Geographic Distribution of Hantaviruses Associated with Neotomine and Sigmodontine Rodents, Mexico

El Moro Canyon virus and Limestone Canyon virus are widely distributed and may cause hantavirus pulmonary syndrome.

To increase our knowledge of the geographic distribution of hantaviruses associated with neotomine or sigmodontine rodents in Mexico, we tested 876 cricetid rodents captured in 18 Mexican states (representing at least 44 species in the subfamily Neotominae and 10 species in the subfamily Sigmodontinae) for anti-hantavirus IgG. We found antibodies against hantavirus in 35 (4.0%) rodents. Nucleotide sequence data from 5 antibody-positive rodents indicated that Sin Nombre virus (the major cause of hantavirus pulmonary syndrome [HPS] in the United States) is enzootic in the Mexican states of Nuevo León, San Luis Potosí, Tamaulipas, and Veracruz. However, HPS has not been reported from these states, which suggests that in northeastern Mexico, HPS has been confused with other rapidly progressive, life-threatening respiratory diseases. Analyses of nucleotide sequence data from 19 other antibody-positive rodents indicated that El Moro Canyon virus and Limestone Canyon virus are geographically widely distributed in Mexico.
Specifi c rodents (usually 1 or 2 closely related species) are the principal hosts of the hantaviruses, for which natural host relationships have been well characterized. The current principal host relationships of some hantaviruses seem to represent a long-term association between viruses in the genus Hantavirus and rodents in the family Cricetidae. Evidence for this ancient relationship includes the association of phylogenetically closely related hantavirus species with phylogenetically closely related allopatric rodent species. For example, Catacamas virus is associated with Coues's rice rat (Oryzomys couesi) in Honduras, and Bayou virus is associated with the marsh rice rat (Oryzomys palustris) in the southeastern United States (9,15).
Our knowledge of the rodent-associated hantaviruses in Mexico includes the following fi ndings: HUIV RNA in a western harvest mouse (R. megalotis) captured in Morelos (8); CARV RNA in a Sumichrast's harvest mouse (R. sumichrasti) and MTNV RNA in an Orizaba deer mouse (P. beatae) from Guerrero (8); Playa de Oro virus RNA in a Mexican oryzomys (Oryzomys mexicanus) and Jaliscan cotton rat (S. mascotensis) from Colima (13); ELMCV RNA and SNV RNA in western harvest mice from Zacatecas (14); antibody against hantavirus in nimble-footed mice (P. levipes) captured in Tamaulipas (19); and antibody against hantavirus in a North American deer mouse (P. maniculatus), transvolcanic mice (P. hylocetes), blackeared mice (P. melanotis), and Sumichrast's harvest mouse captured in the state of Mexico (20,21). The purpose of this study was to extend our knowledge of the geographic distribution of hantaviruses associated with neotomine or sigmodontine rodents in Mexico.

Materials and Methods
Blood samples from 876 rodents, representing at least 44 species in the Neotominae and 10 species in the Sigmodontinae, were tested for anti-hantavirus IgG. The 876 rodents were captured during 1998-2008 at 43 localities in 18 states in Mexico (online Appendix Table  1, wwwnc.cdc.gov/EID/article/18/4/11-1028-TA1.htm). Blood samples from all of the rodents and lung samples from the antibody-positive rodents were acquired from the Natural Science Research Laboratory, Museum of Texas Tech University, Lubbock, Texas, USA.
The blood samples were tested for IgG to Caño Delgadito virus (CADV) strain VHV-574 by using an ELISA in which CADV can be highly cross-reactive with SNV, BCCV, and other North American hantaviruses (22). The antibody titers in the antibody-positive blood samples were recorded as 320, 1,280, or >5,120.
Samples of lung tissue from the antibody-positive rodents were tested for hantavirus nucleocapsid (N) protein gene RNA. Subsequently, we determined the nucleotide sequences of a 1,078-nt fragment of the glycoprotein precursor (GPC) genes of 11 of the hantaviruses associated with the antibody-positive rodents. We chose these 11 viruses to represent the geographic distribution and natural host associations of the hantaviruses in Mexico included in this study. Total RNA was isolated from 30 mg to 45 mg of lung tissue by using Tri Reagent (Sigma-Aldrich, St. Louis, MO, USA). First-strand cDNA was synthesized from small (S) segment and medium (M) segment RNA by using SuperScript II RNase H -Reverse Transcriptase (Invitrogen Life Technologies, Inc., Carlsbad, CA, USA) in conjunction with oligonucleotide 5′-GGTGGTTGTGGTAGTAGTAGACTCC-3′ (23). The fi rst-round and second-round (hemi-nested) PCR assays used the MasterTaq Kit (Eppendorf North America, Inc., Westbury, NY, USA). (The sequences of the oligonucleotides that were used to prime the PCR are available from the corresponding author.) The sizes of the N protein gene amplicons from the second-round assays ranged from 377 to 545 bp, the sizes of the GPC gene amplicons from the second-round assays ranged from 607 to 631 bp (M1 amplicon) and 571 to 618 bp (M2 amplicon), and the lengths of the overlaps between the M1 and M2 amplicons ranged from 125 to 134 bp. Together, the nucleotide sequences of the M1 and M2 amplicons encoded a 359-aa fragment of the G C glycoprotein.
Sequence nonidentities were equivalent to uncorrected (p) distances. The phylogenetic analyses of nucleotide sequences were conducted with MRBAYES 3.1.2 (25) and programs in the computer software package PAUP* (26). The Bayesian analyses used the general time reversible + proportion invariant + Γ model and the following options in MRBAYES 3.1.2: two simultaneous runs of 4 Markov chains, 2 million generations, and sample frequency = every 1,000th generation. The fi rst 1,000 trees were discarded after review of the likelihood scores, convergence statistics, and potential scale reduction factors; and a consensus tree (50% majority rule) was constructed from the remaining trees. Probability values in support of the clades were calculated a posteriori, and clades with probability values >0.95 were considered supported by the data (27).
Hantavirus N protein gene RNA was detected in samples of lung from 24 (68.6%) of the 35 antibodypositive rodents (Table). The Bayesian analyses of the N protein gene sequences separated the 24 Mexican viruses in this study into 4 groups (Figure 2). Group I included CARV, group 2 included HUIV, group III included LSCV and MTNV, and group IV included SNV strains Convict Creek 74, Convict Creek 107, and NM H10.
Hantavirus GPC gene RNA was detected in each of the 11 rodents assayed for GPC gene RNA (Table). The topology of the GPC gene tree (Figure 3) was essentially identical to the topology of the N protein gene tree ( Figure  2) with respect to relationships between the viruses from Mexico in this study, CARV, HUIV, MTNV, and the other hantaviruses found in North America. M0040008, CARV, M0040059, H0460041, HUIV, and ELMCV were monophyletic in the Bayesian analyses of the GPC gene sequence data ( Figure 3) and N protein gene sequence data ( Figure 2).
Nonidentities among the amino acid sequences of the 359-aa fragment of the G C glycoproteins of the 11 hantaviruses from Mexico in this study, CARV, HUIV, and MTNV ranged from 0% to 18.4% (online Appendix Table 3, wwwnc.cdc.gov/EID/article /18/4/11-1028-TA3. htm). Nonidentities between the sequences of the 359-aa fragment of the G C glycoproteins of these 14 hantaviruses and the sequences of the homologous fragment of the G C glycoproteins of the other hantaviruses found in North America ranged from 1.1% (H0030073 and SNV strain Blue River-Oklahoma) to 18.9% (M0040049 and BCCV strain SPB 9408076).

Discussion
The Eighth Report of the International Committee on Taxonomy of Viruses sets forth the criteria for species demarcation in the genus Hantavirus (1). One of these criteria is that strains of different species must exhibit at least a 7% difference in amino acid sequence identity in comparisons of complete N protein sequences and in comparisons of complete GPC sequences.
ELMCV was fi rst described in 1994 (10); LSCV was described in 2001 (11); and CARV, HUIV, and MTNV were described in 2011 (8). In a previous study (8), the amino acid sequence of the N protein of MTNV was 5.8% different from the amino acid sequence of the N protein of LSCV strain 68273, and the amino acid sequence of the N protein of CARV was 3.7% different from the amino acid sequence of the N protein of ELMCV strain RM·97 and 8.4% different from the amino acid sequence of the N protein of RIOSV strain RMx·Costa·1. The amino acid sequence of the N protein of HUIV was 1.4% different from the amino acid sequence of the N protein of ELMCV strain RM·97 and 8.4% different from the amino acid sequence of the N protein of RIOSV strain RMx·Costa·1. Accordingly, MTNV should be considered a strain of LSCV, and CARV and HUIV could be considered strains of ELMCV or RIOSV. Alternatively, CARV and HUIV could be considered members of a species complex that includes ELMCV, RIOSV, and other hantaviruses that are naturally associated with harvest mice (Reithrodontomys spp.). There is presumptive evidence for RIOSV or hantavirus(es) that are genetically closely related to RIOSV in Sumichrast's harvest mice, a Mexican harvest mouse (R. mexicanus), and a Chiriqui harvest mouse (R. creper) captured in Panama (29).
Collectively, the results of the Bayesian analyses of N protein gene sequence data (Figure 2), Bayesian analyses of the GPC gene sequence data (Figure 3), and pairwise comparisons of G C sequences (online Appendix Table  3) indicate that H0020015, H0030065, H0030073, and H0460038 are strains of SNV. The results of these analyses also indicate that M0040008, M0040059, and H0460041 are strains of ELMCV or RIOSV and that B0030008, H0460008, H0460023, and M0040049 are strains of LSCV. Accordingly, the N protein gene RNA-positive rodents in this study (Table) were infected with SNV, ELMCV, RIOSV, or LSCV.
Specifi c knowledge of the natural host relationships of LSCV previously was limited to LSCV RNA in 6 brush mice (P. boylii) captured in northern Arizona (11). The results of this study indicate that the transvolcanic mouse (P. hylocetes) and the black-eared mouse (P. melanotis) in the state of México, the nimble-footed mouse (P. levipes) in Nuevo León, the El Carrizo deer mouse (P. ochraventer) in San Luis Potosí, and the gleaning mouse (P. spicilegus) in Jalisco are natural but not necessarily principal hosts of LSCV.
HPS was fi rst recognized as a distinct clinical entity in the southwestern United States in 1993 (30). Through 2009, a total of 510 HPS cases were reported to the National Notifi able Diseases Surveillance System or registered by the Centers for Disease Control and Prevention (31). Most of these cases occurred in the southwestern United States, 92 (33.7%) of 273 HPS cases that occurred in the southwestern United States before 2010 were fatal, and all of the cases from the southwestern United States through 2009 were attributed to SNV.
The results of this study indicate that SNV is widely distributed in northeastern Mexico. The geographic distribution of deer mice (P. maniculatus) in Mexico includes 23 states (28), and ≈20 million persons lived in rural areas in this 23-state region in 2010 (32). Yet, to our knowledge, no cases of HPS have been reported from northeastern Mexico or elsewhere in Mexico.
We hypothesize that HPS caused by SNV in Mexico has been confused with other rapidly progressive, lifethreatening respiratory diseases (e.g., plague, tularemia, pneumococcal pneumonia, infl uenza). Alternatively, SNV in Mexico is substantially less virulent than SNV in the western United States, or human contact with SNV-infected rodents in Mexico is less frequent or less intimate than human contact with SNV-infected rodents in the western United States.
Laboratory confi rmation of the diagnoses of most HPS cases in the United States before 2010 was based on the results of serologic assays in which ELMCV and LSCV can be highly cross-reactive with SNV (31). Thus, in all likelihood, some of the HPS cases from the western United States were actually caused by ELMCV or LSCV, and these viruses as well as SNV are etiologic agents of HPS in Mexico.