Nosema apis Zander, is a microsporidian pathogen of honey bees, Apis mellifera L., causing nosemosis in its host (Bailey & Ball 1991; Fries 1997). Nosema apis was first observed in the United States in 1910 (White 1914), and was assumed to be the only Nosema species infecting honey bees in the country. Another Nosema species, N. ceranae Fries, was observed in the Asian honey bee, Apis cerana F., by Fries et al. (1996), and then first reported in A. mellifera in 2004 (Higes et al. 2006). Further studies have revealed N. ceranae has been present in the United States since at least 1995 (Chen et al. 2008), and in Europe since 1998 (Paxton et al. 2007). Nosema ceranae is highly pathogenic in honey bees (Higes et al. 2008) and is associated with reduced honey production and increased winter mortality (Higes et al. 2006). Molecular diagnostics using Poly merase Chain Reaction (PCR) primers specific for Nosema ceranae and N. apis are well established and have been conducted on managed honey bee populations from Turkey (Whitaker et al. 2011), Argentina (Medici et al. 2012), Crotia (Gajger et al. 2010) the United States (Chen et al. 2008; Chapon et al. 2009; Traver & Fell 2011; Szalanski et al. in press), and Canada (Williams et al. 2008). Despite these studies on the occurrence of Nosema apis and N. ceranae in managed honey bees in the United States, it is unknown if there is a different rate of infection with feral populations of European honey bees (EHB) or with Africanized honey bees (AHB) in the United States. Feral honey bees can be defined as those sourced from established, unmanaged colonies in a tree, building, or other cavities (Schiff et al. 1994). Only one study has been conducted on Nosema from feral honey bees in the United States. Gilliam & Tabler (1991) found N. apis from one out of seven feral honey bee colonies sampled in Arizona. A recent study by Teixeira et al. (2013) has documented that N. ceranae has been present in AHB in Brazil since 1979, confirming its presence in A. mellifera considerably earlier than was previously assumed. The objectives of this study were to determine how widespread Nosema ceranae and N. apis are in feral honey bee populations in south central and southwestern United States, and if there are any differences in Nosema infection rate between feral AHB and EHB populations.

Materials and Methods

Honey bee specimens were collected from Arkansas, Mississippi, Oklahoma, New Mexico, Texas, and Utah as part of other studies on genetic variation of feral EHB and AHB populations (Szalanski & McKern 2007; Szalanski & Magnus 2010) and for this study (Table 1, Fig. 1). Worker honey bees were collected from established colonies in trees, buildings, or other cavities from 1991 to 2010 and were preserved in 70-90% ethanol. The sample unit in this study was a single colony, with the number of worker honey bees collected for each sample ranging from 10 to 50 individuals. DNA was extracted separately from individual worker honey bee thoraces (AHB/EHB testing) and abdomens (Nosema testing) using a salting-out procedure with in-house reagents (Sambrook & Russell 2001). Molecular diagnostics using multiplex Polymerase Chain Reaction (PCR) of a mitochondrial DNA cytochrome b marker (Szalanski & McKern 2007) was used to determine if the sample was AHB or EHB. Two individuals were tested from each sample.

Table 1.

Number of samples from six stat es and year of sample for this study on the molecular detection of Nosema from feral Apis mellifera.

Molecular detection of Nosema was conducted using the associated abdominal DNA extractions, with PCR primers that were designed by finding conserved regions in aligned Nosema small subunit (SSU) sequences obtained from GenBank (Geneious v6.1.6, Biomatters, Auckland, New Zealand). The PCR primers that were developed were NosemaSSU-1F (5′-ACAATATGTATTAGATCTGATATA-3′) and NosemaSSU-1R (5′-TAATGATATGCTTAAGTTCAAAG-3′). These primers amplify a 222 bp (N. apis) or 237bp (N. ceranae) region of a SSU marker for Nosema. Thermocycler conditions were: an initial denaturation of 94 °C for 2 min, followed by 40 cycles of 94 °C for 45s, 50 °C for 60s and 72 °C for 60 s, and a final extension of 5 min at 72 °C. PCR products were subjected to electrophoresis in 2% agarose gels, and PCR products were visualized under UV light (BioDocit Imaging System, UVP, LLC, Upland, California).

Geneious software was then used to evaluate restriction enzymes which would distinguish N. ceranae from N. apis using PCR-Restriction Fragment Length Polymorphism (RFLP). Samples which exhibited an amplicon approximately 230 bp in length were subjected to PCR-RFLP using the restriction enzymes Dra I, which exclusively cuts N. ceranae at 79 bp, and Rsa I, which only cuts N. apis at 130 bp. For each digestion, 2 µl of PCR product, 0.4 µL restriction enzyme, 1 µL enzyme buffer, and 6.2 µl ultrapure water were incubated at 37 °C overnight. An enzyme-negative control was included by substituting ultrapure water for the enzyme. Resulting products were visualized as above. Estimates of amplicon size were made by visual comparison to a 100 bp ladder. Amplified DNA from 2 N. ceranae and 3 N. apis positive PCR-RFLP samples were purified, concentrated with PES 30k filter devices (VWR, Radnor, Pennsylvania), and sent for direct sequencing in both directions (University of Arkansas Medical School, Little Rock, Arkansas). Consensus sequences were constructed using Geneious v6.1.6. Identification of samples was then confirmed using N. ceranae and N. apis SSU sequences from GenBank. Statistical analysis was conducted using pairwise comparisons via Fisher's exact tests (Sokal & Rohlf 1969).

Fig. 1.

Feral Africanized (AHB) and European honey bee (EHB) sample locations.

Results and Discussion

A total of 517 feral A. mellifera samples (245 AHB, 272 EHB), collected from Arkansas, Mississippi, Texas, Utah, Oklahoma, and New Mexico from 1991 to 2010, were subjected to PCR-RFLP analysis for molecular identification of Nosema apis and N. ceranae (Table 1, Fig. 1). Of these, 43 samples (8.3%) were positive for Nosema, with 82.1% being N. ceranae, and the remainder were N. apis (Table 2). No mixed samples were observed. For the AHB samples, Nosema was detected in 9.0% of the samples with 89.5% of the Nosema identified as N. ceranae, and 10.5% as N. apis. With the EHB samples, 7.7% had Nosema, with 75.0% of these with N. ceranae, and 25.0% with N. apis. There was no significant difference in Nosema infection rate (P = 0.635), or the frequency of N. apis and N. ceranae infections (P = 0.407) between AHB and EHB populations.

Table 2.

Occurrence of Nosema ceranae and N. apis in Africanized Honey Bee (ahb) and European Honey Bee (ehb) colonies from six states.

Prevalence of Nosema from AHB was highest in Texas (12%) and lowest in New Mexico (3%), although there was no significant difference (P = 0.183). From the EHB, results were similar to the AHB, with Nosema most common in Texas (22%) and lowest in New Mexico (0% infection) (P = 0.003). There was a significant difference in overall Nosema infection for feral honey bees in Oklahoma compared to Texas (P = 0.007), and between Texas and New Mexico (P = 0.010), but not between Oklahoma and Utah (P = 0.302), and Texas and Utah (P =0.126).

This is the first study to document the occurrence of both N. ceranae and N. apis from feral EHB and AHB in the United States. The only previous study which observed N. apis from feral honey bees in the United States was limited in scope and only found 1 colony out of 7 in Arizona infected with N. apis (Gilliam & Tabler 1991). We also found evidence that N. ceranae has been present in AHB colonies in Texas since 2005. A recent study by Teixeira et al. (2013) documented the existence of N. ceranae in a sample of AHB drones from Brazil that were collected 34 yr ago. This suggests that N. ceranae may have followed AHB into the United States as AHB expanded northward. We did not detect Nosema from the 1991 and 2004 Texas samples, but the small sample sizes (17 and 10, respectively) may be the reason for not detecting it. It may also be that, as Teixeira et al. (2013) suggest, microscopy-based detection of Nosema prior to the molecular age did not distinguish species as well as modern methods.

Several statewide Nosema surveys have been conducted on managed honey bee colonies in the United States. Traver & Fell (2011) studied 586 managed A. mellifera colonies from Virginia in 2009 using real time PCR to detect Nosema. They found 37% of the colonies to have Nosema with 97.3% infected with N. ceranae and 2.7% with N. apis. Szalanski et al. (2013) found 31% of 1,200 apiaries in New York and 44% of 1,009 apiaries in South Dakota to have Nosema based on spore counts. Of the samples subjected to molecular diagnostics they found 96% of the New York spore positive samples to have N. ceranae, 3% co-infected with N. ceranae and N. apis, and 1% with N. apis. All of their South Dakota samples subjected to molecular diagnostics were positive for N. ceranae. In the present study, the percentage of feral honey bees infected with Nosema (8.3%) was considerably lower than the levels observed in managed honey bees (31–44%). This is not surprising, since it has been documented that management of honey bee colonies can increase the infection rate of Nosema in apiaries. For example, Oertel (1967) observed that opening honey bee colonies can increase the presence of N. apis, and Moeller (1978) noted that honey bee equipment from colonies that died over winter can be a source of Nosema for new colonies established by package bees or divisions. Nosema-contaminated honey supers or brood chambers can also inoculate uninfected colonies in the summer (Moeller 1978).

Among all 6 sampled states, Nosema sp. occurrence was lowest in the Spring and highest in the Winter (Table 3). Studies in Germany (Gisder et al. 2010) and in Virginia (Traver et al. 2012) have revealed that Nosema infection levels were highest in the Spring and lower in the Fall and Winter. It is difficult to draw conclusions from our dataset on the seasonality of Nosema prevalence given the broad geographical distribution of our samples, uneven sampling throughout the years, and the low incidence of Nosema-positive samples. Robust, regular sampling throughout the region, as conducted by Gisder et al. (2010) in Germany, would be required to determine if there are seasonal differences in Nosema infection rates that are driven by climate.

We found that Nosema was widespread throughout the southwestern and south central United States. Nosema ceranae was present in all sampled states, yet N. apis was only present in Mississippi, Texas, and Oklahoma. Despite N. ceranae being first detected in the United States as recently as 1995 (Chen et al. 2008), N. ceranae seems more common in both managed and feral populations of honey bees than N. apis, although increased sample sizes are needed to determine if this is truly the case. Our study is the first to report that N. ceranae occurs in feral AHB populations within the United States. Although our work did not show a difference in Nosema species occurrence between AHB and EHB populations, most (98%) of our samples were collected after the known occurrence of N. ceranae in the United States. If samples collected in the past become available, the molecular identification techniques employed in this study should prove useful in further investigations of the invasion history of Nosema and AHB into the United States.

Table 3.

Seasonal occurance of Nosema in Africanized Honey Bee (ahb) and European Honey Bee (ehb) colonies from six states.