HPV vaccines are highly effective at preventing anogenital HPV infections and the neoplastic diseases that they cause (Herrero et al., 2015). Like other licensed anti-viral prophylactic vaccines, the HPV vaccines are thought to function primarily by inducing antibodies that bind the virus, thereby preventing infection (Schiller and Lowy, 2012). In the case of the three licensed HPV vaccines, these antibodies are induced by antigens comprised of L1 virus-like particles (VLPs), which morphologically and immunologically resemble the outer shell of the authentic virus. Cervarix, Gardasil, and Gardasil-9 are based on VLPs of two, four, and nine HPV types, respectively. The ability of the VLPs to consistently induce robust and durable systemic virus neutralizing antibody responses after intramuscular injection almost certainly explains their high type-specific efficacy against infection in clinical trials and effectiveness in national vaccination programs. In contrast to vaccination, natural HPV infections inconsistently induce virion antibody responses, usually after a delay of several months. The responses are typically several orders of magnitude lower than after vaccination, and are detected only transiently in some individually (Carter et al., 2000). The effectiveness of the antibodies induced by natural infection in preventing reinfection by the same HPV type remains controversial, although there is some evidence for protection in women, at least those with relative high titers of circulating antibodies (Beachler et al., 2016).

Given this background, the study of Scherer et al. reported in this issue (Scherer et al., 2016) is of considerable interest, because it provides the first in-depth analysis of the change in the VLP/virion antibody repertoire in individual women who were naturally infected by HPV, and who are then vaccinated with VLPs. The authors used stateof-the-art technologies to isolate circulating virion-specific memory B cells (Bmem). They then sequence the antibody-encoding genes, and finally clone, express, and functionally test the antibodies they encode. The primary, and perhaps surprising, conclusion of the study is that the antibodies in the Bmem repertoire after natural infection are predominantly low avidity and non-neutralizing. In contrast, the Bmem repertoire in women who were boosted by a single dose of Gardasil mostly contains higher avidity and neutralizing antibodies. The authors previously reported that women who were seronegative