All affected larvae examined were pharate instars. While the old cuticle was unshed , the new cuticle was thin and undulated . Segmentally, the infected epidermis was 2-to-5 times thicker than the unaffected epidermis. Infected epidermal cells had cytoplasmic elongation. Their nuclei were at different levels, giving the epidermis an appearance of a pseudostratification or hyperplasia . On hematoxylin and eosin–stained preparations, InIs varied from basophilic to eosinophilic . The cells’ nuclei were severely enlarged, and the inclusions replaced or displaced the chromatin to the nuclear margin . There was individualization of epidermal cells and fragmentation of the nucleus and cytoplasm with formation of apoptotic bodies as later confirmed by transmission electron microscopy . Eosinophilic inclusions had variable textures. While the homogeneous, basophilic inclusions correlated with strong, Feulgen-positive reactions, the eosinophilic inclusions presented faint Feulgen staining . Large numbers of basophilic intracytoplasmic granulations, interpreted as autophagy and heterophagy, were present in a significant number of epidermal cells, as later confirmed by TEM. Multifocal, epidermal cells of all segments of the body were hypertrophied with InI and cell death. In advanced stages of infection, this lesion progressed to formation of vesicles between the endocuticle and basal lamina or subcuticular epidermal cell layer discontinuity . Vesicles contained individual epidermal cells, cellular and nuclear debris, and hemocytes floating in a proteinaceous fluid. When epidermal cells that specialize in the skeletal muscle/ cuticular attachment were infected and damaged by the virus, square plastic pot the skeletal muscle separated from its insertion with the cuticle—“muscle avulsion” .

The detached skeletal muscle cell had variable degrees of degeneration with hyperstretching, hypercontraction, and necrosis and was observed in the muscle of the head, thorax, and legs, which included the craniomandibular muscle, dorsal and ventral longitudinal muscles, the dorso-ventral muscle, and muscles of the legs . Infected epithelial cells of the tracheae and tracheoles were hypertrophied and contained InIs, and some cells were undergoing cell death with fragmentation of the nucleus and cytoplasm . Similar to the cuticular epidermis, tracheal and tracheolar epithelial cells of all body segments and organs or systems, such as skeletal muscle and the nervous system, were affected . Other cells infected included pharyngeal, esophageal, and rectal epithelial cells . There was a variable influx of hemocytes and proteinaceous fluid filling the pseudocoelom and legs . The viral infection was variable, but, in general, the head, thorax, and podal appendages were more severely affected. Other findings in infected YMW mealworm larvae not directly attributed to the viral infection were loss of mass and necrosis of fat body cells, and multiple hemocyte nodules.Negative-contrasted virions were nonenveloped and icosahedral . Densovirus virions varied in diameter from 23.79 to 26.99 nm, averaging 25.19 nm. Frequently, hollow particles were observed . Approximately 7 × 1012 densovirus virions per gram of tissue were recovered from T. molitor larvae. This estimation is based on an average of 361.5 viral particles enumerated per 0.35 μm2 over a total of 20 individual 0.35 μm2 areas and is provided under the assumption of even viral distribution on the grid and in the viral pellet suspension, as well as the assumption that the totality of viral particles was extracted from the sample tissue using the viral extraction methods employed. Ultrastructurally, infected epidermal and epithelial cells of respiratory tubules presented with disintegration of the cuticular and intercellular anchoring junctions .

Cells were rounded and displayed electron-dense mitochondria, laminar-to-circular ribosomal-lamellar complexes, and tubular-reticular complexes . The nuclei were enlarged to double or triple the normal size, and the chromatin was replaced by a densovirus replication and assembly complex . Infected epithelial cells at the “cuticle-epithelial cell-skeletal muscle attachment” detached from the endocuticle due to the disintegration of the muscle attachment fiberhemidesmosome junction and disintegration of the desmosome at the epithelial cell-skeletal muscle attachment. Degeneration of the tonofibrils was also present . Respiratory epithelial cells presented similar ultrastructural changes to epidermal cells. However, DRAC in these cells was denoted by homogeneous accumulation of nucleocapsids compared with the DRAC of the cuticular epidermal cells . In epidermal cells, DRAC most commonly presented itself as a pleomorphic structure composed of a viral matrix that displaced and replaced the chromatin . The virus matrices were fibrillar-granular with variable densities. The denser segment of the virus matrix, known as “virogenic stroma,”13 was composed of short-to-long anastomosing streams, in which presumably maturing and mature virions were aggregating at the interface with the looser granular matrix spaces . DRACs evolved to form paracrystalline arrays , although these paracrystalline arrays were less common. In general, there was high variability of the DRAC morphology, which seemed dependent on both the stage of the cell infection and the proportion of dense and loose viral matrix. Figure 7 displays the most distinct features observed. Occasionally, nuclear fragments containing chromatin, virions, or paracrystalline arrays were found in the cytoplasm of infected and unaffected epithelial cells . Cell death was characterized by condensation of the cytoplasm and organelles, as well as nuclear envelope breakdown, blebbing, and fragmentation, which was interpreted as apoptosis . However, lysis of cytoplasm and nuclei was also observed .

Nucleoli were displaced to the periphery and fibrillar and granular components were distinguishable . Free, small fragments of epidermal cell cytoplasm containing chromatin and/or viroplasm were interpreted as apoptotic bodies .Whole-genome sequencing of the mealworm densovirus generated 296,000 reads, of which 71,338 reads mapped to a bird associated densovirus reference strain . The entire 5579 bp genome was detected at more than 10,000× depth of coverage. The consensus sequence was submitted to GenBank and given the accession number: MW628494. The genome contained 5 open reading frames , 2 on one strand and 3 on the other . ORF 1 and ORF 2 encode predicted structural proteins of 593 and 303 amino acids, respectively, while ORF 3 , ORF 4 , and ORF 5 encode predicted nonstructural proteins of 533, 265, and 170 amino acids, respectively. A phylogenetic analysis of the mealworm densovirus showed it to be closely related to several bird- and bat associated densovirus, sharing 97% to 98% identity. Meanwhile, the nucleotide similarity to mosquito, cockroach, and cricket densoviruses was 55%, 52%, and 41%, respectively .An outbreak of densovirus infection with high mortality in a mealworm farm is described. The initial diagnosis was based on cellular hypertrophy with InI bodies and death of epithelial cells of the epidermis, pharynx, esophagus, rectum, and tracheal system on histological examination, as well as the detection of virions consistent with viruses of the family Parvoviridae by direct TEM. Further genomic characterization shows that the detected YMW virus belonged to the genus Ambidensovirus of the subfamily Densovirinae, a group characterized by its unique bidirectional transcription. A phylogenetic comparison of the whole genome placed this virus among several closely related densoviruses detected in the feces of various birds and bats. Unfortunately, given the broad diets of these insectivorous hosts, square pot it is impossible to ascertain whether meal worms or other unrelated insect species may have been the sources of the described densovirus. However, a densovirus recently identified as the causative agent of a superworm mortality event at a Moscow zoo showed 97% nucleotide identity across 774 bp of the ORF 3 region to the mealworm densovirus in this study. The next most closely related densovirus of insect origin is a mosquito densovirus , which only shares 55% nucleotide identity with the mealworm densovirus. In comparison, the mosquito densovirus genome is significantly smaller , and its organization is vastly different, containing a truncated structural protein gene and completely lacking a nonstructural protein . There are currently relatively few complete insect densovirus genomes available in Genbank. This makes it difficult to draw firm conclusions about their relationships, and it is likely that the taxonomic groupings within the subfamily will be further refined as more genomes are described. As this is the first whole-genome characterization of a YMW densovirus, we propose the name TmDNV. The current outbreak of densovirus in T. molitor larvae achieved 100% mortality. This is consistent with 90% to 100% mortality reported in densovirus outbreaks affecting Z. morio39 and house crickets in Europe and in the United States. Introduction of diseased insects or persistently infected individuals is the most probable origin of the current outbreak, although the exact mechanism of the virus’ introduction to the mealworm rearing facility and disease predisposing factors are impossible to determine as epidemiological information remains largely anecdotal.

Experimentally, it has been demonstrated that a colony of mid-age-to-young Z. morio larvae can collapse within 1 to 2 months after the initial infection. Larvae usually die within a few days of infection; however, surviving larvae can evolve to the pupal and adult stages of their life cycle. Densoviruses are actively infectious for more than 18 months after the death of their host insect. We estimated by TEM that perished larvae of T. molitor contain 7 × 1012 particles per gram of larvae tissue . This is an extraordinary number of virus particles per larval carcass that, if aerosolized, could contaminate the surrounding environment and fomites, leading to effective dissemination in and across rearing facilities for extended periods of time. The large quantity of viral particles generated in dead larvae permits the use of direct TEM as a rapid diagnostic screening method, preceding molecular tests or due to their unavailability. One inconsistency, which may limit the diagnosis of the virus on phosphotungstic acid–contrasted direct TEM preparation, is the large dimensional range reported for Densovirinae at 17 to 27 nm in diameter. These dimensions overlap in the higher range with viruses of the order Picornaviruses and in the lower range with viruses of the family Circoviridae.Densovirus particles in the current outbreak in T. molitor averaged 25.19 nm in diameter, ranging from 23.79 to 26.99 nm. A significant contribution of our report is the detailed histological and ultrastructural characterization, which confirms densovirus as the cause of high mortality in T. molitor. The TmDNV primarily affects epithelial cells of the cuticular epidermis, tracheae, and tracheoles, in addition to the pharynx, esophagus, and rectum. These are all cuticular-secreting epithelial cells, which suggests that this TmDNV is primarily epitheliotropic. On hematoxylin and eosin preparations, cytomegaly and karyomegaly with basophilic and eosinophilic inclusions of these cells are widely found in the head, thorax, and abdomen of early and late larval instars. Because tracheoles are distributed throughout the body of the larva, densovirus inclusions can beobserved in the skeletal muscles, nervous systems, fat bodies, Malpighian tubules, gastrointestinal tracts, dorsal vessels , and silk glands of severely infected individuals. These findings diverge with polytrophic members of the genus Ambidensovirus, which can efficiently replicate in most larval, nymphal, and adult tissues. Nevertheless, similar to polytrophic Ambidensovirus, TmDNV does not affect the midgut glandular epithelium of its host. Generally, clinical signs reported in lethally infected natural hosts are anorexia and lethargy followed by flaccidity, which progresses to paralysis, slow “melanization,” and death.38 These clinical signs are consistent with those reported in the current densovirus outbreak in T. molitor; however, the terminology used should be revised to avoid imprecisions. Thus, based on the nature and extension of the lesions in densovirusinfected T. molitor, it can be suggested that clinical signs should primarily reflect an impairment of the locomotion, food ingestion, respiration, and body water balance as a consequence of the infection of epithelial cells of the epidermis, foregut and hindgut, and tracheal system. Locomotion impairment due to significant lesion was the result of the death of epidermal cells at the skeletal muscle-cuticular attachment, causing intrasegmental and intersegmental muscle detachment. Detached muscle fibers presented hyperstretching and hypercontraction followed by degeneration and necrosis. Skeletal muscle detachment affecting craniomandibular muscles results in an inability to prehend food rather than “anorexia,” which would imply a lack of appetite. Furthermore, multifocal infection of dorso-ventral muscles, dorsal and ventral longitudinal muscles, and intrinsic and extrinsic leg muscles will cause asymmetrical and asynchronous locomotion that generates swirling, rolling, and non-ambulatory larva rather than paralysis and/or ataxia, which are terms usually associated with neurological impairment. Respiration impairment is due to skeletal muscle lesions that interfere with larval breathing by affecting convective gas transport, which variably depends on muscle contractions. In addition, the death of tracheolar epithelial cells may further impair larval respiration by interfering with O2/CO2 transport and exchange. In terrestrial insects, transpirational water loss through the integumentary and respiratory systems accounts for more than 60% of the body’s water loss.