Recent studies have revealed that a number of insect baculoviruses (Baculoviridae) exist as a mixture of genotypes that can simultaneously infect their hosts (5, 10, 12, 26). Baculoviruses are ideal model organisms for studies on host-virus interactions because they often naturally survive as mixed-genotype populations, they are highly amenable to laboratory manipulation, both in insect cell lines and in insects, a great deal of genomic sequence information is available in data-bases, they are easily quantified and have an out-standing safety record.
The genome of these arthropod pathogenic viruses is a circular, double-stranded DNA molecule that is included in a protein capsid, constituting the nucleoca-psid. Nucleocapsids are enveloped by a lipoprotein membrane to form the infectious particles or virions. Outside their hosts, baculoviruses are found as large proteinaceous occlusion bodies (OBs) in which one or more virions are embedded. The usual route of infection is the ingestion of OBs by susceptible larvae. OBs dissolve in the alkaline environment of the host midgut, releasing enzymes and occlusion derived virions (ODVs) that cross the peritrophic matrix and initiate a characteristic biphasic infection cycle. The ODVs set off the primary infection by fusion with the columnar epithelial cells of the larvae. This first step is enabled by the presence of a group of viral encoded proteins such as P74 (16, 30, 32), PIF1 (14), PIF2 (23), PIF3 (21) and PIF4 (Ac150) (17, 31). Infected cells give rise to a second morphotype, the budded virus (BVs), which are responsible for disseminating a systemic infection within the host.
Lepidopteran nucleopolyhedroviruses (NPVs) (genus Alphabaculovirus) are characterized by having multiple virions occluded within each OB. Commonly observed features of NPVs are the coexistence of multiple genot-ypes within the same virus isolate (5, 15, 18). This genetic diversity can be observed at the level of single larvae (4, 5, 22). It has been demonstrated that in larvae, individual cells are usually infected by mul-tiple BVs (2, 9, 24), and the genotypes in these BVs might be diverse. The importance of such diversity in the insecticidal properties of the virus population can be studied using experimental mixtures of genotypes, in addition to the observation of the behavior of natural virus populations.
In a Nicaraguan isolate of Spodoptera frugiperda multiple nucleopolyhedrovirus (named SfNIC), eight out of nine plaque purified variants were identified as genotypes that contained deletions (named SfNIC-A, and SfNIC-C to SfNIC-I). These deletions represented between 4 and 11% of the genome compared to the genome of the single complete genotype, named SfNIC-B (26). SfNIC-C possess the largest deletion (16.4 kb) that affects a number of genes including pif1 and pif2. For this reason it is unable to infect larvae per os (26). However, this genotype can survive by complementation with the complete genotype (SfNIC-B). In cells that are simultaneously coinfected by both genotypes, both PIF1 and PIF2 proteins are produced by the complete genotype rendering the ODVs in-fectious per os, irrespective of the genotypic com-position of the progeny nucleocapsids. Moreover, remarkable cooperative behavior has been demons-trated to occur in mixtures comprising SfNIC-B and SfNIC-C. The LC50 value of OBs comprising SfNIC-B is ~2.5-fold higher than that of wild-type SfNIC OBs, indicating reduced potency of SfNIC-B OBs. How-ever, when ODVs were released from mixtures com-prising ~75% of SfNIC-B OBs and ~25% of SfNIC-C OBs, and injected into insects that subsequently died of polyhedrosis disease, the resulting OBs comprising co-occluded genotype mixtures had an insecticidal potency similar to that of the wild-type isolate (19). Recent studies with pif1/pif2 deletion recombinants have revealed that (ⅰ) it is the deletion of this region that is responsible for the observed potentiation of insecticidal potency in mixtures of complete and deletion genotypes and, (ⅱ) the proportions of com-plete and deletion genotypes present in mixtures subjected to serial passage per os rapidly converge to a genotype composition that results in the highest virus transmissibility (3).
Whether this selection for specific ratios of genotypes occurs mainly during the establishment of the primary infection process in insect midgut epithelial cells or during the later development of the systemic infection of the host is unknown. Recently, Zwart et al. (33) attributed negative selection for defective genotypes to changes in the average multiplicity of infection (MOI) during the systemic infection process. In an attempt to elucidate this question, we analyzed the frequency of SfNIC-B and SfNIC-C genotypes at different time points throughout the infection process in larvae inoculated with different co-occluded mixtures of these two genotypes.
Spodoptera frugiperda larvae are a laboratory colony maintained at 25℃, 75% RH, and 16 h light: 8 h dark photoperiod on a semi-synthetic diet (7). Two distinctive genotypes, SfNIC-B and SfNIC-C, were used in this study. These genotypes were purified in vitro from a natural population of S. frugiperda nucleopolyhedrovirus (SfMNPV) originally isolated in Nicaragua (SfNIC) (6). SfNIC-B genotype was the only genotype with a complete genome; all other genotypes presented deletions of varying length in one region of the genome (26). The SfNIC-C genotype presents a 16 kb deletion and is not infectious per os but is capable of replication following injection or when co-infected with a genotype that is infectious per os. OBs of SfNIC (wild-type), SfNIC-B and SfNIC-C were amplified by droplet-feeding (13) fourth instar larvae in the case of SfNIC and SfNIC-B or by intra-hemocelic injection in the case of SfNIC-C. OBs were collected from dead diseased insects, filtered through muslin and centrifuged to eliminate insect debris. OB concentrations were determined using an improved Neubauer hemocytometer (Hawksley, Lancing, UK) under phase-contrast microscopy. Viral DNA extraction from a suspension containing ~109 OBs was performed following the method described by Simon et al. (27). Restriction endonuclease analysis was used to verify the DNA profiles of both purified genotypes and the wild-type isolate.
To obtain co-occluded genotype mixtures, purified OB suspensions of the SfNIC-B and SfNIC-C genot-ypes were diluted to a concentration of 5×108OBs/mL, and dissolved in alkaline buffer (1 vol. OBs: 1 vol. Na2CO3 0.1mol/L: 5 vol. H2O) at 60℃ for 10 min. to release ODVs. Undissolved OBs were pelleted at 2 655×g for 5 min and discarded. Each ODV-containing supernatant was then mixed in one of three different proportions: 90:10, 50:50 and 10:90. These mixtures were named the co-injection samples (INJ), as they were injected into groups of 50 S. frugiperda fourth instars (~8 μL/larva) from the laboratory colony. These larvae were individually reared until death. The OBs obtained from these larvae were purified and pooled, and each sample, named Day zero samples (D0), was used as inoculum to infect second instars per os by the droplet feeding method. Groups of 50 larvae were inoculated with one of the three co-occluded mixtures at a concentration of 1.2×106 OBs/ mL, which was estimated to result in ~90% mortality. Groups of eight larvae were sampled every 24 h during a five day infection period, anesthetized on ice, surface-sterilized with ethanol and their prolegs were cut to obtain the hemolymph, which was transferred im-mediately into ice-cold, sterile eppendorf tubes con-taining 5 μmol/L L-cysteine to prevent melanization. A haemocyte-free suspension was obtained by centrifu-gation at 664×g for 5 min. at 4℃. Supernatants, consti-tuting the samples taken at 1 to 5 days post-infection (D1-D5), were stored at -20℃ until further use. The whole procedure was performed three times.
The relative proportions of SfNIC-B and SfNIC-C present in: ⅰ) the ODV mixtures that constituted the co-injected samples (INJ), ⅱ) the D0 samples of OB inocula and, ⅲ) the BVs from hemolymph samples taken at 1 to 5 d post-infection were analyzed by PCR using primers designed by Simon et al. (26) that differentially amplify the two genotypes (Fig. 1). The lengths of the amplified fragments were estimated to be 763 bp for SfNIC-B and 646 bp for SfNIC-C (Table 1). PCR reactions were stopped in the linear phase of amplification (19 cycles) as previously described (19). Each amplification was performed three times and amplicons were then purified using a PCR purification kit (Roche, Sant Cugat del Valles, Spain). PCR products were separated by 1% agarose gel electrophoresis and the relative proportions of individual genotypes estimated by densitometric analysis of the intensities of their respective PCR products using QuantityOne (BioRad, Alcobendas, Spain) image analysis program as described by López-Ferber et al. (19).
Figure 1. Illustration of the locations of primers used to differentiate SfNIC-B and SfNIC-C genotypes present in mixtures by semi-quantitative PCR within the hypervariable region of the SfNIC genome.
Table 1. PCR primer combinations used for the amplification of SfNIC-B and SfNIC-C and sizes of the products.