Porcine SWC1 is CD52--final determination by the use of a retroviral cDNA expression library.

During the last decades for several species--e.g. swine--many mAb to leukocyte-specific molecules have been developed and clusters of differentiation corresponding to human CD could be established. However, for a significant amount of the raised mAb the corresponding antigens were not characterized on the molecular level and therefore preliminary clusters--in swine so-called Swine workshop clusters (SWC)--were established. These clusters contain antigens with currently no obvious orthologs to human leukocyte differentiation antigens. In this study, we describe the generation of a eukaryotic cDNA expression library from in vitro activated porcine peripheral blood mononuclear cells. Screening of this library with an antibody recognizing SWC1 enabled isolation and sequencing of cDNAs coding for the porcine SWC1 molecule. A BLAST search of the obtained sequence revealed that SWC1 is the orthologous molecule of human CD52. Therefore, our study provides the basis for comparative studies on the role of CD52 in different mammalian species. In addition, the established cDNA library can be used for investigation of additional SWC-defined molecules.

Introduction

Our current understanding of immunity is largely based on studies on the immune system in humans and mice. Studying immune responses in additional species will reveal similarities and differences in immune function and will deepen our understanding how the immune system has evolved to efficiently cope with pathogens. Furthermore, research on the immune system of domestic animals has significant economical impact as it should lead to measures to avoid diseases, e.g. by aiding the development of novel or improvement of existing vaccines and vaccination strategies (Saalmuller, 2006). The importance of swine in livestock production and biomedical research has additionally prompted immunological research on porcine immune cells.

During the last decades numerous mAb against distinct porcine leukocyte-specific surface molecules have been generated and several porcine clusters of differentiation (CD) could be identified. Molecular data of the respective antigens served as basis for their identification and classification. This allowed then the use of the nomenclature adapted from human leukocyte differentiation antigen workshops (HLDA) defining the gold standards of leukocyte antigens. In swine several CD could be defined on the molecular level (Lunney, 1993; Saalmuller, 1996) and immunological research is now greatly facilitated by the availability of these well-characterized monoclonal antibodies to established CD.

Besides these well characterized mAb a huge panel of mAb exists, which has been sorted without detailed knowledge of molecular data to swine workshop cluster (SWC) summarizing less characterized antigens with no obvious human orthologs. These SWC enable scientists working in porcine immunology a uniform determination of the respective antigens, but still represent a preliminary determination (Piriou-Guzylack and Salmon, 2008). To overcome this preliminary determination and to sort these molecules into the existing human CD nomenclature molecular approaches leading to sequence data of the molecules have to be engaged.

Aruffo and Seed have pioneered the use of eukaryotic expression cloning for the identification of cDNAs encoding human leukocyte antigens with antibodies clustered to CD2, CD7 and CD28 (Aruffo and Seed, 1987a,b; Seed and Aruffo, 1987).

Whereas these studies used cDNA libraries that were transiently expressed in COS cells, the advent of retroviral expression technologies allows a stable expression of cDNA libraries in the target cells thereby considerably facilitating the screening procedure. We have previously employed retroviral expression cloning to isolate the cDNA encoding human CD93 and generated a retroviral expression library from human dendritic cells (DC) to identify the antigens of a panel of DC-reactive mAbs (Kirchberger et al., 2005; Pfistershammer et al., 2004; Steinberger et al., 2002, 2004). To characterize porcine leukocyte differentiation antigens, especially those classified as SWC on the molecular level, a retroviral cDNA library derived from activated porcine peripheral blood mononuclear cells (PBMC) was generated to identify antigens formerly classified into the SWC nomenclature. By the use of this expression system we were able to identify SWC1 as the porcine ortholog of human CD52.

Materials and methods

Isolation of porcine peripheral blood mononuclear cells

Heparinized whole blood samples from six-month-old healthy pigs were obtained from an abattoir. Animals were subjected to electric high voltage anesthesia followed by exsanguination. This procedure is in accordance to the Austrian Animal Welfare Slaughter Regulation. PBMC were isolated by gradient centrifugation using lymphocyte separation medium (PAA Laboratories, Pasching, Austria) as described previously (Saalmuller et al., 1987). For cultivation cells were re-suspended in cell culture medium consisting of RPMI 1640 supplemented with 10% heat-inactivated FCS, 100 IU/ml penicillin and 0.1 mg/ml streptomycin (all PAA) or frozen in freezing medium consisting of 10% dimethyl sulfoxide (Sigma–Aldrich, St. Louis, MO), 40% (v/v) FCS and 50% (v/v) RPMI 1640 at −150 °C prior to use.

Cell culture, antibodies and flow cytometry

HEK 293T cells and the mouse thymoma cell line Bw5147 (short designation within this work Bw) were cultured as described (Leitner et al., 2009). Anti-human CD52 antibody Campath-1H (Alemtuzumab, humanized IgG1) was obtained from Genzyme Corporation (Cambridge, MA). Mouse IgG1 control antibody was purchased from Biolegend (San Diego, CA). MAbs against porcine CD2 (MSA4, mIgG2a) (Hammerberg and Schurig, 1986), CD3 (PPT3, mIgG1) (Pescovitz et al., 1998), CD5 (b53b7, mIgG1) (Saalmuller et al., 1994b), CD8α (11/295/33, mIgG2a) (Jonjic and Koszinowski, 1984), CD25 (3B2, mIgG1) (Bailey et al., 1992), CD172a (74-22-15, mIgG1) (Pescovitz et al., 1984) and SWC1 (11/8/1 and 11/305/44, both mIgG2b) (Saalmuller et al., 1994a, 1987), were produced by hybridoma cultures at the Institute of Immunology, Department for Pathobiology, University of Veterinary Medicine Vienna. For cell sorting and staining of transductants binding of primary antibodies was detected with PE-conjugated goat anti-mouse IgG-Fcγ specific Abs or goat-anti-human IgG-Fcγ specific Abs (both Jackson ImmunoResearch Laboratories, West Grove, PA).

To investigate SWC1 expression on PBMC cells were thawed and re-suspended in PBS (without Ca2+/Mg2+, PAA) containing 10% (v/v) porcine plasma. PBMC (6 × 105 cells/sample) were seeded in a volume of 50 μl in 96-well round-bottom microtiter plates (Greiner, Kremsmünster, Austria). All incubation steps for flow cytometry staining were performed for 20 min on ice.

PBMC were incubated in a first round with mAb against CD172a and SWC1 followed by staining with isotype-specific conjugates (anti-mouse-IgG1-Alexa488 and anti-mouse-IgG2b-Alexa647, Invitrogen, Carlsbad, CA). Free binding sites of the secondary antibodies were blocked with mouse IgG molecules (2 μg per sample; Jackson ImmunoResearch Laboratories). After an intermediate washing step, PBMC were incubated with biotin-labeled anti-CD3 followed by streptavidin-eFluor450 conjugate (eBioscience, San Diego, CA). Subsequently, cells were fixed, permeabilized and stained for CD79α expression using anti-CD79α-PE (clone HM57, mIgG1, Dako, Glostrup, Denmark) as described previously (Gerner et al., 2008). Isotype-matched non-binding antibodies (mIgG1, mIgG2b, Dianova, Hamburg, Germany, and mIgG1-PE, clone DAK-GO1, Dako) served as negative controls.

Flow cytometric analyses were done using a FACSCalibur flow cytometer supported by CELLQUEST software (BD Biosciences, San Jose, CA) or using a FACSCanto™ II flow cytometer (BD Biosciences). Data were then analyzed using FACSDiva software, version 6.0 (BD Biosciences). Fluorescence intensity is shown on a standard logarithmic scale.

Construction and expression of a porcine cDNA library

Porcine PBMC (4.8 × 108 total) were stimulated with PMA/Ionomycin (100 nM each, Sigma–Aldrich) or with 5 μg/ml Concanavalin A (GE Healthcare, Waukesha, WI) for 24, 48 and 72 h. According to manufacturer's instructions, total RNA was isolated from pooled PBMC preparations using TRI REAGENT™ (Sigma–Aldrich). Poly A+ mRNA was isolated using Oligotex (Qiagen, Hilden, Germany) and converted into 1st strand cDNA using oligo-(dT)18-Primers and RevertAid™ Premium Reverse Transcriptase (Fermentas, Ontario, Canada). Second strand synthesis was done using DNA Polymerase I and Ribonuclease H from Fermentas according the manufacturers’ protocol. T4 DNA Polymerase (Fermentas) was used for blunting DNA ends and BstXI-adapters (5′-PO4-CTTTCCAGCACA-3′ and 5′-PO4-CTGGAAAG-3′) were ligated to the double stranded cDNA. The resulting product was size-fractioned by agarose gel electrophoresis and three DNA fractions (low: 0.7–1.5 kbp; middle: 1.5–2.5 kbp and high: >2.5 kbp) were gel-purified, cloned into the retroviral expression vector pBMN and transformed into electrocompetent E. coli (ElectroMaxDH10B, Invitrogen). Transformed E. coli cells were grown on 150 mm LB-ampicillin plates and subjected to plasmid preparation. To assess the size of our library, dilutions of the E. coli cells transformed with the ligation reactions were plated on separate plates.

For quality control purposes library plasmid DNAs (10 ng/reaction) were subjected to PCR analysis using primer pairs specific for the coding regions of porcine CD2, CD11a, and CD54.

For the expression of the library, plasmid DNA representing the retroviral cDNA library and pEAK12-MLV-env-gag-pol were co-transfected into HEK 293T cells. Retrovirus containing cell culture supernatant was used for the transduction of the target cell line, Bw5147, as described (Leitner et al., 2010). The cell pool expressing the library was used for screening 72 h post-transduction.

Screening procedure of the porcine cDNA expression library and PCR recovery of the retroviral inserts

2 × 107 Bw cells expressing the library were incubated with a cocktail of anti-porcine mAb as described in Section 2.2, washed and bound antibodies were detected with appropriate secondary reagents. The cell pool representing the porcine cDNA library was subjected to two rounds of sorting using a FACSAria cell sorter (BD Biosciences).

Subsequently, single cell clones were established by limiting dilution culturing and reactivity with the antibodies used for screening was confirmed. Genomic DNA was prepared from reactive clones using a Puregene genomic DNA isolation kit (Qiagen) according the manufacturer's protocol. The retroviral inserts were retrieved by PCR as described (Leitner et al., 2011; Pfistershammer et al., 2009). The PCR products were gel-purified and PCR-reamplified using the primers pBMN-Sfi-F: 5′-GCGCCCGGCCATTACGGCCGCCGGATCCCAGTGTGG-3′ and pBMN-SFi-rev: 5′-GCGCCCGGCCGAGGCGGCCCGTCGACCACTGTGGTGG-3′. The PCR products harboured SfiI sites for directional cloning into the retroviral expression vector pCJK2 generated in our laboratory. The resulting construct was retrovirally expressed in the Bw cell line and reactivity with the antibodies was confirmed. Plasmid DNA was subjected to sequence analysis (Eurofins MWG, Ebersberg, Germany). The nucleotide sequence encoding porcine CD52 has been submitted to the GenBank (JN_544166).

GPI-anchorage of porcine CD52

The predicted amino acid sequence of porcine CD52 was analyzed for presence of a GPI-Modification Site (“big-PI Predictor”: http://mendel.imp.ac.at/gpi/gpi_server.html) (Eisenhaber et al., 1998). To experimentally assess whether porcine CD52 is expressed as a GPI anchored molecule, CD52 transductants were mock-treated or treated with phosphatidylinositol-specific phospholipase C (PI-PLC, final concentration 0.1 U/ml; American Radiolabeld Chemicals, St. Louis, MO) followed by flow cytometric analysis as described (Kueng et al., 2007).

Results

Expression pattern of the SWC1 antigen on porcine PBMC

The SWC1 antigen defined in the 1st International swine cluster of differentiation workshop 1992 is recognized by ten different mAb, raised in various laboratories (Lunney, 1993; Saalmuller, 1996). Four out of the ten mAb clearly recognize the same epitope detected by mAb 11/8/1 (Saalmuller et al., 1994a). They were distributed in the 2nd International swine cluster of differentiation workshop to the SWC1a cluster (Saalmuller, 1996). Five other mAb were distributed to SWC1 showing the same labeling pattern as 11/8/1, but no blocking of binding of the biotinylated prototype mAb 11/8/1 (Saalmuller, 1996; Saalmuller et al., 1998). Molecular analyses of the SWC1 antigen gave inconsistent results. Magyar and Mihalik (1997) characterized the SWC1 antigen as a dimeric protein with a molecular mass of 41 kDa and 15 kDa. Aasted et al. (1998) described the molecular mass of SWC1 with either 16 kDa (mAb 76-7-6) or 19 kDa (mAb 335-2). SWC1 expression was found on monocytes, granulocytes and resting T-lymphocytes, while it was absent on erythrocytes, platelets and B-lymphocytes (Pescovitz et al., 1984; Saalmuller et al., 1994a). In vitro stimulation of T-lymphocytes led to a down-regulation of the SWC1 expression on the cell surface (Saalmuller et al., 1987).

To confirm these data, we stained PBMC with mAb against CD3, CD79 as well as CD172a together with mAb against SWC1 (Fig. 1). CD3 positive T cells showed a clear SWC1 expression, whereas the majority of CD79-positive B cells was negative for SWC1. CD172a monocytes demonstrated the highest SWC1 expression as described in previous publications (Saalmuller and Reddehase, 1988).

Fig. 1

Expression of the SWC1 antigen on porcine PBMCs. By the use of four-color flow cytometry monocytes (CD172a+, M), B cells (CD79a+, B) and T cells (CD3+, T) were analyzed for SWC1 antigen expression. Data are representative of five individuals.

Construction of a representative retroviral cDNA expression library from in vitro activated porcine PBMC

To identify the antigen recognized by SWC1 antibody, we generated a retroviral cDNA expression library from porcine PBMC. The RNA was converted into double stranded cDNA, BstXI-adaptors were added, and the resulting product was size-fractioned (Fig. 2A, B). The fractions (high, middle and low) were kept separated during the subsequent steps of library generation. This was done to prevent that small cDNA species will be over-represented in the library as they tend to out-compete larger fragments during the ligation and the transformation steps. Based on the number of transformed bacteria, we estimated that our library contained over 107 independent clones. In order to confirm that the library contains cDNA species encoding representative cell surface proteins expressed in porcine PBMC, we used primer pairs specific for porcine CD2, CD11a and CD54. Small amounts of plasmid DNA representing the three fractions of the library as well as a pool of these fractions were PCR-probed with these primer pairs. High amounts of CD2, CD11a and CD54 PCR product were obtained with plasmid DNA harbouring cDNA inserts of the corresponding size as well as with the plasmid pool representing the final library (Fig. 2C). The library pool was retrovirally expressed in the murine cell line Bw. These cells were chosen as target cells, since they allow an efficient retroviral transduction. Furthermore, murine cells are especially suited for screening with murine mAb, which usually do not react with mouse cells. In order to validate the efficient expression of our library in the target cell pool, we first screened our library for the presence of transductants expressing common porcine leukocyte surface antigens using mAbs reacting with porcine CD2, CD5, CD8α, and CD25. After two rounds of sorting the majority of the isolated cells strongly reacted with the antibody employed in the second round of sorting (Fig. 2D).

Fig. 2

Construction and quality control of a porcine PBMC cDNA expression library. (A) Agarose-gel showing size-fraction of the double stranded cDNA after adapter ligation. (B) Analysis of the size-fractionated cDNA by agarose-gel electrophoresis. First lane: DNA ladder, second to fourth lane: double stranded cDNA after size-fractioning, high (H), middle (M), low (L). (C) Porcine CD11a, CD54 and CD2 cDNAs were PCR-amplified from 10 ng of plasmid DNA containing the library fractions high (H), middle (M), low (L) and a plasmid pool representing the final library. (D) Quality control screening of the library. Bw cells expressing the porcine library were incubated with the indicated mAbs (CD2, CD5, CD8α, CD25) (upper panel). Bound antibodies were detected by PE-labeled anti-mouse IgG antibodies and PE-positive cells were gated for sorting. The sorted cell pool was probed with the mAb used for sorting (lower panels). Bound antibodies were detected by PE-labeled anti-mouse IgG antibodies (Y-axis: FL-2).

Isolation of cDNAs encoding the SWC1 antigen

Next we used our porcine expression library to characterize the SWC1 antigen on the molecular level. Bw cells expressing the porcine library were subjected to two rounds of sorting using mAb SWC1 clone 11/305/44 (Saalmuller et al., 1994a, 1987) (Fig. 3A + B). From the cell pool obtained after the second round of sorting, which strongly reacted with the mAb 11/305/44 directed against SWC1, single cell clones were established by limiting dilution culturing and tested for their SWC1 expression (Fig. 3C). Genomic DNA was prepared from SWC1 positive single cell clones and retroviral inserts were retrieved using vector specific primers. A 500 bp band was present in all single cell clones reacting with mAb 11/305/44 (Fig. 3D). The band was isolated, cloned and re-expressed in Bw cells. The resultant transductants again strongly reacted with mAb 11/305/44. Furthermore, specific reactivity was also obtained with another SWC1 antibody mAb 11/8/1 (Saalmuller et al., 1987) confirming that the 500 bp insert indeed encoded the SWC1 antigen (Fig. 3E).

Fig. 3

Isolation of cDNAs encoding the SWC1 antigen. (A + B) Cells expressing the porcine library were incubated with SWC1 mAb11/305/44. Bound antibodies were detected by PE-labeled anti-mouse IgG antibodies and PE-positive cells were gated for sorting (A) 1st sorting round, (B) 2nd sorting round. Reactive cells were enriched by sorting, expanded and probed with mAb11/305/44 (right panels in A and B). Bound antibodies were detected by PE-labeled anti-mouse IgG antibodies (Y-axis: FL-2). (C) FACS staining of single cell clones (SCC) established after two rounds of sorting with mAb11/305/44 (grey histograms). Bound antibodies were detected by PE-labeled anti-mouse IgG antibodies (X-axis: FL-2). Open histograms represent reactivity of this antibody with Bw control cells. (D) PCR-recovery of retroviral inserts from the genomic DNA isolated from mAb11/305/44 reactive single cell clones. A 0.5 kb band was present in all SCC as indicated. First and last lanes: DNA ladder. (E) The 0.5 kb cDNA was cloned in a retroviral expression vector and re-expressed in Bw cells. The resultant cells were probed with mAb11/305/44 (left panel, grey histogram) and with another SWC1 antibody, mAb11/8/1 (right panel, grey histogram). Bound antibodies were detected by PE-labeled anti-mouse IgG antibodies. Open histograms represent reactivity of these antibodies with Bw control cells.

Identification of the SWC1 antigen as porcine CD52

The insert was sequenced and the obtained DNA sequence and deduced amino acid sequence were compared with known sequences by BLAST analysis. High sequence homologies to CD52 sequences of other mammalian species unequivocally identified the SWC1 antigen as porcine CD52. While the amino acid sequence of the signal peptide and the GPI linkage motif was highly homologous to other CD52 molecules, it was found that the mature porcine CD52 has only significant sequence similarities with the CD52 molecules of dog and horse (Fig. 4A). In line with these data, it was observed that the anti-human CD52 antibody Campath-1H and the anti-porcine CD52 antibody 11/305/44 did not react with porcine and human CD52, respectively (Fig. 4B). Human CD52 is a GPI anchored molecule and to confirm this also for porcine CD52, we used an algorithm described in (Eisenhaber et al., 1998). Since this analysis did not unequivocally score the Glycine at position 12 of the mature protein as GPI anchor, we treated Bw cells expressing porcine CD52 with PI-PLC. We found that this treatment reduced surface density of CD52 molecules, strongly indicating that porcine CD52, like its human ortholog, is a GPI-anchored molecule (Fig. 4C).

Fig. 4

Identification of the SWC1 antigen as porcine CD52. (A) CD52 amino acid sequence alignments of different species. (B) Bw control cells (left panel), Bw cells expressing human (middle panel) and porcine CD52 (right panel) were probed with anti-human CD52 antibody (Campath-1H; open histograms) or with porcine SWC1 antibody (mAb 11/305/44; grey histograms). (C) Phospholipase C (PI-PLC) treatment of CD52 expressing cells. Bw cells expressing porcine CD52 were mock-treated (left histogram) or probed with PI-PLC (right histogram). Cells were probed with control antibody (white histograms) or porcine SWC1 antibody (mAb 11/305/44; grey histograms). Data are representative for three independently performed experiments.

Discussion

Eukaryotic expression cloning is a powerful method that is especially suited for the characterization of surface antigens. Many studies have described its successful use for the identification of membrane-resident binding partners for cell surface molecules or soluble ligands (Pfistershammer et al., 2008; Sedy et al., 2005; Yamanishi et al., 2010). Another important application of this technology is the characterization of surface antigens defined by mAbs and more recently by polyclonal antibody preparations like anti-thymocyte globulins or human sera (Leitner et al., 2011; Pfistershammer et al., 2009). Although numerous expression libraries established from different tissues and cells have been described in the literature, almost all of them were of human or mouse origin. For swine different approaches have been performed with the expression of single nucleotide sequences for distinct porcine molecules as SWC9 (CD203a) and CD14 (Petersen et al., 2007a,b), but the expression system described in this manuscript is to our knowledge the first example for a retroviral-based eukaryotic expression library that was generated from porcine PBMC.

The aim of this library was to characterize antigens of porcine leukocyte-reactive mAbs on the molecular level and allow then a clear CD classification for porcine orthologs of known human molecules. Although, proteomics-based approaches are also a powerful and well established strategy to identify antibody-defined antigens, we have chosen expression cloning over proteomics-based methods for several reasons: (i) this approach does not require antibodies that are suitable for immunoprecipitation or immunoblotting, (ii) it leads to the isolation of cDNAs encoding the antigens of interest (iii) which can subsequently be used to generate cell lines that express the antigen at high levels. Thus, this method is especially suited for our purpose since for many porcine leukocyte antigens there is currently no sequence information available in the databases. Sequence data are the most important criteria to establish an orthologous CD nomenclature also in less characterized species (Saalmuller, 1996). Up to date no sequence data were available for porcine CD52, the first antigen that was characterized by screening our library with a mAb initially described more than twenty years ago (Saalmuller et al., 1994a, 1987). Sequence comparison of CD52 molecules from different species revealed that the leader sequence and the C-terminal GPI-attachment signal sequence of the porcine orthologue is highly homologous to other CD52 molecules, whereas the mature extra-cellular portions of CD52 molecules derived from different species have little sequence homologies. In line with this, mAb 11/305/44 did not recognize human CD52 and likewise the anti-human CD52 antibody CAMPATH-1H did not recognize porcine CD52. Furthermore, there are differences regarding the expression pattern of this molecule: CD52 is broadly expressed on human lymphocytes including B cells, whereas in porcine B cells it is only expressed on a small subset.

This different expression pattern underlines the need for molecular data for a final characterization of porcine CD orthologs and for additional strategies to establish well-defined antibodies to non-human leukocyte differentiation antigens. Although the generation and validation of a retroviral library requires some effort, once established, it facilitates the characterization of antigens of unknown specificity. This is an effective and direct approach for a molecular characterization of antigens with unknown or less known specificity being expressed on the cells that were used as a mRNA source for library generation. In the meantime, we have been able to identify additional porcine antigens defined by antibodies raised in our laboratory (unpublished data) and we have initiated collaborations with other groups to characterize antigens recognized by additional porcine leukocyte reactive mAbs. Thus, the data of the current study demonstrate that retroviral expression cloning is a useful approach to identify mAb-defined porcine leukocyte differentiation antigens and to characterize them on the molecular level. This enables classification of these antigens as porcine orthologs in the existing human-based CD nomenclature. Therefore, our approach is an important step for the validation and standardization of reagents with importance in veterinary immunology as well as for studies dealing with comparative immunology. The methodology described herein could readily be adopted by immunologists working on other model species.

Conflict of interest

The authors declare no conflict of interest.