An IHNV-based reverse genetics system (5) has been used to generate a recombinant IHNV that expresses the Renilla reniformis luciferase reporter gene (20), rIHNV_LUC (Fig. ​(Fig.1A,1A, middle). A full-length pIHNV cDNA clone (4) was modified such that an additional expression cassette, containing the M gene-derived start and stop transcription sequences, was inserted into the unique EagI restriction enzyme site located between the M and G genes. The Renilla luciferase open reading frame was inserted into the expression cassette, leading to a pIHNV_LUC construct. Recovery of rIHNV_LUC was achieved as previously described (3, 4). Expression of the LUC gene was first monitored in rIHNV_LUC-infected EPC cell lysates by using a luminometer. In a time course experiment, we found that luciferase expression started as early as 2 hours postinfection, and the light intensity progressively increased to the highest level 36 h postinfection (data not shown).

In order to test whether the insertion of the luciferase expression cassette into the IHNV genome had led to a modification in the pathogenic potential of the virus, juvenile trout (n= 75; mean weight, 0.5 g) were infected, in three separate experiments, by bath immersion with the viral isolate IHNV 32-87 and either the wild-type (wt) rIHNV or rIHNV_LUC (virus titer, 5 × 10⁴ PFU/ml), and mortalities were recorded for 20 days. All three viruses, rIHNV_LUC, wt rIHNV, and IHNV 32-87, proved to be highly pathogenic in yearling trout (Fig. ​(Fig.1B).1B). Experimental infection was repeated in order to sample infected fish during the first 12 days of infection. Various fish organs (spleen, kidney, heart, and liver) were taken every day. Organ samples were homogenized, and luciferase activity was monitored. The most affected organs exhibiting active virus replication appear to be the spleen and kidney at day 3. In the very early stages of infection (days 0 to 2), no significant luciferase activity was detectable in the fish organs (data not shown). In order to follow the spread of the virus in a noninvasive way, we first investigated the possibility of direct visualization of virus replication in the fish by infecting trout and adding the EnduRen live cell substrate (Promega) into the water tank. As a first try, infected fish were taken at 4 days postinfection, a time at which extensive virus replication is expected. Anesthetized fish were subjected to imaging using a CCD camera (Fig. ​(Fig.2A).2A). As visualized in Fig. ​Fig.2B,2B, active replication of rIHNV_LUC was observed most abundantly in the oral cavity, the esophagus/cardiac stomach region, the pyloric caeca, the kidney, and the spleen, but also in the dorsal fin. One of the infected fish was sacrificed, and the skin was removed. The autopsied fish was subjected to CCD imaging to demonstrate that the detected light was readily emitted from inside the body, indicating that imaging can be achieved through the mucus and the skin on live fish (data not shown). The kinetics of virus replication following waterborne exposure was analyzed starting from 8 h postinfection. To do that, the fish were exposed to virus infection by bath immersion (5 × 10⁴ PFU/ml), and at the different time points, the fish were sampled and transferred in a tank containing the water-soluble luciferase substrate. The results of the different observation periods (Fig. ​(Fig.3)3) were unexpected, since at the longer times post-virus exposure, 40 and 24 h, replication of the virus was seen in various fish organs but virus replication was also evidenced at the fin bases. It was even more demonstrative at 16 and 8 h, times at which viral replication was readily restricted to the fins. No virus replication could be observed in the gills at the times considered.

IHNV produces an acute, lethal infection in rainbow trout. However, between 10 and 20% of infected trout do survive the infection (14, 16). In order to investigate the spread of the virus after 3 weeks, we subjected four trout that survived the infection and one mock-infected trout to BLI (Fig. ​(Fig.4A).4A). Contrary to the findings at the earlier time point (when all the analyzed trout showed similar patterns of luciferase activity), we noticed at this late time point a large heterogeneity among the infected trout both in the level and in the distribution of the luciferase activity. Some trout showed a systemic dissemination of the virus, while the ones that succeeded in containing the viral infection demonstrated a localized bioluminescence signal in the fins, which supports the idea of particularly permissive cells for IHNV being localized in the fins.

We previously described a recombinant rIHNV-ΔNV in which the NV gene encoding a nonstructural protein of unknown role was deleted and replaced by the reporter green fluorescent protein (GFP) gene (4). This virus is nonpathogenic for trout (28). rIHNV-ΔNV was further engineered to add the luciferase expression cassette between the M and G genes, rIHNV_LUC-ΔNV (Fig. ​(Fig.1A).1A). The fish were infected by bath immersion and left for 3 weeks, with no signs of disease and no mortalities. Bioluminescence assays at 3 weeks postinfection on rIHNV_LUC-ΔNV-infected fish showed a limited viral replication exclusively localized in the fin bases (Fig. ​(Fig.4B),4B), reminiscent of the observations at the early times (8 h) of infection with rIHNV_LUC (Fig. ​(Fig.33).