How natural infection by Nosema ceranae causeshoneybee colony collapse

Mariano Higes1 Raquel Martiacuten-Hernaacutendez1

Cristina Botiacuteas1 Encarna Garrido Bailoacuten1

Amelia V Gonzaacutelez-Porto2 Laura Barrios3

M Jesuacutes del Nozal4 Joseacute L Bernal4

Juan J Jimeacutenez4 Pilar Garciacutea Palencia5 andAraacutenzazu Meana61Bee Pathology laboratory Centro Apiacutecola RegionalJCCM 19180 Marchamalo Spain2Hive Products laboratory Centro Apiacutecola RegionalJCCM 19180 Marchamalo Spain3Statistics Department CTI Consejo SuperiorInvestigaciones Cientiacuteficas 28006 Madrid Spain4Analytical Chemistry Department Facultad de CienciasUniversidad de Valladolid 47005 Valladolid Spain5Animal Medicine and Surgery Department Facultad deVeterinaria Universidad Complutense de Madrid28040 Madrid Spain6Animal Health Department Facultad de VeterinariaUniversidad Complutense de Madrid 28040 MadridSpain

Summary

In recent years honeybees (Apis mellifera) have beenstrangely disappearing from their hives and strongcolonies have suddenly become weak and died Theprecise aetiology underlying the disappearance ofthe bees remains a mystery However during thesame period Nosema ceranae a microsporidium ofthe Asian bee Apis cerana seems to have colonized Amellifera and itrsquos now frequently detected all over theworld in both healthy and weak honeybee coloniesFor first time we show that natural N ceranae infec-tion can cause the sudden collapse of bee coloniesestablishing a direct correlation between N ceranaeinfection and the death of honeybee colonies underfield conditions Signs of colony weakness were notevident until the queen could no longer replace theloss of the infected bees The long asymptomaticincubation period can explain the absence of evidentsymptoms prior to colony collapse Furthermore ourresults demonstrate that healthy colonies near to an

infected one can also become infected and that Nceranae infection can be controlled with a specificantibiotic fumagillin Moreover the administrationof 120 mg of fumagillin has proven to eliminate theinfection but it cannot avoid reinfection after6 months We provide Kochrsquos postulates between Nceranae infection and a syndrome with a long incuba-tion period involving continuous death of adult beesnon-stop brood rearing by the bees and colony loss inwinter or early spring despite the presence of suffi-cient remaining pollen and honey

Introduction

As a bee colony can be considered as a complex livingsystem of individuals that functions as a whole diseasepathology of an individual bee is different to the pathologyat the colony level Indeed a particular pathogen can belethal to bees but the colony may be able to compensatefor their loss In this sense the queen is essential tomaintain the population of the colony stable Depopulationsyndrome or honeybee colony collapse has recently beenthe focus of many reports that have tried to determine whystrong colonies suddenly or at least thatrsquos the way itseems become weak and occasionally die (Faucon et al2002 Higes et al 2005 Ortiz 2005 Gonzaacutelez 2007Stokstad 2007 Molga 2008) Many colonies are lostcompletely without any evident signs of a disorder Yetthese observations could be just the tip of the iceberg ascolony breakdown may just be the final phase of a longprocess or of chronic infection by a silent pathogenDuring the same period that this syndrome has beendetected all over the world Nosema ceranae seems tohave colonized Apis mellifera (Martiacuten-Hernaacutendez et al2007) and it is now frequently detected in both healthyand weak honeybee colonies (Cox-Foster et al 2007)

Nosema ceranae is a widespread microsporidium thatseems to have recently jumped from its host the Asianhoneybee Apis cerana to the worldwide honey producerA mellifera As it was first detected outside of Asia (Higeset al 2006) its presence has been confirmed in fourcontinents although there is no available data from Africa(Chauzat et al 2007 Cox-Foster et al 2007 Klee et al2007 Martiacuten-Hernaacutendez et al 2007 Chen et al 2008)Nosema ceranae is very pathogenic when experimentallyinoculated into A mellifera (Higes et al 2007 Paxton

Received 14 March 2008 accepted 13 May 2008 Forcorrespondence E-mail ameanavetucmes Tel (+34) 913943903 Fax (+34) 91 3943908

Environmental Microbiology (2008) doi101111j1462-2920200801687x

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd

et al 2007) and natural infection has been associatedwith a syndrome of gradual depopulation copious colonydeath in autumn or winter and poor honey production Infact the risk of colony depopulation is six times higher incolonies infected with N ceranae than in uninfected ones(Martiacuten-Hernaacutendez et al 2007)

Here we show the pathological repercussion of Nceranae infection of colonies and its clinical evolution untilthe colony dies We have not only analysed the naturalinfection of N ceranae but also demonstrate the trans-mission of the disease to nearby uninfected hives andthe effect of fumagillin (as Fumidil B) in controlling theinfection

Results and discussion

Clinical monitoring

Once the presence of spores from Nosema spp wasdetected microscopically in a productive honeybee colony(B255) in the experimental apiaries of CAR the specieswas identified as N ceranae and this was confirmedthroughout the study (Accession Number DQ 329034)Infection was transmitted to another uninfected colony(B111) and 13 uninfected nuclei within 3 months whileanother nucleus was infected after 5 months of exposure

From the outset it was difficult to define the health statusof a colony following infection as no classic signs of nose-mosis were evident (OIE 2004) As the mean spore countfluctuates greatly from the start to the end of the disease ininterior bees (Fig 1A) this is not a reliable measure of acolonyrsquos health when bees are infected with N ceranae(CATPCA analysis Fig S1) Indeed Nosema infectioncannot be assessed at the colony level on the basis of beessampled from the brood nest (El-Shemy and Pickard1989) In fact the proportion of forager bees infected withN ceranae was the only useful indicator of the extent ofdisease in the colony (Pearson r = -08 P = 0001) and likethe spore count foragers were always more infected thaninterior bees The more foragers infected the smaller thenumber of brood combs and the fewer frames full of bees(Pearson P = 0001 CATPCA Fig S1) This relationshipdid not hold for the proportion of infected interior bees orthe spore counts The dynamics of bees and brood combswere related to the proportion of infected foragers andthe maximum mean temperatures (regression stepwise

adjusted R2 0702) No statistical differences could beestablished with other meteorological parameters

Four Phases of natural N ceranae infection were deter-mined (Fig 1A) Each of the phases of clinical nosemosiswere observed in the whole colonies (number ofcombs = 10 or higher if honey is going to be collected andan upper box is on the brood chamber box then n = 20)although they were milder in the B111 hive probablybecause of its infection in summer and not in spring Nucleionly displayed three phases probably because of the factthat the nuclei had fewer combs and consequently asmaller bee population At the outset infected colonieslooked similar to those of other uninfected hives at leastfrom the spring to early autumn months without any signof disease This was defined as Phase 1 or the lsquoasymp-tomaticrsquo phase Fewer than 60 of foragers were infected(n = 30) and mean spore counts never rose above onemillion Next phase was considered as Phase 2 or thelsquoreplacementrsquo phase and started when unusual colonybehaviours were detected (for this region) when the queencontinued to lay eggs throughout the winter months (broodarea around 125 cm2) The fact that this phase took placein winter when the number of newborn bees is low couldexplain why the percentage of foragers and mean sporecounts were always higher than in Phase 1

The third phase started the following spring when thecolonies built up rapidly and the queens laid so manyeggs that almost every comb was full of brood Clinicalparameters (forager infection spore counts) were similarto Phase 1 and as a result of the increasing bee popula-tion this Phase 3 was named as the lsquofalse recoveryrsquophase The bee population had built up in such a way thatswarming seemed imminent but it never happened Theegg-laying lsquofeverrsquo went on until the autumn when suddendepopulation (fewer than 40 of frames covered withbees) was observed and the brood reached a minimumdefining the fourth lsquodepopulationrsquo phase In this phase thebees were very active but 2 months later the queen wasfound dead and surrounded by only young bees thatprobably died because of the cold (breakdown or col-lapse Table 1) Pollen and honey stores were presentand even then a tiny brood spot of capped brood cellswas detected in most colonies

Only in the fourth phase were more than 40 of theinterior bees infected with a much higher proportion ofinfected foragers (like Phase 2)

Fig 1 Clinical monitoring of N ceranae natural infectionA Dynamics of natural N ceranae infection of a model colony and the phases of nosemosis Percentage of N ceranae-infected bees (n = 30)mean spore count (n = 30) and number of combs covered by brood or bees in the B255 colonyB Dynamics of the natural infected N ceranae B111 colony and the phases of nosemosisC Colony collapse of untreated colonies (B255 B111) and nuclei (CN1ndashCN5 NSN1ndashNSN4) could occur in the cold or warm season Boxplots display differences in the percentage of infection and spore counts between dead bees collected when colonies collapsed during thewarm (April to May n = 6) or cold (November to February n = 5) months in the Spanish climatic conditions The horizontal line within each boxmarks the median indicates collapse of the colony FB forager bees IB interior bees

2 M Higes et al

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

Natural Nosema ceranae infection 3

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

The dynamics of the B111 and B255 colonies were verysimilar Hive B111 (Fig 1B) was infected in summer andso the first asymptomatic phase lasted for the 3 monthsof autumn and with fewer than 50 of the foragersinfected the second phase started in winter with fewerforagers infected than in hive B255 and more brood (threebrood combs versus 1 in B255) This delayed infectioncoupled with enhanced replacement could explain thelower proportion of infected interior or forager beesDuring the next spring and early summer (about a yearafter infection) false recovery was evident through thepresence of brood in the combs of the upper box in JuneHowever again the colony did not swarm and from this3rd phase of false recovery the colony displayed clearsigns of depopulation (Phase 4) and death was imminent

All the untreated nuclei (CN6ndashCN10 NSN1ndashNSN4) hadbrood combs throughout the study until they collapsed

(Fig 1C) and at least one comb with brood was found evenduring the coldest months (DecemberndashFebruary Fig 2A)Breakdown occurred in all the untreated nuclei (CN6ndashCN10 NSN1-NSN4) and none of the nine nuclei winteredMaybe because of the smaller population (the nuclei hadonly 5 combs instead of 10) Phase 3 of false recovery wasnot seen and from Phase 2 they went directly to Phase 4 ofdepopulation and death While some died in the winter theremaining nuclei did so in the early spring generally after amuch shorter incubation period than in the whole coloniesand reaching the depopulation phase 3ndash5 months afterinfection Like whole colonies collapse occurred in a2 month period and again more than 80 of the foragers(Table S1) were infected and fewer than 40 of the combs(n = 5) were full of bees

Two ways of collapse were established that could berelated to the moment in which the colonies or nuclei died

Table 1 Characteristics of colony breakdown in the cold or warm season

Colony Brood cells Bees Infection onworker bees

N ceranae (sporesbee)Worker bees (mean) Queen

Cold season breakdownB255 100 133 50 1275 yen 106 02 yen 106

B111 241 231 51 108 yen 106 01 yen 106

CN6 50 13 90 21 yen 106 021 yen 106

CN9 250 33 65 25 yen 106 Not foundCN10 350 65 50 119 yen 106 04 yen 106

Warm season breakdownCN7 0 350 44 11 yen 106 lt 50 000

(PCR+)CN8 400 150 30 1 yen 106 NegativeNSN1 200 350 15 025 yen 106 lt 50 000

(PCR+)NSN2 1200 1081 35 35 yen 106 NegativeNSN3 500 512 30 2 yen 106 NegativeNSN4 0 0 Not found Not found Not found

Cold season breakdown from December to February warm season breakdown from April to June B255 and B111 hives and CN6ndashCN10 nucleiwere treated and sampled monthly while NSN1ndashNSN4 were nuclei not treated and not sampled till they were dead

TREATED

6

8

10 Bees

Brood

NO TREATED

6

8

10

Bees

Brood

A B

0

2

4

0

2

4Com

bs

infe

ction

Com

bs

infe

ction

0

20

40

60

80

100

IB FB

0

20

40

60

80

100

IB FB

0

June-06 Aug-06 Oct -06 Dec-06 Feb-07 Apr-07

0

June-06 Aug-06 Oct-06 Dec-06 Feb-07 Apr-07

Fig 2 Experiment 1 response trials with Fumidil B treatment in natural N ceranae-infected nuclei Mean dynamics of brood combs and beecombs and the mean percentage infection of interior bees (n = 30) and foragers (n = 30) in (A) untreated (n = 5) and (B) fumagillin-treated(n = 5) nuclei (error bars as mean SD) Shaded areas indicate the moment of application (fumagillin in treated nuclei and syrup in untreatedones) indicates collapse of the untreated colonies

4 M Higes et al

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

(Fig 1C Table 1) When collapse occurred during thecold months more than 50 of the bees found deadinside the hive were infected the mean spore count inthese bees was always higher than 10 million and thequeens (when found) were infected However when col-lapse occurred later in early spring the percentage ofinfection and mean spore counts were lower Moreoverunder these circumstances the queens were usuallyuninfected or infection could only be detected by PCRProbably the differences between these two cases reflectthe quantity of old and young bees in each season Inearly spring the proportion of newborn uninfected beeswill reduce the infection parameters thereby delaying theinfection of the queen (Table 1)

Similar pathological lesions were seen in samplesfrom dead or dying bees collected from the ground of aLavandula latifolia crop and in infected bees collectedfrom hives throughout the year even in the summermonths Cells all along the ventriculus epithelium wereinfected (Fig 3) showing comparable alterations to thosedescribed experimentally (Higes et al 2007) and intrac-ellular germination was also seen Indeed there was evi-dence of epithelial cell degeneration and extensive lysisThe heavily infected cells may either be dead or dyingwhich will eventually lead to the early death of the infectedworker bees as a result of starvation (Liu 1984) Althoughpathological features due to N ceranae closely resembledthose of N apis some clinical signs usually associatedwith N apis infection were not observed such as crawling

bees dysentery evidenced by the presence of fecal spotsin the hive or supersedure of the queen Foragers collectedin lavender crop were infected with a mean of 21 million Nceranae spores per bee This demonstrates that someheavily infected foragers did not return to the hive

The age at which worker honeybees begin foragingvaries under different colony conditions and the age atwhich foraging commences seems to be delayed in thepresence of foragers (Huang and Robinson 1996) Thuswhen the number of foragers is depleted the age of onsetof this behaviour drops (Huang and Robinson 1996Amdam and Omholt 2003) Infestation by Varroa destruc-tor mite was recently proposed to influence the flightbehaviour of forager bees such that they might not returnto the colony (Kralj and Fuchs 2006) This effect wasinterpreted as an adaptative behaviour of the bees toexpel the pathogens from the colony a process known aslsquosuicidal pathogen removalrsquo In a other interpretation ofdata the death due to N ceranae could be a suddenevent or a debilitating process by which bees cannotreturn to the hive This population must be replaced byinterior bees which presumably start flying some daysearlier than usual as previously associated to N apisinfection (Hassanein 1953) Once the queen cannot com-pensate for the loss of foragers depopulation becomesevident and death is forthcoming

Throughout the period studied no adult or brood dis-eases were diagnosed no other parasites were detectedand Israeli acute paralysis virus (IAPV) a virus that was

Fig 3 Microscopy images of an infected bee ventriculus and N ceranae stagesA Representative example of infected epithelial cells filled with different stage parasites A predominance of basophilic mature spores in theapical cytoplasm is seen the bottom of the cytoplasm is full of more vegetative parasites (MS mature spore N nucleus IS immature stageV vacuole) Toluidine blue staining (20yen magnification)B Infected epithelial cells with breakage of the cell plasma membrane at the base of the brush border Brightly stained nucleus displacedapically and vacuolar basal degeneration were evident in the enlarged parasitized cells (4000yen magnification)C Detail of an extruded spore inside an infected cell indicating that intracellular germination of spores takes place (60 000yen magnification)

Natural Nosema ceranae infection 5

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

The dynamics of the B111 and B255 colonies were verysimilar Hive B111 (Fig 1B) was infected in summer andso the first asymptomatic phase lasted for the 3 monthsof autumn and with fewer than 50 of the foragersinfected the second phase started in winter with fewerforagers infected than in hive B255 and more brood (threebrood combs versus 1 in B255) This delayed infectioncoupled with enhanced replacement could explain thelower proportion of infected interior or forager beesDuring the next spring and early summer (about a yearafter infection) false recovery was evident through thepresence of brood in the combs of the upper box in JuneHowever again the colony did not swarm and from this3rd phase of false recovery the colony displayed clearsigns of depopulation (Phase 4) and death was imminent

All the untreated nuclei (CN6ndashCN10 NSN1ndashNSN4) hadbrood combs throughout the study until they collapsed

(Fig 1C) and at least one comb with brood was found evenduring the coldest months (DecemberndashFebruary Fig 2A)Breakdown occurred in all the untreated nuclei (CN6ndashCN10 NSN1-NSN4) and none of the nine nuclei winteredMaybe because of the smaller population (the nuclei hadonly 5 combs instead of 10) Phase 3 of false recovery wasnot seen and from Phase 2 they went directly to Phase 4 ofdepopulation and death While some died in the winter theremaining nuclei did so in the early spring generally after amuch shorter incubation period than in the whole coloniesand reaching the depopulation phase 3ndash5 months afterinfection Like whole colonies collapse occurred in a2 month period and again more than 80 of the foragers(Table S1) were infected and fewer than 40 of the combs(n = 5) were full of bees

Two ways of collapse were established that could berelated to the moment in which the colonies or nuclei died

Table 1 Characteristics of colony breakdown in the cold or warm season

Colony Brood cells Bees Infection onworker bees

N ceranae (sporesbee)Worker bees (mean) Queen

Cold season breakdownB255 100 133 50 1275 yen 106 02 yen 106

B111 241 231 51 108 yen 106 01 yen 106

CN6 50 13 90 21 yen 106 021 yen 106

CN9 250 33 65 25 yen 106 Not foundCN10 350 65 50 119 yen 106 04 yen 106

Warm season breakdownCN7 0 350 44 11 yen 106 lt 50 000

(PCR+)CN8 400 150 30 1 yen 106 NegativeNSN1 200 350 15 025 yen 106 lt 50 000

(PCR+)NSN2 1200 1081 35 35 yen 106 NegativeNSN3 500 512 30 2 yen 106 NegativeNSN4 0 0 Not found Not found Not found

Cold season breakdown from December to February warm season breakdown from April to June B255 and B111 hives and CN6ndashCN10 nucleiwere treated and sampled monthly while NSN1ndashNSN4 were nuclei not treated and not sampled till they were dead

TREATED

6

8

10 Bees

Brood

NO TREATED

6

8

10

Bees

Brood

A B

0

2

4

0

2

4Com

bs

infe

ction

Com

bs

infe

ction

0

20

40

60

80

100

IB FB

0

20

40

60

80

100

IB FB

0

June-06 Aug-06 Oct -06 Dec-06 Feb-07 Apr-07

0

June-06 Aug-06 Oct-06 Dec-06 Feb-07 Apr-07

Fig 2 Experiment 1 response trials with Fumidil B treatment in natural N ceranae-infected nuclei Mean dynamics of brood combs and beecombs and the mean percentage infection of interior bees (n = 30) and foragers (n = 30) in (A) untreated (n = 5) and (B) fumagillin-treated(n = 5) nuclei (error bars as mean SD) Shaded areas indicate the moment of application (fumagillin in treated nuclei and syrup in untreatedones) indicates collapse of the untreated colonies

4 M Higes et al

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

(Fig 1C Table 1) When collapse occurred during thecold months more than 50 of the bees found deadinside the hive were infected the mean spore count inthese bees was always higher than 10 million and thequeens (when found) were infected However when col-lapse occurred later in early spring the percentage ofinfection and mean spore counts were lower Moreoverunder these circumstances the queens were usuallyuninfected or infection could only be detected by PCRProbably the differences between these two cases reflectthe quantity of old and young bees in each season Inearly spring the proportion of newborn uninfected beeswill reduce the infection parameters thereby delaying theinfection of the queen (Table 1)

Similar pathological lesions were seen in samplesfrom dead or dying bees collected from the ground of aLavandula latifolia crop and in infected bees collectedfrom hives throughout the year even in the summermonths Cells all along the ventriculus epithelium wereinfected (Fig 3) showing comparable alterations to thosedescribed experimentally (Higes et al 2007) and intrac-ellular germination was also seen Indeed there was evi-dence of epithelial cell degeneration and extensive lysisThe heavily infected cells may either be dead or dyingwhich will eventually lead to the early death of the infectedworker bees as a result of starvation (Liu 1984) Althoughpathological features due to N ceranae closely resembledthose of N apis some clinical signs usually associatedwith N apis infection were not observed such as crawling

bees dysentery evidenced by the presence of fecal spotsin the hive or supersedure of the queen Foragers collectedin lavender crop were infected with a mean of 21 million Nceranae spores per bee This demonstrates that someheavily infected foragers did not return to the hive

The age at which worker honeybees begin foragingvaries under different colony conditions and the age atwhich foraging commences seems to be delayed in thepresence of foragers (Huang and Robinson 1996) Thuswhen the number of foragers is depleted the age of onsetof this behaviour drops (Huang and Robinson 1996Amdam and Omholt 2003) Infestation by Varroa destruc-tor mite was recently proposed to influence the flightbehaviour of forager bees such that they might not returnto the colony (Kralj and Fuchs 2006) This effect wasinterpreted as an adaptative behaviour of the bees toexpel the pathogens from the colony a process known aslsquosuicidal pathogen removalrsquo In a other interpretation ofdata the death due to N ceranae could be a suddenevent or a debilitating process by which bees cannotreturn to the hive This population must be replaced byinterior bees which presumably start flying some daysearlier than usual as previously associated to N apisinfection (Hassanein 1953) Once the queen cannot com-pensate for the loss of foragers depopulation becomesevident and death is forthcoming

Throughout the period studied no adult or brood dis-eases were diagnosed no other parasites were detectedand Israeli acute paralysis virus (IAPV) a virus that was

Fig 3 Microscopy images of an infected bee ventriculus and N ceranae stagesA Representative example of infected epithelial cells filled with different stage parasites A predominance of basophilic mature spores in theapical cytoplasm is seen the bottom of the cytoplasm is full of more vegetative parasites (MS mature spore N nucleus IS immature stageV vacuole) Toluidine blue staining (20yen magnification)B Infected epithelial cells with breakage of the cell plasma membrane at the base of the brush border Brightly stained nucleus displacedapically and vacuolar basal degeneration were evident in the enlarged parasitized cells (4000yen magnification)C Detail of an extruded spore inside an infected cell indicating that intracellular germination of spores takes place (60 000yen magnification)

Natural Nosema ceranae infection 5

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

(Fig 1C Table 1) When collapse occurred during thecold months more than 50 of the bees found deadinside the hive were infected the mean spore count inthese bees was always higher than 10 million and thequeens (when found) were infected However when col-lapse occurred later in early spring the percentage ofinfection and mean spore counts were lower Moreoverunder these circumstances the queens were usuallyuninfected or infection could only be detected by PCRProbably the differences between these two cases reflectthe quantity of old and young bees in each season Inearly spring the proportion of newborn uninfected beeswill reduce the infection parameters thereby delaying theinfection of the queen (Table 1)

Similar pathological lesions were seen in samplesfrom dead or dying bees collected from the ground of aLavandula latifolia crop and in infected bees collectedfrom hives throughout the year even in the summermonths Cells all along the ventriculus epithelium wereinfected (Fig 3) showing comparable alterations to thosedescribed experimentally (Higes et al 2007) and intrac-ellular germination was also seen Indeed there was evi-dence of epithelial cell degeneration and extensive lysisThe heavily infected cells may either be dead or dyingwhich will eventually lead to the early death of the infectedworker bees as a result of starvation (Liu 1984) Althoughpathological features due to N ceranae closely resembledthose of N apis some clinical signs usually associatedwith N apis infection were not observed such as crawling

bees dysentery evidenced by the presence of fecal spotsin the hive or supersedure of the queen Foragers collectedin lavender crop were infected with a mean of 21 million Nceranae spores per bee This demonstrates that someheavily infected foragers did not return to the hive

The age at which worker honeybees begin foragingvaries under different colony conditions and the age atwhich foraging commences seems to be delayed in thepresence of foragers (Huang and Robinson 1996) Thuswhen the number of foragers is depleted the age of onsetof this behaviour drops (Huang and Robinson 1996Amdam and Omholt 2003) Infestation by Varroa destruc-tor mite was recently proposed to influence the flightbehaviour of forager bees such that they might not returnto the colony (Kralj and Fuchs 2006) This effect wasinterpreted as an adaptative behaviour of the bees toexpel the pathogens from the colony a process known aslsquosuicidal pathogen removalrsquo In a other interpretation ofdata the death due to N ceranae could be a suddenevent or a debilitating process by which bees cannotreturn to the hive This population must be replaced byinterior bees which presumably start flying some daysearlier than usual as previously associated to N apisinfection (Hassanein 1953) Once the queen cannot com-pensate for the loss of foragers depopulation becomesevident and death is forthcoming

Throughout the period studied no adult or brood dis-eases were diagnosed no other parasites were detectedand Israeli acute paralysis virus (IAPV) a virus that was

Fig 3 Microscopy images of an infected bee ventriculus and N ceranae stagesA Representative example of infected epithelial cells filled with different stage parasites A predominance of basophilic mature spores in theapical cytoplasm is seen the bottom of the cytoplasm is full of more vegetative parasites (MS mature spore N nucleus IS immature stageV vacuole) Toluidine blue staining (20yen magnification)B Infected epithelial cells with breakage of the cell plasma membrane at the base of the brush border Brightly stained nucleus displacedapically and vacuolar basal degeneration were evident in the enlarged parasitized cells (4000yen magnification)C Detail of an extruded spore inside an infected cell indicating that intracellular germination of spores takes place (60 000yen magnification)

Natural Nosema ceranae infection 5

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

only identified in one of the nuclei studied was not isolatedin B255 bees (Table 2) Indeed this is the first report ofIAPV in Spain This virus was previously proposed as aCCD marker (Cox-Foster et al 2007) however it is notprevalent in our studies and it was not associated withcolony collapse The microsporidium may have some influ-ence on other viruses as the epithelial lesions caused bythe parasite reduce the natural resistance that the intesti-nal barrier provides against viral infection Many viruseshave been associated with N apis infection (revised byChen and Siede 2007) possibly because of the cyto-pathogenic effect that facilitates transenteric viral infectionThe absence of clinical signs related to viruses suggeststhat infection was covert or latent

Although some viruses were detected by PCR in B255and in six nuclei B111 and four nuclei were negative andno relationship could be established between the pres-ence of these viruses and collapse (Table 2) Chronic beeparalysis virus (CBPV) and deformed wing virus (DWV)viruses were always detected in B255 although therewere no phenotypic signs of these infections at anymoment Indeed the genomic load of the CBPV virusafter colony death was about 20 million times lower thanthat necessary to produce symptoms in the nervoussystem (Blanchard et al 2007) Furthermore CBPVDWV and BQCV were also detected in another six nuclei

studied While gut tissue may be a major reservoirfor DWV (Chen et al 2006) virus particles were notdetected in any intestinal sample studied by transmissionelectron microscopy (TEM)

The residue analysis of 40 compounds (Table S2)demonstrated that the bees from all the colonies werenot exposed to any pesticides The pollen availability itsdiversity and the presence of more than enough storedpollen at the time of breakdown (Table S3) were demon-strated by palynological studies Moreover neither sun-flower nor corn crops previously related with largedecreases in the population of bee colonies (Laurent andRathahao 2003) were close to the location of the hiveas reflected by the complete absence of these pollens inthe samples analysed (Table S3) The presence of pollenis directly related with the beekeeping flora and indicatesthe botanic resources available monthly throughout thestudy Twenty different taxa were found indicating highpollen diversity and availability in foraging areas Theblooming period of the different taxa identified indicatedthat the bees had access to sufficient flowers throughoutthe year except in January

Experiment 1 response trial in nuclei

In Experiment 1 fumagillin proved to be 100 effective incontrolling N ceranae when the bee colonies are nuclei

Table 2 Samples from untreated nuclei were taken on the collapse date and samples from the treated ones were collected when the lastuntreated nuclei collapsed (April 2007)

Colony CBPVa ABPV DWV SBV KBV BQCV IAPV

CN1 ND ND ND ND ND ND NDCN2 ND ND Positive ND ND ND NDCN3 Positive

726 yen 106 ND ND ND ND ND NDCN4 Positive

201 yen 107 ND ND ND ND ND NDCN5 ND ND ND ND ND ND NDCN6 Positive

605 yen 106 ND Positive ND ND Positive NDCN7 ND ND ND ND ND ND NDCN8 ND ND Positive ND ND ND NDCN9 ND ND Positive ND ND ND NDCN10 ND ND ND ND ND ND NDB255

July 2006 Positive ND Positive ND ND ND NDOctober 2006 Positive

229 yen 105 ND Positive ND ND ND NDDecember 2006 Positive

605 yen 106 ND Positive ND ND ND NDB111

October 2006 ND ND ND ND ND ND NDJune 2007 ND ND ND ND ND ND NDNovember 2007 ND ND ND ND ND ND ND

ColonyIAPV

NSN1 NSN2 NSN3 NSN4ND Positive ND ND

a Quantitatively analysedSamples were taken from B255 and B111 at different timesND not detected KBV Kashmir bee virus SBV sacbrood virus ABPV acute bee paralysis virus BQCV black queen cell virus

6 M Higes et al

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

All treated nuclei (Fig 2B) wintered in good health and thequeens stopped laying eggs for 1 or 2 months Indeedthese nuclei became colonies with 10 combs full of beesthat developed until the next spring when they were againinfected (6 months after treatment) By contrast allthe untreated nuclei had brood combs throughout thestudy until they died (CN6ndashCN10 Fig 2A) Breakdownoccurred in all the untreated nuclei (CN6ndashCN10 NSN1ndashNSN4) Furthermore although the queen continuedlaying eggs during the cold months these nuclei neverdeveloped into a whole colony Significant differenceswere observed from November 2006 (brood combs) andDecember 2006 (bee combs) None of the treated nucleihad brood for a month or two during the cold months

Experiment 2 colony response trial

In the Experiment 2 (Table 3) fumagillin treatment againcontrolled N ceranae infection in standard size coloniesTreated hives were infected a year after the treatment hadbeen applied and in that moment they were clinicallyin Phase 1 (Fig S2) Fumagillin treatment significantlyreduced the risk of depopulation (Chi-squared P lt 0001)in a 1 year period and untreated colonies were always ina more advanced clinical phase or dead (Chi-squaredP lt 0001)

Kochrsquos postulates have been shown here for coloniesas previously confirmed in individual bees In essence wehave extracted the pathogen confirmed that it can betransmitted to healthy colonies inducing the disease andcolony breakdown and we have recovered N ceranaefrom these newly infected colonies Multiplication of theparasite occurs throughout the year with no standstill in itslife cycle Moreover no differences in the pathologicalalterations to infected bees were observed in differentseasons Fumagillin treatment was a useful methodto avoid colony collapse although it doesnrsquot preventreinfection

Bee biology and colony dynamics are undoubtedly influ-enced by multiple facts such as regional beekeepingmanagement health status or climate The increasingprevalence of N ceranae in the last 20 years its spread(Cox-Foster et al 2007 Klee et al 2007 Martiacuten-Hernaacutendez et al 2007 Chen et al 2008) and its capac-ity to cause colony collapse lead us to consider it as anemergent disease Moreover for the first time our dataprovide experimental evidence of its aetiological role inthe death of bees and colonies under field conditions

Experimental procedure

Honeybee colonies

The studies were performed on the experimental apiaries atthe lsquoCentro Apiacutecola de Marchamalorsquo (CAR Spain 40deg40prime5747primeN 03deg12prime308primeW Fig 4) No genetically modified agricul-tural crops are found in the surrounding areas nor areinsecticides such as fipronil or imidacloprid used

All the colonies (10 bee combs) and nuclei packages(5 bee combs) were analysed to confirm the absence of anybee pathology prior to their introduction into hives at theCAR apiaries Varroa was controlled during the study bytwice yearly treatment with amitraz (Apivareg) and flumetrin(Bayvarolreg) alternating the products each year Theabsence of any brood pathology was also confirmed

Experimental design

i Clinical monitoring The first N ceranae-infected colonydetected was a successful and productive colony namedB255 previously used as a healthy control in various studiesThe presence of the parasite was detected microscopicallyon the 25 May 2005 and the species was later confirmed byPCR and sequencing of the SSUrRNA gene This colony wasleft untreated and from the date of detection to the final deathof the colony (11 December 2006) it was monitored monthly(about colony strength parasitic burdens pathologicalstudies and virological palynological and pesticide determi-nations as described below)

Table 3 Experiment 2 colony response trial

Group n Intervention

Colonies positives to N ceranae or death

Nosemosis status 111107301006a 151206 260407 260907

Positive treated 18 Fumagillin (120 mg) 18 0 5 18 18 in Phase 1Positive untreated 15 Syrup 15 13b 6 2 1 in Phase 2

2 death 7 deathc 4 deathc 1 in Phase 4Negative untreated 17 Syrup 0 5 17 14 4 in Phase 2

5 death

a Pre-treatmentb Two colonies infected by N ceranae and N apisc One colony infected by N ceranae and N apisTreatments (fumagillin as Fumidilmiddot B in treated colonies and syrup in non-treated ones) were applied as described in the Experimental proceduresThe first sampling of apiaries was made on the 30 October 2006 a day before the application of treatment Nosemosis status represents thenumber of colonies according to the phases described in Fig 1

Natural Nosema ceranae infection 7

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

One year after N ceranae was first detected in B255 (June2006) 14 uninfected nuclei (NSN1ndashNSN4 CN1ndashCN10) andanother uninfected hive (B111) were placed close to the origi-nal hive (B255) All but four nuclei (NSN1ndashNSN4) werechecked monthly analysing samples in a similar manner tothose of B255 These colonies (10 nuclei and B111 hive) wereused to establish differences in the evolution of the clinicalsigns of natural infection The remaining four nuclei were onlyobserved visually to avoid any negative effect of manipulationon bee health and behaviour

Once infection was confirmed nuclei were used in anexperiment designed to determine the effect of an antibioticon the clinical evolution of nuclei through a response totreatment trial (see below)

In summer 2006 and 2007 when the bees were foragingintensively dead bee samples were also collected from theground from a L latifolia crop situated 700 m from the apiar-ies 1 and 2 (Fig 4)

ii Experiment 1 was carried out in 10 nuclei (CN1ndashCN10)five of which (CN1ndashCN5) were treated with Fumidil B (fum-agillin ai 15 g colony-1 in a 250 cc syrup of sugardistilledwater 50 four treatments a week so a total of 120 mg offumagillin per hive) Nuclei CN6ndashCN10 were offered the

same syrup without the antibiotic (250 cc yen 4 applications aweek) and the treatments were applied from October 2006 toNovember 2006 Syrup with or without Fumidil B was placedin a plastic bag over the upper surface of brood combs andthe entire content was consumed by the bees in fewer than48 h All colonies were checked monthly

iii Experiment 2 The same administration regime as inExperiment 1 was used for the entire 50 colonies located inthree experimental apiaries to reproduce the real situation ofbeekeeping under field conditions The distance that sepa-rated the three apiaries was ~500 m (Fig 4) Nosema infec-tion was diagnosed in October 2006 and 18 N ceranae-parasitized colonies were treated with Fumidil B in November2006 while the other 32 colonies were left untreated as acontrol (13 N ceranae-positive 2 N apis- and N ceranae-positive and 17 microsporidia-negative)

Sample collection and determination of colony strength

Throughout the study forager and interior adult bee sampleswere taken monthly from hives and nuclei at 1200 at noonInitially the hive entrance was closed and forager beescoming from the outside were taken for 30 min Afterwards a

1 Apiary 1 Two colonies (B255 B111) and 14 nuclei

Apiary 2 Sixteen untreated colonies (8 infected and 8 negatives)

3 Apiary 3 Eighteen infected and treated colonies

2

444

p y g

4 Apiary 4 Sixteen untreated colonies (7 infected and 9 negatives)

3

1

2

33

11

22

Lavender crop

Clinical monitoringn=11

B255 and B111 colonies Apiary 1

Experiment 1n=10

Fumagillin treated (n=5) Apiary 1Control ( n=5) Apiary 1

Experiment 2 n=50

Fumagillin treated (n=18) Apiary 3Control infected colonies (n=15) Apiary 249 nuclei Apiary 1

Control uninfected colonies (n=17) Apiary 24

From May 05 to Dec 07 From June 06 to May 07 From Oct 06 to Nov 07

Colony strengthParasite determinations

PathologyPesticides

PalynologicalVirological

Colony strengthParasite determinations

PesticidesPalynological

Virological

Colony strengthParasite determinationsDay 0 +15+180 +365

Virological

5 of this colonies were used as control in Experiment 1

Fig 4 Diagram summarizing the distribution and location of colonies and nuclei between the different apiaries the timecourse of samplingand infection status of N ceranae

8 M Higes et al

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

sample of interior bees was taken Stored pollen samplesfrom the combs were also taken monthly from the nuclei andcolonies These samples were used for multi-residue andpalynological analysis in order to confirm the absence ofpesticides in foraged crops and the availability of pollenreserves Honey was only sampled monthly from hive B255

Samples of dead adult foraging honeybees were picked upfrom the ground in L latifolia crop

Colony strength was determined monthly by visually count-ing the number of combs covered with bees or brood

Parasite determination

All samples from hives and the nuclei were studied to detectthe presence of Nosema spores Malpinghamoeba cystV destructor and Acarapis woodi mites in forager andinterior bees according to Office International des Epizootiesmethods (OIE 2004) The percentage infection and the meanN ceranae spore burden were calculated in 30 interior beesand 30 foragers analysed individually and checked eachmonth (Doull and Cellier 1961 Pickard and El-Shemy 1989)After dissecting the bees the gut was placed in a sterileEpperdorf and 200 ml of ddH2O was added After thoroughgrinding (Eppendorf) a drop was visualized under a micro-scope to detect the presence of spores (Higes et al 2007)Afterwards the 30 samples were mixed homogenized cen-trifuged and the spore count obtained The remaining homo-genate was used for PCR analysis

Molecular diagnosis was performed monthly to identify thespecies in all Nosema-positive colonies The analysis wasperformed in B255 from the 25 May 2005 to collapse on the11 December 2006 and the nuclei were monitored monthlyfrom June 2006 to April 2007 Determinations in B111 startedin June 2006 until collapse day on the 13 December 2007Infection was confirmed by PCR in all cases (Higes et al2006 Martiacuten-Hernaacutendez et al 2007)

The identification of Nosema was carried out when a posi-tive spore count was identified except for the treated nucleiwhere it was carried out monthly even when spores were notidentified DNA was extracted from 500 ml of the remaininghomogenate after microscopy analysis of interior and forag-ers bees Samples were shaken with ceramic beads (MagNALyser Green Beads Roche 03 358941 001) in MagNA Lyser(Roche) at 9500 rpm for 95 s and the DNA extracted withMagNA Pure Compact Nucleic Acid Isolation Kit I (Roche 03730964 001) in a MagNA Pure Compact instrument (Roche)From May 2005 to June 2006 PCR was performed with NOSFORNOS REV primers and the products were subjected toSSUrRNA sequencing (Higes et al 2006 2007) Morerecent samples were analysed by multiplex PCR (Martiacuten-Hernaacutendez et al 2007)

Pathological studies

Samples for pathological studies were taken from B255 andon dying bees collected from the surrounding area Two adultinterior bees and two foragers were analysed monthly byoptical microscopy After dissecting out the gut the ventricu-lus was divided into sections fixed in 10 buffered formalinembedded in paraffin wax sectioned at 4 mm and stained

with haematoxylin and eosin A complete histopathologicalstudy was performed in each case The ventriculus fromanother three interior and forager bees from colony B255were also analysed monthly in semithin (05 mm) and TEMstudies (Higes et al 2007) Representative tissue areas forTEM analysis were selected after examining the semithinsections For thin and semi-thin sections an Olympus BX 50microscope with an Olympus television camera was usedconnected to a Phillips 50yen max Pentium computer with microImage Version 40 for Windows 98 image analysis softwareThe TEM images were obtained and photographed with aJeol 1010 electron microscope at an accelerating voltage of80ndash100 kV

Virological determinations

Virus analysis was performed on interior and forager bees byDr Ribiere (AFSSA) and Dr Sela (University of Jerusalem) asdescribed previously (Stoltz et al 1995 Hung 2000 Benjed-dou et al 2001 Grabensteiner et al 2001 Bakonyi et al2002 Blanchard et al 2007 Maori et al 2007) The samplesfor virological analysis were collected from the B255 colony inJuly 2006 October 2006 and December 2006 and on the dayof colony collapse Samples from B111 were collected inOctober 2006 June 2007 and November 2007 Samples fromuntreated nuclei were collected on the day when each nucleuscollapsed Samples from treated nuclei were also collected inApril 2007 on the same day when the last untreated nucleuscollapsed In the unmanaged control nuclei (NS1ndashNS4) IAPVwas only analysed when they collapsed

Palynological determinations

To confirm the type of foraging crop and the nutritive avail-ability of pollen in colonies (B255 and nuclei CN1ndashCN10)05 g of stored pollen was diluted in 10 ml of acidulated H2O(05 H2SO4) and 1 ml was taken for pollen extraction(Erdtman 1969) Species were identified using a photo-graphic atlas (Valdeacutes et al 1987 Faegri and Iversen 1989)and the reference collection of pollen slides from the CAR

Presence of pesticides

Honey pollen and bee samples from B255 and pollensamples from B111 and nuclei were stored at -20degC in darkrecipients until they were analysed All samples were sub-jected to a multi-residue analysis (Table S2) at the ChemicalLaboratory (University of Valladolid) following slight modifica-tions of previously reported methods Briefly pollen and beesamples were extracted with acetonitrile water was added tothe extract and the mixture was subjected to solid-phaseextraction on octadecyl silane cartridge (Jimeacutenez et al2007) Honey samples were diluted with methanol and solid-phase extraction on a Florisil-packed column was performed(Jimeacutenez et al 1998) Finally the analysis of residues in theextracts was carried out by gas chromatography with massspectrometric detection in combination with a matrix-matchedcalibration

Meteorological data

Meteorological data were obtained from the CREA-SIAR ofthe lsquoConsejeriacutea de Agricultura de la JCCMrsquo (httpcreauclm

Natural Nosema ceranae infection 9

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

sample of interior bees was taken Stored pollen samplesfrom the combs were also taken monthly from the nuclei andcolonies These samples were used for multi-residue andpalynological analysis in order to confirm the absence ofpesticides in foraged crops and the availability of pollenreserves Honey was only sampled monthly from hive B255

Samples of dead adult foraging honeybees were picked upfrom the ground in L latifolia crop

Colony strength was determined monthly by visually count-ing the number of combs covered with bees or brood

Parasite determination

All samples from hives and the nuclei were studied to detectthe presence of Nosema spores Malpinghamoeba cystV destructor and Acarapis woodi mites in forager andinterior bees according to Office International des Epizootiesmethods (OIE 2004) The percentage infection and the meanN ceranae spore burden were calculated in 30 interior beesand 30 foragers analysed individually and checked eachmonth (Doull and Cellier 1961 Pickard and El-Shemy 1989)After dissecting the bees the gut was placed in a sterileEpperdorf and 200 ml of ddH2O was added After thoroughgrinding (Eppendorf) a drop was visualized under a micro-scope to detect the presence of spores (Higes et al 2007)Afterwards the 30 samples were mixed homogenized cen-trifuged and the spore count obtained The remaining homo-genate was used for PCR analysis

Molecular diagnosis was performed monthly to identify thespecies in all Nosema-positive colonies The analysis wasperformed in B255 from the 25 May 2005 to collapse on the11 December 2006 and the nuclei were monitored monthlyfrom June 2006 to April 2007 Determinations in B111 startedin June 2006 until collapse day on the 13 December 2007Infection was confirmed by PCR in all cases (Higes et al2006 Martiacuten-Hernaacutendez et al 2007)

The identification of Nosema was carried out when a posi-tive spore count was identified except for the treated nucleiwhere it was carried out monthly even when spores were notidentified DNA was extracted from 500 ml of the remaininghomogenate after microscopy analysis of interior and forag-ers bees Samples were shaken with ceramic beads (MagNALyser Green Beads Roche 03 358941 001) in MagNA Lyser(Roche) at 9500 rpm for 95 s and the DNA extracted withMagNA Pure Compact Nucleic Acid Isolation Kit I (Roche 03730964 001) in a MagNA Pure Compact instrument (Roche)From May 2005 to June 2006 PCR was performed with NOSFORNOS REV primers and the products were subjected toSSUrRNA sequencing (Higes et al 2006 2007) Morerecent samples were analysed by multiplex PCR (Martiacuten-Hernaacutendez et al 2007)

Pathological studies

Samples for pathological studies were taken from B255 andon dying bees collected from the surrounding area Two adultinterior bees and two foragers were analysed monthly byoptical microscopy After dissecting out the gut the ventricu-lus was divided into sections fixed in 10 buffered formalinembedded in paraffin wax sectioned at 4 mm and stained

with haematoxylin and eosin A complete histopathologicalstudy was performed in each case The ventriculus fromanother three interior and forager bees from colony B255were also analysed monthly in semithin (05 mm) and TEMstudies (Higes et al 2007) Representative tissue areas forTEM analysis were selected after examining the semithinsections For thin and semi-thin sections an Olympus BX 50microscope with an Olympus television camera was usedconnected to a Phillips 50yen max Pentium computer with microImage Version 40 for Windows 98 image analysis softwareThe TEM images were obtained and photographed with aJeol 1010 electron microscope at an accelerating voltage of80ndash100 kV

Virological determinations

Virus analysis was performed on interior and forager bees byDr Ribiere (AFSSA) and Dr Sela (University of Jerusalem) asdescribed previously (Stoltz et al 1995 Hung 2000 Benjed-dou et al 2001 Grabensteiner et al 2001 Bakonyi et al2002 Blanchard et al 2007 Maori et al 2007) The samplesfor virological analysis were collected from the B255 colony inJuly 2006 October 2006 and December 2006 and on the dayof colony collapse Samples from B111 were collected inOctober 2006 June 2007 and November 2007 Samples fromuntreated nuclei were collected on the day when each nucleuscollapsed Samples from treated nuclei were also collected inApril 2007 on the same day when the last untreated nucleuscollapsed In the unmanaged control nuclei (NS1ndashNS4) IAPVwas only analysed when they collapsed

Palynological determinations

To confirm the type of foraging crop and the nutritive avail-ability of pollen in colonies (B255 and nuclei CN1ndashCN10)05 g of stored pollen was diluted in 10 ml of acidulated H2O(05 H2SO4) and 1 ml was taken for pollen extraction(Erdtman 1969) Species were identified using a photo-graphic atlas (Valdeacutes et al 1987 Faegri and Iversen 1989)and the reference collection of pollen slides from the CAR

Presence of pesticides

Honey pollen and bee samples from B255 and pollensamples from B111 and nuclei were stored at -20degC in darkrecipients until they were analysed All samples were sub-jected to a multi-residue analysis (Table S2) at the ChemicalLaboratory (University of Valladolid) following slight modifica-tions of previously reported methods Briefly pollen and beesamples were extracted with acetonitrile water was added tothe extract and the mixture was subjected to solid-phaseextraction on octadecyl silane cartridge (Jimeacutenez et al2007) Honey samples were diluted with methanol and solid-phase extraction on a Florisil-packed column was performed(Jimeacutenez et al 1998) Finally the analysis of residues in theextracts was carried out by gas chromatography with massspectrometric detection in combination with a matrix-matchedcalibration

Meteorological data

Meteorological data were obtained from the CREA-SIAR ofthe lsquoConsejeriacutea de Agricultura de la JCCMrsquo (httpcreauclm

Natural Nosema ceranae infection 9

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

es~siardatmeteoresu_histphp) Monthly rainfall and tem-peratures were recorded by standard procedures

Statistical analysis

Pearson correlations (two-tailed P lt 001 level) wereobtained to determine the relationship between the percent-age infection and the mean spore counts in forager andinterior bees the brood combs the combs with bees and themeteorological data in B255 colony A principal componentsanalysis for categorical data CATPCA was used to deter-mine the relationship between multiple parameters derivedfrom measurements on B255 and the variables were trans-formed with an optimal scaling process and subsequentlycorrelated The relationships between the data are repre-sented in a component loading graph A multiple linearregression analysis was performed in which the choice ofpredictive variables was carried out automatically selectingthe best-fit model The relationships were establishedbetween a criterion variable (frames full of bees and broodcombs) and a set of predictors (percentage infection of inte-rior and forager bees temperatures and monthly rainfall)Statistical analyses were performed using SPSS 1501 andSAS 91 for Windows XP

To compare colonies from the second response trial datawere grouped as uninfected colonies (n = 17) infected andtreated (n = 18) and infected and untreated (n = 15) at themoment the trial commenced To study the dependence ofthese groups upon bee health the statistical analysis wascross-tabulated Chi-squared tests (P lt 001) and a corre-spondence analysis were used to calculate the exactprobabilities To perform these analyses a variable related tobee health at the end of the trial was established with differentlevels

Acknowledgements

We thank M Ribiegravere F Schurr and I Sela for the virologicalanalysis F A Rojo Vaacutezquez N Azpiazu A Alonso C delAacuteguila L Prieto and M Spivak for reviewing the manuscriptML Garciacutea and A Fernaacutendez for image-processing RM-Hand AVG-P were co-financed by the Junta de Comu-nidades de Castilla-La Mancha (JCCM) and INIA-EuropeanSocial Fund The work reported here was supported byJCCM and Ministerio de Agricultura Pesca y Alimentacioacuten(APIFEGA-MAPYA FOUNDS)

References

Amdam GA and Omholt W (2003) The hive bee to foragertransition in honeybee colonies the double repressorhypothesis J Theor Biol 223 451ndash464

Bakonyi T Frakas R Szendroi A Dobos-Kovacs M andRusvai M (2002) Detection of acute bee paralysis virus byRT-PCR in honeybee and Varroa destructor field samplesrapid screening of representative Hungarian apiaries Api-dologie 33 63ndash74

Benjeddou M Leat N Allsopp M and Davison S (2001)Detection of acute paralysis virus and black queen cell

virus from honeybees by reverse trasciptase PCR ApplEnviron Microbiol 67 2384ndash2387

Blanchard P Ribiere M Celle O Lallemand P SchurrF Olivier V et al (2007) Evaluation of a real-time two-step RT-PCR assay for quantitation of Chronic bee paraly-sis virus (CBPV) genome in experimentally-infected beetissues and in life stages of a symptomatic colony J VirolMethods 141 7ndash13

Chauzat MP Higes M Martiacuten-Hernaacutendez R Meana ACougoule N and Faucon JP (2007) Presence ofNosema ceranae in French honeybee colonies J Apic Res46 127ndash128

Chen Y Pettis JS Collins AM and Feldlaufer MF(2006) Prevalence and transmission routes of honey beeviruses App Environ Microbiol 72 606ndash611

Chen Y Evans JD Smith IB and Pettis JS (2008)Nosema ceranae is a long-present and wide-spreadmicrosporidian infection of the European honey bee (Apismellifera) in the United States J Invertebr Pathol 97 186ndash188

Chen YP and Siede R (2007) Honey bee virus Adv VirusRes 70 33ndash80

Cox-Foster DL Conlan S Holmes E Palacios GEvans JD Moran NA et al (2007) A metagenomicsurvey of microbes in honey bee colony collapse disorderScience 318 283ndash287

Doull KM and Cellier KM (1961) A survey of the inci-dence of Nosema disease (Nosema apis Zander) of thehoney bee in South Australia J Insect Pathol 3 280ndash288

El-Shemy AAM and Pickard RS (1989) Nosema apisZander infection levels in honeybees of known age J ApicRes 28 101ndash106

Erdtman G (1969) Handbook of Palynology An Introductionto the Study of Pollen Grains and Spores CopenhagenDenmark Munksgaard

Faegri K and Iversen J (1989) Textbook of Pollen AnalysisIV Edition Faegri K Kaland PE Krzywinski K (eds)Chichester USA John Wiley amp Sons

Faucon JP Mathieu L Ribiere M Martel AC Drajnu-del P Zeggane S et al (2002) Honey bee wintermortality in France in 1999 and 2000 Bee World 8313ndash23

Gonzaacutelez R (2007) Crisis Apiacutecola Argentina se estima maacutesde 1 450 000 colmenas muertas [WWW document] URLhttpwwwnoticiasapicolascomar

Grabensteiner E Ritter W Carter MJ Davison S Pech-hacker H Kolodziejek J et al (2001) Sacbrood virus ofthe honeybee (Apis mellifera) rapid identification and phy-logenetic analysis using reverse transcription-PCR ClinDiagn Lab Immunol 8 93ndash104

Hassanein MH (1953) The influence of infection withNosema apis on the activities and longevity of the workerhoneybee Ann Appl Biol 40 418ndash423

Higes M Martiacuten R Sanz A Alvarez N Sanz A GarciaMP and Meana A (2005) El siacutendrome de despo-blamiento de las colmenas en Espantildea Consideracionessobre su origen Vida Apiacutecola 133 15ndash21

Higes M Martiacuten R and Meana A (2006) Nosemaceranae a new microsporidian parasite in honeybees inEurope J Invertebr Pathol 92 93ndash95

Higes M Garciacutea-Palencia P Martiacuten-Hernaacutendez R and

10 M Higes et al

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology

Meana A (2007) Experimental infection of Apis melliferahoneybees with the Microsporidia Nosema ceranaeJ Invertebr Pathol 94 211ndash217

Huang ZY and Robinson GE (1996) Regulation of honeybee division of labor by colony age demography BehavEcol Sociobiol 39 147ndash158

Hung ACF (2000) PCR detection of Kashmir bee virus inhoney bee excreta J Apic Res 39 103ndash106

Jimeacutenez JJ Bernal JL del Nozal MJ Toribio L andMartiacuten MT (1998) Gas chromatography with electron-capture and nitrogen-phosphorus detection in the analysisof pesticides in honey after elution from a Florisil columnInfluence of the honey matrix on the quantitative resultsJ Chromatogr A 823 381ndash387

Jimeacutenez JJ Bernal JL del Nozal MJ Martiacuten MT andMayo R (2007) Comparative study of sample preparationprocedures to determine fipronil in pollen by gas chroma-tography with mass spectrometric and electron capturedetection J Chromatogr A 1146 8ndash16

Klee J Besana AM Genersch E Gisder S Nanetti ATam DQ et al (2007) Widespread dispersal of theMicrosporidian Nosema ceranae an emergent pathogen ofthe western honey bee Apis mellifera J Invertebr Pathol96 1ndash10

Kralj J and Fuchs S (2006) Parasitic Varroa destructormites influence flight duration and homing ability of infestedApis mellifera foragers Apidologie 37 577ndash587

Laurent FM and Rathahao E (2003) Distribution of(C-14)imidacloprid in sunflowers (Helianthus annuus L)following seed treatment J Agric Food Chem 51 8005ndash8010

Liu TP (1984) Ultrastructure of the midgut of the workerhoney bee Apis mellifera heavily infected with Nosemaapis J Invertebr Pathol 44 282ndash291

Maori E Tanne E and Sela I (2007) Reciprocal sequenceexchange between non-retro viruses and hosts leading tothe appearance of new host phenotypes Virology 362342ndash349

Martiacuten-Hernaacutendez R Meana A Prieto L Martiacutenez Salva-dor A Garrido-Bailoacuten E and Higes M (2007) Outcomeof colonization of Apis mellifera by Nosema ceranae ApplEnviron Microbiol 73 6331ndash6338

Molga P (2008) La mort des abeilles met la planegravete endanger Les echos [WWW document] URL httpwwwlesechosfrinfoenergie4611614htm

Office International des Epizooties (OIE) (2004) Manual ofStandards for Diagnostic Test and Vaccines [WWWdocument] URL httpwwwoieintengnormesmmanualA_00123htm

Ortiz AM (2005) la Extrantildea Muerte de Las Abejas [WWWdocument] URL httpwwwelmundoespapel20050320cronica1772969html

Paxton R Klee J Korpela S and Fries I (2007) Nosemaceranae has infected Apis mellifera Europe in at least 1998and may be more virulent than Nosema apis Apidologie38 558ndash565

Pickard RS and El-Shemy AAM (1989) Seasonal varia-

tion in the infection of honeybee colonies with Nosema apisZander J Apic Res 28 93ndash100

Stokstad E (2007) The case of empty hives Science Maga-zine [WWW document] URL httpwwwsciencemagorgcgireprint3165827970pdf DOI 101126science3165827970

Stoltz D Shen XR Boggis C and Sisson G (1995)Molecular diagnosis of Kashmir bee virus infection J ApicRes 34 153ndash160

Valdeacutes B Diez MJ and Fernaacutendez I (1987) AtlasPoliacutenico de Andaluciacutea Occidental Sevilla Spain Univer-sidad de Sevilla (eds)

Supplementary material

The following supplementary material is available for thisarticle onlineFig S1 Component loading graph of measurements on theB255 colony Correlated transformed variables matrix inter-pretation is angle 0deg related with absolute positive depen-dency angle 180deg with absolute negative dependency andangle 90deg with independenceFig S2 Experiment 2 Component analysis of treated anduntreated hives of groups of treatment (NEG free ofinfection POS + TREATED infection and treated POS +UNTREATED infection and non-treated) and bee healthdegrees (Phase 1 Phase 2 Phase 3 Phase 4 Phase 5_d1colony breakdown in 121506 Phase 5_d2 colony break-down in 042607 Phase 5_d3 colony breakdown in 092607 Phase 5_d4 colony breakdown in 111107)Table S1 A Average N ceranae spore counts in treated(CN1ndashCN5) and untreated (CN6ndashCN10) nucleiB Percentage of N ceranae-infected bees in treated (CN1ndashCN5) and untreated (CN6ndashCN10) nucleiTable S2 Compounds studied and detection limits (LOD) bymulti-residue analysis None of them were detected in anystored pollen samples from B225 and the nuclei analysedAdult bee and honey analysed from B255 were alsonegativesTable S3 Palynological analysis of stored pollen collectedfrom cells Percentage of taxa present in stored pollen Thepercentage pollen in the nuclei was calculated as the averageand standard deviation of the nuclei CN1ndashCN10 (treated anduntreated) Spring stored pollen samples were collected inJune 2006 Collapse samples from B255 on the 11 December2006 and untreated nuclei (CN6ndashCN10) from February toApril 2007 on the collapse date Pollen samples stored intreated nuclei (CN1ndashCN5) were collected when the lastuntreated nuclei was found dead (April 2007)

This material is available as part of the online article fromhttpwwwblackwell-synergycom

Please note Blackwell Publishing is not responsible forthe content or functionality of any supplementary materialssupplied by the authors Any queries (other than missingmaterial) should be directed to the corresponding authorfor the article

Natural Nosema ceranae infection 11

copy 2008 The AuthorsJournal compilation copy 2008 Society for Applied Microbiology and Blackwell Publishing Ltd Environmental Microbiology