About 38.6 million people worldwide are now living with the Human Immunodeficiency Virus (HIV), which causes AIDS 1. Heterosexual contact is the predominant mode of transmission of HIV infections worldwide. Women are at particularly increased risk of acquiring HIV through heterosexual contact. Despite this gender disparity, there are to date only limited options by which women may actively protect themselves against HIV. 2. Recent studies have defined factors that are associated with increased susceptibility to HIV-1 34. Among these, genital infection with herpes simplex virus type 2 (HSV-2) is considered a major cofactor for both sexual transmission and acquisition of HIV-1 5. HSV-2 is a member of the genus of double-stranded DNA viruses called

simplexvirus

In the light of the above evidence, developing biomedical strategies for the prevention of sexual transmission of HSV-2 has become recognized as a critical strategy in the control of sexual transmission of HIV-1 16. We recently described pre-integration viral genome slicing [PRINT_GSX] as a novel model for devising antiviral gene-based therapies using a retrovirus replication model (HIV cDNA) 17. This approach explores the natural antiviral defense model inherent in bacteria through a nucleic-acid enzymatic system called the restriction modification (R-M) system 18. Bacteria endowed with R-M systems have been shown to be remarkably resistant to tropism by bacteriophages. Four taxonomic classes of R-M systems are recognized to day, with type II being the most widespread 18. Type II R-M systems comprise two distinct peptides functioning respectively as restriction endonuclease (REase) and cognate methyltransferase (MTase). As a model illustration of function, class I RMS systems, the evolutionary ancestors of R-M systems, are employed here. The class I RMS of Escherichia coli strain K-12 comprises 6 enzymes, of which the respective genes are located on the bacterial chromosome in a region called an immigration island: the hsdS gene, hsdR gene, hsdM gene, mcrB/C genes and mrr gene. Products of the first two genes play the central antiviral defense function (by recognizing and splicing the exogenous DNA through recognizing 4–12 base pair palindromes; that is nucleotide sequences that read the same in both directions). The site specific subunit hsdS product serves to recognize the specific 4–12 " palindromic" base pair sequence in the genome of the invading phage, while the hsdR restriction subunit product cleaves the DNA if this site is unmethylated. The other 4 gene products serve to protect the host genome as follows: the hsdM gene product is a methyltransferase that transfers a methyl group from S-adenosylmethionine (SAM) to the DNA at the indicated A residues; the mcrBC system restricts DNA containing methylcytosine residues; while the mrr system restricts DNA with m6-methyladenine or m6-methylcytosine 1920.

The aim of this study is to extend our previous work on viral genome slicing (GSX) to HSV-2 by identifying REases (DNases) with potent ability to cleave the HSV-2 genome. Although the replicative cycles of some eukaryotic viruses such as HSV-2 do not necessary involve viral genome integration into the host nuclear DNA as occurs for retroviruses, we propose that these REases are equally worth exploring for the development of novel HSV-2 microbicides. Two models are proposed for using the REase EcoRII to target HSV-2: first, by cross-linking the enzyme through the formation of C31G (Savvy) and EcoRII PLGA-loaded nanoparticles (nano-C31G-EcoRII); second, by expressing EcoRII in Lactobacillus that also expresses a truncated recombinant form of the receptor nectin-1 (xREPLAB-tN1). The former are nanoparticles that may be explored to develop a model combinational microbicide, while the latter is a model "live" microbicide strategy for diverting and disrupting infectious HSV-2 particles.

This work extends the concept of viral genome slicing (GSX), previously described for human retroviruses as a module for research and development of novel antivirals at the genome level 17, to HSV-2. Because HSV-2 has been noted as a major cofactor in the sexual acquisition and transmission of HIV-1 56789101112131415, preventing HSV-2 infection in this way may be a potential strategy for reducing the sexual transmission and acquisition of HIV-1.

Here, we detail the first focused effort to identify REases with potential splicing activity against the HSV-2 genome (more than 700 sites) – BmyI, Bsp1286I, Bst2UI, BstNI, BstOI, EcoRII, HgaI, MvaI and SduI – which may be applied to research and the development of HSV-2 biomedical prevention strategies. All 9 of these REase are Type II restriction enzyme subtype P, derived respectively from the bacteria Bacillus mycoides 21, Bacillus sphaericus 22, Bacillus stearothermophilus 2U, Bacillus stearothermophilus 23, Bacillus stearothermophilus O22, Escherichia coli R245 24, Haemophilus gallinarum 25Micrococcus varians RFL19 26 and Streptococcus durans RFL3 27 (see Table 3; details of other cutting enzymes and frequency of splices are shown in Tables 1, 2 and [Additional file 1]). However, it should be noted that some of these enzymes are isoschizomers that are not significantly active under human physiological conditions. For instance, the three REases derived from Bacillus stearothermophilus have optimal activity at 60°C 212223. Such characteristics make them impractical for use in the design of microbicides. Therefore, not all these suggested restriction enzymes may actually be successfully applied in both approaches modeled. The enzyme EcoRII was selected because: (1) it is metabolically stable at temperature ranges inclusive of normal human body temperature(see Table 4 and Additional file 2) 24; (2) its source, the bacterium Escherichia coli, is similarly a Gram positive bacteria of which the cell wall anchoring system can be modified to express heterologous proteins as in Lactobacillus strains; (3) it exhibits one of the highest slicing potentials against the HSV-2 genome (a strategy that may be beneficial in avoiding spontaneous ligation-see tables 1, 2 and 3); (4) The REase is encoded on plasmids rather than the bacterial chromosome, making its transfer to other bacterial strains possible.

Several questions remain to be answered about the two proposed models. However, many of them can be addressed fully through in situ experimentation rather than modeling approaches. In both proposed models, it is possible to question whether the additional modifications – (i) cross linking EcoRII to N-9 or C31G (ii) expressing EcoRII in HveCt-expressing Lactobacilli – are relevant. For instance, while it is reasonable to propose that the EcoRII and N-9 or C31G PLGA-loaded nanoparticles may disrupt the viral envelope and possibly the viral capsid, bringing the naked genome into contact with the REase, one could nevertheless argue that the virus is no longer infectious by the time the genome is released from the virion, which would make the REase redundant. A similar argument could be made for the Lactobacillus approach. Once the virus has infected Lactobacillus, it cannot infect the vaginal epithelium, so destruction of the genome by REase appears unnecessary. Moreover, the N-9 comprised nanoparticles are used here for theoretical purposes, as their use in humans is bound to raise safety concerns emanating from the previous evidence of mucosal irritation and enhancement of both HIV and STI transmission 28. Never the less, in the absence of experimental evidence based on such nanoparticles, one could still argue their case from the fact that chemotherapeutic agents with noted in-vivo toxicity have been observed to exhibit extensively reduced such adverse effects when complexed into nanoparticles. For instance, DiJoseph et al have recently shown that conjugation of calicheamicin to rituximab with an acid-labile or acid stable linker vastly enhances its growth inhibitory activity against BCL in vitro, has no deleterious effect on the effector functional activity of rituximab, and exhibited greater anti-tumor activity against B cell lymphoma(BCL) xenografts and improved survival of mice with disseminated BCL over that of unconjugated rituximab. Such demonstrated reduced adverse effects of a calicheamicin immunoconjugate of rituximab demonstrate the safety advantage nanoparticles confer to initially unsafe bioactive agents 53.

In the case of the proposed nanoparticle model, it is not fully known by which bonds the REase will combine with the polymer (whether convalent or hydrogen bonds, as shown in Figure 4). Such bonds would presumably influence or affect the pattern of release of the components (covalent bonds are stronger and harder to break than hydrogen bonds). Moreover, the chemical models of "N-9 or C13G and EcoRII" PLGA-loaded nanoparticles shown in Figure 4 propose a single nonoxynol-9 or C31G molecule per REase. However, that may not be the case in the resultant nanoparticles (in situ evaluation of the composition of the nanoparticles is required). In addition, whether the molar concentrations of the respective active ingredients (N-9 or C31G and EcoRII) are sufficient to destabilize the viral envelope and genome, respectively, can only be decided by in situ experiments. Because of its previously demonstrated unsafe profiles in humans 29, any attempts to employ N-9 in such nanoparticles strategies are likely to exploit much lesser concentrations so as to achieve safety. In so doing, that may compromise efficacy for viral envelope disruption. Further still, it is not known whether such polymerization may affect enzyme or surfactant activity. Enzyme activities depend on active site conformations, and any changes in the 3D structure will probably influence activity. We have assumed that, since REases are stored in the simple ester construct glycerol, and PLGA is in essence a poly-ester, EcoRII may remain active despite copolymerization. Also, in the proposed nanoparticle model, the involvement of the hydrophilic hydroxyl group of N-9 or C31G or any other detergents in the interaction with PLGA could possibly affect the amphiphatic properties required to disrupt the viral envelope and capsid.

Irrespective of the answers to these questions, such nanoparticles would have advantages of their own. For instance: (i) they help to increase the stability of drugs and possess useful release-control properties; (ii) they offer an increased surface area of action for the drug iii) and enhance efficacy considerably; thereby involve use of lower concentrations of the bioactive agent relative to when used alone 535455. Nano-properties i-iii may avail one reason for experimental re-trial of agents like N-9 which has been previously found unsafe for use to prevent HIV or other STI 29. For such nanoparticles to be applicable in human conditions, it is imperative that we not only determine their size and Zeta potential but safety. In the past, dynamic laser light scattering from the Malvern Zetasizer 3000HAs system (Malvern Instruments, Worcestershire, UK) at 25°C at a 90° angle using PCS 1.61 software has been used to determine both nanoparticle size and Zeta potential 5455.

The "live microbicide" model also raises unique questions that can only be answered experimentally. First, there is still a need for in situ experiments to evaluate the efficacy of surface anchored HveCt expression by xREPLAB-tN1 in the same way that Liu et al have for 2D CD4 38. Previous expression of HveCt in insect line lines does not guarantee that it will be successfully expressed in Lactobacillus. Therefore, the efficiency of xREPLAB-tN1 engineering in respective to HveCt surface expression needs be determined by either (i) Partial purification of HveCt(tN1), (ii) Western analysis of HveCt expression in xREPLAB-N1, (iii) growth phase evaluation of HveCt productivity, or (iv) HSV-2 gD binding assays using whole-cell Lactobacillus extracts and affinity-purified anti-nectin1 antibodies (R7), as has been done elsewhere 3852. In situ experiments are also required to evaluate potential EcoRII expression, say by Phage (λ) DNA digestion assays following REase elution from L. jensenni whole cell extracts using electrophoresis, as described elsewhere 56. Lastly, testing the in vitro safety and efficacy of "xREPLAB-tN1" is mandatory prior to clinical application in humans. We have found no example of a eukaryotic virus infecting a bacterium, so it cannot be guaranteed outright that surface anchoring of HveCt would enable HSV-2 to be diverted into Lactobacilli.

Finally, many genomes of bacteriophages contain unusual nucleic acids bases 1920. For example, the T-even coliphage DNA contains not cytosine but 5-hydroxymethylcytosine, and most of the hydroxymethylcytosine residues in these DNAs are glycosylated as well 20. The genome of the B. subtilis phage contains a diversity of thymidine replacements, including uracil, 5-hydroxymethylcytosine, glycosylated or phosphorylated 5 uracil and alpha-glutamyl thymine. These unusual bases serve to render the phage genome resistant to degradation by host restriction enzymes 1920. It is likely that HSV-2 may become resistant to REase cleavage through similar variations in the viral genomes. This is a likely mechanism for the evolution of resistance to REase-based microbicides. Moreover, R-M systems do not operate with 100% efficiency, and a small number of phages have been noted to survive and produce progeny in bacteria 1920. This too may be a shortcoming of REase-based microbicides. We believe that such resistance may be overcome in future by altering the specificity of EcoRII. This concept is based on the fact that among R-M systems of the same class, transfer of the hsdS specificity gene (or protein) occurs naturally and serves to alter the specificity of the "R-M progeny" 1920. Similar alterations may be achieved through recombinant engineering, which implies application of the other 8 REases with potent cleavage potential against the HSV-2 genome, but with characteristics that make them less than ideal for use in either proposed model. Again, whether the transfer of specificity subunits from REase such as those derived from the Bacillus spp. would entail the persistence of unfavorable characteristics, such as functioning best at temperature ranges outside the normal human physiological range, can only be answered by experiments in situ.

We identify the REase EcoRII as a potential ingredient of HSV-2 microbicides. Modeled for the first time ever are (i) a nanoparticle for use in research and development of microbicides against HSV-2, and (ii) a "live microbicide" for diverting primary HSV-2 infection from genital mucosal cells coupled to genome disruption. Surfactants with safer profiles may form better candidates for conjugating to EcoRII.

All authors are affiliated to Restrizymes Biotherapeutics, a Ugandan biotech pioneering PRINT_GSX for antiviral therapy R&D.

WM conceived of the study, carried out the boinformatics analysis and participated in writing the draft manuscript. WM, BW and KH participated in the modeling, coordinating and writing the final manuscript. All authors read and approved the final manuscript.

HSV-2 whole genome = NCBI|

NC_001798

http://rna.lundberg.gu.se/cutter2/

http://www.expasy.ch/tools/protparam.html

http://www.ebi.ac.uk/pdbsum/1NA6