Animal Reproduction Science 145 (2014) 47–53

Contents lists available at ScienceDirect

Animal Reproduction Science

journal homepage: www.elsevier.com/locate/anireprosci

Removal of bacteria from stallion semen by colloid

centrifugation
J.M. Morrell a,∗ , C. Klein b , N. Lundeheim a , E. Erol c , M.H.T. Troedsson b
a
Swedish University of Agricultural Sciences, Uppsala, Sweden
b
Maxwell H. Gluck Equine Research Center, University of Kentucky, Lexington, USA
c
University of Kentucky Veterinary Diagnostic Laboratory, Lexington, USA

a r t i c l e i n f o a b s t r a c t

Article history: Bacteria (environmental contaminants and occasionally potential pathogens) are found in
Received 21 October 2013 most stallion ejaculates and may negatively affect sperm quality during storage. Since the
Received in revised form
use of antibiotics can lead to the development of resistance, an alternative means of micro-
19 December 2013
bial control is desirable. The removal of bacteria from stallion semen using Single Layer
Accepted 3 January 2014
Available online 17 January 2014 Centrifugation through Androcoll-E was investigated. Known doses of cultured bacteria
were added to freshly collected ejaculates (15 mL aliquots) before processing by Single Layer
Keywords: Centrifugation. The resulting sperm pellets and controls (not processed by Single Layer Cen-
SLC trifugation) were cultured and the bacteria identified. In experiment 1, doses of E. coli from
Androcoll-E 2 × 102 to 2 × 107 colony forming units were added to aliquots of semen. In experiment 2,
Pathogens Taylorella equigenitalis or a mix of E. coli, Klebsiella pneumoniae and Streptococcus equi subsp.
Contaminating bacteria zooepidemicus (approximately 7 × 106 , 5 × 106 , and 6 × 106 cfu, respectively) were added to
15 mL aliquots of semen. In experiment 1, more than 90% of the bacteria were removed
where loading doses were > × 104 cfu/mL. In experiment 2, varying proportions of different
bacteria were removed, ranging from 68% for naturally occurring Corynebacterium spp. to
>97% for added cultured E. coli. Thus, Single Layer Centrifugation can separate spermatozoa
from many, but not all bacteria in stallion ejaculates and could be a useful alternative to
adding antibiotics to semen extenders to control bacterial contamination. However, further
research is needed to determine the effect of small numbers of bacteria on sperm quality.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction example tetracycline-resistant strains of Clostridium per-

Since the publication of the Swann report (Anon, 1969),
there has been considerable concern about the emer-
gence of antibiotic-resistant strains of bacteria, particularly
methicillin-resistant bacteria (Catry et al., 2010). Even a
small amount of antibiotic usage can result in consider-
able antibiotic resistance within an animal species, for
∗ Corresponding author at: Clinical Sciences, Swedish University of
Agricultural Sciences, Box 7054, SE-75007 Uppsala, Sweden.
Tel.: +4618671152; fax: +4618673545.
E-mail address: jane.morrell@slu.se (J.M. Morrell).
fringens in Swedish broilers (Johansson et al., 2004).
Furthermore, genes for antimicrobial resistance are readily
exchanged between reservoirs in humans, farm animals
and companion animals (Duijkeren et al., 2005). Antibiotics
may be used prophylactically, as in the case of extenders
added to semen when preparing sperm doses for artificial
insemination (AI). However such usage could, theoretically,
cause antimicrobial resistance in bacteria found in horses
with subsequent transfer to bacteria in human beings.
Contaminating bacteria are commonly found in stal-
lion ejaculates after semen collection (Malmgren et al.,
1998; Rota et al., 2011). Most of these bacteria are com-
mensals but opportunistic pathogens such as Pseudomonas

0378-4320/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.anireprosci.2014.01.005
48 J.M. Morrell et al. / Animal Reproduction Science 145 (2014) 47–53
aeruginosa, Klebsiella pneumonia and Streptococcus zooepi-
demicus, as well as true pathogens such as T. equigenitalis,
may cause infection and increase inflammation after arti-
ficial insemination (AI) leading to infertility in mares
(Parlevliet et al., 1997; Metcalf, 2001; Samper and
Tibury, 2006). Bacterial contamination may also nega-
tively affect semen quality during storage, regardless of
their pathogenicity (Aurich and Spergser, 2007). Since
semen may be stored under conditions that potentially per-
mit bacterial growth and multiplication, the presence of
antibacterial agents may be essential to prevent the induc-
tion of disease. Although antibiotics are added routinely
to semen extenders to control the growth of contaminat-
ing bacteria their use and efficacy has been questioned.
For example, gentamicin was reported to have an adverse
effect on sperm motility and function (Aurich and Spergser,
2007).
An alternative to the use of antibiotics could be to
remove the contaminating bacteria physically instead
of controlling their growth. Previously it was shown
that bacterial contamination in boar semen could be
removed or reduced (depending on the bacterial species
and circumstances) by Single Layer Centrifugation (SLC)
through a species-specific colloid, Androcoll-P (Morrell
and Wallgren, 2011). These results were similar to those
reported for density gradient centrifugation of human
semen (Nicholson et al., 2000). The SLC method was
reported to improve sperm quality in a variety of species
e.g., stallion, boar, bull, dog, etc. (Morrell and Rodriguez-
Martinez, 2009). However, its use to remove bacteria from
stallion semen has not been studied previously. The objec-
tives of the present study were (i) to determine whether
stallion spermatozoa can be separated from added bacte-
ria by SLC; and (ii) to investigate whether the separation
effect is dependent on factors such as bacterial species and
bacterial load.
2. Materials and methods
2.1. Animals
Ejaculates from three stallions (two Quarter Horses and
one Tennessee Walking Horse) were collected over a 4-
week period in March 2011, at the start of the breeding
season. The animals were aged 16, 12 and 10 years, respec-
tively. They were kept on pasture at the University of
Kentucky Equine Research Farm. All research procedures
were approved by and were in accordance with IACUC reg-
ulations (IACUC # 2010-0771).
2.2. Semen samples
Ejaculates (n = 12; four from each stallion) were col-
lected using an artificial vagina (Missouri type) fitted with
an in-line filter to remove gel. Each stallion had its own arti-
ficial vagina. The penis was washed with warm water and
dried prior to allowing the stallion to mount the phantom.
One ejaculate was discarded due to contamination with
sand.
2.3. Bacterial culture and enumeration
All the culturing and identifications were performed
at the University of Kentucky Veterinary Diagnostic Lab-
oratory (UKVDL); this laboratory is accredited by the
American Association of Veterinary Laboratory Diagnosti-
cians for veterinary diagnostics. The isolates of Taylorella
equigenitalis, E. coli, Klebsiella pneumonia, S. equi subsp.
zooepidemicus to be added to the semen (see experimen-
tal design) were originally recovered from stallions at the
UKVDL. Taylorella equigenitalis was cultured on choco-
late agar. E. coli, Klebsiella pneumonia and S. equi subsp.
zooepidemicus (and various other bacteria) were cultured
on blood agar (with 5% horse blood), Columbia nalidixic
acid (CNA) agar with 10% sheep blood and eosin methy-
lene blue (EMB) agar plates. Plates were incubated at 37 ◦ C
microaerophilically (with 7% CO2 ). For preparation of “spik-
ing” experiments, fresh colonies (24–48 h) of bacteria were
picked and standard 10-fold serial dilutions (in semen
extender -EquiPro) and plating were performed to prepare
the “spiking” doses (Baumann, 2011). “Spiked” samples
before and after the SLC were subjected to standard 10-fold
serial dilution and plating again to determine the bacte-
rial numbers (colony forming units, cfu) in the samples
(Baumann, 2011). Taylorella equigenitalis, E. coli, Klebsiella
pneumonia, S. equi subsp. zooepidemicus and other bacte-
ria were identified using colony morphology, dark field
microscopic examination, gram staining and standard bio-
chemical tests including catalase, oxidase, various sugar
fermentation, triple sugar iron (TSI), sulfide indole motility
(SIM), urease and citrate.
2.4. Experiment 1
An ejaculate from each of the three stallions, pooled
and extended to a sperm concentration of 100 × 106 /mL
in warm (37 ◦ C) EquiPro without antibiotics (Minitube of
America, Verona, WI, USA), was divided into seven aliquots
(15 mL) for “spiking” with E. coli cultures (Ecc) at the fol-
lowing levels: 2 × 107 , 2 × 106 , 2 × 105 , 2 × 104 , 2 × 103 ,
2 × 102 colony-forming units (cfu/mL) and a control sample
with no “spiking”. The “spiked” aliquots were subsequently
prepared by Single Layer Centrifugation using Androcoll-E
(Morrell et al., 2009), modified by the inclusion of a tube
insert to facilitate collection of the sperm pellet (Morrell
et al., 2013). The resulting sperm pellets were harvested
into 1.5 mL EquiPro by passing a long 1.5 mL EquiPro by
passing a long Pasteur pipette down the central tube insert
to aspirate the sperm pellet from beneath the colloid and
supernatant (Morrell et al., 2013). All samples, together
with some extender as a control, were transported to the
laboratory on ice for bacterial culture. The proportion of
bacteria removed by SLC was calculated as follows: (num-
ber of bacteria in SLC sample/number of bacteria in “spiked”
sample) × 100.
2.5. Experiment 2
Three ejaculates were collected from each of the three
stallions and were extended in warm (37 ◦ C) EquiPro
without antibiotics (Minitüb of America, Verona, WI,
 J.M. Morrell et al. / Animal Reproduction Science 145 (2014) 47–53
Table 1
Bacterial load (cfu/mL) in “spiked” semen samples before and after Single Layer Centrifugation through Androcoll-E.
Bacteria from control Bacteria added
sample (cfu/mL) (cfu/mL)
3.3 × 104 2 × 107
3.3 × 104
2 × 106
3.3 × 104 2 × 105
3.3 × 104 2 × 104
3.3 × 104 2 × 103
3.3 × 104 2 × 102
a
Different type of E. coli than the one added to the ejaculate. Extender: 2 × 103 E. coli; bacteria in control sample (semen + extender): 3 × 104 E. coli;
2 × 103 Bacillus; 1 × 103 ␣-hemolytic streptococcus.
b
Calculated as a proportion of the original load i.e., (number of bacteria in SLC sample/number of bacteria in “spiked” sample) × 100.
USA) to a sperm concentration of 100 × 106 /mL. Aliquots
(15 mL) of the semen were “spiked” by adding approx-
imately 3.3 × 106 cfu/mL cultured Taylorella equigenitalis,
further aliquots were “spiked” with a mixture of E. coli
(mean 7 ± 1.8 × 106 cfu/mL), Klebsiella pneumoniae (mean
5 ± 0.8 × 106 cfu/mL) and S. equi subsp. zooepidemicus (S.
zooepidemicus) (mean 6 ± 1 × 106 cfu/mL). The ejaculate
remaining after removal of the aforementioned aliquots
was placed on ice to serve as a control that did not undergo
centrifugation. The 15 mL aliquots of the “spiked” samples
were prepared by SLC using the modified tube containing
the insert to facilitate harvesting of the sperm pellet, as
described above (Morrell et al., 2013). After centrifuging at
300 × g for 20 min, the supernatant (semen extender plus
seminal plasma, and colloid) was removed and the sperm
pellet was resuspended in fresh EquiPro without antibi-
otics (1.5 mL). These sperm samples were immediately
placed on ice and transported to the Veterinary Diagnostic
Laboratory, Lexington, Kentucky, for bacterial culture. The
proportion of bacteria removed by SLC was calculated as
described for experiment 1.
2.6. Statistics
Data was analyzed using PROC MIXED (analysis of vari-
ance) in the SAS software (ver. 9.2, SAS Inst. Inc., Cary, NC).
For each of the two experiments, the ratio ‘after SLC/before’
was analyzed according to a statistical model including the
fixed effect of bacterial flora (T. equi and normal flora or
E. coli, S. zooepidemicus, K. pneumoniae and normal flora,
respectively). The statistical model also included the ran-
dom effect of ejaculate, nested within stallion.
3. Results
Semen volume and sperm concentration (mean ± SD)
for the three stallions were as follows: Stallion 1:
108 ± 29 mL and 196 × 106 ± 30 × 106 ; Stallion 2:
33 ± 15 mL and 423 × 106 ± 41 × 106 ; stallion 3: 43 ± 22 mL
and 309 × 106 ± 77 × 106 .
3.1. Experiment 1
The bacterial content of the control pooled semen
sample was 3 × 104 cfu/mL E. coli; 2 × 103 cfu/mL Bacil-
lus; 1 × 103 cfu ␣-hemolytic streptococcus. The EquiPro
extender contained 2 × 103 cfu/mL E. coli, which was the
49
After SLC (cfu/mL) % Removedb
3.0 × 105 “spiking” E. coli; 2 × 104 E. colia ; 98.4
1 × 103 bacillus; 1 × 103 ␣-hemolytic streptococcus
3.0 × 104 “spiking” E. coli; 0.7 × 104 E. colia 98.2
3.0 × 103 “spiking” E. coli; 0.9 × 104 E. colia 94.8
3.0 × 102 “spiking” E. coli; 0.5 × 104 E. colia 90.0
3.0 × 101 “spiking” E. coli; 1.1 × 104 E. colia 68.0
3.0 “spiking” E. coli; 0.7 × 104 E. colia 79.0
same as the E. coli in the semen sample. In the SLC samples
prepared from the “spiked” semen samples, the content of
“spiked” E. coli (Ecc) was 3 × 105 , 3 × 104 , 3 × 103 , 3 × 102 ,
3 × 10 and 3 cfu/mL respectively, plus 0.5–2.0 × 104 cfu/mL
of the E. coli species from the original semen samples and
extender (Table 1). In addition, 1 × 103 cfu/mL Bacillus spp.
and 1 × 103 cfu ␣-hemolytic streptococcus were identified
in the SLC-sample from the most heavily “spiked” semen
sample. Thus, SLC decreased the concentration of “spiked”
E. coli (Ecc) by 98.5% so that only 1.5% of the original load
remained. The total bacterial content (bacteria originating
in the semen and the extender plus the added bacteria) was
reduced to 1.5%, 2%, 5%, 10%, 33% and 21%, respectively.
3.2. Experiment 2
The original ejaculates all contained contaminat-
ing bacteria but these differed in number and species
between the stallions and between ejaculates for each
stallion (Table 2). Overall, the mean “spiking” dose was
18 × 106 cfu/mL for the mixed load, 3.3 × 106 cfu/mL
for the samples “spiked” with Taylorella equigeni-
talis, and the mean contaminating bacterial load was
2 × 103 –8.4 × 103 cfu/mL. All stallions had ␣-hemolytic
Streptococci in at least one ejaculate, in all three ejaculates
for Stallion 1 and one ejaculate each for Stallions 2 and
3. Coliform bacteria were present in two of the stallions
(two ejaculates for Stallion 1 and one for Stallion 2);
coagulase negative staphylococci were present in one
ejaculate from Stallion 1 and two from Stallion 2; Bacillus
spp. were present in one ejaculate from Stallion 1 and two
from Stallion 2; Corynebacterium spp. were present in one
ejaculate from Stallion 2 and one from Stallion 3. Other
bacterial species were present in only one ejaculate, com-
prising Streptococcus spp. in one ejaculate from Stallion 1,
Enterococcus spp. and Pseudomonas spp. in one ejaculate
each from Stallion 2, and Citrobacter spp. and Pantoea spp.
in one ejaculate from Stallion 3.
The ejaculates varied between and within stallions
according to the contaminating bacterial species and load.
Mean values of the different species of bacteria before and
after SLC are shown in Table 2, and per stallion in Table 3.
The overall reduction in the load of contaminating bacte-
ria varied from 78% to 100% with the median reduction in
bacterial load being 88%. The mean removal of individual
species varied from 68% for Corynebacterium spp. to 91% for
Bacillus spp.
50 J.M. Morrell et al. / Animal Reproduction Science 145 (2014) 47–53

Table 2
Effect of SLC on mean load of contaminating bacterial species.

Bacterium Classificationa Number of ejaculates Before SLC (cfu/mL) After SLC (cfu/mL) Proportion removedb (%)

Alpha hemolytic streptococci Gram positive cocci 4 1460 ± 780 324 ± 200 78
Bacillus Gram positive rods 3 825 ± 850 73 ± 88 91
Coagulase negative staphylococci Gram positive cocci 3 4567 ± 3855 610 ± 512 87
Coliform bacteria Gram negative rods 3 717 ± 701 107 ± 168 85
Pseudomonas spp. Gram negative rods 2 1110 ± 1273 175 ± 176 84
Corynebacterium Gram positive rods 2 6000 ± 0 1900 ± 1556 68
Citrobacter Gram negative rods 1 460 50 89
Enterococcus Gram positive cocci 1 100 0 100
Pantoea Gram negative rods 1 300 50 83
Streptococcus spp Gram positive cocci 1 2000 500 75
a
Gram positive, Gram negative refer to Gram staining.
b
Calculated as a proportion of the original load i.e., (number of bacteria in SLC sample/number of bacteria in “spiked” sample) × 100.

Table 3 of bacteria left after SLC and the original load (r = 0.8,
Effect of SLC on mean load of contaminating bacterial species (cfu/mL) for
P < 0.001), indicating that the effectiveness of removal of
each stallion.
bacteria was also dependent on the original load.
Stallion 1 Stallion 2 Stallion 3 Concerning the “spiking” bacteria, there was a reduction
Extended semen 2107 ± 759 8433 ± 1069 3480 ± 3988 in the number of bacteria after SLC that differed according
SLC 373 ± 155 1883 ± 1269 605 ± 700 to bacterial species (Fig. 2). Thus, 97% E. coli, 93% Klebsiella
Mean proportion 82% 78% 83% pneumoniae, 93% S. zooepidemicus and 81% of T. equi were
removeda
removed by SLC.
a
Calculated as a proportion of the original load i.e., (number of bacteria
in SLC sample/number of bacteria in “spiked” sample) × 100.
4. Discussion

The removal of the contaminating bacteria was corre-
lated to the original load (Fig. 1) but was more effective
where there were more contaminating bacteria present
originally: for all samples r = 0.74, P < 0.001; for sam-
ples with <2000 cfu/mL, r = 0.83, P < 0.001. Corynebacterium
were the most difficult to remove. For the “spiked” sam-
ples, the correlation between the numbers of bacteria
added and those remaining after SLC varied according to
bacterial species i.e., for T. equi, r = 0.355, P > 0.5 (not sig-
nificant) whereas for samples with the mixture of E. coli, S.
zooepidemicus and K. pneumoniae added, r = 0.73, P < 0.001.
However, if the total bacterial flora (i.e., added bacteria
plus contaminating bacteria) was taken into consider-
ation, there was a strong correlation between the number
Fig. 1. Number of contaminating bacteria (colony forming units)
remaining after SLC compared to the number in the original sample (n = 8
ejaculates).
The present study was designed to investigate whether
stallion spermatozoa can be separated from added bacte-
ria by SLC with Androcoll-E and to determine whether
the separation effect is dependent on bacterial species and
load. All ejaculates were contaminated with bacteria, in
agreement with the findings of others (Ortega-Ferrusola
et al., 2009; Rota et al., 2011). However, the bacterial
species cultured varied between stallions and ejaculates,
although the species cultured (␣-hemolytic streptococci,
Bacillus, coagulase negative staphylococcus, coliforms, Pseu-
domonas spp., Corynebacteria, Citrobacter spp.) were similar
to those reported by Ortega-Ferrusola et al. (2009) and Rota
et al. (2011).
When semen was “spiked” with various doses of cul-
tured E. coli (experiment 1), it was found that stallion
spermatozoa can be separated from bacteria by colloid cen-
trifugation in a dose-dependent fashion such that more
than 90% of a bacterial load of approximately ≥5 × 104
could be removed. However, relatively lower proportions
of the bacterial species were removed (68% and 79%) where
the initial E. coli load was less (3.5 × 104 and 3.3 × 104 ,
respectively). Removal of the bacteria contaminating the
ejaculate varied considerably, with all of the Bacillus and
␣-hemolytic streptococcus being removed from most of the
samples. The exception was in the sample that received
the greatest “spiking” dose where small numbers of Bacil-
lus and ␣-hemolytic streptococcus were found in the SLC
sample.
In the second experiment, where semen samples were
“spiked” with various bacterial species, more than 90% of
the “spiking” bacterial load was removed for E. coli, K.
pneumoniae and S. zooepidemicus, and 81% of T. equigen-
italis. Of the bacteria contaminating the ejaculate during
semen collection, 68– 91% were removed by SLC depend-
ing on species. Removal of Corynebacterium, however, was
 J.M. Morrell et al. / Animal Reproduction Science 145 (2014) 47–53
Fig. 2. Bacterial load (“spiked” bacteria) before and after Single Layer Centrifugation (n = 8).
less than the others. These results are in broad agreement
with those reported by Morrell and Wallgren (2011) where
90% of Burkholderi and 95% of E. coli were removed from
boar semen samples by SLC, although Pantoea were not
removed.
Previous reports have quantified the bacterial counts
in fresh equine semen e.g. 1.1–2.6 × 104 cfu/mL (Clément
et al., 1995) but the latter authors commented that it
was not possible to state a typical average value because
of the large variation between ejaculates. In another
study in which bacterial loads were reported for fresh
and frozen semen from five stallions (Ortega-Ferrusola
et al., 2009), fresh semen was found to contain the
following bacterial species: Staphylococcus spp. was
seen in ejaculates from all five stallions ranging from
0.7 × 106 ± 0.69 × 106 to 37.1 × 106 ± 63.13 × 106 cfu/mL;
Micrococcus spp. were also found in all five stallions
(0.9 × 106 ± 0.8 × 106 to16.0 × 106 ± 21.1 × 106 cfu/mL);
␤-hemolytic streptococci were found in three stallions
(0.7 × 106 ± 0.7 × 106 –4.0 × 106 ± 5.3 × 106 cfu/mL);
␣-hemolytic streptococcus in two stallions (23.3 ×
106 ± 40.1 × 106 to 66.6 × 106 ± 115.5 × 106 cfu/mL);
Corynebacterium spp. in four stallions (85.6 ×
106 ± 99.02 × 106 –224.0 × 106 ± 222.7 × 106 cfu/ml);
Rhodococcus spp. in one stallion 100.0 × 106 ± 0.0 ×
106 cfu/mL); Bacillus spp. in two stallions (0.1 × 106 ±
0.2 × 106 cfu/mL); Pseudomonas spp. in one stal-
lion (9.7 × 106 ± 14.02 × 106 cfu/mL); Klebsiella spp. in
three stallions (0.1 × 106 ± 0.1 × 106 –7.3 × 106 ± 12.7 ×
106 cfu/mL). Since these bacterial loads are similar to
the greatest “spiking” doses used in the SLC experiment
reported in the present study, by extrapolation it could
be expected that 90% or more of these bacteria would
be removed by SLC. From other experiments it has been
shown that the sperm pellet from each SLC can be used for
one to four insemination doses depending on the sperm
quality in the original ejaculate and the sperm dose used
(e.g. Lindahl et al., 2012). Thus the bacterial content in
51
Note: SLC = Single Layer Centrifugation.
theoretical insemination doses could be reduced down
to 25% of the values shown in the present study, giving
potential counts that are much lower than the mean value
of 6.9 × 103 cfu/mL reported in standard AI doses (Althouse
et al., 2010). Since natural mating is itself not a sterile
procedure, the female reproductive tract has developed
its own defense mechanisms to cope with a bacterial
onslaught at mating. Thus it remains to be established
what bacterial load will cause pathology in any given
inseminated mare and whether it is the total bacterial load
or the number of cfu/mL that is important for any given
bacterium.
Three species of spiking bacteria (Streptococci, E. coli
and Klebsiella) were removed more efficiently than simi-
lar types of contaminating bacteria (93–96% for “spiking”
bacteria compared to 75 to 85% for similar classes of
contaminating bacteria). T. equi was removed at approx-
imately the same efficiency as contaminating coliform
species (81% compared with 85%). Therefore, there may
be subtle differences between cultured bacteria (as used
for “spiking”) and naturally occurring bacteria in the man-
ner in which they interact with spermatozoa. This should
be considered when planning future studies on removal of
bacteria by SLC. These issues are interesting and are the
subject of ongoing investigations.
With regard to T. equigenitalis, similar proportions of
the “spiking” bacteria (81%) were removed as for the con-
taminating coliforms (85%). Previous studies reported a
bacterial load of 9.98 × 104 cfu/mL in raw semen from a
stallion naturally infected with T. equigenitalis. By extrapo-
lation, approximately 1.9 × 104 T. equigenitalis would have
been left in the sperm pellet had SLC been carried out on
such a semen sample, which would probably be sufficient
to cause infection in inseminated mares if no antibiotics
were included in the semen extender (Klein et al., 2012).
Removing contaminating bacteria from semen by physical
means could be an attractive alternative to the use of antibi-
otics. In this respect, complete removal of contaminating
52 J.M. Morrell et al. / Animal Reproduction Science 145 (2014) 47–53
bacteria in the present study would have been desirable.
However, if only small numbers of bacteria are left after
SLC and the semen is stored under anaerobic conditions at
4 ◦ C, it might be possible to avoid the addition of antibiotics
altogether, provide that no infection with T equigenitalis is
suspected. In fact, from the point of view of development
of antibiotic resistance, it could be advantageous not to
use antibiotics in the semen extender if only small num-
bers of bacteria are present. Further research is underway
to quantify bacterial growth in SLC-selected sperm sam-
ples during cooled storage and to identify any risks to the
mare. The effect on sperm quality of such bacterial loads
during storage of AI doses is currently being quantified,
as are the interactions between bacteria and also between
spermatozoa and bacteria. It is intended to include sperm
ultrastructural studies in this project as well as more con-
ventional assays of sperm quality.
In a previous study with boar semen (Morrell and
Wallgren, 2011), similar bacterial species were seen as in
the present study. The bacteria were removed with 100%
efficiency when SLC was carried out immediately after
semen collection. Where the boar semen doses (from boars
at a different site) were transported for several hours before
SLC, the efficiency of removal was reduced. Since different
bacteria were involved in the two sets of boar samples, it
is not clear whether the efficiency of removal was related
to the species of bacteria concerned, whether they were
themselves motile because of the possession of a flagellum,
or whether the bacteria were able to form associations with
spermatozoa over time which enable them to pass through
the colloid together with the sperm. In the present study
with stallion semen, both cocci and rod-shaped bacteria
were present in the semen and were removed with approx-
imately the same efficiency. The two types of bacteria that
were least easily removed, Corynebacterium and Strepto-
cocci, both have the ability to group together, the former
as palisades and the latter in chains or pairs, which may
influence their behavior during colloid centrifugation by
altering the density of the bacterial unit as a whole. Further-
more, Corynebacterium is a rod-shaped bacterium with
thickened poles, which conceivably may affect its density
and therefore its behavior during colloid centrifugation.
Another factor potentially influencing removal could
the type of semen extender used. In one of the boar samples
in the previous study, the bacterial load was 10,000 cfu/mL
before SLC; 100% removal was achieved (Morrell and
Wallgren, 2011). In the present study, the contaminating
bacteria did not reach this level; the highest contamina-
tion was 6000 cfu/mL Corynebacterium, 69% of which were
removed. However, since the “spiking” doses were much
greater than the contaminants (18 × 106 cfu/mL compared
to 2 × 103 –8.4 × 103 cfu/mL), the removal of bacteria by SLC
can be considered to be equal to that of the “spiking” bacte-
ria i.e. 81% to >90%.
Whatever method is used to control the growth of
contaminating bacteria, it is essential to prevent con-
tamination during collection as far as possible. Bacterial
contamination can be influenced by husbandry and hous-
ing. There is considerable debate over whether or not
the penis should be washed before semen collection to
reduce bacterial contamination of the ejaculate. Bowen
et al. (1982) reported that the systematic washing of a
stallion’s penis, even where only water is used, will cause
the normal flora to be replaced with pathogens and poten-
tial pathogens, and these findings were later confirmed by
others (e.g. Varner et al., 1998; Dietz et al., 2007). Since
Pseudomonas spp. may be present in the water supply,
washing may result in contamination of the penis (Allen
et al., 2011). In the present study, even a bottle of sterile
semen extender, opened in the laboratory where the semen
was brought immediately after collection, became contam-
inated with the same species of E. coli as in the semen,
showing the importance of hygiene in the laboratory dur-
ing semen processing.
In conclusion, >SLC with Androcoll-E is capable of reduc-
ing the bacterial load in stallion semen samples although a
proportion of some bacteria can accompany the spermato-
zoa through the colloid. Both bacterial species and bacterial
load appear to be implicated in the efficiency of removal.
Conflict of interest
JM Morrell is one of the patent holders for Androcoll-E.
Acknowledgements
We thank the farm personnel, Department of Veteri-
nary Science for the management of the stallions and for
facilitating semen collection. The study was financed by
a travel fellowship from the Swedish Research Council for
the environment, agricultural sciences and spatial planning
(FORMAS) and by the Albert and Lorraine Clay Fellowship
Endowment Fund. The funding bodies played no part in the
study design, in the collection, analysis and interpretation
of data, in the writing of the report, or in the decision to
submit the article for publication
References
Althouse, G.C., Skaife, J., Loomis, P., 2010. Prevalence and types of contam-
inating bacteria in extended chilled equine semen. Anim. Reprod. Sci.
1215, S24–S225.
Allen, J.L., Begg, A.P., Browning, G.F., 2011. Outbreak of equine endometri-
tis caused by a genotypically identical strain of Pseudomonas
aeruginosa. J. Vet. Diagn. Invest. 23, 1236–1239.
Anon., 1969. Report of the Joint Committee on the Use of Antibiotics in
Animal Husbandry and Veterinary Medicine. HMSO, London, UK.
Aurich, C., Spergser, J., 2007. Influence of bacteria and gentamicin on
cooled-stored stallion spermatozoa. Theriogenology 67, 912–918.
Bowen, J.M., Tobin, N., Simpson, R.B., 1982. Effect of washing on the bac-
terial flora of the stallion’s penis. J. Reprod. Fertil. Suppl. 32, 41–45.
Baumann, R.W., 2011. Microbiology with diseases by taxonomy. In: Ber-
riman, L. (Ed.), Microbial Nutrition and Growth, third ed. Pearson
Education Inc., Glenview, IL, pp. 185–186 (Chapter 6).
Catry, B., van Duijkere, E., Pomba, M.C., Greko, C., Moreno, M.A., Pyörälä,
S., Ruzauskas, M., Sanders, P., Threlfall, E.J., Ungemach, F., Törneke,
K., Munoz-Madero, C., Torren-Edo, J., Scientific Advisory Group on
Antimicrobials (SAGAM), 2010. Reflection paper on MRSA in food-
producing and companion animals: epidemiology and control options
for human and animal health. Epidemiol. Infect. 138, 626–644.
Clément, F., Vidament, M., Guérin, B., 1995. Microbial contamination of
stallion semen. Biol. Reprod. Monogr. 1, 779–786.
Dietz, J.P., Sertich, P.L., Boston, R.C., Benson, C.E., 2007. Comparison of ticar-
cillin and piperacillin in Kenney’s semen extender. Theriogenology 68,
848–852.
Duijkeren, E.V., Box, A.T., Schelle n, P., Houwers, D.J., Fluit, A.C., 2005. Class
1 integrons in Enterobacteriaceae isolated from clinical infections of
horses and dogs in the Netherlands. Microb. Drug Resist. 11, 383–386.
 J.M. Morrell et al. / Animal Reproduction Science 145 (2014) 47–53
Johansson, A., Greko, C., Engström, B.E., Karlsson, M., 2004. Antimicro-
bial susceptibility of Swedish, Norwegian and Danish isolates of
Clostridium perfringens from poultry, and distribution of tetracycline
resistance genes. Vet. Microbiol. 99, 251–257.
Klein, C., Donahue, J.M., Sells, S.F., Squires, E.L., Timoney, P.J., Troedsson,
M.H., 2012. Effect of antimicrobial-containing semen extender on risk
of dissemination of contagious equine metritis. J. Am. Vet. Med. Assoc.
241, 916–921.
Malmgren, L., Olsson Engvall, E., Engvall, A., Albihn, A., 1998. Aero-
bic bacterial flora of semen and stallion reproductive tract and
its relation to fertility under field conditions. Acta Vet. Scand. 39,
173–182.
Metcalf, E.S., 2001. The role of international transport of equine semen on
disease transmission. Anim. Reprod. Sci. 68, 229–237.
Morrell, J.M., Rodriguez-Martinez, H., 2009. Biomimetic techniques for
improving sperm quality in animal breeding: a review. Open Androl.
J. 1, 1–9.
Morrell, J.M., Wallgren, M., 2011. Removal of bacteria from boar ejacu-
lates by single layer centrifugation can reduce the use of antibiotics
in semen extenders. Anim. Reprod. Sci. 123, 64–69.
Morrell, J.M., Johannisson, A., Dalin, A-M., Rodriguez-Martinez, H., 2009.
Single layer centrifugation with AndrocollTM -E can be scaled-up to
allow large volumes of stallion ejaculate to be processed easily. The-
riogenology 72, 879–884.
53
Morrell, J.M., Timoney, P., Klein, C., Shuck, K., Campos, J., Troedsson, M.,
2013. Single layer centrifugation reduces equine arteritis virus titer in
the semen of shedding stallions. Reprod. Domest. Anim. 48, 604–612.
Nicholson, C.M., Abramsson, L., Holm, S.E., Bjurulf, E., 2000. Bacterial con-
tamination and sperm recovery after semen preparation by density
gradient centrifugation using silane-coated silica particles at different
g forces. Hum. Reprod. 15, 662–666.
Ortega-Ferrusola, C., Gonzàlez-Fernàndez, L., Muriel, A., Macias-Garcia, B.,
Rodriguez-Martinez, H., Tapia, J.A., Alonso, J.M., Pena, F.J., 2009. Does
the microbial flora in the ejaculate affect the freezeability of stallion
sperm? Reprod. Domest. Anim. 44, 518–522.
Parlevliet, J.M., Bleumink-Pluyn, N.M., Houwers, D.J., Remmen, J.L., Slui-
jter, F.J., Colenbrander, B., 1997. Epidemiological aspects of Taylorella
equigenitalis. Theriogenology 47, 1169–1177.
Rota, A., Calicchio, E., Nardoni, S., Fratini, F., Ebani, V.V., Sgorbini, M., Pan-
zani, D., Camillo, F., Mancianti, F., 2011. Presence and distribution
of fungi and bacteria in the reproductive tract of healthy stallions.
Theriogenology 76, 464–470.
Samper, J.C., Tibury, A., 2006. Disease transmission in horses. Theriogenol-
ogy 66, 551–559.
Varner, D.D., Scanlan, C.M., Thompson, J.A., Brumbaugh, G.W., Blanchard,
T.L., Carlton, C.M., Johnson, L., 1998. Bacteriology of preserved stal-
lion semen and antibiotics in semen extenders. Theriogenology 50,
559–573.