A linearised PCV-1 genome inserted in either orientation immediately upstream of the resident CMV I/E promoter (Pcmv) in the proven DNA vaccine pTHgrttnC significantly improved expression of the HIV-1C-derived polyantigen GrttnC. Early results suggested that genome orientation influenced expression through PCV-1 rep (pTHRepgrttnC) or cap (pTHCapgrttnC) promoter effects. Plasmid pTHRepgrttnC was unstable, suggesting that whole-genome orientation influences stability. Therefore, given that the effect was the same with either orientation of the PCV genome, we focused on elucidating the capacity of the PCV-1 capsid promoter Pcap to act as an expression enhancing element in conjunction with strong promoters in mammalian expression plasmids. Our chosen PCV-1 Pcap fragment was larger than the previously identified minimal Pcap [22, 23] and included more putative TF binding sites than those previously put forward to explain Pcap activity . When the Pcap fragment was cloned in reverse orientation (PcapR) upstream of Pcmv in pTHgrttnC (pTHPcapRgrttnC), GrttnC expression was enhanced to the same level as when the equivalent whole genome PCV-1 construct (pTHCapgrttnC) was used. This result is compatible with the possibility that PcapR may act as an upstream transcriptional enhancer for Prep in the native circular viral genome: circoviruses have highly compact ambisense genomes and some regions encode more than one function. In particular, in PCV-1 one sequence encodes Pcap, part of the rep gene intron, and the coding sequence for the full-length Rep protein  and lies approximately 1 Kb upstream of Prep - a feasible distance from Prep for PcapR sequence to act as an upstream transcription enhancing sequence for the rep gene . We speculate that the PcapR sequence in pTHRepgrttnC could have contributed to expression enhancement in pTHRepgrttnC, although this was not investigated in this study.
We identified two sequences within Pcap that enhanced GrttnC expression from Pcmv: these were the "core" sequence and the CLE (Figure 1 and 2). The active Pcap "core" sequence included closely situated consensus DNA binding sites for the transcription factors c/EBPb, the CCAAT/enhancer-binding protein which controls cell cycle progression , GATA-1, a constitutive Zn finger transcription factor , and CREB, a cAMP response element-binding protein  binding sites without AP2, SP1 and AP3 binding sites (Figure 2B). The nucleotide distances between these sites are consistent with them making up a composite TF binding site, as identified in other genes such as rat CYP2D5, with a composite c/EPB-beta/SP1 site , and human gonadotropin alpha-subunit, with a composite CREB/GATA-1 site . This putative C/EBPb/GATA-1/CREB composite TF binding site was not identified in a 76 bp minimal Pcap fragment , where Pcap activity was tentatively attributed to putative binding sites for AP2, SP1 and AP3 transcription factors. However, our active Pcap "core" fragment excluded AP2, SP1 and AP3 binding sites (Figure 2B), and Pcap promoter activity was previously seen to drop off almost completely , concomitant with the deletion of a fragment that we have currently identified as containing the putative GATA-1 and CREB sites. This strongly suggests that the GATA-1 and CREB binding sites, at least, are crucial to Pcap activity.
We tested constructs carrying other circovirus Pcap-cognate core sequences from PCV-2, a pathogenic strain of PCV, and BFDV-AFG-ZA, an avian circovirus. The PCV-1 and PCV-2 Pcap core fragments share 80% sequence identity overall and identical consensus c/EBPb/GATA-1/CREB binding site sequence and spacing. The c/EBPb, GATA-1 and CREB TF consensus sequence and site arrangement in the BFDV-AFG-ZA Pcap core is typical of avian circoviruses but different from that of PCV-1 and PCV-2. Equally improved expression occurred with pTHCRgrttnC (PCV-1) and pTHCR2grttnC (PCV-2). It thus appears the order and spacing of TF binding sites in the composite TF binding site, but not the specific nucleotide sequence between sites, was important to overall Pcap core activity in mammalian cells, at least in the current in vitro expression system.
Lying adjacent to the "core" sequence in our chosen Pcap fragment was a direct repeat of a consensus sequence AGTGGGCCCG separated by 19 nt and capable of forming a hairpin structure. This was identical to the transcription enhancing Conserved Late Element (CLE), first identified as an element in the geminiviruses, which are distantly evolutionarily related to circoviruses, but which infect plants . The CLE was later found to be a native upstream transcription-enhancing element in a number of bacterial and plant genes, showing additive increase in activity with number of CLE units [31, 32]. In a sequence database search performed in the current study we made the novel observation that CLE consensus sequences may be found in the DNA binding sites of zinc finger containing proteins in a number of genes from all phyla.
We showed that a concatenated PCV-1 CLE trimer enhanced GrttnC expression over that of pTHPcapgrttnC, the first demonstration that an identified CLE enhances gene expression in a mammalian expression system. It has been observed elsewhere  that maximal Pcap expression activity was associated with a sequence that included putative SP1 and AP3 binding sites (to which the maximal activity was attributed), but which coincidentally included the complete CLE. The 47 bp CLE - containing fragment tested and found active in the present study encompasses the putative SP1 and AP3 binding sites previously indicated , and so a contribution from these TFs cannot be ruled out.
To exclude the possibility that the observed expression enhancement in plasmids containing Pcap or its derivatives was fortuitous, we tested three other PCV-1 sub-genomic DNA fragments in pTHgrttnC and pGL3-promoter for transgene expression enhancement. These fragments, ranging in size from 266 - 316 bp, included putative TF binding sites as identified via Transcompel, other than any previously identified rep or cap promoter sequences. None of these fragments in either orientation showed any transgene expression enhancing ability in pTHgrttnC or pGL3-promoter (not shown).
On testing PCV-1 promoters in a different transgene or promoter context, we observed that PCV-1 promoters in whole genome inserts in pGL3-promoter appeared to be too distant from the SV40 promoter to affect luciferase expression, but the isolated Pcap fragment brought Pcap into closer proximity with the SV40 promoter, permitting enhanced luciferase expression. The stronger CMV I/E promoter [5, 8] gave approximately 25 fold more expression in the pTH context than the SV40 promoter in the pGL3-promoter context. Luciferase expression in pTHluc was not increased by a Pcap or PcapR insert. Our experience with other expression systems [28, 40] is that one gene insert may be expressed at significantly higher levels than another from the same gene control elements. Thus, the specific contribution of the PCV-1 Pcap element should be empirically determined for each new transgene tested.
The PCV-1-containing vaccine plasmids pTHCapgrttnC, pTHPcapgrttnC and pTHPcapRgrttnC showed significantly improved immunogenicity compared with the parent plasmid, pTHgrttnC. Use of pTHCapgrttnC clearly showed the potential for dose sparing, in that 10 ug doses elicited CTL responses in mice that were comparable to 100 ug doses of pTHgrttnC. This effect was further improved with an MVA-grttnC boost. Thus, even at low doses, these DNA vaccines elicited memory cells capable of responding to a matched booster vaccine.
A potential safety concern with using full length PCV-1 genome-containing plasmids as DNA vaccines is that the Rep protein encoded in the PCV-1 genome might bind to essential host factors involved in regulation of basic cellular functions. For example, Rep of the related PCV strain, PCV-2, was shown in vitro to bind c-myc, a multi-functional transcription factor involved with aspects of cell regulation, including cell proliferation and apoptosis . This concern could potentially apply to pTHCapgrttnC. Thus it is encouraging that pTHPcapgrttnC and pTHPcapRgrttnC showed similar immunogenic potential to pTHCapgrttnC. The 172 bp PCV-1 Pcap sequence included in these constructs has no involvement with c-myc binding, and so these plasmids are free of the above potential safety concern.