Pigs and mosquitoes were sampled at 4 sites. One site was a commercial farm in Xmatkuil, a suburb 16 km south of Merida, the largest city in Yucatan. The site contained a herd of Yucatan black hairless pigs and a commercial genetic line of breeding pigs. The other sites were Mayan villages to the east and southeast of Merida: Tzucacab (148 km southeast), Valladolid (159 km east), and Xkalakdzonot (155 km southeast). Each village maintained herds of Yucatan black hairless pigs as a food source for residents. We visited each site 1–3 times during 2018 and 2019, and no pigs were sampled more than once. An unusually high number of porcine fetal deaths occurred in Xmatkuil and Xkalakdzonot several weeks before our initial visits. The stillborn pigs displayed signs of mummification but no apparent neurologic malformations, according to their owners. During each visit, we searched human-made structures and vegetation for resting mosquitoes, which were collected by manual aspiration.

Serum samples were assayed by plaque-reduction neutralization test (PRNT) using dengue virus (DENV) serotype 1 (strain Hawaii), DENV serotype 2 (strain NGC), DENV serotype 3 (strain H-87), DENV serotype 4 (strain 241), Ilheus virus (original strain), St. Louis encephalitis virus (strain TBH-28), West Nile virus (strain NY99–35261–11), and Zika virus (strain PRVABC59). Serum specimens were initially screened at a dilution of 1:20 by using Zika virus. Positive samples were further diluted, then assayed using all 8 viruses. Titers were expressed as the reciprocal of serum dilutions yielding >90% reduction in the number of plaques (PRNT₉₀). For etiologic diagnosis, the PRNT₉₀ antibody titer to the respective virus was required to be >4-fold that of other flaviviruses tested.

Mosquitoes were transported alive to the arbovirus laboratory at the Universidad Autonoma de Yucatan and sorted into pools of <50 according to species, sex, date, study site, and location within the study site. Mosquitoes were transported in RNAlater (Sigma-Aldrich, https://www.sigmaaldrich.com) to Iowa State University, then homogenized by using mortars and pestles. Total RNA was extracted by using Trizol Reagent (ThermoFisher Scientific, https://www.thermofisher.com) and tested for Zika virus RNA by using reverse transcription PCR and Sanger sequencing using primers that amplify a 667-nt region of the envelope protein gene.

Serum specimens were collected from 297 pigs (20 from Tzucacab, 73 from Valladolid, 74 from Xkalakdzonot, and 130 from Xmatkuil). Thirty-eight (12.8%) pigs were positive for flavivirus-specific antibodies. Thirteen (4.8%) pigs were seropositive for Zika virus, 1 (0.3%) pig was seropositive for West Nile virus, and 24 (8.1%) pigs had antibodies to an undetermined flavivirus. Zika virus PRNT₉₀ titers ranged from 40 to 320 (Table 1). Eleven pigs seropositive for Zika virus were from Xmatkuil, and 1 each was from Tzucacab and Valladolid.

The entomologic investigation yielded 1,870 mosquitoes of 8 species that were sorted into 190 pools. Of these, 381 mosquitoes were collected inside pigpens, and >50% were engorged (Table 2). Mosquitoes were tested for Zika virus RNA by reverse transcription PCR, and resulting amplification products were analyzed by Sanger sequencing. Five pools, all of which contained >1 engorged mosquito, were positive for Zika virus sequence, and all consisted of mosquitoes collected inside pigpens in Xmatkuil (Genbank accession nos. MT309004–309008). One pool each of the following mosquito species tested positive: Aedes aegypti, Ae. taeniorhynchus, Culex lactator, Cx. nigripalpus, and Cx. thriambus. All sequences were identical and differed from the positive control, an isolate from the state of Chiapas, Mexico, in 2016 (Genbank accession no. KX446950.2) in 1 nucleotide position, a C→T substitution at genomic position 1893.

We detected Zika virus RNA sequence in Ae. aegypti, Ae. taeniorhynchus, Cx. lactator, Cx. nigripalpus, and Cx. thriambus mosquitoes that were temporally and spatially associated with pigs seropositive for this virus. The role of Culex spp. mosquitoes in Zika virus transmission has been debated, but the consensus among the arbovirus community is that they are inefficient vectors (5,6). Culex spp. mosquitoes and Zika virus were first linked after experimental infection studies demonstrated that the Cx. quinquefasciatus mosquito is a competent vector of this virus (7). Many other studies have shown otherwise, including a study that demonstrated that Cx. quinquefasciatus mosquitoes in the state of Jalisco, Mexico, were refractory to Zika virus (5,6,8).

Vector competence experiments have also evaluated mosquitoes from >6 other Culex spp., although Cx. lactator, Cx. nigripalpus, and Cx. thriambus mosquitoes are not among them, and none were able to transmit Zika virus (9). We add to the small number of studies that have detected Zika virus nucleic acid in field-collected Culex spp. mosquitoes (7,10), but we did not isolate virus or provide evidence of a disseminated infection. We cannot dismiss the possibility that the Zika virus RNA–positive Culex spp. mosquitoes had recently fed upon a viremic host but virus replication had not occurred within the mosquito. Therefore, the link between Culex spp. mosquitoes and Zika virus remains tenuous. The Ae. taeniorhynchus mosquito is also considered an inefficient vector of Zika virus (11). In contrast, the Ae. aegypti mosquito is the principal urban vector of Zika virus in the Americas (12). Ae. taeniorhynchus and Ae. aegypti mosquitoes are not known to have a strong preference for porcine blood, although 2.4% of engorged Ae. taeniorhynchus mosquitoes in the Galapagos Islands had acquired blood from pigs, and the Cx. nigripalpus mosquito shifts seasonally to opportunistic feeding behavior (13,14). Porcine blood has occasionally been detected in Ae. aegypti mosquitoes (15,16).

The mosquito infection rates in our study are high. All Zika virus RNA–positive mosquitoes and most seropositive pigs were sampled at the same site (Xmatkuil) on the same date (June 5, 2018). We speculate that these pigs were infected with Zika virus just before our visit and that some mosquitoes then bit them, without virus disseminating from the midguts of Culex spp. mosquitoes. Recent studies have demonstrated that pigs are susceptible to experimental Zika virus infection (1–4). We provide serologic evidence that pigs are also susceptible to natural Zika virus infection. A high number of stillbirths occurred at 2 study sites before sampling, but none displayed malformations typical of Zika virus infection.

We provide additional evidence that pigs produce neutralizing antibodies upon Zika virus exposure and are potential sentinels. This information will be useful for investigators and public and veterinary health personnel conducting surveillance in Zika virus–endemic areas where pigs are common and usually raised outdoors. One limitation of our study is that pig farmers were not tested for evidence of flavivirus infection. Future studies should investigate whether those persons are at increased risk for Zika disease.