Chikungunya virus (CHIKV) is an arthropod-borne virus primarily transmitted by the peridomestic mosquito, Aedes (Ae.) aegypti
]. It causes a febrile illness accompanied by arthralgia of the joints, with occasional chronic arthralgia after virus clearance [2
]. Unfortunately, there is no vaccine against CHIKV, leaving naïve human populations at risk of an epidemic [2
]. Indeed, a recent outbreak in the Caribbean and the Americas resulted in 2.6 million confirmed cases, causing significant strain to both the healthcare system and economy in affected countries [3
]. Due to the nature of CHIKV’s error-prone RNA-dependent RNA polymerase (2–4 mutations/104
nucleotides in vivo and 600–1300 mutations/104
nucleotides in vitro [5
]) CHIKV has a high mutation rate, enabling it to sample many genotypes and adapt to novel environments. A prime example of CHIKV’s adaptability is the 2005 outbreak on La Réunion Island where over 300,000 people fell ill [4
]. The success of this particular CHIKV strain was pinpointed to a single amino acid substitution that allowed for more efficient transmission by Ae. albopictus
, yet did not affect transmission by Ae. aegypti
, the original primary vector [7
As CHIKV continues to extend beyond its endemic region, it is particularly disconcerting that certain CHIKV strains are transmitted by both tropical Ae. aegypti
and temperate-tolerant Ae. albopictus
. This expands the potential range of CHIKV beyond the equatorial region. Accordingly, a number of studies have been published attempting to characterize the vector competence of Ae. aegypti
and Ae. albopictus
mosquitoes in regions at risk of a CHIKV introduction [10
]. Most populations of Ae. albopictus
and Ae. aegypti
tested from Europe, South America, and the United States (USA) have readily detectable CHIKV virions in saliva after feeding on an infectious blood meal. However, there exist both species level (between Ae. aegypti
and Ae. albopictus
) and population level differences for overall infectability, as well as degree and ease of dissemination throughout the mosquito body to the salivary gland [10
]. Remarkably, even Aedes
populations within the same city can differ in their transmission efficiencies of CHIKV [11
], suggesting differences in either vector genetics or ecology across relatively short geographical distances.
The primary goal of the aforementioned studies is to describe the risk of transmission to humans. The metrics used are transmission rate and efficiency, which are a culmination of the virus’ traversal through the mosquito body. Both the midgut and the salivary gland present barriers the virus must overcome to successfully enter the saliva for transmission to mammals [18
]. These barriers contain numerous obstacles (e.g., antiviral proteins, bacteria) that contribute to their ability to prevent or promote infection, which may differ between Aedes
populations. For example, a recent predictive model based on empirical data shows that the number of viral particles needed for successful dissemination and eventual transmission varies greatly between mosquito populations found in Europe and China [10
]. This would result in significant differences in ease of viral transmission in one region of the world versus another. One study focusing specifically on the salivary gland found differences at the level of the salivary gland exit barrier between Ae. aegypti
and different U.S.A populations of Ae. albopictus
]. Interestingly, for certain CHIKV strains where the salivary gland exit barrier was greatest in Ae. aegypti
, the opposite was found in Ae. albopictus
, where the salivary exit was not as great a barrier [14
]. Together, these studies are evidence that the midgut and salivary gland can act as potent blocks, and categorizing an entire species as competent for CHIKV is not necessarily accurate. However, the genetic, anatomical, and microbial differences between mosquito populations are not completely understood. Elucidating this gap in knowledge would significantly advance our understanding of how specific mosquito populations control viral infections.
Therefore, it is essential to test specific mosquito populations for competency to assess the potential ease of spread of arboviruses in a non-endemic area, as well as to probe differences between mosquito populations [20
]. Because New York City is a hub through which many people travel, there is the possibility of disease spread from distant locations. Here, we set out to test whether Ae. albopictus
populations in New York City are competent for CHIKV in a laboratory setting. We also probe if these populations exhibit differences in the D7 long form salivary protein, a host factor known to influence viral infections [27
]. In addition, as evidenced by the La Réunion outbreak, if novel viral genotypes emerge that are potentially advantageous, this can create a foothold from which an epidemic can be initiated. It can be argued that observing which genotypes emerge and are potentially transmitted to human hosts is an important aspect of vector capacity. However, there are few studies that have surveyed the CHIKV genotypes found in mosquito saliva as part of competence characterization. Here, we extracted RNA from mosquito saliva and Sanger sequenced the structural region of CHIKV to determine which mutations arose during viral replication in the vector. We found that indeed mutants arose with known virulence phenotypes in the saliva of local New York City Ae. albopictus
. This is highly relevant for predicting which viruses will emerge in a non-endemic region.
2. Materials and Methods
2.1. Cell Culture and Generation of Viral Stocks
BHK-21 cells (ATCC CCL-10) were maintained in Dulbecco’s Modified Eagle Media (DMEM; Corning, Corning, NY, USA) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals, Minneapolis, MN, USA), 1% nonessential amino acids (NEAA; Corning), and 1% penicillin/streptomycin (P/S; Corning) at 37 °C with 5% CO2. Vero cells (ATCC CCL-81) were maintained in DMEM supplemented with 10% newborn calf serum (NBCS) and 1% P/S at 37 °C with 5% CO2. All cell lines were confirmed to be mycoplasma free (Lookout PCR Detection Kit; Sigma-Aldrich, St. Louis, MO, USA).
The chikungunya virus (CHIKV) strain 06-409 (AM258994) was rescued from the CHIKV infectious clone [5
]. Briefly, 10 μg of plasmid encoding the entire CHIKV genome was linearized using the restriction enzyme NotI. The linearized product was phenol-chloroform extracted, ethanol precipitated, and used as the template for in vitro transcription (SP6 mMessage mMachine, Ambion, Austin, TX, USA). After in vitro transcription, these nucleic acids were extracted using phenol–chloroform followed by ethanol precipitation. Subsequently, 10 μg of purified RNA was mixed with 3.9 × 106
BHK-21 cells in a 2 mm electroporation cuvette and electroporated using 1 pulse at 1.2 kV, 25 mF and infinite resistance. Cells were transferred into 6 mL of complete media (DMEM with 10% FBS and 1% NEAA) and placed in a T25 flask at 37 °C for 72 h. The passage “zero” (P0) virus was collected after 72 h and clarified via centrifugation at 1200× g
for 5 min. This was then used to infect a monolayer of BHK-21 cells (MOI ~ 0.01), and virus containing supernatants were collected 24 h later to generate the P1 stocks used in this study. Virus stocks were aliquoted and frozen at −80 °C, and viral titers were determined by plaque assay as described below.
eggs were collected from three locations in Queens, New York City (Figure 1
) (132nd St, Whitestone, Queens: 40.78521, −73.83621; Juniper Valley Park, Middle Village, Queens: 40.72043, −73.87437; Powell’s Cove Blvd, College Point, Queens (Tallman Island Wastewater Treatment Plant): 40.792268, −73.83826) by the New York Department of Health.
(Poza Rica, Mexico; F20 plus) eggs were obtained from Dr. Gregory Ebel (Colorado State University) [29
]. Mosquitoes were hatched and reared at 28 °C with 70% humidity and a 12 h diurnal light cycle in a climate controlled chamber (Memmert HPP750). Ae. albopictus
mosquitoes used for infection did not exceed generation F8.
2.3. Mosquito Infections
Prior to infection with chikungunya virus, females from the three Ae. albopictus
populations (Juniper, Tallman Island, and 132nd St, New York City, NY, USA) and Ae. aegypti
(Poza Rica, Mexico) were sorted into pint cups and starved for 8 to 12 h. These mosquitoes were then exposed to an artificial blood meal containing between 1 and 4 × 106
(high dose infection) infectious viral particles or containing 2 × 105
and 7.5 × 105
(low dose infection) infectious viral particles/mL diluted in PBS-washed sheep blood (Fisher Scientific, Waltham, MA, USA) supplemented with 5 mM ATP using a hemotek membrane feeding system for one hour. After feeding, engorged females were sorted into pint cups and incubated at 28 °C and 70% humidity with 10% sucrose ad libitum. Mosquitoes were collected and dissected at end time points—7 and 14 days post infection. At designated endpoints, mosquito legs and wings were removed and placed in a 2 mL roundbottom tube filled with 300 μL PBS and a 5 mm stainless steel ball (Qiagen, Germantown, MD, USA). Mosquitoes were then salivated by placing their proboscis into a 200 μL pipette tip filled with 5 μL of FBS. After one hour of salivation, the FBS was diluted in 45 μL of DMEM. The bodies were collected and placed in 300 μL of PBS in a 2 mL tube with a single 5 mm stainless steel ball. Legs and wings and bodies were homogenized with a Tissue Lyser II (Qiagen), and clarified by centrifugation at 8000 rpm for 8 min. Viral titers for saliva, bodies, and legs and wings were quantified by plaque assay (see below). Numbers of mosquitoes used in each replicate from each population can be found in Table S3 in the Supplementary Materials
Infection rate was calculated as the total mosquitoes from which we detected virus using a plaque assay out of those that were engorged post feeding. Similarly, the dissemination rate was calculated as the total number of mosquitoes with detectable virus in the legs and wings of those mosquitoes that had virus-positive bodies. Finally, the transmission rate was calculated as the total number of mosquitoes with detectable viral particles in their saliva (using force salivation) out of those that had virus in their legs and wings.
2.4. Mouse Transmission Studies
Five to six week old male and female C57BL/6J were bred and reared in-house. Animal experiments were performed in accordance with all NYU School of Medicine Institutional Animal Care and Use Committee guidelines (IACUC). All mouse studies were performed using biosafety level 3 conditions. Ae. albopictus
mosquitoes from Tallman (F7) were exposed to an artificial blood meal containing 1 × 106
CHIKV infectious particles/mL [30
]. Mosquitoes exposed to non-infectious bloodmeals were used as non-infected controls. After feeding, engorged females were sorted into pint cups and incubated at 28 °C and 70% humidity with 10% sucrose ad libitum. Seven and 11 days after the blood meal, mosquitoes were food-deprived for 12 h. Then, 5 to 6 week old C57BL6 mice were immobilized over a mesh covered pint cup containing the previously exposed to virus or exposed to blood only mosquitoes, and mosquitoes were allowed to feed for 40 min [31
]. Each mouse was exposed to 1 to 5 mosquitoes. Afterwards, mice were returned to their cages, mosquitoes were killed and homogenized, and viral titers were determined by plaque assay. Mice were sacrificed at 2 and 3 days post transmission. Mice were euthanized by CO2
inhalation, and calf and quadriceps muscles were collected. Muscles were placed in a round-bottomed 2 mL tube containing 500 µL of PBS and two 5 mm stainless steel beads (Qiagen). Tissues were homogenized with a Tissue-Lyser II (Qiagen), and debris was pulled down through centrifugation at 8000 rpm for 8 min. Viral titers in tissue homogenates were determined by RT-qPCR.
2.5. Plaque Assay
Clarified virus containing supernatants were applied to a monolayer of Vero cells at 10-fold dilutions in order to determine plaque-forming units per milliliter (PFU/mL). Briefly, media was removed from cells, and virus diluted in DMEM was placed on the Vero monolayer for one hour at 37 °C and 5% CO2. Post incubation, cells were overlaid with 0.8% agarose in DMEM with 2% NBCS and incubated at 37 °C and 5% CO2 for 72 h. Cells were then fixed with 4% formalin, agarose plugs were removed, and wells were stained with crystal violet to quantify PFU/mL. Infectious titers were determined from the lowest dilution where plaques could confidently be counted.
2.6. RNA Extraction and RT-qPCR
RNA extractions were performed using TRIzolTM
reagent (Invitrogen, 250 μL of clarified tissue homogenate was added to 500 μL of TRIzolTM
) following the manufacturer’s guidelines. The number of viral genomes/mL was quantified by RT-qPCR using the Taqman RNA-to-CT One-Step RT-PCR kit (Applied BiosystemsTM
, Beverly, MA, USA) and CHIKV primers to amplify a small region of nsp4 (primers in Table S1, Supplementary Materials
) and a probe (5′-(6-carboxyfluorescein)-AGGTACGCGCTTCAAGTTCGGCG-(black-holequencher)-3′) targeting an amplicon in nonstructural protein 4 (nsP4) [5
]. A standard curve was generated for each dataset using in vitro transcribed CHIKV RNAs.
2.7. Filter Paper Assay
Engorged female mosquitoes were individually housed in 50 mL conical tubes post blood feed [33
]. Each individual mosquito was provided a 0.5 cm2
filter paper square soaked with 10% sucrose on which to feed. Filter papers were collected every 24 h for 4 days and placed directly into TRIzolTM
reagent. A new sucrose soaked filter paper square was provided each day. Filter papers were vortexed in TRIzolTM
reagent for approximately 30 s, and RNA was extracted following the manufacturer’s instructions and re-solubilized in water. This RNA was used as a template for cDNA synthesis (Maxima H Minus First Strand cDNA synthesis kit; Thermo Scientific, Waltham, MA, USA). cDNA served as a template for PCR (Phusion HF, Thermo Scientific) to amplify a 500 bp fragment of the 18S gene as well as a 500 bp fragment of the CHIKV E1 glycoprotein (primer sequences can be found in Table S1 in the Supplementary Materials
). Individual PCR fragments were mixed with 6x DNA loading dye, separated on a 1% agarose-TAE (Tris, acetic acid, EDTA) gel, stained with ethidium bromide, and visualized on a Bio-Rad gel doc system.
2.8. Mosquito Saliva Sequencing
Mosquito saliva was collected as described above (Mosquito infection—force salivation). The DMEM plus saliva mixture was mixed with TRIzolTM
reagent, RNA extracted, and cDNA synthesized as described above (filter paper assay). We amplified the structural region of the CHIKV genome using two overlapping PCR products (Fragment 1 and Fragment 2—primer sequences found in Table S1, Supplementary Materials
). Resulting PCR products of the correct size were purified using a PCR cleanup kit (Macherey–Nagel, Bethlehem, PA, USA) and Sanger sequenced (Genewiz, South Plainfield, NJ, USA) (sequencing primer sequences can be found in Table S1, Supplementary Materials
2.9. D7 Sequencing from Mosquito Salivary Glands
Mosquito salivary glands were extracted from individual mosquitos and placed in TRIzolTM
reagent. RNA was extracted and used for cDNA synthesis as described above. The D7 long form transcript was amplified by PCR and Sanger sequenced (Genewiz) using the same primers (primer sequences can be found in Table S1 in the Supplementary Materials
2.10. Virus and D7 Sequence Analysis
Virus Sanger sequencing results were aligned to the reference CHIKV sequence, strain 06-409. A mutation was considered real and significant if the peak of the non-WT nucleotide was at least half the amplitude of the original nucleotide. D7 Sanger sequencing results were aligned to the Ae. albopictus
reference AALF024477 (VectorBase; www.vectorbase.org
]. Aedes albopictus
FOSHAN.AaloF1.2). A single nucleotide polymorphism (SNP) was called when the non-reference peak either replaced the reference peak or was approximately 50% of the reference peak (heterozygote).
2.11. Protein Structure Analysis
The CHIKV E1 protein (PDB: 3J2W) structure was analyzed using PyMOL (version 2.2.2).
2.12. Linear Modeling
The lm function in R (versions 3.3.3 and 3.6.3) was used to fit a linear regression to the mosquito infection data. Titers from body infection and saliva infection were all compared to one another to visualize and determine if a significant correlation existed. The with function was used to determine the direction of the correlation, i.e., negative or positive.
2.13. Data Availability
D7 sequences have been deposited in GenBank with accession numbers MT353980–MT354023.
2.14. Data analysis and Statistics
All statistical analysis and data visualization and editing were done in either GraphPad Prism (version 7.0b) or R Studio (version 1.2.5001). Agarose gels were analyzed using Image Lab and Photoshop. All experiments were completed in at least two independent biological replicates or using multiple individual mosquitoes. The specific statistical test and experimental N can be found in the figure legends. Tests for normal distribution (D’Agostino and Pearson, Shapiro–Wilk and Kolmogorov–Smirnov) were applied prior to choosing a statistical test to compare means. This influenced whether a parametric or non-parametric test was chosen.
Chikungunya virus (CHIKV) is an emerging virus, whose range has rapidly expanded over the past two decades. In 2013 the Asian lineage of CHIKV reached the Caribbean, where it had not previously sustained autochthonous transmission, and spread to Central and South America resulting in 2.6 million reported cases of disease [3
]. CHIKV’s rapid dispersal necessitates a competent vector. Competence describes the ability of a vector to become infected, maintain, and potentially transmit the infectious agent [47
]. In addition to competence, a number of other intrinsic (daily blood feeding rate, the extrinsic incubation period for the virus, and probability of survival [18
]) and extrinsic (environmental [25
] and microbial [49
]) qualities affect the capacity of the vector to launch an epidemic. Previous work has shown that Aedes
populations around the globe are relatively susceptible to numerous non-endemic arboviruses [47
]. However, empirically derived threshold values describing the number of viral particles required for dissemination in mosquitoes have revealed differences between European and Asian mosquitoes, with some populations requiring a lot more virus particles for transmission [10
]. In addition, some studies have shown that dynamics of infection, i.e., how quickly CHIKV appears in saliva across a 10 to 14 day period, varies [16
]. Therefore, we set out to specifically study the degree of infection, infection rate, dissemination rate, and transmission rate of three populations of Ae. albopictus
found in Queens, New York City, infected with the Indian Ocean Lineage strain of CHIKV.
To both characterize the salivary microenvironment as well as determine potential genetic divergence at an antiviral locus in these populations, we sequenced the D7 long form salivary messenger RNA transcript. We found that each population harbors unique genetic variants. Further sampling would be required to determine the absolute haplotype networks of these mosquito populations, and mechanistic validation is required to determine whether these variants function differentially with respect to viral inhibition. However, these initial studies highlight the genetic differences amongst individuals within and between populations of a single mosquito species residing in one city.
We found significant differences in CHIKV titers in both bodies and legs and wings between populations (Figure 3
). These data support the hypothesis that these populations are likely distinct, and that there are unique obstacles to the number of disseminated viral particles in individuals from discrete locations. In general, we found that all three populations of Ae. albopictus
isolated from Queens, New York City are susceptible to the Indian Ocean Lineage strain of CHIKV (Figure 3
, Figure 4
and Figure 5
). In addition, NYC Ae. albopictus
continue to harbor virus 14 days after infection, in-line with previous work showing that Aedes
species are able to sustain CHIKV replication for weeks after the onset of the infection [31
]. Yet, unlike previous work where titers in bodies and legs and wings were maintained at 14 days, we find titers drop by a log after 14 days compared to 7 days post infection. This may suggest certain barriers to viral replication or an eventual clearing of virus over time, which is common to all three populations. It is possible that experimental setup influenced our results, leading to lower rates of infection and therefore dissemination. However, replicate infections were done at similar times of day, weeks apart, and with independently reared batches of mosquito eggs, and yet still resulted in overlapping titers (Figure 3
). It should be noted that infection and dissemination rates remained stable and above 72% 14 days after infection (Figure 3
), albeit a drop in titers. It is possible NYC Ae. albopictus
are better able to control levels of infection, yet remain productively infected.
Importantly, for competence, virus must enter the saliva in order to be transmitted. We used a non-invasive filter paper method that allowed us to monitor viral RNA in individual mosquito saliva over a 4-day period. We observed transmission of viral RNA as early as two days post infection for two of the NYC Ae. albopictus
populations, suggesting that these mosquitoes potentially transmit virus early after infection. In order to directly determine possible transmission rates, we used two methods: (i) salivation into a pipette tip (force salivation) and (ii) an in vivo transmission model. Using the pipette tip method, we calculated transmission rates between 3 and 10% for 7 days post infection, which while low, would still constitute a large number of mosquitoes given their population size (Figure 3
). Importantly, we also evaluated transmission rates using an in vivo transmission model where infected mosquitoes were able to feed on mice. We found Ae. albopictus
from Tallman island could transmit virus to mice at 7 and 11 days post infection, with a transmission rate ranging around 60 to 80% (Figure 5
). This discrepancy between methods is likely technical due to the fact that (1) the pipette method does not allow for us to assess whether the mosquito has salivated into the pipette tip used for collection, and (2) mice were fed on by multiple mosquitoes. Additionally, each method is quantifying virus at different points in a transmission cycle. When using force salivation we are attempting to assess the initial inoculum that would be transmitted to a mammal. In the mouse transmission experiment, the initial inoculum is transmitted and the infected host tissue collected 2 or 3 days later, which assesses whether the virus successfully replicated in the host target organ. Taken together, these transmission rates are lower than those reported in studies elsewhere that used a force salivation method for saliva collection. Those studies report a range between 30% and 80% [11
]. Accordingly, this may suggest NYC Ae. albopictus
populations have a more robust salivary gland exit barrier or a saliva specific biochemical barrier. Finally, our data did not suggest an association between high body titers or leg and wing titers and presence of virus in the saliva (Figure S2
). Again, this potentially supports a barrier to transmission at the level of the salivary gland. No specific molecules have been described as salivary gland exit barriers for CHIKV in Ae. albopictus.
Studies of Ae. aegypti
salivary transcripts have identified specific transcripts expressed in separate lobes of the salivary gland, potentially influencing where virus can replicate and is found [51
]. Additionally, changes in protein expression occur in the salivary gland of Ae. aegypti
post CHIKV infection, suggesting a role for specific proteins [53
]. Further work is necessary to pinpoint which proteins influence viral infection, their mechanisms, and whether they impact viral selection.
Another important aspect of emerging virus dynamics is the potential to adapt to a new vector, or acquire novel traits that aid in transmission. Here, we find the emergence of unique viral variants in saliva after 7 days of replication in NYC mosquitoes (Figure 7
, Table 1
). Previous work characterizing adaptive mutations in the E1 glycoprotein has shown that mutations at sites V80 and A129 can result in increased virulence and transmission efficiency in mice [30
]. These mutations arose during experimental evolution in the laboratory, similar to the setting in which we conducted this study. Interestingly, we find the same exact substitutions at site E1-80, a valine to an isoleucine, in mosquitoes from each population (Figure 7
B). In addition, we find substitutions at site E1-129 and nearby at E1-162 (Figure 7
A). This suggests that these substitutions arise in viral populations replicating in Aedes
species regardless of location, potential differences in host genetics, and microbiota. Given that we sampled only a few mosquitoes, it is remarkable that these evolutionary events are recapitulated. And while virulent phenotypes have been associated with mutations at sites E1-80 and 129, it is still unknown why they emerge in saliva. Therefore, NYC Ae. albopictus
can potentially serve as a model to study the mechanism driving the emergence of these CHIKV variants. Taken together, these studies emphasize the need to proactively study arbovirus infections in naïve mosquito populations. This allows us to better understand how distinct mosquito populations control arbovirus infections, and how arboviruses may evolve when introduced to a novel host during an epidemic.