Reprod Dom Anim 45, 142146 (2010); doi: 10.1111/j.1439-0531.2008.01287.

x ISSN 0936-6768

Establishment of an ELISA for Measuring Bovine Pregnancy-Associated Glycoprotein in Serum or Milk and Its Application for Early Pregnancy Detection
M Friedrich and W Holtz
Institute of Animal Breeding and Genetics, Go ttingen, Germany

Contents
It has been shown in several mammalian species that during pregnancy, trophoblast cells express a range of pregnancyassociated glycoproteins (PAG). The presence of PAG in the maternal serum of cows may serve as an indicator of pregnancy from day 28 after AI onward. The present study addresses (1) conversion of an existing PAG-RIA to a competitive double antibody ELISA using a polyclonal antibPAG-IgG and an anti-rabbit-IgG raised in sheep for coating and (2) application of newly established ELISA to test its suitability for pregnancy detection by measuring PAG in serum or milk. The intraassay coecients of variation (CV) for the PAG-ELISA were 214% for serum and 1012% for milk; the corresponding interassay CVs were 822% and 1222%, respectively. Pregnancy-associated glycoprotein proles established in milk and serum of 12 pregnant cows showed a characteristic pattern with measurable amounts from approximately day 20 onwards in serum and from day 60 onwards in milk. In a eld trial, serum PAG was determined in 397 cows sampled between 20 and 50 days after insemination. The outcome was that, pregnancy could reliably be diagnosed from day 28 onwards in serum and from day 150 onwards in milk. In conclusion, it may be stated that the established ELISA provides an ecient and reliable means of pregnancy diagnosis that will, in our judgement, gain in popularity with cattle breeding. The ELISA proved to be an adequate and ecient way of measuring PAG in maternal serum or milk and will be a useful means of pregnancy detection in cows.

associated glycoprotein is also detectable in milk (Tainturier et al. 1996; Metelo et al. 2004), but to date, attempts to base a pregnancy diagnosis on the analysis of a milk sample have only been reported for goats (Gonzalez et al. 2001). This paper addresses the establishment of an ELISA and its applicability to detect pregnancy in cows by analysing blood or milk samples.

Materials and Methods

Establishment of a PAG-ELISA On the basis of an existing RIA (Zoli et al. 1992), a PAG-ELISA was established as described by Miyai et al. (1976), Mutayoba et al. (1990) and Moller (1991). The competitive two-step immunometric assay uses polyclonal anti-rabbit-IgG sheep antiserum (coating antiserum, cAS) and anti-bPAG1-IgG polyclonal rabbit antiserum (sAS) for specic binding of PAG. Pure stock bPAG-1 and specic antiserum were provided by Prof. J.F. Beckers (ULG, Belgium). The extraction of PAG from placental tissue was described by Zoli et al. (1991). The cAS was extracted and puried according to Moller (1991) following the procedure described by Harlow and Lane (1988). Microtitre plates (96-well, NUNC MaxiSorb; NALGENE Nunc, Langenselbold, Germany) were pre-incubated with cAS and stored at )20C until use. Puried bPAG-1 was biotinylated with 20-fold concentration of BcapNHS (Sigma-Aldrich, Seelze, Germany) to serve as tracer (BbPAG-1) (Hennies 1994). For validation, the PAGELISA was compared with an established RIA (Zoli et al. 1992). The minimum detectable level was assessed by conducting 20 measurements of pooled PAG-free serum. Cross-reactivity (CR) of the antisera has been tested for foetal proteins already (Perenyi et al. 2002). In this approach, CR was examined for pregnancyrelevant hormones such as bovine FSH, LH and progesterone which are expected to be very low. Anti-bPAG1-IgG polyclonal rabbit antiserum was diluted 1 : 200 000 in assay buer [0.1% casein, 0.005 M NaOH, 0.12 M NaCl, 0.02 M Na2HPO4, 0.01 M EDTA, 0.002% phenol red, 0.005% chlorhexidine-digluconat (20%), pH 7.3 (24)] and incubated at +4C on coated plates (100 ll per well) for at least 10 h to maximize sensitivity of the assay. Standard curves and control samples were prepared from puried bPAG in pooled PAG-free serum from heifers and bulls with concentrations of 12.5, 6.25, 3.125, 1.56, 0.78, 0.39 and 0.0 ng ml. For the determination of PAG in milk, consumable skimmed ultra-high-temperature (UHT) milk, shown to be devoid of PAG by RIA and ELISA, was used for the preparation of standard curves and control samples.

Introduction
Pregnancy is traditionally diagnosed by manual or ultrasonographic examination per rectum. Alternatively, immunological detection of pregnancy-specic substances in maternal serum such as pregnancy-specic protein B (PSPB) (Sasser et al. 1986), pregnancyassociated glycoprotein (PAG) (Zoli et al. 1992) or pregnancy serum protein (PSP60) (Mialon et al. 1994) may be conducted. In a number of mammals, these substances, belonging to the highly polymorphic family of aspartic proteinases (Green et al. 1998), originate in placental cells (Hughes et al. 2000). They are prospectively named PAG (Beckers et al. 1999). Various PAGs dier in molecular weight, amino acid sequence (Xie et al. 1991) and degree of glycosylation (Klisch et al. 2005). Detection of PAG by way of RIA (Sasser et al. 1986; Zoli et al. 1991) or ELISA (Friedrich and Holtz 2004; Friedrich et al. 2005; Green and Roberts 2006; Gabor et al. 2007) in maternal serum has been shown to be a useful means of early pregnancy detection in cows (Sasser et al. 1986), goats (Gonzalez et al. 1999) and sheep (Karen et al. 2003; Ledezma-Torres et al. 2006). In cows, it was found to be reliable from day 30 after insemination onwards (Szenci et al. 1998). Pregnancy-

2008 The Authors. Journal compilation 2008 Blackwell Verlag

Establishment of an ELISA for Measuring Bovine PAG and Its Application

143

Standards and samples of serum and milk were diluted 1 : 1 with assay buer and incubated on preincubated 96-well plates before adding the tracer, as described by Zoli et al. (1992). The tracer (BbPAG-1) was diluted 1 : 800 with assay buer. To each well, 50 ll of the diluted tracer was added, followed by a 90-min incubation at room temperature. After two washing steps (washer Columbus Plus, Tecan, Crailsheim, Germany), streptavidin-peroxidase (50 lg ml) and TMB (12.5 mg 3,3,5,5-Tetramethylbenzidine per 1 ml DMSO, SIGMA) were added and incubated as described elsewhere (Moller 1991). Optical density was measured photometrically at a wave length of 450 nm using a TECAN SUNRISE photometer with software MAGELLAN 4.0 (Tecan). Concentrations were calculated using a logit-log transformation according to Rodbard (1974). For further validation, serum samples obtained from 171 cows at various stages of pregnancy were analysed with both RIA (laboratory of J.F. Beckers in Liege) and ELISA (own laboratory). Accuracy and recovery rates (RR) were assessed for 0, 10 and 20 h of pre-incubation of specic antisera and 0, 10 and 20 h of sample incubation before addition of biotinylated tracer. Application of the PAG-ELISA Blood samples were collected from the tail vein (Vena caudalis mediana) of 20 non-pregnant cows and twenty 35 weeks-pregnant cows. After 12 h, at room temperature, the blood was centrifuged at 1500 g for 10 min to obtain serum. Milk samples were stripped from a healthy quarter after milking on days 40, 60, 80 and 100 of pregnancy. They were immediately cooled to +4C and centrifuged at 620 g for 30 min to remove the lipid fraction. The skimmed milk was frozen within 1 h after sample collection. For the establishment of PAG proles, blood and milk samples were collected as described before from 12 pregnant dairy cows from 3 weeks after AI until 2 weeks postpartum. Until 10 weeks after insemination, samples were collected weekly, from 10 weeks onwards biweekly. For technical reasons, milk samples collected in the course of the eld study had to be stored at +4C for 8 h before centrifugation. Blood serum and skimmed milk were stored at )20C until being assayed for PAG content. To test the suitability of the PAG-ELISA in the eld, serum samples were obtained from 397 cows 1050 days after AI. In these animals, pregnancy was conrmed by way of manual (274 cows) or ultrasonographic examination (38 cows) per rectum or by an additional PAG examination 4 weeks later (110 cows). In 25 cows, both rectal examination and the additional PAG examination were conducted. Data were analysed using SPSS 12.0.1 (SPSS GmbH Software, Munchen, Germany) for one-sided ANOVA and SAS (PROC NLIN) (SAS Institute, Cary, NC, USA) for non-linear analysis of regression.

accuracy was achieved after an incubation of 10 h with both antiserum and serum or milk samples. The correlation between RIA and ELISA was r = 0.97 (p < 0.01). Doseresponse curves for serum samples and puried PAG (standard curves) ran parallel. The mean concentration in PAG-free serum plus two SDs amounted to 0.4 ng ml. Cross-reactions between PAG and FSH, LH and progesterone ran to less than 0.5% (Fig. 1) as expected, since a displacement of biotinylated tracer at the stage of 50% bound unbound was not reached by the concentrations employed. Concentrations in diluted bPAG-1 standard (STD) and PAG in pooled serum from pregnant cows ran parallel. Coecients of variation (CV) and RR were tested in the range of 0.412.5 ng ml (Table 1). Mean RRs were 104% for serum and 112% for milk. Inter- and intraassay CV at 0.8 ng ml were 13% and 16% for serum and 12% and 22% for milk, respectively. As to be expected, the CV increased with decreasing PAG concentration. The characteristics of the ELISA are summarized in Table 1.

Fig. 1. Doseresponse curves of puried PAG standard (STD) and serum bPAG-1 (PAG) and CR of PAG with LH, FSH and progesterone (P4) (the dotted lines signify 50% of tracer bound unbound)

Table 1. Properties of the competitive double antibody assay with preincubation of the specic antibody for measuring bPAG-1 in blood and milk
Feature Dilution of antisera Tracer Standard Sample volume (ll) Detection range (ng ml) 50% Binding (ng ml) Mean recovery rates (%) Coecient of variation (%) Intraassay 0.8 ng ml 6.2 ng ml Interassay 0.8 ng ml 6.2 ng ml Blood 1 : 200 000 Biotinylated bPAG1 Puried bPAG1 50 0.412.5 2.1 104.1 Milk 1 : 200 000 Biotinylated bPAG1 Puried bPAG1 50 0.412.5 3.2 112.3

Results
The standard curves for competitive direct capture PAG-ELISA showed a sigmoidal pattern. Maximum

13.2 6.5 16.3 8.2

11.8 9.8 22.1 12.0

2008 The Authors. Journal compilation 2008 Blackwell Verlag

144

M Friedrich and W Holtz

Fig. 2. Pregnancy-associated glycoprotein concentrations during pregnancy in serum (left) and skimmed milk (right)

Table 2. Accuracy of a pregnancy diagnosis based on the serum PAG level assessed with the newly established ELISA on dierent days after AI at various threshold values
Correct diagnoses PAG-threshold (ng ml) 1.5 2.0 2.5 1.5 2.0 2.5 1.5 2.0 2.5 1.5 2.0 2.5 Pregnant (%) 87 94 100 89 95 94 95 95 100 97 100 100 Non-pregnant (%) 93 77 74 96 93 82 93 93 89 91 92 63 Overall (%) 89 87 88 91 94 90 95 95 96 96 99 92

Days after AI 2630 (n = 106)

3135 (n = 88)

3640 (n = 57)

>40 (n = 128)

Fig. 3. Serum PAG concentrations (means SEM) in 267 pregnant and 130 non-pregnant cows between days 10 and 50 after AI

As shown in Fig. 2, PAG was detectable in serum well before day 20. Signicant dierences between pregnant and non-pregnant cows were manifest from day 22 after insemination onward. In milk, the PAG concentration amounted to 3.2% of that in serum and from day 60 onwards, the dierence between pregnant and nonpregnant cows was signicant. With milk samples collected in the eld, however, a reliable diagnosis of pregnancy was not possible before day 150 (Fig. 2). The serum PAG concentrations of samples collected from 397 cows 1050 days after AI are presented in Fig. 3. The overall mean for 130 non-pregnant cows was 1.1 ng ml (SEM 0.5, range 0.32.3 ng ml). Deviating individual means result from the odd cow that had conceived but was detected as non-pregnant by the second pregnancy test conducted at a relatively late stage. Pregnancy-associated glycoprotein values for pregnant cows were signicantly increased over those of non-pregnant cows from day 25 onwards (p < 0.01, Dunnetts t-test). From the eld data (Table 2), it may be deduced that, for an early pregnancy diagnosis, a threshold value of 2 ng ml is appropriate. At

3135 days after AI, the overall reliability for the identication of pregnant and non-pregnant animals was 94%.

Discussion
The possibility to detect pregnancies in cows by way of PAG determination in blood has been available for almost 20 years (Sasser et al. 1986). Yet the technique is not utilized to any extent. This might change with increasing herd size and a higher degree of automation in the dairy business. With the ELISA being more ecient than the RIA, the practicability of the PAG test is greatly enhanced. By the use of a heterogeneous two species-competitive-indirect capture-assay (Porstmann and Kiessig 1992), the double antibody approach with pre-incubation of the specic antibody brings about an enhancement of assay sensitivity at low concentrations and reduces the amount of tracer needed (Moller 1991) while providing for higher specicity. The assay, established in 2003 (Friedrich and Holtz 2004), is in constant use in our laboratory every since, initially on an

2008 The Authors. Journal compilation 2008 Blackwell Verlag

Establishment of an ELISA for Measuring Bovine PAG and Its Application

145

experimental basis, more recently as a routine service to the dairy industry. There are more sensitive and rapid detection systems, yet the ELISA is a proven and inexpensive means of quantitative verication of a single substance in biological uids (Revoltella et al. 1998). Markers other than the streptavidin-peroxidase conjugate used by us might have advantages, however, sensitivity, variation coecients and RRs observed are in the range of established assays (Zoli et al. 1991; Ledezma-Torres et al. 2006) and are suitable for reliable pregnancy detection. In USA, a PAG-ELISA operating on the basis of monoclonal antibodies (Green et al. 2005) is utilized for routine pregnancy detection. The earliest possible detection of PAG with the aid of our ELISA turned out to be day 22 of pregnancy in blood and day 60 in milk. The variability of the PAG concentration in milk was much higher than that found in serum (Zoli et al. 1991) and is presumably associated with proteolytic processes taking place during storage of samples (Le Roux et al. 1995), although every attempt was made to chill or freeze the samples immediately after collection. Milk samples were handled in a standardized routine, yet apparently minute amounts of remaining fat may act as a source of interference. Factors, such as milk yield (Lopez-Gatius et al. 2007), milk composition or contamination of milk samples (Friedrich et al. 2007), conceivably even dietary manipulations (Jenkins and McGuire 2006) or stage of lactation (Gaspardy et al. 2004), might contribute to variability. Consequently, at the present stage, under farming conditions, reliable pregnancy detection on the basis of a milk sample is not possible before day 150 of pregnancy. The unusually high PAG concentrations observed in some non-pregnant animals during the eld trial may be explained either by residual PAG in cows inseminated less than 50 days postpartum (Zoli et al. 1992; Sousa et al. 2003) or by cows that had conceived but early embryonic death had occurred (Breukelman et al. 2005). In high-yielding dairy cows, early insemination is rather uncommon as they tend to have prolonged postpartum service periods (Dobson et al. 2007). According to Zoli et al. (1992), another possible source of error could be the (rare) occurrence of accessory PAG-formation. The proportion of false negative diagnoses in cows sampled more than 30 days after AI amounted to 7%. This, according to Perenyi (2002), might be the result of an occasionally occurring delay in the PAG-increase. Therefore, in cases with a PAG concentration of between 1.5 and 2.0 ng ml, another test on a sample taken 1 week later is recommended. Advantages of the PAG determination over a progesterone test as a means of pregnancy detection are (a) that the former can be conducted at any time from day 30 onward, whereas, for the latter, the time of sampling is restricted to the days 1923 after insemination and (b) that, with the progesterone test, the reliability of a positive result is, at best, 80% (Holtz et al. 1986; Sasser and Ruder 1987) when compared with 95100% in case of the PAG test (Sasser et al. 1986; Zoli et al. 1992; Szenci et al. 1998). The necessity for sending samples to a laboratory may delay management decisions; on the other contrary, low

input in cost and labour and high reliability easily compensate for this disadvantage. Since the introduction of the PAG pregnancy test, the demand has steadily risen and exceeds that for progesterone determinations by a wide margin. Farmers appear to be satised with the test and there is an indication that the demand will continue to grow. In conclusion, the serum PAG-ELISA employed in the present investigation proved to be a convenient and reliable means of diagnosing pregnancy in cows. From 30 days, after AI onward, its reliability amounts to more than 94% and, therefore, matches or excels conventional approaches of pregnancy detection. Pregnancy-associated glycoprotein determination in milk is possible, but, due to the low concentrations, demands a more sensitive assay before being of practical relevance. Acknowledgements
The authors would like to thank J.F. Beckers and J. Sulon (University of Liege, Belgium) for their support and advice.

Author contributions
The establishment of the assay, the collection and analysation of the samples, the statistical analysis and writing of the article was conducted by Morten Friedrich. The idea and the design of the study was developed by Wolfgang Holtz who also took part in the data analysis and writing of the article.

References
Beckers JF, Drion PV, Garbayo JM, Perenyi Z, Zarrouk A, Sulon J, Remy B, Szenci O, 1999: Pregnancy associated glycoproteins in ruminants: inactive members of the aspartic proteinase family. Acta Vet Hung 47, 461469. Breukelman SP, Szenci O, Beckers JF, Kindahl H, Mulder EJ, Jonker FH, van der Weijden B, Revy D, Pogany K, Sulon J, Nemedi I, Taverne MA, 2005: Ultrasonographic appearance of the conceptus, fetal heart rate and proles of pregnancyassociated glycoproteins (PAG) and prostaglandin F2alphametabolite (PGF2alpha-metabolite) after induction of fetal death with aglepristone during early gestation in cattle. Theriogenology 64, 917933. Dobson H, Smith RF, Royal MD, Knight CH, Sheldon IM, 2007: The high-producing dairy cow and its reproductive performance. Reprod Domest Anim 42, 1723. Friedrich M, Holtz W, 2004: Establishment of an ELISA to assess PAG-concentrations in blood and milk of dairy cows. Reprod Abstr Ser 31, 21. Friedrich M, Kuwer A, Beckers JF, Holtz W, 2005: Suitability of bovine pregnancy-associated glycoprotein (bPAG-1) ELISA for early pregnancy detection in cows. Reprod Domest Anim 40, 407. Friedrich M, Kalscheuer E, Hoedemaker M, Holtz W, 2007: Pregnancy detection in dairy cows by measuring pregnancyassociated glycoprotein (PAG) in milk. Reprod Domest Anim 42(Suppl. 2), 70. Gabor G, Toth F, Ozsvari L, Abonyi-Toth Z, Sasser RG, 2007: Early detection of pregnancy and embryonic loss in dairy cattle by ELISA tests. Reprod Domest Anim 42, 633636. Gaspardy A, Schwartz Z, Zoldag L, Veresegyhazy T, Fekete S, 2004: Changes in daily energy amounts of main milk components (lactose, protein and fat) during the lactation of high-yielding dairy cows. Acta Vet Hung 52, 457467. Gonzalez F, Sulon J, Garbayo JM, Batista M, Cabrera F, Calero P, Gracia A, Beckers JF, 1999: Early pregnancy

2008 The Authors. Journal compilation 2008 Blackwell Verlag

146 diagnosis in goats by determination of pregnancy-associated glycoprotein concentrations in plasma samples. Theriogenology 52, 717725. Gonzalez F, Sulon J, Calero P, Batista M, Gracia A, Beckers JF, 2001: Pregnancy-associated glycoproteins (PAG) detection in milk samples for pregnancy diagnosis in dairy goats. Theriogenology 56, 671676. Green JA, Parks TE, Avalle MP, Telugu MP, McLain AL, Peterson AJ, McMillan W, Mathialagan N, Hook RR, Xie S, Roberts RM, 2005: The establishment of an ELISA for the detection of pregnancy-associated glycoproteins (PAGs) in the serum of pregnant cows and heifers. Theriogenology 63, 14811503. Green JA, Roberts RM, 2006: Establishment of an ELISA for the detection of native bovine pregnancy-associated glycoproteins secreted by trophoblast binucleate cells. Methods Mol Med 122, 321330. Green JA, Xie S, Roberts RM, 1998: Pepsin-related molecules secreted by trophoblast. Rev Reprod 3, 6269. Harlow E, Lane D, 1988: Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Hennies M, 1994: Enzymimmunologische Untersuchungen am Somatotropin bei Fleischziegen. (Enzyme Immunological Assay for Measuring Somatotropin in Meat Goats) Thesis, University of Goettingen, Goettingen, Germany (German, with English abstract). Holtz W, Brackel A, Kuster J, 1986: Der Milchprogester ontest: Instrument zur Fruchtbarkeitspege beim Rind. (Milk progesterone determination: an aid for fertility monitoring in cattle.) Zentralbl Veterinarmed 33, 321336 (German). Hughes AL, Green JA, Garbayo JM, Roberts RM, 2000: Adaptive diversication within a large family of recently duplicated, placentally expressed genes. Proc Nat Acad Sci USA 97, 33193323. Jenkins TC, McGuire MA, 2006: Major advances in nutrition: impact on milk composition. J Dairy Sci 89, 13021310. Karen A, Beckers JF, Sulon J, el Amiri B, Szabados K, Ismail S, Reiczigel J, Szenci O, 2003: Evaluation of false transrectal ultrasonographic pregnancy diagnoses in sheep by measuring the plasma level of pregnancy-associated glycoproteins. Reprod Nutr Dev 43, 577586. Klisch K, De Sousa NM, Beckers JF, Leiser R, Pich A, 2005: Pregnancy associated glycoprotein-1, -6, -7, and -17 are major products of bovine binucleate trophoblast giant cells at midpregnancy. Mol Reprod Dev 71, 453460. Le Roux Y, Colin O, Laurent F, 1995: Proteolysis in samples of quarter milk with varying somatic cell counts. 1. Comparison of some indicators of endogenous proteolysis in milk. J Dairy Sci 78, 12891297. Ledezma-Torres RA, Beckers JF, Holtz W, 2006: Assessment of plasma prole of pregnancy-associated glycoprotein (PAG) in sheep with a heterologous (anti-caPAG(55 + 59)) RIA and its potential for diagnosing pregnancy. Theriogenology 66, 906912. Lopez-Gatius F, Garbayo JM, Santolaria P, Yaniz J, Ayad A, de Sousa NM, Beckers JF, 2007: Milk production correlates negatively with plasma levels of pregnancy-associated glycoprotein (PAG) during the early fetal period in high producing dairy cows with live fetuses. Domest Anim Endocrinol 32, 2942. Metelo R, Silva S, Beckers JF, Moreira DA, Silva F, 2004: Determination of pregnancy-associated glycoproteins (bPAG) in cows milk. Twenty-third World Buiatrics Congress, Quebec City, Canada. CABV, Saskatoon, Canada. Mialon MM, Renand G, Camous S, Martal J, Menissier F, 1994: Detection of pregnancy by radioimmunoassay of a pregnancy serum protein (PSP60) in cattle. Reprod Nutr Dev 34, 6572.

M Friedrich and W Holtz Miyai K, Ishibashi K, Kumahara Y, 1976: Enzyme-linked immunoassay of thyrotropin. Clin Chim Acta 67, 263268. Moller R, 1991: Enzymimmunoassays mit zweitem Antikorper fur Progesteron und porcinem LH. Thesis, University of Goettingen, Goettingen, Germany (German, with English abstract). Mutayoba BM, Meyer HH, Schams D, Schallenberger E, 1990: Development of a sensitive enzyme immunoassay for LH determination in bovine plasma using the streptavidinbiotin technique. Acta Endocrinol (Copenh) 122, 227232. Perenyi Z, 2002: Investigations on Pregnancy-Associated Glycoproteins in the Cow, Thesis. University of Liege, Liege, Belgium. Perenyi ZS, Szenci O, Drion PV, Banga-Mboko H, Sousa NM, El Amiri B, Beckers JF, 2002: Aspartic proteinase members secreted by the ruminant placenta: specicity of three radioimmunoassay systems for the measurement of pregnancyassociated glycoproteins. Reprod Domest Anim 37, 324329. Porstmann T, Kiessig ST, 1992: Enzyme immunoassay techniques: an overview. J Immunol Methods 150, 521. Revoltella RP, Laricchia Robbio L, Liedberg B, 1998: Comparison of conventional immunoassays (RIA, ELISA) with surface plasmon resonance for pesticide detection and monitoring. Biotherapy 11, 135145. Rodbard D, 1974: Statistical quality control and routine data processing for radioimmunoassays and immunoradiometric assays. Clin Chem 20, 12551270. Sasser RG, Ruder CA, 1987: Detection of early pregnancy in domestic ruminants. J Reprod Fertil Suppl 34, 261271. Sasser RG, Ruder CA, Ivani KA, Butler JE, Hamilton WC, 1986: Detection of pregnancy by radioimmunoassay of a novel pregnancy-specic protein in serum of cows and a prole of serum concentrations during gestation. Biol Reprod 35, 936942. Sousa NM, Zongo M, Pitala W, Boly H, Sawadogo L, Sanon M, de Figueiredo JR, Goncalves PB, El Amiri B, Perenyi Z, Beckers JF, 2003: Pregnancy-associated glycoprotein concentrations during pregnancy and the postpartum period in Azawak Zebu cattle. Theriogenology 59, 11311142. Szenci O, Beckers JF, Humblot P, Sulon J, Sasser G, Taverne MA, Varga J, Baltusen R, Schekk G, 1998: Comparison of ultrasonography, bovine pregnancy-specic protein B, and bovine pregnancy-associated glycoprotein 1 tests for pregnancy detection in dairy cows. Theriogenology 50, 7788. Tainturier D, Bedel M, Beckers JF, Fieni F, Bruyas JF, 1996: Cinetique de la bPAG (bovine pregnancy associated glycoprotein) dans le plasma et dans le lait au cours des trois ` semaines suivant le part chez la vache laitiere. Reprod Prod ` Laitiere 1996, 129133.(French). Xie SC, Low BG, Nagel RJ, Kramer KK, Anthony RV, Zoli AP, Beckers JF, Roberts RM, 1991: Identication of the major pregnancy-specic antigens of cattle and sheep as inactive members of the aspartic proteinase family. Proc Natl Acad Sci USA 88, 1024710251. Zoli AP, Beckers JF, Wouters-Ballman P, Closset J, Falmagne P, Ectors F, 1991: Purication and characterization of a bovine pregnancy-associated glycoprotein. Biol Reprod 45, 110. Zoli AP, Guilbault LA, Delahaut P, Ortiz WB, Beckers JF, 1992: Radioimmunoassay of a bovine pregnancy-associated glycoprotein in serum: its application for pregnancy diagnosis. Biol Reprod 46, 8392.

Submitted: 11 Sep 2008 Authors address (for correspondence): Morten Friedrich, Institute of Animal Breeding and Genetics, Albrecht-Thaer-Weg 3, 37075 Gottin gen, Germany. E-mail: Morten.Friedrich@agr.uni-goettingen.de

2008 The Authors. Journal compilation 2008 Blackwell Verlag