Membrane phase behavior during cooling of stallion sperm and its correlation with freezability

Abstract

Stallion sperm exhibits great male-to-male variability in survival after cryopreservation. In this study, we have investigated if differences in sperm freezability can be attributed to membrane phase and permeability properties. Fourier transform infrared spectroscopy (FTIR) was used to determine supra and subzero membrane phase transitions and characteristic subzero membrane hydraulic permeability parameters. Sperm was obtained from stallions that show differences in sperm viability after cryopreservation. Stallion sperm undergoes a broad and gradual phase transition at suprazero temperatures, from 30–10°C, whereas freezing-induced dehydration of the cells causes a more severe phase transition to a highly ordered gel phase. Sperm from individual stallions showed significant differences in post-thaw progressive motility, percentages of sperm with abnormal cell morphology, and chromatin stability. The biophysical membrane properties evaluated in this study, however, did not show clear differences amongst stallions with differences in sperm freezability. Cyclodextrin treatment to remove cholesterol from the cellular membranes increased the cooperativity of the suprazero phase transition, but had little effects on the subzero membrane phase behavior. In contrast, freezing of sperm in the presence of protective agents decreased the rate of membrane dehydration and increased the total extent of dehydration. Cryoprotective agents such as glycerol decrease the amount of energy needed to transport water across cellular membranes during freezing.

Keywords::

• Cryopreservation
• Fourier transform infrared spectroscopy
• membrane hydraulic permeability

Introduction

A large part of the stallion population does not qualify for semen freezing programs, because of unsatisfactory post-thaw sperm quality and fertility rates. Stallions are designated as ‘good’ or ‘poor’ freezers based on post-thaw motility characteristics (Tischner 1979, Vidament et al. 1997, Loomis and Graham 2008). The mechanisms underlying the large male-to-male variation in survival rates after freezing and thawing of stallion sperm, however, are largely unknown. Identification of cellular properties that correlate with sperm freezability may help designing customized freezing protocols for individuals (Loomis and Graham 2008, Mazur et al. 2008).

When semen is cryopreserved, cells are exposed to cold shock, ice crystal formation and cellular dehydration, which all can cause irreversible damage (Mazur 1984, Amann and Pickett 1987, Hammerstedt et al. 1990). In the suprazero temperature regime, thermotropic membrane lipid phase changes have been implicated in chilling injury of a variety of cells. Passage through the so-called gel-to-fluid membrane phase transition is associated with leakage of solutes to the extracellular environment, which is detrimental to cells (Crowe et al. 1989, Drobnis et al. 1993). Furthermore, alterations in the structural organization of membrane components are often irreversible (Steponkus 1984, Gousset et al. 2002). At subzero temperatures, ice nucleation causes cellular dehydration, due to an increased solute concentration in the unfrozen fraction. In addition, cellular membranes undergo a highly cooperative transition to a more ordered gel phase during freezing (Wolkers et al. 2007, Balasubramanian et al. 2009, Oldenhof et al. 2010). This is likely caused by the removal of bound water from the phospholipid head groups, which increases van der Waals interactions between the lipid acyl chains resulting in a denser packing of the bilayer (Crowe et al. 1990, Popova and Hincha 2003). The extent and rate of freezing-induced cellular and membrane dehydration is dependent on the cooling rate and the temperature at which ice nucleation takes place, and is influenced by cryoprotective agents (Mazur 1984, Amann and Pickett 1987, Oldenhof et al. 2010). Upon re-warming and thawing, extracellular ice melts and the dehydrated cells are exposed to hypotonic conditions.

The osmotic response of a cell upon freezing and thawing is dependent on the membrane permeability to water as well as the permeability of the membrane to cryoprotective agents. The subzero membrane permeability to water is one of the key parameters determining survival after freezing and thawing of cells (Mazur 1963, Levin et al. 1976). Membrane hydraulic permeability is cell specific and is dependent on osmotic properties as well as on membrane composition. In addition, the membrane permeability for water changes in the presence of cryoprotective agents and is thus dependent on the composition of the freezing extender (reviewed in Sieme et al. 2008).

We recently showed that Fourier transform infrared spectroscopy (FTIR) can be used to derive subzero hydraulic permeability properties of cells, from freezing-induced changes in membrane phase state (Balasubramanian et al. 2009, Oldenhof et al. 2010, Akhoondi et al. 2011). With this technique, changes in membrane conformational disorder (fluidity) during heating or cooling can be studied by following the peak position of the symmetric CH₂ stretching vibration absorption band that arises from the lipid acyl chains as a function of temperature. The water-to-ice phase change during freezing can be determined by following the band area of the H₂O-libration and -bending combination vibration band as a function of temperature. The subzero membrane hydraulic permeability can be estimated by assuming that freezing-induced changes in membrane phase state are proportional to cellular dehydration. Cryomicroscopy and differential scanning calorimetry (DSC) are alternative methods to assess subzero water transport parameters. Cryomicroscopy can be used to directly study freezing-induced cellular dehydration, but cannot be used for small non-spherical cells such as sperm. DSC, which has been used to determine subzero membrane properties of sperm (Devireddy et al. 1998, 2002), has the disadvantage that cellular dehydration is measured indirectly and that it cannot discriminate between transport of free and bound water water.

We hypothesized that differences in sperm freezability are related to differences in membrane phase and permeability properties. In order to test this, we used FTIR to determine supra and subzero membrane phase behavior and characteristic subzero membrane hydraulic permeability parameters for sperm from stallions that show differences in survival after cryopreservation. Biophysical membrane properties (Tm, membrane phase transition temperature; E_Lp, activation energy for membrane hydraulic permeability; Lpg, membrane hydraulic permeability at 0°C) were correlated with pre- and post-freeze sperm characteristics, including motility and membrane integrity. In addition, FTIR was used to establish the effects of removal of cholesterol from the cellular membranes and effects of cryoprotective agents on membrane phase behavior and subzero membrane hydraulic permeability parameters.

Materials and methods

Semen collection and cryopreservation

Semen was collected from 10 stallions (ages 4–17 years) of the Hanoverian warmblood breed, that were held at the National Stud of Lower Saxony in Celle and the Unit for Reproductive Medicine of the University of Veterinary Medicine Hannover. The horses were kept in box stalls bedded with straw, were fed oats and hay three times a day and water was freely available. Semen was collected using an artificial vagina (Model Hanover; Minitube, Tiefenbach, Germany). A breeding phantom was used, and a mare was fixed in front of the breeding phantom. Sterile gauze filtration was performed to remove the gel portion. Semen was evaluated for volume, cell concentration and motility, after which it was diluted with pre-warmed skim milk extender (INRA82; Vidament et al. 2000) at 50 × 10⁶ cells ml^-1, and stored at room temperature until further processing.

Cryopreservation of semen was carried out as previously described (Sieme et al. 2003). For each of 10 stallions three different ejaculates were frozen. In short, extended semen was centrifuged for 10 min at 600 g after which the cell pellet was resuspended in skim milk extender supplemented with 2.5% clarified egg yolk and glycerol, resulting in final concentrations of 2% glycerol and 200 × 10⁶ cells ml^-1. This suspension was cooled to 5°C during 2 h, after which it was packaged in 0.5 ml straws, cooled down to −140°C at 60°C min^-1 using a controlled rate freezer (Minidigitcool; IMV-Technologies, L'Aigle, France) and stored in liquid nitrogen. Thawing of straws was performed in a water bath of 37°C for 30 s.

Assessment of sperm motility, morphology, and chromatin and membrane integrity

The percentages of motile and progressively motile sperm were determined for extended and freeze-thawed sperm samples using computer assisted sperm analysis (Spermvison; Minitube, Tiefenbach, Germany). Samples were incubated for 10 min at 37°C, after which one chamber of a four-chamber counting slide (Leja, Nieuw Vennep, The Netherlands) was filled with 3 μl sample. The slide was maintained at 37°C, and 10 microscopic fields were analyzed per sample, with 60 frames per second, for which averages were calculated.

Evaluation of sperm morphology was performed for 200 cells per sample according to Brito (2007); 100 μl semen was stained with 300 μl nigrosin-eosin solution (10% nigrosin, 0.7% eosin, 3.75 mM Na₂HPO₄, 1.88 mM KH₂PO₄, 5.78 mM NaK tartrate, 3.75 mM glucose).

The sperm chromatin structure assay (SCSA) was used to evaluate chromatin integrity. In this assay, sperm chromatin is treated with acid after which the extent of DNA fragmentation is determined (Evenson et al. 1980, 2002). In short, thawed samples were diluted in TNE (0.15 M NaCl, 0.01 M TRIS-HCl, 1 mM disodium EDTA, pH 7.4), at approximately 2 × 10⁶ cells ml^-1. From this 200 μl was taken, diluted with 400 μl acid solution (0.08 N HCl, 0.15 M NaCl, 0.1% Triton X-100, pH 1.2), and mixed for 30 s. Then, 1.2 ml acridine orange (Polysciences, Warrington, PA, USA) staining solution (0.15 M NaCl, 0.037 M citric acid, 0.126 M Na₂HPO₄, 0.0011 M disodium EDTA, pH 6.0, containing 6 μg ml^-1 acridine orange) was added. Samples were incubated on ice for 3 min, after which 10,000 cells were analyzed using a FACScan flow cytometer (Becton-Dickinson, Heidelberg, Germany). This flow cytometer is equipped with a 488 nm argon ion laser of 15 mw for excitation, and BP 530/30, BP 582/42 and LP 650 nm filters for green, orange and red fluorescence, respectively. Acridine orange stains normal double-stranded DNA green, and denatured single-stranded DNA red; the acid treatment potentially denatures damaged DNA. The DNA fragmentation index (DFI) was calculated from the fractions of cells with single and double stranded DNA.

For assessment of plasma and acrosomal membrane integrity, a double staining with propidium iodide (PI; Sigma-Aldrich, St Louis, MO, USA) and fluorescein labeled peanut agglutinin (FITC-PNA; Vector Laboratories, Burlingame, CA, USA) was performed, and samples were analyzed using a flow cytometer (Cell Lab Quanta SC MPL; Beckman-Coulter, Fullerton, CA, USA). This flow cytometer is equipped with a 488 nm argon ion laser of 22 mw for excitation, and BP 525/30, BP 590/30 and LP 670 nm filters for green, orange and red fluorescence, respectively. HBS of 300 mOsm kg^-1 (20 mM HEPES pH 7.4, 137 mM NaCl, 10 mM glucose, 2.5 mM KOH) was used as sheath fluid. Ten μl sperm sample (50 × 10⁶ cells ml^-1 in extender) was diluted in 480 μl HBS, and 2.5 μl 500 μg ml^-1 PI and 7.5 μl 600 μg ml^-1 FITC-PNAwere added; resulting in 1 × 10⁶ cells ml^-1, 2.5 μg ml^-1 PI and 9 μg ml^-1 FITC-PNA Samples were incubated for 10 min at room temperature, and 10,000 cells were measured. Sperm were selected based on their side scatter and electronic volume properties. Cells that have a plasma membrane that is permeable for PI will show red fluorescence of PI bound to DNA; cells that have a damaged acrosome will show green fluorescence of FITC-PNA bound to glycoproteins that are present in the inner site of the outer acrosomal membrane. Sperm that were both PI- and FITC-PNA-negative were considered viable with intact plasma and acrosomal membranes.

For each stallion, averages with standard deviations were calculated from three ejaculates. This was done for all of the above-mentioned sperm characteristics, before and after freezing and thawing.

Fourier transform infrared spectroscopy studies

FTIR studies were performed as described in detail elsewhere (Oldenhof et al. 2010). For studies in which membrane properties of individual stallions were studied, gel-free semen was used, and at least two different ejaculates. About 1 ml of semen was centrifuged for 20 s at 14,000 g, the supernatant was removed and the pellet was resuspended in 1 ml HBS, after which a pellet was obtained by centrifugation as described above and mounted between two CaF₂ windows. For depletion of cholesterol from cellular membranes, sperm pellets were resuspended in HBS supplemented with 10 mM methyl-β-cyclodextrin (MβCD; Sigma-Aldrich, St Louis, MO, USA) at 50 × 10⁶ cells ml^-1 and incubated for 1 h at 37°C. For testing the effects of protectants, semen diluted with INRA82 was used, or INRA82 supplemented with 5% glycerol. In such cases, sperm pellets obtained after centrifugation were directly used for FTIR analysis, unless otherwise stated.

Infrared absorption measurements were carried out with a Perkin Elmer 100 FTIR spectrometer (Perkin Elmer, Norwalk, CT, USA), equipped with a liquid nitrogen cooled mercury/cadmium/telluride-detector. The optical bench was continuously purged with dry air from an FTIR purge gas generator (Whatman, Clifton, NJ, USA). The acquisition parameters were: 4 cm^-1 resolution, 8 co-added interferograms, 4000–900 cm^-1 wavenumber range. Spectra analysis and display were carried out using Perkin Elmer software (Perkin Elmer, Norwalk, CT, USA) and Omnic software (Thermo-Nicolet, Madison, WI, USA). For temperature-dependent FTIR measurements, a variable temperature sample holder was used (Harrick Scientific Products, Pleasantville, NY, USA) together with a pump system for using liquid nitrogen as a coolant (Linkam Scientific Instruments, Tadworth, Surrey, UK). The sample temperature was monitored using a thermocouple that was located close to the sample (touching the CaF₂ windows). Suprazero membrane phase behavior was studied by cooling the sample from 38–0°C at 2°C min^-1, followed by heating to 70°C at 2°C min^-1. For studying subzero membrane phase behavior, the temperature was decreased from 20°C to −40°C at 1°C min^-1. Spectra were acquired every 20 s. During cooling scans, ice nucleation was induced by touching the sample edge using a liquid nitrogen-cooled copper wire.

Spectra analysis was done as described in detail elsewhere (Wolkers and Hoekstra 2003, Wolkers and Oldenhof 2010). Membrane conformational disorder, which is dependent on the lipid composition and membrane organization, was monitored by observing the position of the CH₂ symmetric stretching band at around 2850 cm^-1 (νCH₂). For each spectrum, acquired at a particular temperature, the inverted second derivative spectrum was calculated using a 9-point smoothing factor, and the 2865–2835 cm^-1 region was selected and normalized such that the νCH₂ lipid band maximum was scaled to one. The position of the lipid band was calculated as the average of the spectral positions at 80%. The lipid band position (νCH₂) was plotted against the sample temperature to reveal the temperature range in which cells undergo membrane phase transitions.

To assess the effect of ice formation on membrane dehydration, wavenumber versus temperature plots were constructed for the different ice nucleation temperatures that were tested. From this, freezing-induced membrane dehydration was calculated as the slope at the onset of ice nucleation (dνCH₂/dT in cm^-1 °C^-1). The freezing-induced membrane dehydration rate (dνCH₂/dt in cm^-1 s^-1) was obtained by multiplying this slope with the cooling rate (0.0167 °C s^-1). From the same data sets, ice formation was monitored by following the H₂O-libration and -bending band at approximately 2200 cm^-1 (νH₂O). Ice formation results in a sudden shift in the position of this band, as well as an increase in area (A). The area under this band (2683–1965 cm^-1 region) was calculated and plotted as a function of temperature. Ice nucleation temperatures (Tn) were determined as the onset-point at which the A(νH₂O) sharply increased.

Water transport model during freezing

The cellular volume response upon freezing has been described by the water transport model (Mazur 1963, Levin et al. 1976, Toner et al. 1990). The water transport model was further modified to incorporate the effect of cryoprotectants on the volume response (Karlsson et al. 1993, Devireddy et al. 2002):

1

Parameters and variables are summarized in Table I. In this equation, the permeability of the plasma membrane to water (Lp in m s^-1 Pa^-1) is defined by an Arrhenius relationship. An Arrhenius plot displays the logarithm of a reaction rate plotted against the inverse of the temperature, and gives a straight line for a single rate-limited thermally activated process:

Table I. Water transport model parameters and variables from equations (1) and (2).

Symbol	Description	Value	Unit
V	Volume	(Variable)	m^3
T	Temperature	(Variable)	K
TR	Reference temperature	273.15	K
Lp	Membrane hydraulic permeability	(Variable)	m s^-1 Pa^-1
Lpg	Permeability of the membrane to water at TR	(Parameter)	m s^-1 Pa^-1
ELp	Activation energy for the permeation process	(Parameter)	J mol^-1
A	Membrane surface area of the cell	150 10^-12	m^2
R	Universal gas constant	8.314	J K^-1 mol^-1, or m^3 Pa K^-1 mol^-1
B	Cooling rate	0.0167	K s^-1
vw	Molar volume of water	18 10^-6	m^3 mol^-1
Vo	Isotonic volume	50 10^-18	m^3
Vb	Osmotically inactive volume	30 10^-18	m^3
vs	Dissociation constant of salt	2	(-)
ns	Number of moles of solutes in the cell	5.694 10^-15	mol
▵Hf	Heat of fusion of water	6011	J mol^-1
ncpa	Number of moles of CPA (5% glycerol) in the cell	1.4 10^-14	mol
vcpa	Molar volume of CPA (5% glycerol)	7.303 10^-5	m^3 mol^-1

2

where Lpg is the membrane hydraulic permeability at 0°C (273.15 K), and E_Lp the activation energy for water transport through the membranes during freezing (in J mol^-1). Thus, E_Lp expresses the temperature-dependence of subzero water transport through the cellular membrane. Dimensions of stallion sperm were taken from Devireddy et al. (2002).

Assessment of membrane permeability parameters from FTIR data

Methods to derive membrane permeability parameters (Lp, E_Lp, Lpg) from freezing-induced membrane phase behavior data as determined using FTIR has been described in detail elsewhere (Oldenhof et al. 2010, Akhoondi et al. 2011). In short, it was assumed that the freezing-induced shift in νCH₂ is proportional to the reduction in cellular volume that occurs due to water transport out of the cell in response to freezing. This assumption allows to construct a normalized volume versus temperature plot, needed to derive the membrane permeability parameters using Equations 1 and 2.

The νCH₂ versus temperature plots obtained at different nucleation temperatures were first baseline corrected for the thermotropic decrease in νCH₂ with decreasing temperature. The relative shift in νCH₂ was scaled such that the normalized volume was one, just prior to nucleation. The maximum shift in νCH₂, as obtained at the highest nucleation temperature, was scaled to correspond to maximum cellular dehydration to the osmotic inactive volume V_b (0.6 × V_o). Partial dehydrating conditions, as obtained for lower nucleation temperatures, were scaled relative to this. The obtained normalized volume (V/V_o) versus temperature plots were used to calculate Lp at different nucleation temperatures. The slope at nucleation that was determined from such plots was multiplied with V_o to obtain dV/dT, which in turn was used to calculate the membrane hydraulic permeability (Lp) using Equation (1).

The natural logarithm of Lp was plotted as a function of the inverse of the nucleation temperature (Tn) in an Arrhenius plot, from which the activation energy (E_Lp) and the membrane hydraulic permeability at 0°C (Lpg) were derived. Alternatively, ln(dνCH₂/dt) was plotted versus 1/Tn to derive the activation energy for the membrane dehydration rate without performing a volume normalization procedure (E_a).

Assessment of correlations between subzero membrane hydraulic permeability parameters and sperm motility and membrane integrity was performed using linear regression analysis.

Results

Membrane phase transitions in stallion sperm during cooling

FTIR was used to study membrane phase behavior of sperm at supra as well as subzero temperatures, to determine fluid-to-gel phase transitions taking place during cooling. Stallion sperm pellets were cooled, while acquiring FTIR spectra. The wavenumber position of the symmetric CH₂ stretching vibration band (νCH₂ at 2850 cm^-1), arising from the membrane lipids, was plotted as a function of the temperature. Stallion sperm membranes are in the fluid phase at 37°C, and fluidity decreases with decreasing temperature. Cellular membranes undergo a broad phase transition from about 30–10°C (Figure 1A), which most likely is the result of reorganization of membrane components. Upon ice nucleation, which initiates freezing-induced cellular dehydration, membranes undergo a sharp phase transition and form a highly ordered gel phase (Figure 1B).

Figure 1. Membrane phase behavior of stallion sperm, as determined using FTIR. Sperm pellets were cooled from 40 down to 0°C with 2°C min^-1 (A) and from 20 to −30°C with 1°C min^-1 (B). Ice nucleation was induced at −2°C. The band position of the symmetric CH₂ stretching vibration band arising from endogenous lipids was determined and plotted as a function of the sample temperature. Stronger packing of phospholipids results in a decrease in the wavenumber position of this band. The temperature-dependent decrease in membrane conformational disorder or fluidity is illustrated with lines. Discontinuities from this indicate when cellular membranes undergo phase transitions, which are indicated with arrows.

Suprazero membrane phase behavior for sperm from stallions that differ in freezability

Membrane phase behavior of cellular membranes is dependent on their composition and organization. Figure 2 illustrates the effect of methyl-β-cyclodextrin (MβCD) treatment on membrane phase behavior of stallion sperm. Removal of cholesterol from the cellular membranes results in a more cooperative phase transition, and makes membranes more fluid, particularly at higher temperatures. The more pronounced phase transition after cholesterol depletion by MβCD is illustrated more clearly after first derivative analysis (Figure 2B). The main phase transition temperature (Tm) occurring at 24°C in untreated sperm increases to 32°C after MβCD-treatment.

Figure 2. Membrane phase behavior of stallion sperm (closed circles, solid line), as well as sperm that was MβCD-treated for removal of cholesterol from cellular membranes (open circles, dashed line). FTIR spectra were collected from sperm pellets that were warmed from 0–60°C with 2°C min^-1. In panel A, νCH₂ arising from endogenous lipids is plotted as a function of the sample temperature. Panel B shows first derivatives of these curves, to visualize membrane phase transitions more clearly.

Pre-freeze and post-freeze characteristics were determined for sperm from 10 stallions. Characteristic parameters that were determined included sperm morphology, chromatin stability, plasma and acrosomal membrane integrity, and progressive motility (Table II). Sperm obtained from different stallions showed differences in sperm survival after cryopreservation. Based on this, the 10 stallions were divided into two groups: Five that showed more than 26% post-thaw progressive motility (A–E) and five that showed less than 26% (F–J).

Table II. Pre- and post-freeze characteristics for sperm from 10 stallions, which differ in freezability (A–J). Evaluation of sperm morphology (cells with abnormal morphology) was done by microscopic observations, chromatin integrity (DNA fragmentation index) and plasma and acrosomal membrane integrity (viable PI- and PNA-negative cells) were determined using flow cytometry, and motility (progressive motility) was determined using computer assisted sperm analysis. For each stallion, averages ± standard deviations were determined from three ejaculates. Subzero membrane hydraulic permeability parameters (E_a, E_Lp, Lpg) were determined by FTIR, from (V/V_o-normalized) νCH₂ band analysis. Parameters were derived from Arrhenius plots in which ln(dνCH₂/dt) or ln(Lp) was plotted as a function of the ice nucleation temperature. Arrhenius plots were constructed from data obtained for 8–10 nucleation temperatures, using two ejaculates per stallion, and regression coefficients (R²) are indicated.

	Pre-freeze characteristics				Post-freeze		FTIR -νCH2 data analysis
	Cells with abnormal morphology (%)	DNA fragmentation index (%)	PI- and PNA-negative cells(%)	Progressively motile cells (%)	PI- and PNA-negative cells	Progressively motile cells (%)	Ea (R^2) (kcal mol^-1)	ELp (R^2) (kcal mol^-1)	Lpg (μm min^-1 atm^-1)
Stallion A	31 ± 3.7	7.1 ± 0.1	85 ± 3.7	59 ± 4.4	28 ± 6.0	36 ± 5.7	17 (0.73)	59 (0.95)	0.0034
Stallion B	20 ± 4.8	6.7 ± 1.0	82 ± 3.1	61 ± 3.0	45 ± 11	34 ± 4.3	19 (0.82)	49 (0.91)	0.0029
Stallion C	39 ± 13	7.0 ± 0.8	81 ± 3.0	41 ± 5.2	46 ± 2.5	32 ± 2.8	16 (0.81)	57 (0.85)	0.0029
Stallion D	16 ± 1.6	9.3 ± 0.7	66 ± 8.8	60 ± 6.8	26 ± 16	31 ± 6.6	20 (0.92)	63 (0.97)	0.0031
Stallion E	33 ± 7.3	11 ± 1.6	67 ± 6.4	56 ± 7.6	40 ± 13	30 ± 5.1	15 (0.94)	48 (0.98)	0.0021
Stallion F	9.3 ± 2.7	5.9 ± 0.4	71 ± 9.1	63 ± 5.4	15 ± 5.1	26 ± 5.4	24 (0.81)	79 (0.93)	0.0042
Stallion G	44 ± 3.3	6.1 ± 0.6	79 ± 4.8	66 ± 6.6	25 ± 8.6	20 ± 2.1	21 (0.88)	53 (0.96)	0.0025
Stallion H	67 ± 4.3	13 ± 0.5	70 ± 4.6	29 ± 11	43 ± 13	16 ± 2.4	25 (0.96)	76 (0.93)	0.0037
Stallion I	58 ± 5.5	9.1 ± 1.7	77 ± 3.1	43 ± 4.2	45 ± 3.2	13 ± 1.3	14 (0.83)	62 (0.95)	0.0066
Stallion J	29 ± 3.1	30 ± 6.6	58 ± 4.4	35 ± 12	32 ± 1.9	3.8 ± 0.9	20 (0.91)	54 (0.91)	0.0015

Suprazero membrane phase behavior of sperm from these 10 stallions was determined using FTIR, with the results shown in Figure 3. First derivatives of the νCH₂ vs. temperature plots were calculated to visualize phase transitions more clearly. In all cases, sperm undergo a broad gel-to-fluid phase transition from about 10–30°C. The 10 stallions did not show clear differences in membrane phase behavior that could explain individual differences in sperm freezability.

Figure 3. Membrane phase behavior of sperm from 10 stallions, which differ in freezability (A–J: with decreasing post-thaw progressive motility, as listed in Table II). FTIR spectra were recorded for sperm pellets that were warmed from 0–60°C at 2°C min^-1. The band position of the symmetric CH₂ stretching vibration band arising from endogenous lipids was determined and plotted as a function of the temperature (left y-axis, filled and open squares). Sperm from stallions that had a post-thaw progressive motility higher or lower than 26% after cryopreservation are indicated with filled (A–E) and open squares (F–J), respectively. Average membrane phase behavior curves ± standard deviation were calculated from two to three ejaculates per stallion. Single measurements are indicated with grey lines. First derivatives were calculated to visualize the phase transition more clearly (right y-axis, black lines).

Extent and rate of freezing-induced membrane dehydration are affected by nucleation temperature and protectants

Freezing-induced membrane-phase behavior was studied for different ice nucleation temperatures that result in differences in the extent of membrane dehydration. Upon extracellular ice formation at high nucleation temperatures, which result in cellular dehydration, membranes undergo a profound membrane phase transition to a more ordered gel phase. By contrast, at low nucleation temperatures that likely result in supercooling of the intracellular fluid, membranes only undergo a small and gradual decrease in membrane conformational disorder and remain relatively fluid (Figure 4A).

Figure 4. Membrane phase behavior of stallion sperm pellets that were cooled from 20°C to −30°C at 1°C min^-1, as determined using FTIR. Panel A shows freezing-induced membrane dehydration for semen washed in HBS and then exposed to various ice nucleation temperatures. Panel B shows freezing-induced membrane dehydration for semen that was diluted in INRA82 (black squares) or INRA82 supplemented with 5% glycerol (open squares), as well as semen that was diluted in INRA82 after which it was washed using HBS (gray filled squares). The data points reflect the relative shift in position of the symmetric CH₂ stretching band (▵νCH₂). Panel C shows Arrhenius plots in which the natural logarithm of the membrane dehydration rate (dνCH₂/dt) is plotted as a function of the inverse of the nucleation temperature (Tn). From this, via linear regression analysis, activation energies (E_a) were determined that describe the temperature dependence of the cell membrane dehydration rate at subzero temperatures in the absence and presence of INRA82 as well as INRA82 supplemented with 5% glycerol.

Addition of protectants affects the extent and rate of freezing-induced membrane dehydration, visualized as the wavenumber shift (▵νCH₂) at low subzero temperatures and the slope at ice nucleation, respectively (Figure 4B). Use of skim milk extender (INRA82) results in more severe membrane dehydration compared to isotonic buffer (HBS). This is likely due to the non-permeable components that are present in the skim milk extender that result in a more hypertonic non-frozen fraction upon extracellular ice formation. Addition of the cryoprotective agent glycerol, results in a more gradual membrane dehydration. The membrane dehydration rate at the onset of ice nucleation exhibits Arrhenius behavior between −2 and −12°C (Figure 4C).

Subzero membrane hydraulic permeability parameters, and their correlation with freezability

Subzero membrane hydraulic permeability values (Lp) were estimated from freezing-induced membrane phase behavior data. Figure 5A shows Arrhenius plots of Lp in isotonic buffer and in skim milk extender supplemented with protectants. Freezing in skim milk extender lowers the activation energy of subzero water transport, as compared to native sperm washed and frozen in buffer. Addition of the cryoprotectant glycerol further decreases the activation energy. This means that in the presence of glycerol, the energy needed for transport of water across the membrane decreases. Cholesterol depletion from sperm membranes by MβCD did not drastically affect subzero water transport (Figure 5B), despite its major effect on suprazero membrane phase behavior (Figure 2).

Figure 5. Arrhenius plots of the membrane hydraulic permeability (Lp) as derived from volume-normalized FTIR-νCH₂ data, as a function of the nucleation temperature (Tn). Panel A shows Arrhenius plots for stallion sperm frozen in the presence (black filled squares) and absence (gray filled squares) of INRA82 as well as INRA82 supplemented with 5% glycerol (open squares). Panel B shows plots for stallion sperm that was MβCD-treated for removal of cholesterol from cellular membranes (filled circles: non-treated, open circles: MβCD-treated). Panel C shows Arrhenius plots for sperm from stallions that exhibited differences in survival after cryopreservation. Sperm from stallions that had a post-thaw progressive motility higher or lower than 26% after cryopreservation are indicated with filled and open squares, respectively (five stallions in each group, 8–10 nucleation temperatures on two ejaculates per stallion). Activation energies (E_Lp) derived from these data for presence and absence of protectants as well as MβCD-treated sperm are indicated in the plots. For the individual stallions they are listed in Table II, while averages ± standard deviation for the five ‘good’ and ‘poor’ freezer stallions are indicated in panel C.

In order to determine if sperm from stallions that exhibited differences in cryosurvival exhibited differences in subzero water transport parameters, E_Lp and Lpg values were calculated from Arrhenius plots of Lp in the absence of protectants (Table II). Figure 5B shows Arrhenius plots of sperm from stallions that were grouped as either having higher or lower than 26% post-thaw progressive motility after cryopreservation using a conventional protocol (five stallions in each group), with E_Lp values of 55 ± 3.3 and 65 ± 12 kcal mol^-1, respectively. Figure 6 shows correlation plots of pre- and post-thaw progressive motility (Figure 6A), post-thaw progressive motility and post-thaw membrane integrity (Figure 6B), as well as post-thaw progressive motility and subzero membrane permeability parameters (Figure 6C and D). For the membrane parameters, no clear differences were observed that correlate with sperm freezability.

Figure 6. Correlation between pre- and post-freeze progressive motility (A) and post-thaw progressive motility and post-thaw membrane integrity (B) for 10 stallions, as well as the correlation between the membrane permeability parameters E_Lp (C) and Lpg (D) and post-thaw progressive motility. Sperm from stallions that had a post-thaw progressive motility higher or lower than 26% after cryopreservation are indicated with filled and open squares, respectively. Averages ± standard deviation for motility and membrane integrity percentages were calculated from three ejaculates per stallion. Membrane hydraulic permeability parameters were derived from freezing-induced membrane dehydration analysis for 8–10 nucleation temperatures from two ejaculates per stallion.

Discussion

In this study, FTIR was used for in situ determination of supra- and subzero membrane phase and permeability properties of sperm from stallions that exhibited differences in survival after cryopreservation. It was found that stallion sperm undergoes a broad and gradual phase transition at suprazero temperatures, whereas freezing-induced dehydration results in a more severe fluid-to-gel phase transition. The extent and rate of freezing-induced membrane dehydration is affected by the solution in which sperm are frozen. Sperm from individual stallions show clear differences in post-thaw progressive motility, percentages of sperm with abnormal cell morphology, and chromatin stability. The biophysical membrane properties (Tm, E_Lp, Lpg) that were evaluated in this study, however, did not show clear differences for sperm from individuals that exhibit differences in post-thaw progressive motility after cryopreservation.

Effects of temperature on the structure, organization and phase behavior of sperm membranes have been studied in some detail (Holt and North 1986, De Leeuw et al. 1990, Drobnis et al. 1993, Ricker et al. 2006). Cooling can cause phase separation and rearrangement of membrane components, events that appear to be irreversible on warming and cause membranes to become leaky. Phospholipids that were isolated from stallion spermatozoa have a phase transition temperature of 21°C, as determined by DSC (Parks and Lynch 1992). This is in good agreement with our FTIR membrane phase measurements on intact sperm showing a main phase transition at 24°C. Our studies show reduced membrane phase behavior transition temperatures and overall higher membrane fluidity as compared to Ricker et al. (2006), who determined a main lipid phase transition at 35°C for stallion sperm. We show that stallion sperm membranes are predominantly in the fluid phase at physiological temperatures.

We have previously shown that stallion sperm membranes display a highly cooperative transition to a more ordered gel phase during freezing, from which subzero hydraulic membrane permeability properties that describe water transport across the cellular membrane during freezing can be derived (Oldenhof et al. 2010). Here we have followed up on this study, and correlated subzero hydraulic permeability parameters (E_Lp, Lpg) of stallion sperm with freezability (post-thaw progressive motility). Hydraulic permeability parameters are clearly changed in the presence of cryoprotective agents. Glycerol for example, decreases the activation energy for water transport. The subzero hydraulic permeability parameters, however, do not display a clear male-to-male variation correlating with sperm freezability. It should be noted that in contrast to post-thaw motility, membrane integrity also did not show large variation, which could explain the lack of differences in membrane phase behavior and permeability parameters amongst stallions. In order to focus on intrinsic cellular properties, the current FTIR studies on individual stallions were performed in the absence of protectants. We recently found that sperm from ‘good’ and ‘poor’ freezer stallions that were cryopreserved with increasing concentrations of various cryoprotective agents exhibited different post-thaw osmotic tolerance limits (Hoffmann et al. 2011).

Interspecies variation in sperm freezability has previously been associated with differences in membrane phospholipid composition and cholesterol-lipid ratios (Parks and Lynch 1992). Cholesterol content greatly affects membrane fluidity and the cooperativity of the phase transition, as can be detected using FTIR (McMullen et al. 1994, Wolkers et al. 2002). Also, differences in stallion sperm quality and susceptibility to lipid peroxidation have been correlated with membrane lipid composition (Brinsko et al. 2007, Marcías García et al. 2011). Membrane modification strategies have been implicated to improve cryosurvival of sperm. Such strategies include incubation of sperm with liposomes to modify the membrane lipid composition (De Leeuw et al. 1993, Zeron et al. 2002, Ricker et al. 2006, Röpke et al. 2011) or the use of cholesterol-loaded cyclodextrin to enrich the membrane cholesterol content (Müller et al. 2008, Moore et al. 2005). Use of cyclodextrin to remove cholesterol was shown here to have a drastic effect on the suprazero membrane phase behavior, whereas relatively small effects were observed on subzero membrane phase behavior. The phase transition occurring at 32°C after removal of cholesterol is characteristic for sphingolipids (Gousset et al. 2002, Wolkers et al. 2002).

In summary, in this study the correlation between cryosurvival and biophysical membrane properties of stallion sperm has been evaluated. Cryoprotectants have a distinct effect on freezing-induced membrane phase behavior and subzero water transport processes. The intrinsic membrane phase and hydraulic permeability parameters, however, did not show clear differences amongst sperm from ‘good’ and ‘poor’ freezers. We postulate that differences in freezability of semen may be related to differences in the cellular or membrane response to addition of cryoprotectants.

Declaration of interest: This work was financially supported by the German Research Foundation (Deutsche Forschungsgemeinschaft [DFG]), Cluster of Excellence ‘From regenerative biology to reconstructive therapy’ (REBIRTH). The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.