Located in the equatorial zone of South America, Colombia has a long-time tradition of sheep production with strong links with its culture. In the last decade, the number of herds as well as animals increased with a significant impact in local economy. An economic advantage of sheep production in the equatorial countries and other tropical and sub-tropical environments is that reproduction is unseasonal and meat production is stable throughout the year ( Rosa & Bryant, 2003 ). However, ambient temperature, humidity, solar radiation, and rainfall can affect livestock production systems by environmental stressors on sheep physiology ( Marai et al. , 2007 ; Abecia et al. , 2016 ) and the annual rain-drought cycle can modify the food availability ( Rosa & Bryant, 2003 ). The rainfall effect on fertility rate has been studied in ewes on temperate zones ( Galina et al. , 1996 ; Arrebola et al. , 2009 , 2016 ). Low rainfall has a detrimental effect on the offspring after artificial insemination in spring ( Abecia et al. , 2016 ) summer and autumn ( Palacios & Abecia, 2014 ), and a possible explanation for the effect of annual rainfall was associated with the nutritional status of animals ( Arrebola et al. , 2009 ). Therefore, in this study we have investigated the possible sperm quality differences among breeds in the raining season. Creole is a local sheep breed adapted to environmental conditions that constitutes an important genetic resource for the small-scale agriculture in certain ecoregions ( Soma et al. , 2012 ; Bravo et al. , 2015 ). Hampshire and Romney Marsh are genetic-selected breeds introduced in Colombia by direct importation from European countries since 1960 to improve the local sheep breed. All these breeds display an unseasonal reproduction under equatorial environment. Research related to the diversity and adaptation of local and foreign breeds may help to identify the most appropriate animal genetic resources for each region ( Bravo et al. , 2015 ).

Sperm quality is the main factor that limit the male reproductive efficiency. In addition, the influence of breed, geographical location and differences among individual rams have been reported ( Karagiannidis et al. , 2000 ; Kasimanickam et al. , 2007 ). Likewise, differences in sperm functionality, progressive motility and volume of the ejaculate have also been observed between genetically selected breeds in temperate zones ( Kasimanickam et al. , 2007 ) and between local breed in arid zones ( Ibrahim, 1997 ). Conversely, no significant differences have been found in sperm quality parameters between local breeds in the Mediterranean area ( Karagiannidis et al. , 2000 ).

Mammalian sperm cell physiology is supported by seminal plasma (SP) ( Desnoyers & Manjunath, 1992 ; Manjunath & Thérien, 2002 ; Muiño-Blanco et al. , 2008), a complex biological fluid composed of multiple secretions from the testis, epididymis, prostate, seminal vesicles and bulbourethral glands ( Domínguez et al. , 2008 ). SP provides metabolic maintenance as an energy source for the sperm cell, and modulates sperm functionality in terms of motility, viability ( Maxwell et al. , 1998 ) and resistance to damage by keeping spermatozoa in a decapacitated-state both in fresh ( Luna et al. , 2015 ) and frozen-thawed ( Domínguez et al. , 2008 ; Bernardini et al. , 2011 ) semen. The beneficial components of SP are difficult to define because of its complexity and the presence of both stimulating and inhibitory factors ( Muiño-Blanco et al. , 2008 ; Leahy, et al. , 2010a ). Certain proteins isolated from SP stabilise the sperm membrane and reduce stress damage, increasing the membrane resistance to cold-shock ( Barrios et al. , 2000 ; Perez-Pe et al. , 2001 ; Colás et al. , 2009 ) and cryopreservation ( Leahy et al. , 2009 ; Bernardini et al. , 2011 ). Other SP proteins are enzymes, which are very important for the maintenance of the sperm fertilizing ability after ejaculation. Among these enzymes, the antioxidant defence system plays a critical role given that, under physiological conditions, spermatozoa generate small amounts of reactive oxygen species (ROS). Increased ROS and decreased antioxidant activity in SP have been related with male infertility ( de Lamirande et al. , 1997 ; Aitken & Baker 2004 ). The SP antioxidant enzyme defence system includes glutathione peroxidase (GPx), glutathione reductase (GR), superoxide dismutase (SOD) and catalase (CAT), and plays an essential role in the sperm protection from the oxidative stress ( Aitken & Roman, 2008 ; Agarwal et al. , 2014 ), working synergistically to remove the oxidizing components ( Marti et al. , 2007 ; Shiva et al. , 2011 ).

Taking into account the wide range of roles of SP protein components in sperm functions, it might be assumed that the absence, presence or differences in content of specific SP proteins and enzymes might be reflected in altered sperm functionality and fertilizing ability ( Agarwal et al. , 2014 ; Sharma et al. , 2015 ). Techniques such as enzyme kinetics and proteomics allow the characterization or identification of these SP proteins at molecular level and are useful to reveal differences between breeds. It has been shown that SP composition is related to sperm membrane-binding proteins and is highly conserved between seasonal breeds in the southern hemisphere ( Bernardini et al. , 2011 ). Conversely, differences in the protein composition of SP have been found in crossbred rams, and it has been suggested that they could be influenced by season, breeds, age and SP management ( Asadpour, 2012 ). These analyses have provided knowledge of the biological pathways involved in fertility ( du Plessis et al. , 2011 ) and are the basis for the identification of fertility biomarkers ( Agarwal et al. , 2014 ; Muhammad Aslam et al. , 2014 ; Soleilhavoup et al. , 2014 ).

Information on characteristics and differences of sperm functionality ( Ibrahim, 1997 ; Karagiannidis et al. , 2000 ; Kasimanickam et al. , 2007 ) and SP composition ( Bernardini et al. , 2011 ; Asadpour, 2012 ) of several breeds of ram has been reported. Nevertheless, little is known about sperm quality, SP composition and antioxidant defense of the Creole (a local breed) and other introduced sheep breeds under high altitude and tropical conditions in Colombia. In this study, we tested the hypothesis that sheep breed can influence SP composition. Thus, the objectives of this study were 1) to evaluate differences in sperm quality; 2) to evidence changes in the SP protein content and antioxidant enzyme activity; and 3) to determine whether there is any relationship between the antioxidant enzyme activity and SP proteins with sperm quality. The study was carried out on three sheep breeds under the high altitude and tropical conditions in Colombia.

Although information on the semen characteristics of rams is widely available, little is known about the semen of ovine breeds adapted to tropical conditions. This study is focused on three sheep breeds during the rainy season in high altitude tropical conditions in Colombia.

The reproductive capacity and seminal quality of rams may be affected by several factors such as breed ( Kasimanickam et al. , 2007 ), seasonality ( Ibrahim, 1997 ; Avdi et al. , 2004 ), photoperiod ( Pérez & Mateos, 1996 ), nutrition and environmental variables ( Galina et al. , 1996 ). Our results showed that, although the semen characteristics of the three breeds were considered satisfactory, Creole and Hampshire breeds had a higher sperm quality compared with Romney Marsh. The ejaculate volume of Hampshire was higher than that of Creole, which had the highest sperm concentration. A possible physiological reason for the higher ejaculate volume rather than concentration in Hampshire is due to the volume of SP in the ejaculate ( Ibrahim, 1997 ). It has been reported that the amount of SP surrounding each spermatozoon in an ejaculate varies depending on the sperm concentration in ram ( Gundogan et al. , 2010 ) and boar ( Kommisrud et al. , 2002 ). Furthermore, it is well known that, in mammalian species, SP and its components play essential roles in the regulation of sperm functionality ( Rodríguez-Martínez et al. , 2011 ). Improve motility characteristics ( El-Hajj Ghaoui et al. , 2007 ), increase the sperm resistance to stress ( Barrios et al. , 2000 , 2005 ; Muiño-Blanco et al. , 2008 ; Mendoza et al. , 2013 ) and to aging during storage ( Gundogan et al. , 2010 ), and reduce the deleterious effects of cryopreservation ( El-Hajj Ghaoui et al. , 2007 ; Leahy et al. , 2011 ). Therefore, we could suggest that a higher volume of SP in the ejaculate might be one of the reasons for the better sperm quality of the Hampshire breed.

Regarding the kinematic parameters, our data showed that the differences between breeds were generally not significant, similarly to what has been reported for the kinematic values of Suffolk and Katahdin rams ( Kasimanickam et al. , 2007 ). Although the ALH (amplitude of lateral head displacement) was significantly higher in the Creole breed than in Romney Marsh and Hampshire, no difference in LIN was found. It is worth noting that Hamilton Thorne instruments use ALH _max for kinematic analysis ( Mortimer, 2000 ). Therefore, based on the criteria defined by Mortimer & Maxwell (1999) for hypermotility of ram spermatozoa (ALH _max ≥9 μm, LIN 30 %), the ALH values were lower and the LIN values were higher in the three breeds than the established threshold values, which indicates that these spermatozoa do not have a hyperactivated movement.

Comparison among rams within breeds revealed no significant differences in sperm quality in Creole, while Romney Marsh and Hampshire breeds presented significant differences among males. These results would indicate that the introduced breeds would be less adapted to the environmental conditions.

Concerning morphology, the Romney Marsh semen had the lowest sperm rate with normal characteristics; additionally, a significantly higher value of bent tail and distal cytoplasmic droplet was found. This might be related to the presence of immature cells in the ejaculate ( Belleannée et al. , 2011 ), which is indicative of oxidative stress ( Marti et al. , 2007 ) and the generation of free radicals, as already reported in human semen ( Aitken et al. , 1998 ). Specifically, a positive correlation between GR activity and the sperm percentages with bent and coiled tail was observed in Romney Marsh breed. Likewise, a negative correlation between GR and ALH (head lateral amplitude) and VAP (path velocity) was observed in the Creole breed. The fact that excessive antioxidant activity may be detrimental to sperm functions ( Maxwell & Stojanov, 1996 ; Calamera et al. , 2003 ) must be taken into account. The GR enzyme is part of the GR/GPx system and its substrate (GSH), which functions as a defence system against the production of hydrogen peroxide ( Alvarez & Storey, 1989 ; Storey et al. , 1998 ). The action of this system implies a balance between the production of ROS and scavengers ( Câmara et al. , 2011 ), which involves preserving the sperm metabolic activity ( De Lamirande & Gagnon, 1995 ). This toxicity would be related to the action of the antioxidant enzymes, given that high levels of GSH cause an early decrease in motility and kinetics parameters in ovine sperm cooled at 5 °C ( Câmara et al. , 2011 ) and can affect cell homeostasis ( Sikka, 1996 ). This would be the explanation for the negative correlation between GR activity and the VAP and ALH variables in the Creole breed. Our findings would indicate that GR may be utilized to counter the possible oxidative damage due to spermatozoa with cytoplasmic droplet retention ( Kasimanickam et al. , 2006 ).

Several studies have shown a positive relationship between sperm quality parameters and ram seminal plasma protein levels ( Taha et al. , 2000 ; Almadaly et al. , 2016 ), and that the semen collection procedure has no significant effect on protein concentration ( Marco-Jiménez et al. , 2008 ; Ledesma et al. , 2014 ) although differential protein profile was observed between artificial vagina and electroejaculation ( Ledesma et al. , 2014 ). A correlation between seminal plasma proteins and male fertility has been widely shown ( Almadaly et al. , 2016 ), and although information regarding breed comparisons is scarce, differences between hair and wool breeds have not been associated with fertility ( Taha et al. , 2000 ; Marco-Jiménez et al. , 2008 ). The SP total protein concentration found in the three breeds analysed here was higher than that described for Rasa Aragonesa rams in breeding season ( Marti et al. , 2007 ), and lower than that of Corriedale in the southern hemisphere (synthetic breed created in New Zealand) ( Ledesma et al. , 2014 ). These differences would suggest that the SP total protein levels of local breeds are not usually very high, as already described for Awassi and Barki Egyptian breeds ( Taha et al. , 2000 ). Environmental and seasonal factors can affect both ram physiology ( Marai et al. , 2007 ) and SP total protein concentration ( Smith et al. , 1999 ) and profile ( Cardozo et al. , 2006 ).

It has been shown that SP proteins are adsorbed onto the mammalian sperm surface and modify its properties ( Desnoyers & Manjunath, 1992 ; Manjunath & Thérien, 2002 ), stabilize the plasma membrane and increase the sperm fertilizing potential ( Mendoza et al. , 2013 ; Luna et al. , 2015 ). Furthermore, it has been suggested that some of the coating proteins are essential for maintaining the stability of the membrane up to the process of capacitation (decapacitating factors) ( Desnoyers & Manjunath, 1992 ; Manjunath et al. , 2002 ). However, the role of SP in ram sperm function still remains controversial ( Leahy & de Graaf, 2012 ) as both inhibitory ( de Graaf et al. , 2007 ; Leahy, et al. , 2010b ) and stimulating ( Barrios et al. , 2000 ; Pérez-Pé et al. , 2002 ) effects have been found. Additionally, specific protein fractions have been isolated from ram SP and identified as responsible for preventing ( Barrios et al. , 2005 ; Cardozo et al. , 2008 ; Rickard et al. , 2015 ) and repairing ( Barrios et al. , 2000 ; Bernardini et al. , 2011 ) cold-shock damage. These specific fractions ( Barrios et al. , 2005 ) have a similar protein profile to those identified by other authors in ram SP of different breeds ( Bergeron et al. , 2005 ; Jobim et al. , 2005 ; Bernardini et al. , 2011 ; Soleilhavoup et al. , 2014 ), and are named RSVP-14, RSVP-20 and RSVP-22 ( Barrios et al. , 2005 ; Fernandez-Juan, 2006 ; Cardozo et al. , 2008 ) given that they are exclusively synthesized by the seminal vesicles. These proteins belong to the binding of sperm protein (BSP's) family ( Manjunath et al. , 2008 ; Serrano et al. , 2015 ) due to the presence of two fibronectin type II domains (FN2 domain) which confer many binding properties on them ( Desnoyers & Manjunath, 1992 ). BSP's play an important function in the stabilization of the sperm membrane as they bind to the sperm surface during ejaculation ( Manjunath & Thérien, 2002 ). The gene codification for these BSP proteins is BSP5a (25 kDa band protein), BSP5c (RSVP-22), BSP5b (RSVP-20) and BSP1 (RSVP-14) ( Serrano et al. , 2015 ). RSVP-20 and -22 function as decapacitating factors binding to the membrane domains in a first step ( Barrios et al. , 2005 ; Cardozo et al. , 2008 ). Both proteins are released from the membrane once capacitation is initiated, contributing to the process in a similar way to RSVP-14 ( Barrios et al. , 2005 ). In this study, we found significant differences between breeds in the content of several SP proteins. Three protein bands of approximately 25, 22 and 20 kDa that we named P25, P22 and P20, respectively presented the lowest levels in Creole breed. It is likely that P22 and P20 could match with the RSVP's described in ovine, and thus they may play a role in membrane stabilization. Likewise, the 43 kDa protein band (P43) showed the lowest concentration in Romney Marsh, while Hampshire showed the highest value. One might speculate that P43 match with clusterin, which has been described as a glycoprotein in mammalian SP ( Blaschuk et al. , 1983 ). This is the most plentiful glycoprotein found in the testis network and has the characteristics of an acidic protein ( Jobim et al. , 2005 ). In denaturing conditions, clusterin has an approximate molecular weight of 40-43 kDa. Its function in the male reproductive tract varies depending on its tertiary structure and posttranslational modifications (glycosilations/sulphatations-state) ( Ibrahim et al. , 1999 ), origin (testicular or epididymis) ( Tung & Fritz, 1985 ; O'Bryan et al. , 1990 ), and location (fluid or cell) ( Sensibar et al. , 1993 ; Ibrahim et al. , 1999 ). Clusterin has been reported to be involved in spermatogenesis ( Fritz et al. , 1983 ; Sylvester et al. , 1991 ), protection of spermatozoa against the complement attack from epididymal and uterine secretions ( O'Bryan et al. , 1990 ; Ibrahim et al. , 1999 ), or agglutination of abnormal spermatozoa ( O'Bryan et al. , 1994 ), and acting like an extracellular chaperone against heat shock on membrane molecules ( Humphreys et al. , 1999 ) and against apoptosis ( Bailey et al. , 2002 ). However, clusterin has been identified in morphologically abnormal human ( Muciaccia et al. , 2012 ) and bull ( Ibrahim et al. , 2000 ; Shojaei Saadi et al. , 2013 ) spermatozoa, while it has been associated with all ram spermatozoa in fresh ejaculates, showing different staining patterns ( Ibrahim et al. , 2001 ). SP clusterin (sCLU) negatively correlates with human sperm abnormal morphology, protamine deficiency, and DNA fragmentation as lower sCLU concentrations have been detected in the SP of infertile ( Salehi et al. , 2013 ) and vasectomized ( O'Bryan et al. , 1990 ) men. Therefore, the low relative presence of this potential clusterin in the SP of Romney Marsh is consistent with a lower sperm quality and a high number of spermatozoa with membrane alterations. Therefore, differences in the presence of these proteins in the SP would be related to differences in sperm quality in the three studied breeds and, possibly, to other molecular events related to fertilization in which these proteins are involved.

In conclusion, our findings show significant differences in the sperm quality of fresh ejaculates from Creole, Romney Marsh and Hampshire breeds. The antioxidant enzymatic defence system in SP is related with semen quality and can affect kinematic sperm parameters in the Creole and Romney Marsh breeds. Certain proteins that might play an important role in sperm physiology and quality have also been identified. These results can be useful for improving the identification of SP proteins able to act as biomarkers of fertilizing potential and the semen-based reproductive technologies in mammalian species.

Special thanks to Dr. Cesar Terán for the provision of the environmental information for the development of the study.