for malaria or for Aedes albopictus, warming acts directly to increase dengue transmission:

1. Tropical cities are moving toward the optimum. Most cities in Southeast Asia, South Asia, Latin America, and Africa had baseline temperatures of 24 to 27 °C, on the rising part of the curve. Warming, combined with the two to five degrees added by urban surfaces, moves them toward the peak and roughly halves the time the virus needs to develop in the mosquito.
2. Transmission is moving upslope. As temperature falls with altitude, warming lifts the lower limit several hundred metres. Dengue now causes epidemics in cities formerly considered too cool, including Kathmandu and highland cities in the Andes and Mexico.
3. Transmission is moving toward the poles. Warmer winters and hotter summers have established transmission across central Argentina, produced recurrent local cases in Florida, Texas, Hawaii, and recently southern California, and allowed locally acquired cases in Italy, France, Spain, and Croatia, transmitted mainly by Aedes albopictus.
4. El Niño amplifies both wet and dry pathways. Warm phases raise temperatures across the tropics and shift rainfall. Heavy rain fills outdoor containers; drought interrupts piped water and drives households to store it indoors. Both increase mosquito breeding. The El Niño of 2023 to 2024 coincided with a sharp rise in cases, the Americas recording about 4.6 million in 2023 and roughly 13 million in 2024.

Molecular and Environmental Surveillance

• Identifying the circulating type. During the first days of illness, PCR testing identifies which of the four types a patient has, and antigen tests detect the secreted NS1 protein. Sequencing whole genomes, now possible on portable instruments, tracks the replacement of one type or genotype by another, often months before hospital admissions rise.
• Testing mosquitoes and wastewater. Pooled mosquitoes caught in traps can be tested for viral RNA, and viral RNA can be detected in municipal wastewater, giving warning of transmission including from the many infections that cause no symptoms.

Direct-Acting Antivirals and Host-Targeted Therapies

No antiviral drug is approved for dengue. Four targets have received most attention.

Antiviral approaches

Tetravalent Vaccine Platforms and Wolbachia Vector Control

Because antibodies against the fusion loop and against prM can enhance rather than prevent infection, vaccine design focuses on avoiding them.

• Live vaccines containing dengue non-structural proteins. The first licensed vaccine, Dengvaxia (CYD-TDV), carried dengue surface proteins (prM and E) on a yellow fever 17D backbone without dengue non-structural proteins, which left seronegative recipients at higher risk of severe dengue upon subsequent natural infection. Second-generation live-attenuated tetravalent vaccines—such as TAK-003 (Qdenga, built on an attenuated DENV-2 backbone) and the single-dose Butantan-DV (TV003/TV005)—include native dengue non-structural proteins to elicit protective CD8⁺ T-cell responses and anti-NS1 antibodies.
• Endosymbiotic Wolbachia replacement. Introgression of Wolbachia pipientis (strains _w_Mel and _w_AlbB) into urban Aedes aegypti populations blocks DENV replication through intracellular lipid competition and innate immune priming, reducing symptomatic human dengue incidence by 77% in cluster-randomized trials.

• Dengue fever
• Aedes aegypti
• Antibody-dependent enhancement
• Arbovirus
• Flaviviridae

How does dengue virus circularize its RNA genome to replicate?

Dengue virus coordinates translation and RNA replication on the same positive-sense genomic strand by alternating between linear and circular conformations. Long-range base-pairing between complementary cyclization sequences (5′ CS, UAR, and DAR) at the 5′ and 3′ ends brings the NS5 RNA-dependent RNA polymerase—which binds exclusively to Stem-Loop A at the 5′ terminus—directly to the 3′ terminus to initiate negative-strand synthesis while simultaneously repressing ribosome initiation.

Why do dengue virus-infected cells release so many immature virions?

The prM cleavage site of dengue virus contains an acidic residue adjacent to the basic furin recognition motif, making proteolytic cleavage by host furin in the trans-Golgi network (pH 6.0) structurally inefficient. Consequently, infected human cells secrete up to 30–40% partially mature and fully immature prM-bearing particles; although fully immature virions cannot fuse on their own, non-neutralizing anti-prM antibodies opsonize them into Fcγ-receptor-bearing monocytes, where endosomal furin completes maturation and renders them infectious.

Why can dengue virus NS5 degrade human STAT2 but not mouse STAT2?

The dengue virus NS5 protein functions as an interferon antagonist by binding the coiled-coil domain of primate STAT2 and recruiting the host E3 ubiquitin ligase UBR4 to target STAT2 for proteasomal degradation. Because murine STAT2 diverges in amino acid sequence within the NS5-binding interface, dengue virus cannot block type I interferon signaling in immunocompetent wild-type mice, restricting natural replication to humans and non-human primates.