Study of the depletion of tylosin residues in honey extracted from treated honeybee (Apis mellifera) colonies and the effect of the shook swarm procedure

Abstract

Bee colonies were dosed with tylosin tartrate 1.1 g per hive (single dose in sucrose solution) and samples of honey were then
collected at intervals over a 20-week period. The samples were analysed for tylosin A and desmycosin (tylosin B) using LC-MS/MS.
The mean concentration of tylosin A in the honey (pooled results) 3 days after dosing was 17 μg/g, declining to 0.9 μg/g after
140 days. The mean concentration of desmycosin was 2.3 μg/g, 3 days after dosing declining to 1.1 μg/g after 140 days. The
shook swarm procedure was investigated and resulted in a tylosin A concentration in brood honey of 10 μg/g, 3 days after dosing
declining to 0.02 μg/g, 140 days after dosing. A corresponding decrease in the mean concentrations of desmycosin in brood
honey, 1.1 fxg/g, 3 days after dosing to 0.03 [μg/g, 140 days after dosing also was observed. Tylosin A depletes to desmycosin
in honey and can still be detected 238 days after dosing. Thus a more accurate residue definition is the sum of tylosin A
and desmycosin.

Tylosin wurde kürzlich in den USA für die Bekämpfung der Amerikanischen Faulbrut in Bienenvölkern zugelassen und stellt somit
ein alternatives Antibiotikum zu Oxytetracyclin dar. Allerdings sind nach EU-Bestimmungen Tylosinrückstände in Honig nicht
erlaubt und Honige aus den USA mit Tylosinrückständen wären auf dem EU-Markt nicht verkehrsfähig. Daher wurde hier die Beziehung
zwischen Tylosin A und dem Abbauprodukt Desmykosin untersucht. Damit sollte eine Markersubstanz etabliert werden, um den Abbau
von Tylosin im Bienenvolk zu erfassen und die Verwendung dieses Wirkstoffes in der Imkerei nachzuweisen.

Bienenvölkern wurde eine Dosis von 1,1g Tylosintartrat pro Volk in Form einer einmaligen Zuckerlösung gegeben. Die Proben
wurden vor der Futtergabe und danach über 20 Wochen in regelmäßigen Abständen und schließlich nach der Überwinterung gezogen.
Die Proben wurden über HPLC-MS auf Tylosin A und Desmycosin analysiert.

Die Konzentration an Tylosin A im Honig nahm im Zeitraum der Probennahmen kontinuierlich ab: Von 17 [μg/g 3 Tage nach Applikation
über 3,3 μg/kg 56 Tage danach bis auf 0,9 μg/kg 140 Tage danach. Die Konzentration von Desmycosin nahm lediglich von 2,3 μg/kg
3 Tage nach Applikation auf 1,1 μg/kg 140 Tage danach ab (Abb. 1, Tab. III und IV). Es gibt einen raschen Abbau an Tylosin
A während der ersten 28 Tage nach Applikation gefolgt von einer geringeren Abbaurate danach. Trotz der Abnahme an Tylosin
bleibt die Konzentration an Desmycosin weitgehend konstant, vermutlich wegen einer kontinuierlichen Umwandlung von Tylosin
A zu Desmycosin. Die Konzentrationen von Desmycosin und Tylosin A gleichen sich 84 Tage nach der Applikation an (Abb. 2).

Die Kunstschwarmbildung auf neues Wabenwerk 7 Tage nach der Applikation reduzierte die Rückstandskonzentrationen von Tylosin
A und Desmycosin um den Faktor 30 zum Ende der Probennahme (140 Tage nach Applikation). Tab. V zeigt vergleichend die Abnahme
der Rückstandskonzentrationen für behandelte Bienenvölker mit und ohne Kunstschwarmbildung. Nach Applikation von Tylosin können
Rückstände auch 238 Tage danach noch nachgewiesen werden selbst wenn zwischenzeitlich die Kunstschwarmmethode angewendet wird.
Tylosin A ist eine geeignete Markersubstanz um den Gebrauch bzw. Missbrauch von Tylosin nachzuweisen. Eine exakte Bestimmung
von Tylosinrückständen sollte allerdings über die Summe von Tylosin A und Desmycosin erfolgen.

Tylosin–desmycosin–honey–veterinary drug–residues–apiculturetylosine–desmycosine–antibiotique–résidu–miel–technique apicoleTylosin–Desmycosin–Honig–Rückstände–Tierarzneimittel–Imkerei

Full text

Page 1

Apidologie 38 (2007) 315–322
c ? INRA/DIB-AGIB/ EDP Sciences, 2007
DOI: 10.1051/apido:2007017
Available online at:
www.apidologie.org
Original article
Study of the depletion of tylosin residues in honey extracted
from treated honeybee (Apis mellifera) colonies
and the effect of the shook swarm procedure*,**
Stuart J. Aa, Katharina Ha, Mike Ha, Richard J. Fa,
Selwyn Wb, Helen M. Tb, Matthew Sa
aFood Safety Group, Central Science Laboratory, Sand Hutton, York, YO41 1LZ, UK
bNational Bee Unit, Central Science Laboratory, Sand Hutton, York, YO41 1LZ, UK
Received 5 July 2006 – Revised 7 December 2006 – Accepted 22 December 2006
Abstract – Bee colonies were dosed with tylosin tartrate 1.1 g per hive (single dose in sucrose solution) and
samples of honey were then collected at intervals over a 20-week period. The samples were analysed for
tylosin A and desmycosin (tylosin B) using LC-MS/MS. The mean concentration of tylosin A in the honey
(pooled results) 3 days after dosing was 17 µg/g, declining to 0.9 µg/g after 140 days. The mean concentra-
tion of desmycosin was 2.3 µg/g, 3 days after dosing declining to 1.1 µg/g after 140 days. The shook swarm
procedure was investigated and resulted in a tylosin A concentration in brood honey of 10 µg/g, 3 days after
dosing declining to 0.02 µg/g, 140 days after dosing. A corresponding decrease in the mean concentrations
of desmycosin in brood honey, 1.1 µg/g, 3 days after dosing to 0.03 µg/g, 140 days after dosing also was
observed. Tylosin A depletes to desmycosin in honey and can still be detected 238 days after dosing. Thus
a more accurate residue definition is the sum of tylosin A and desmycosin.
Tylosin / desmycosin / honey / veterinary drug / residues / apiculture
1. INTRODUCTION
Tylosin along with other macrolides com-
prise a group of antibacterial compounds that
have a wide range of applications in the field
of veterinary medicine, including use as ther-
apeutic agents and as growth-promoting an-
tibiotics. In therapeutic applications tylosin
is used to control certain Gram-positive bac-
teria, mycoplasma and Gram-negative bacte-
ria (EMEA, 1997). Tylosin, along with lin-
comycin, erythromycin and monensin, has
been identified as an effective treatment in
oxytetracycline (OTC) resistant strains of
AmericanFoulbrood(AFB) (Kochanskyet al.,
2001). In October 2005 Tylan Soluble (tylosin
Corresponding author: S.J. Adams,
s.adams@csl.gov.uk
*Manuscript editor: Jean-Noël Tasei
**British Crown Copyright 2007
tartrate) was approved in the USA to control
AFB (FDA, NADA 013-076, 2005).
Current EU legislation does not permit the
use of tylosin or any other antibiotics in bees
for the treatment of either European or Ameri-
can Foulbrood. Residues of tylosin at concen-
trations up to 0.006 µg/g have been detected
in retail samples of honey analysed in the
2004 UK Non-Statutory Surveillance scheme
(VRC Annual Report, 2004). The Canadian
Food Inspection Agency (CFIA) has reported
the presence of tylosin residues in honey,
in the range of 0.0012–0.1156 µg/g (CFIA,
2003−2004 survey). The CFIA currently has
a working residue limit (WRL) of 0.06 µg/g to
account for ‘extra label’ usage (CFIA, 2005).
Recently published methods for the analysis
of tylosin in honey have been established with
limits of quantification as low as 0.0025 µg/g
for a biosensor screening approach (Cal-
dow et al., 2005) and 0.0005 µg/g for a
LC-MS/MS (liquid chromatography coupled
Article published by EDP Sciences and available at http://www.apidologie.org
or http://dx.doi.org/10.1051/apido:2007017

Page 2

316S.J. Adams et al.
to mass spectrometry) confirmation procedure
(Wang, 2004).A method for the determination
of tylosin based on the measurement of zones
of inhibition of the microbial activity has also
been reported (Feldlaufer et al., 2004).
The main purpose of this study was to in-
vestigate the depletion of tylosin A in a hive
system and to ascertain if there is a relation-
ship between tylosin A and desmycosin con-
centrations detected in the collected honey.
Kochansky previously identified desmycosin
as a degradation product of tylosin in honey
(Kochansky, 2004). Therefore it is important
to include the determination of both tylosin
A and desmycosin when monitoring for the
(mis)use of this drug in honey.
Additionally the shook swarm procedure
was investigated as a possible method to re-
duce residue concentrations after treatment
(dosing). Thompson et al. reported the use of
the shook swarm technique to successfully re-
duceconcentrationsofOTCinhoneycollected
from dosed bee colonies. Therefore one of the
aims of the project was to determine the per-
sistence of tylosin and desmycosin with and
without the application of the shook swarm
procedure.
2. METHODS AND MATERIALS
2.1. Bee colonies and treatments
The dosing study was carried out in June
2005–February 2006. Eight standardised free fly-
ing colonies of UK honeybees (Apis mellifera
L.), housed in double Smith brood boxes with 11
British standard frames (33.6 cm by 20.4 cm giv-
ing 685.4 cm2per side of brood frame) per brood
box and at least one super box, with 18–20 frames
of bees in each colony, were used in this study. The
colonies were maintained and owned by the Cen-
tral Science Laboratory (CSL), National Bee Unit.
At the start of the trial these colonies showed no
clinical signs of European or American foulbrood,
sacbrood or baldbrood and had only a low inci-
dence of chalkbrood. Six colonies were treated with
tylosin and were located at an experimental api-
ary approximately 10 km from two undosed con-
trol colonies that were established in parallel at the
CSL site. This was to reduce the risk of cross-
contamination by drifting.
The six treated colonies were dosed with a so-
lution of 1.1 g of tylosin tartrate in 200–250 mL
aqueous sucrose solution (50–60% w/v) by pouring
into the marked top empty brood frame. The treat-
ment comb was placed in the top brood box, two
framesin(usuallyontheedge ofthebrood nest with
the treated side of the frame out). The two control
colonies were fed with untreated sucrose using the
same method of application. Seven days after dos-
ing and honey sampling, two of the treated colonies
were randomly selected and shook swarmed. The
shook swarm treatment involved the transfer of the
adult bees onto clean foundation with the brood and
original frames being removed and destroyed. Dur-
ing winter i.e. after October sampling, the colonies
were fed with 50% w/v sucrose using a rapid tray
feeder
2.2. Sampling
Table I outlines the sampling time points.
In June 2005, two to four days before treatment
(D-2 to D-4) samples of up to 100 g of nectar/honey
were taken from each colony to establish a base-
line residue concentration for the colonies, i.e. to
confirm antibiotic residues were not present. Sam-
ples of brood honey were collected at eight different
time points during the bee keeping season. Samples
of super honey were not available for all colonies
on D3, D58, D84 and D238. Post-wintering sam-
ples were collected in February 2006, at D238.
Each sampling day, four comb samples (approx-
imately 8 cm by 10 cm) were taken from each hive,
two from the brood chamber and two from the su-
per. The four samples were taken from different
frames in the hive. For each colony these samples
were bulked as super sources and brood chamber
sources, except at D28 when each of the individ-
ual samples were extracted separately to enable an
assessment of the distribution of tylosin within each
colony. Thehoney/nectar sampleswereextractedby
filtering the sample through cloth into a clean con-
tainer. All samples were stored at –20◦C prior to
analysis.
At the end of the trial bees from all treated
colonies were shaken onto new foundation and the
brood combs and super combs incinerated.
2.3. Apparatus and reagents
Tylosin tartrate (cell culture tested) was pur-
chased from Sigma Aldrich (Dorset, UK) contain-
ing desmycosin 4% by weight. OASIS HLB SPE
cartridges (6 mL/200 mg) were purchased from
Waters (Manchester, UK). All other reagents were
of analytical grade and obtained from either BDH

Page 3

Depletion of tylosin residues in honey317
Table I. Honey sampling plan from the 8 experimental hives.
Dosed hives
without shook
swarm procedure
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Dosed hives
with shook swarm
procedure
✓
✓
✓
✗
✗
✗
✓
✗*
✓
✓
Time point from dosing
(days)
−4 to −2 (baseline sample)
3
7
14
21
28
56
84
140
238 (over winter)
control hives
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓ = sampling,
✗ = no sampling,
*= insufficient honey available.
(Poole, UK) or Fisher Scientific (Loughborough,
UK). Phosphate buffer (pH 7) was prepared by dis-
solving 13.6 g dipotassium hydrogen orthophos-
phate and 4.0 g potassium dihydrogen orthophos-
phate in 1 litre of water.
2.3.1. Preparation of standard solutions
A stock standard of tylosin A (1 mg/mL) was
prepared in methanol with a shelf life of one month.
Desmycosin was prepared immediately before use
from Tylosin A. A small volume (approximately
200 µL) of formic acid (98%) in water (9.7 mL)
with a pH <4 was added to 100 µL of tylosin stock
standard. The solution was then transferred to a
50 mL falcon tube and placed in a water bath at
50◦C overnight to hydrolyze tylosin A to desmy-
cosin.
2.4. Extraction of control samples
and samples containing residues
at low concentration
Phosphate buffer (10 mL) was added to the sam-
ple (2 g) and placed in a water bath (45◦C) for
10 min. The sample was vortex mixed to dissolve
the honey and allowed to cool to room temperature
(20◦C).
The extract (10 mL) was cleaned up using
an OASIS HLB SPE cartridge; conditioned with
methanol (5 mL) and extraction buffer (5 mL). The
extract was loaded onto the cartridge, which then
was washed with 40% methanol in water (5 mL).
Tylosin A and desmycosin were eluted with ace-
tonitrile (5 mL). The eluate was evaporated to dry-
ness at 40–50◦C under a stream of nitrogen and
reconstituted in 50% methanol in water (1 mL).
Matrix extracted calibration standards (prepared by
spiking honey with tylosin A and desmycosin and
then taken through the extraction procedure) were
prepared over the range 0.001–0.3 µg/g for tylosin
A and 0.01–0.25 µg/g for desmycosin. Method val-
idation data is shown in Table II. Batch recovery
samples spiked with Tylosin A and desmycosin,
gave recoveries in the range 75% to 101% and 69%
to 131%, respectively.
2.5. Dilution of samples containing
residues at high concentration
Honey samples (1.25 g) were dissolved with
H2O (25 mL). An aliquot (0.1 mL) was mixed with
50% methanol in water (0.9 mL). Matrix matched
calibration standards (honey solution spiked with
tylosin A and desmycosin after dilution) were pre-
pared in the range, 0.2–200 µg/g for tylosin A
and 0.2–20 µg/g for desmycosin. The results from
several validation batches gave satisfactory results;
these data are shown in Table II.
2.6. Quantification
The LC-MS/MS system comprised a Quattro
Ultima Platinum Triple Quadrupole (Micromass,
Manchester, UK) coupled to an Alliance 2695 Sep-
arations Module (Waters) controlled by Mass Lynx
version 4.0. An isocratic separation was performed

Page 4

318S.J. Adams et al.
Table II. Method validation information for tylosin A and desmycosin analysed during this study.
Tylosin A Desmycosin
Measured
concentration
(µg/g)
NA
NA
6.1
1.2
0.5
0.18
0.046
NA
SpikeExtraction
method
No. of
replicates
Measured
concentration
(µg/g)
18.5
10.2
4.7
1
0.4
0.18
0.046
0.004
RSD RSD
concentration
(µg/g)
20
10
5
1
0.5
0.2
0.05
0.004
Dilution
Dilution
Dilution
Dilution
Dilution
SPE
SPE
SPE
7
7
7
7
7
7
7
21
6 NA
NA
1.9
3.3
3.6
3.5
5.1
NA
4.8
2.7
3.7
4.3
2
1.4
6.8
SPE = Solid Phase Extraction,
RSD = Relative Standard Deviation,
NA= method not validated at this concentration.
Table III. Results from 4 non shook swarmed colonies; tylosin A residue concentrations obtained from the
34-week tylosin tartrate dosing study (n = number of honey samples analysed from all hives per time point).
Time point
(days from
dosing)
D-4 - D-2
(Baseline)
D3
D7
D14
D21
D28
D56
D84
D140
D238 (over winter)
Mean of pooled results
for brood and super
honey µg /g ± SD
Brood honey mean
residue µg/g (n) ± SD
Super honey mean
residue µg/g (n) ± SD
< 0.006
17 ± 15
12 ± 6.2
9.9 ± 6.6
7.4 ± 6.1
6.1 ± 4.7
3.3 ± 2.3
1.7 ± 1.6
0.92 ± 0.84
0.93 ± 0.77
< 0.004 (4)
8.0 (4) ± 3.7
11 (4) ± 7.3
10 (4) ± 9.3
8.4 (4) ± 8.2
4.8 (8) ± 4.3
3.4 (4) ± 3.0
1.7 (4) ± 1.5
0.65 (4) ± 0.65
0.45 (4) ± 0.26
< 0.002 (2)
29 (3) ± 17
13 (4) ± 5.8
9.5 (4) ± 4.0
6.3 (4) ± 4.3
7.4 (8) ± 5.1
3.2 (3) ± 1.7
1.9 (3) ± 2.2
1.2 (4) ± 1.6
1.4 (3)± 0.93
on a Waters Sunfire C18 (100 mm × 2.1 mm, parti-
cle size 3.5 µm) with a C18 guard column installed.
The mobile phase was 10 mM ammonium ac-
etate in water, 0.5% formic acid in water/methanol,
40/10/50 v/v/v with a flow rate of 0.2 mL/min and
an injection volume of 10 µL. The LC-MS/MS was
operated in electrospray in the positive ion mode
and the transitions monitored were for Tylosin A,
916.5 > 156, 916.5 > 174 and 916.5 > 772.4;
for desmycosin 772.4 > 132, 772.4 > 156 and
772.4 > 174. The limit of quantification for ty-
losin A was 0.001 µg/g and for desmycosin was
0.01 µg/g.
3. RESULTS
The results obtained for tylosin A and
desmycosin for the brood and super honey
are shown in Tables III and IV respectively.
The results obtained for tylosin A D0 con-
trol samples were all <0.001 µg/g. Subse-
quent results for tylosin A in control sam-
ples were all <0.01 µg/g. The control samples
werenotscreenedfordesmycosin.Thehighest
mean concentration of tylosin A found in su-
per honey was 29 µg/g, at D3 and for brood
honey 11 µg/g, at D7. The mean concentra-
tions for tylosin A detected in both brood and
super honey were similar until D84. The mean
concentration for tylosin A in super honey
at D84 was 1.9 µg/g and 1.4 µg/g at D238,
whilst the mean concentration for tylosin A
in brood honey was 1.7 µg/g at D84 and
0.4 µg/g at D238. The highest mean concen-
tration for desmycosin found in super honey

Page 5

Depletion of tylosin residues in honey319
Figure 1. Results from 4 non shook swarmed
colonies; depletion of tylosin A and occurrence of
desmycosin in honey (pooled brood and super re-
sults) 0 to 238 days; after dosing with tylosin tar-
trate.
was 3.7 µg/g, at D3 and for brood honey
2.5 µg/g, at D14. Desmycosin concentrations
were 1.4 µg/g in super honey and 0.6 µg/g in
brood honey, at D238.
The results from brood honey and super
honey collected from the dosed hives were av-
eraged and summarized in Figure 1 (honey
is harvested and bulked from several colonies
prior to bottling for human consumption). The
highest mean (results for the brood and super
honeypooledtogether)concentration,17µg/g,
of tylosin A was detected in samples col-
lected at D3, at D7 was 12 µg/g decreasing to
1.7 µg/g at D84 and 0.9 µg/g at D140 with no
further reduction at D238. The mean (results
forthebroodandsuperhoneypooledtogether)
concentrations of desmycosin remained rela-
tively constant over the period from D3 to
D238.
The desmycosin to tylosin A ratio using
molar concentrations is summarised in Fig-
ure 2. These results show that the ratio of
desmycosin to tylosin A increases at a con-
stant rate over the first 140 days of the dosing
study with the molar concentration of desmy-
cosin becoming equal to tylosin A at 84 days.
The effects of the shook swarm proce-
dure on the concentration of tylosin detected
within a colony’s honey stores are summa-
rized in Table V. Prior to the shook swarm
procedure the initial concentrations of tylosin
A and desmycosin were similar in all dosed
colonies. After the procedure the mean con-
centrations of tylosin A and desmycosin at
D56 were approximately 10 times lower than
Figure 2. Molar concentration ratio of desmycosin
to tylosin A (pooled brood and super results), after
dosing with tylosin tartrate, error bars = S.D.
colonies that did not undergo the procedure.
The over-winter samples from the non shook
swarm and shook swarm colonies taken at
D238 contained detectable residues of tylosin
A (0.65 µg/g and 0.13 µg/g respectively) and
desmycosin (1.1 µg/g and 0.20 µg/g respec-
tively).
The variation of concentrations of tylosin
A and desmycosin in samples taken from dif-
ferent points within the hive are presented in
Table VI. Analysis of the individual samples
confirms that there is variability within the
hives and across hives.
4. DISCUSSION
The degradation of tylosin A to desmy-
cosin in mild acidic conditions (< pH 4) is
thought to be relevant to honey, which has a
pH, range of 3.4–6.1 (Kochansky, 2004). The
rapid depletion of tylosin A in the 28 day
period after dosing can be attributed to di-
lution into the honey and distribution around
the colony as well as degradation to desmy-
cosin. The desmycosin concentration during
this period remains constant as displayed in
Figure 1 whilst it is clearly evident that the
tylosin A concentration is declining. The fact
that the concentration of desmycosin remains
constant indicates that there is some conver-
sion of tylosin A to desmycosin. Figure 2
shows an increasing molar concentration of
desmycosin in the honey relative to tylosin A
and equality of concentrations of both com-
pounds at D84. As the dilution factor is con-
stant for both compounds the equalisation of
concentrations must be due to the conversion