Analysis of transcription of Porcine circovirus type 1

Circoviruses are characterized by a monopartite, circular and single-stranded (ss) DNA molecule of approximately 2000 nucleotides. The family Circoviridae is classified into the two genera Circovirus and Gyrovirus (McNulty et al., 2000 ). Porcine circovirus type 1 (PCV1) (Mankertz et al., 1997 ) and type 2 (PCV2) (Hamel et al., 1998 ), Beak and feather disease virus (BFDV) (Bassami et al., 1998 ), Columbid or Pigeon circovirus (CoCV or PiCV) (Mankertz et al., 2000 ), Canary circovirus (Phenix et al., 2001 ) and Goose circovirus (GCV) (Todd et al., 2001 ) show the typical ambisense genomic structure of the genus Circovirus (Fig. 1), while the genome of Chicken anaemia virus (CAV) (Kato et al., 1995 ), the only member of the genus Gyrovirus, is organized differently. With the exception of PCV1 (Tischer et al., 1986 ), infection with circoviruses results in diseases affecting lymphatic tissues, e.g. depletion of lymphocytes and lymphadenopathy, a typical sign of post-weaning multisystemic wasting syndrome (PMWS), a new emerging disease of swine. A swollen bursa of Fabricius and paracrystalline inclusion bodies are observed in cockatoos and pigeons infected with BFDV and CoCV, respectively (Sanada et al., 1999 ; Soike et al., 2001 ).

(32K): Fig. 1. Linear map of the genome of PCV1. Open reading frames of PCV1 are indicated by boxes. The intergenic region comprising the origin of replication is located between the divergently transcribed rep and cap genes. The location and sequence of the PCV1 promoters Prep and Pcap are given; putative regulatory elements are indicated by grey shaded boxes. In the upper part of the figure, the origin of replication of PCV1 spanning the fragment nt 728838 is enlarged, showing the putative stemloop structure and the four hexamer repeats H1H4.

Circoviruses show a very compact genomic structure. Two open reading frames (ORFs) arranged head-to-head flank an intergenic region, comprising the origin of replication (Fig. 1). The origin of replication is characterized by a putative stemloop structure, displaying a conserved nonamer in its apex and four adjacent hexamer repeats (Mankertz et al., 1997 ). The second largest ORF, cap, is located in an anti-clockwise direction, and is synthesized after infection of the cell, presumably by host-encoded enzymes. It comprises the Cap protein (234 amino acids), which is localized in the nucleus and has an arginine-rich and basic N terminus. The promoter has been mapped within the rep gene on a fragment comprising nt 11681428 (Mankertz et al., 1998b ) and the start of the cap transcript has been mapped to nt 1238. The transcript contains an untranslated leader sequence of 119 nucleotides (nt 12381120), which is joined to exon 2 of the ORF1 transcript at nt 737 immediately adjacent to the ATG. Processing of this RNA may have evolved to avoid synthesis of an alternative protein initiated at an internal start codon in the intron. The Cap protein of PCV2 was expressed in insect cells. The gene product had a molecular mass of 30 kDa, similar to that detected in purified virus particles and it formed capsid-like particles when viewed by electron microscopy (Nawagitgul et al., 2000 ). It is generally agreed that cap encodes the major structural protein of PCV.

The largest ORF of PCV1 is transcribed in a clockwise direction. It encodes the Rep protein, which is highly conserved in all circoviruses (Mankertz et al., 1998a ), and shows homology to the Rep proteins of nanoviruses and geminiviruses. Three motifs, I, II and III, typically found in enzymes initiating replication in the rolling circle replication (RCR) mode, have been identified, as well as a P loop for binding of dNTPs. Phylogenetic analysis has suggested that circoviral Rep proteins may have evolved by a recombination event between the Rep protein of nanoviruses and an RNA-binding protein encoded by picorna-like viruses (Gibbs & Weiller, 1999 ) or a helicase of prokaryotic origin (Nishigawa et al., 2001 ). Two collinear transcripts are synthesized from the rep gene of PCV1: one encodes the full-length reading frame leading to synthesis of a protein of 312 amino acids (Rep); the second is differentially spliced and encodes the Rep' protein, which is 168 amino acids in size. Splicing results in expression of the C-terminal moiety of Rep' in a different reading frame (Mankertz & Hillenbrand, 2001 ). Replication of PCV1 depends on co-expression of both the Rep and Rep' protein. Both proteins bind to double-stranded (ds) DNA fragments comprising the origin of replication (Steinfeldt et al., 2001 ). Using a band shift assay, it could be demonstrated that the left part of the stemloop sequence and the two inner hexamers, H1 and H2, are the minimal binding site of the Rep and Rep' protein. In contrast to the Rep protein of PCV1, Rep' tolerates sequence modification of the stemloop, indicating that the sequence requirements of the two proteins may be different. To analyse further the function and contribution of the two rep gene products, we have mapped and fine-mapped the promoters of the rep and the cap genes and studied their regulation by the viral proteins Rep, Rep' and Cap. Elements necessary for the regulation of P_rep by viral proteins have been identified. In addition, we have analysed the ratio of the two rep transcripts in a kinetic approach using real-time PCR.

Construction of plasmids.
DNA fragments for cloning were PCR-generated as described previously (Mankertz & Hillenbrand, 2001 ). The following plasmids were used in this study. pRP plasmids carry fragments of the putative rep promoter. Plasmid pRP1 carries the PCR-derived and KpnI/BglII-tagged fragment nt 500816 of PCV1 (GenBank accession number: Y09921) cloned into the KpnI/BglII-restricted vector pGL3-basic (Promega); pRP2 carries the fragment nt 541816; pRP3, nt 591816; pRP4, nt 640816; pRP5, nt 692816; pRP6, nt 752816; pRP7 nt 528802; pRP8, nt 528796; pRP9, nt 528727; and pRP1-x is the same as pRP1 but carrying the nonamer sequence altered from 5' TAGTATTAC to 5' CTGTATTAC, which results in inactivation of the PCV1 origin of replication leading to a replication-defective plasmid. pRP16 carries the fragment nt 647819; pRP16-1 is similar to pRP16 but has H1 altered from 5' CGGCAG to 5' AGATCT; pRP16-2 has H2 altered; pRP16-3 has H3 altered; pRP16-4 has H4 altered; pRP16-12 has H1 altered to 5' AGATCT and H2 altered to 5' CCCGGG; and pRP16-34 has H3 and H4 altered. pCP plasmids carry fragments containing the cap promoter. Plasmid pCP1 carries the PCR-derived and KpnI/BglII-tagged fragment nt 15691100 of PCV1 cloned into the KpnI/BglII-restricted vector pGL3-basic; pCP2 carries the fragment nt 15301100; pCP3 nt 14751100; pCP4 nt 14291100; pCP6 nt 14041100; pCP8 nt 13901100; pCP7 nt 13781100; pCP5 nt 13531100; pCP9 nt 13281100; pCP10 nt 12951100; pCP11 nt 12771100; pCP12 nt 12351100; pCP13 nt 13901112; pCP14 nt 13901163; pCP15 nt 13901172; pCP16 nt 13901210; pCP17 nt 13901252; and pCP18 nt 13901298. All plasmids have been sequenced to exclude PCR-acquired misincorporation of nucleotides. Plasmid pGL3-control was used as a positive and plasmid pGL3-basic as a negative control (Promega).

Plasmids used for expression of viral proteins were pORF4A, expressing both rep gene products, pAM9 expressing only the Rep protein due to inactivation of splice donor sites and pAM4 expressing the Rep' protein. Expression of Cap was performed using plasmid pSVL-Cap(PCV1). The rep gene was successively truncated from the 3' end in plasmids pRep(1217), pRep(1186) and pRep(1128), while in pRep-mutI, pRep-mutII, pRep-mutIII and pRep-mutP, the conserved RCR motifs I, II and III and the dNTP binding domain, respectively, have been inactivated by site-directed mutagenesis. All these plasmids have been described previously (Mankertz & Hillenbrand, 2001 ).

Luc/Gal assay.
Transcription activity of viral promoters was investigated using the Dual light kit following the manufacturer's instructions (Applied Biosystems). Luciferase activity was indicative of promoter activity of the investigated fragment, while β-galactosidase activity was determined to standardize for deviation in transfection efficiency. PK15 cells were transfected with Effectene (Qiagen) using 50 ng pRSV- βGal (MacGregor et al., 1987 ) and 200 ng of pRP plasmids carrying the promoter of the rep gene, or pCP plasmids carrying the promoter of the cap gene. The medium was changed after 24 h and cell extracts were prepared and measured after 2 days. Further studies investigated the regulation of P_rep and P_cap by viral proteins. For this purpose, PK15 cells were cotransfected with 100 ng of promoter plasmids pCP or pRP and 100 ng of plasmid pAM4 (expressing the Rep' protein), pAM9 (Rep), pORF4A (Rep+Rep') or pSVL-Cap (Cap), plus 50 ng pRSV- βGal. Luc/Gal activity was measured for 10 s in a Microlumat Plus LB96V (EG&G Berthold) after addition of 1 µl galacton-plus substrate (1:100 diluted in buffer B). Transfections were performed in duplicates; each assay was repeated at least once. Standardized Luc units (SLU) were calculated by division of the Luc units by the Gal units. The presence of putative regulatory elements in the promoter sequences was analysed with the sub-sequence analysis package of the McVector program.

Real-time PCR.
PK15 cells were infected with PCV1 by transfection with the infectious plasmid pIC1 (PCV1 overlength genome nt 11881: 17591098 cloned into vector pUC8; i.e. nt 10981188 are duplicated). Due to the duplication, the viral genome can recombine out of the vector backbone and initiate an infection (K. Hattermann and others, unpublished). RNA was isolated at 12 h intervals and transcribed into cDNA. The cDNA synthesis and real-time PCR was performed as described previously (Mankertz & Hillenbrand, 2001 ), using two TaqMan amplicons discriminating between the spliced and the unspliced rep transcripts. TaqMan probe S2 (nt 15851609; 6FAM-CCCAGGAATGGTACTCCTCAACTGCXTPh) was used to amplify a 114 bp fragment of the unspliced transcript with primer pair T3F (nt 15371558, 5' CTGTTCCTTTTTTGGCTCGCAG) and T2B (nt 16501625, 5' AAGTAGTAATCCTCCGATAGAGAGCT), while 111 bp were amplified from the spliced rep transcript with primer pair T2F (nt 11571175:15591560, 5' CAGCGACCTGTCTACTGCTTA) and T2B.

Mapping of the promoters of the rep gene and of the cap gene
To map the promoters of the rep and cap genes of PCV1, fragments containing the putative promoters were cloned into plasmid pGL3-basic in front of a promoterless luc gene. We assumed that P_rep was located immediately upstream of the rep translation start position at nt 767±10. Consequently, the fragment containing nt 500816 was truncated either from the 5' or the 3' end. The resultant pRP plasmids were investigated for promoter activity using the Luc/Gal assay. Compared with the late SV40 promoter in plasmid pGL3-control, P_rep activity on the initial fragment was strong. The highest activity was observed with plasmid pRP2 (nt 541816). A substantial decrease in Luc activity was obtained when the 5' end of the fragment carrying nt 640816 (pRP4) was further truncated to nt 692 (Fig. 2), indicating that the left border of P_rep must be located between nt 640 and 692. To map the right border of the rep gene promoter, the fragment carrying nt 528816 was shortened from the 3' end in plasmids pRP7 to pRP9: while truncation from nt 816 to 802 or 796 had no effect on Luc expression, further size reduction to nt 727 abolished transcription from P_rep, indicating that the right border of this element must be located between nt 796 and 727. Combining these results, it can be assumed that P_rep is located on the fragment nt 640796. P_rep overlaps the intergenic region and the origin of replication. It contains several elements that may influence its activity (Fig. 1). An SP1 site (nt 693698) is located upstream of the TATA box (nt 739745), which overlaps a putative IFN-stimulated response element (ISRE nt 740753). An AP3 and an AP4 box were found (nt 642653 and 760768), as well as a putative binding site for the upstream stimulating/major later transcription factor (USF/MLTF nt 726737; Sawadogo & Roeder, 1985 ).

(24K): Fig. 2. Mapping of the rep gene promoter, Prep. The PCV1 fragment nt 500816 was cloned into vector pGL3-basic in front of the promoterless luc gene and subsequently truncated from the 3' and the 5' end. Fragment positions are indicated in the legend. The activity of the Prep fragment was determined with the Luc/Gal assay.

A similar approach was followed for fine-mapping of the promoter of the cap gene. Previously published results indicated localization of the anti-clockwise-transcribed P_cap between nt 1168 and 1425 (Mankertz et al., 1998b ). Accordingly, the fragment nt 15691100 was cloned into pGL3-basic and subsequently truncated (Fig. 3). Compared with the positive control pGL3-control and the promoter of the rep gene, activity of P_cap was low. When the initial fragment was shortened from the 5' end to nt 1475 (pCP3), activity of P_cap was lost. Further truncation to nt 1390 in plasmid pCP8 resulted in a restored transcription from P_cap, which was even higher than in the initial fragment. Activity decreased when truncation from the 5' end proceeded further than nt 1328 (plasmid pCP10), indicating that the 5' border of P_cap is located between nt 1328 and 1295. Mapping of the 3' border of P_cap was performed by testing transcription activity of plasmids pCP13 to pCP18, in which the 3' end of the initial fragment was truncated successively from nt 1100 to 1298. Activity of P_cap was diminished in pCP18, indicating that the right border of P_cap is located between nt 1252 and 1298. Taken together, it can be assumed that transcription of the cap gene is directed by a fragment from nt 1328 to 1252. Compared with P_rep, only a few sequences of putative regulatory elements (AP3 site nt 13381349; SP1 site nt 13371332) were identified (Fig. 1).

(29K): Fig. 3. Mapping of the cap gene promoter, Pcap. The PCV1 fragment nt 15691100 was cloned into vector pGL3-basic in front of the promoterless luc gene and subsequently truncated from the 3' or the 5' end. Fragment positions are indicated in the legend. To allow a better resolution, activity of the control promoter in plasmid pGL3-control was reduced to 10%. Localization of Pcap was determined with the Luc/Gal assay.

Regulation of P_rep and P_cap by viral proteins
To test whether viral proteins are involved in regulation of P_rep and P_cap, the transcription activity of plasmids pRP1-x and pCP8 was compared in the presence and absence of the Rep, Rep' and Cap proteins. pRP1-x contains a fragment in which the nonamer TAGTATTAC has been altered to CTGTATTAC, resulting in inactivation of the PCV1 origin (Mankertz et al., 1997 ). This was used to prevent interference of replication and transcription effects. After co-transfection of pRP1-x with plasmids expressing the Rep, Rep' or Cap proteins, P_rep activity was found to be reduced in the presence of Rep. In contrast, no effect was exerted by the Rep' or Cap protein (Fig. 4). A similar experiment was performed with plasmid pCP8 to investigate the regulation of P_cap by viral proteins. Transcriptional activity of P_cap was not influenced by the Rep, Rep' or Cap protein (data not shown).

(33K): Fig. 4. Regulation of Prep by viral proteins. PK15 cells were cotransfected with the reporter plasmid pRP1-x, carrying Prep overlapping the inactivated origin of replication, and plasmids pORF4A (expressing the Rep and Rep' proteins), pAM4 (Rep'), pAM9 (Rep) and pSVL-Cap (Cap). Since pRP1-x is replication-deficient, repression of Prep by viral proteins could not be counteracted by induction of replication by Rep plus Rep'. The effect of the three viral proteins on Prep was determined using the Luc/Gal assay.

Role of the hexamer repeats in regulation of P_rep by the Rep protein
The Rep protein binds alternatively to hexamer repeats H1/H2 and H3/H4 (Fig. 1), respectively, but only H1/H2 has been mapped as an essential part of the minimal binding site (MBS) of the Rep protein (Steinfeldt et al., 2001 ). To determine the effect of Rep binding to the four hexamer repeats on repression of P_rep, mutagenized variants of plasmid pRP16 were constructed. The sequence of each of the four hexamers was altered from 5' CGGCAG to 5' AGATCT in pRP16-1, pRP16-2, pRP16-3 and pRP16-4. Plasmids pRP16-12 and pRP16-34 carried a double mutant, in which a second hexamer sequence was modified to 5' CCCGGG in addition to the sequence alteration in the first hexamer. Luc plasmids were cotransfected with plasmid pAM9 producing only the Rep protein but no Rep' protein. Since the Rep protein by itself is unable to promote replication of PCV1 (Mankertz & Hillenbrand, 2001 ), its influence on P_rep could be tested without producing a replication artefact. A strong repression of P_rep activity in the presence of the Rep protein was seen in the unmodified plasmid pRP16 and in plasmids pRP16-3, pRP16-4 and pRP16-34, carrying mutants in H3 or H4 or in both repeats (Fig. 5). In contrast, only a slight repression of P_rep by the Rep protein was observed with plasmids pRP16-1, pRP16-2 and pRP16-12, in which the sequence of H1, H2 or both was modified. This result indicates that binding of the Rep protein to H1/H2 is involved in down-regulation of P_rep, while binding of Rep to hexamers H3/H4 is not.

(26K): Fig. 5. Mutagenesis of the hexamer repeats H1 and H2 influences regulation of Prep. All the examined pRL16-derived plasmids carried a fragment comprising Prep with the overlapping origin of PCV1. In plasmids pRL16-1, -2, -3 and -4, only one of the four hexamers H1, H2, H3 and H4 was mutagenized, while plasmids pRL16-12 or-34 carried a double mutant in H1/H2 or H3/H4, respectively. pRL16 plasmids were investigated for their promoter activity in the presence or absence of the Rep protein using the Luc/Gal assay. The Rep protein was expressed by cotransfection of plasmid pAM9. Since Rep alone is unable to promote replication of the PCV1 origin, investigation of the regulation of transcription could not be influenced by initiation of replication. To demonstrate the effect of the Rep protein on repression of Prep more clearly, the diagram shows the unrepressed activity of the pRL plasmids normalized to 1.

Binding of Rep mutants to the hexamers
To assess the contribution of the three conserved RCR motifs and the dNTP binding domain of Rep with respect to repression of P_rep, the transcription activity of plasmid pRP1-x was tested in combination with mutant Rep proteins (Fig. 6). In accordance with our previous results, no repression was seen when the Rep protein was omitted (pRP1-x/pSVL), while repression was observed when Rep was supplied by plasmids pORF4A (Rep+Rep') and pAM9 (Rep only) used as positive controls. Repression of P_rep was not altered, or only slightly altered, by Rep mutants expressed by plasmids pRep-mutI and pRep-mutIII, carrying mutated RCR motifs I and III, respectively. When mutants pRep(1217), pRep(1186) and pRep(1128) were tested, which were truncated successively from the C terminus, the capability of these Rep variants to suppress P_rep was reduced remarkably. Loss of repressive function was observed in pRep-mutII and pRep-mutP, in which the RCR motif II and the dNTP binding domain were inactivated, as well as in plasmid pRep(157312) retaining only the C-terminal half of Rep. This result indicates that the repressing function of Rep cannot been attributed to one single element of the Rep protein, but seems to be the result of a coordinated action, in which the C terminus containing the P loop, as well as motif II located in the N terminus, are involved.

(20K): Fig. 6. Repression of Prep by mutant Rep proteins. The replication-defective plasmid pRP1-x carrying the promoter of the rep gene of PCV1 was combined with several plasmids producing different mutants of Rep. The effects of these Rep proteins were analysed using the Luc/Gal assay.

Ratio of rep and rep' transcripts after infection of PK15 cells with PCV1
To analyse further the two transcripts originating from the rep gene, synthesis of rep and rep' transcripts was studied at different time intervals after infection of PK15 cells with PCV1. PK15 cells were transfected with the infectious clone pIC1, which carries the complete genome of PCV1 flanked by a 90 bp duplication (nt 10981188). Using plasmid pIC1, PCV1 infection of PK15 cells occurs via a recombination event, which enables the virus to recombine out of the plasmid backbone and initiate infection. The recombination can be monitored by a PCR reaction that differentiates between the input plasmid pIC1 and the virus (K. Hattermann and others, unpublished). Comparison of the ratio of rep and rep' transcripts was performed with a real-time PCR discriminating between the two transcripts. The result indicated a variation in the ratio of the two transcripts over time (Fig. 7). As early as 12 h after transfection, similar amounts of the two rep transcripts were detected. At 24 and 36 h after infection, rep' was predominant. The number of rep' transcripts then decreased from 48 to 96 h.

(56K): Fig. 7. Ratio of the two rep transcripts after transfection of PK15 cells with the infectious clone pIC1. Plasmid pIC1 carries the complete genome of PCV1 plus a duplication of 90 bp. This arrangement enables the viral genome to recombine out of the plasmid backbone and initiate infection in the cultured cells. RNA was isolated at intervals of 12 h and transcribed into cDNA. cDNA was amplified in a real-time PCR using two amplicons detecting either the unspliced rep or the spliced rep' transcript. The graph shows the ratio of the two transcripts rather than the absolute numbers of detected transcript molecules.

Two major open reading frames have been identified within the genome of PCV1. The cap gene encodes Cap, the major structural protein of PCV1. The rep gene directs the synthesis of the two Rep isoforms, Rep and Rep', which are essential for replication. The promoter of the cap gene has been previously mapped to nt 14251168 (Mankertz et al., 1998b ) and was fine-mapped in this study to nt 13281252. Compared with the late SV40 promoter, activity of P_cap was low. Little information about canonical promoter elements of P_cap was retrieved from database searches. Transactivation of P_cap by the viral proteins Cap, Rep and Rep' was not observed in our experiments. According to our observations, at least two as yet unidentified elements must be located upstream of P_cap. The first element strongly represses transcription originating from P_cap. Since this effect was not observed in plasmids in which nt 14751391 were deleted, this element must be located in close proximity to the cap promoter. A second element counteracts this effect by enhancing transcription of the cap gene. This element is located between nt 1569 and 1476.

The promoter of the rep gene was mapped to nt 640796 and, amongst others, two putative cytokine responsive elements were identified. When the activity of PCV1 promoters was tested in the presence of human IFN- γ and TNF-α, the positive control did not react, indicating that either porcine cytokines cannot be replaced by human analogues or that PK15 cells do not display cytokine receptors on their surface (A. Mankertz, unpublished results). Therefore, conclusions about the regulation of P_rep and P_cap by cytokines could not be drawn.

As demonstrated earlier (Mankertz & Hillenbrand, 2001 ), replication of PCV1 relies on the expression of the Rep and Rep' proteins, indicating that both proteins fulfil different functions in initiation of PCV1 DNA replication. Previous analysis has revealed that both proteins can bind to DNA fragments containing the right part of the stemloop and the inner hexamers H1 and H2 (Steinfeldt et al., 2001 ). Since the origin of replication (nt 728838) overlaps the fragment containing P_rep (nt 640796), we investigated whether Rep/Rep' can transactivate the rep gene promoter. We observed that P_rep was negatively regulated by expression of the Rep protein. Although binding of the Rep' protein to DNA in vitro has been documented, expression of the Rep' protein (as well as of Cap) does not lead to repression of P_rep. This difference in function may be attributed, for example, to the differential capability of Rep and Rep' to interact with transcription factors. Since binding of Rep and Rep' to H1 and H2 is a necessary prerequisite for initiation of replication (T. Steinfeldt, unpublished observation), we were interested in whether these motifs were also responsible for repression of P_rep. When plasmids were used in which one or two of the four hexamers H1/H2 and H3/H4 present in the origin of replication of PCV1 were altered in sequence, the result revealed that repression of P_rep by the Rep protein was mediated by binding of Rep to hexamers H1/H2 but not to H3/H4. Since H1 and H2 are essential elements in initiation of replication and repression of P_rep, this finding may imply a potential link between virus replication and transcription. Additionally, mutant Rep proteins were tested for their competence to repress P_rep. Inactivation of the conserved RCR motif II and the dNTP binding domain as well as the loss of the N-terminal half of the reading frame abolished repression of P_rep by Rep, while truncation of the C terminus led to reduction of the repression function. This implies that the transactivating function of Rep cannot be attributed to one single element or region. Obviously, the C terminus containing the P loop as well as the motif II located in the N terminus are involved in Rep-mediated repression of P_rep. Astoundingly, alteration of motif I, which has been identified as the iteron-recognizing domain of the Rep proteins of geminiviruses (Arguello-Astorga & Ruiz-Medrano, 2001 ), from FTLNN to LTLKN led to a Rep protein still capable of repressing P_rep. This finding has been corroborated by previous results, when the same mutant was found still to support replication to a limited degree (Mankertz & Hillenbrand, 2001 ). Assuming that binding of Rep to P_rep is a necessary event for repression of P_rep, it can be only speculated whether the function of motif I can be adopted by another domain as a back-up mechanism or whether the mutation introduced did not result in complete inactivation of this element.

The investigation of the ratio of the two rep transcripts after transfection with an infectious clone of PCV1 showed that the ratio of the spliced and the full-length transcript was similar at the start of infection. Later on, the number of rep' transcripts increased and then decreased. Although religated DNA of PCV is infectious (Mahe et al., 2000 ), one cannot be sure that transfection mimics the normal events in a PCV1-infected cell. Therefore, this may be seen as a first indication that the rep transcripts are subjected to a delicate and balanced regulation during replication of the virus.

By analysis of the two rep gene products, some functional differences between Rep and Rep' have been found. It has been observed that the two proteins differ with respect to their sequence requirements for binding DNA (Steinfeldt et al., 2001 ). In this study, differences in the functions of Rep and Rep' as transactivating agents of the rep promoter have been seen, as well as a concentration shift of rep and rep' transcripts in cell culture immediately after the initiation of infection with PCV1. The last finding raises new questions, since one can only speculate about the implication of the transcript number variation. Hopefully, future experiments will enable us to answer further questions concerning the functions of the Rep and Rep' proteins. In particular, we would like to know whether the synthesis of more than one rep gene product can be observed in the pathogenic PCV variant. When the DNA sequences of PCV1 and PCV2 are compared, the splice acceptor and donor sites of the rep transcript are conserved in PCV2, and thus we may also expect a Rep' protein in PCV2.

A.M. thanks Ms Petra Kurzendoerfer for technical assistance. This work was supported by the European Union (Project number QLK2-CT-1999-00307) and the Deutsche Forschungsgemeinschaft (MA 2126/2-1).

Footnotes

^a Present address: Medizinische Klinik I Gastroenterologie und Infektiologie, Universitätsklinikum Benjamin Franklin, Freie Universität Berlin, Hindenburgdamm 30, 12200 Berlin, Germany.