Prevalence of Respiratory Adenovirus Species B and C in Sewage Sludge

Human adenoviruses are among the most common human pathogens ¹ and cause variable infectious symptoms based on species, including respiratory, gastrointestinal, and ocular infections, summarized in Table 1 ^{2, 3}. Due to the emerging importance of adenoviruses in the environment, the U.S. Environmental Protection Agency included adenoviruses on the final Contaminant Candidate List (CCL3) ⁴. Sewage sludge represents the concentrated waste from thousands to millions of human inputs and human pathogens can be excreted through either feces or urine ⁵. Human adenoviruses are often considered the most abundant human viral pathogen in sewage sludge ⁶-⁸, and adenovirus has been proposed as an improved indicator of sewage sludge fecal pollution ⁶. The primary disposal method for sewage sludge is through land application as fertilizer, which may expose residents to the pathogens present in the sewage sludge through inhalation, ingestion, or direct contact ⁹. In a previous risk assessment of residential exposure to biosolids-derived aerosols, adenoviruses accounted for significantly higher predicted risk than enteroviruses ⁹, even though enteroviruses are the only viruses considered in U.S. land application regulations. Thus, understanding the infectious properties of adenoviruses present in sewage sludge is necessary to ensure the protection of public health. The goal of this study was to determine the distribution of adenovirus species in untreated and anaerobically digested sewage sludge (commonly termed biosolids), using an amplicon sequencing approach. A comprehensive review of sequencing approaches for viral pathogen identification is provided within reference ¹⁰. Prior adenovirus wastewater research has focused primarily on HAdV-F and gastroenteritis ^{6, 11, 12}, or adenoviruses as a whole, with no discrimination of individual species ^{6, 8}. Previous sequencing efforts of the adenovirus hexon gene in wastewater, while informative, have been limited by number of sequences and have also cultured viruses prior to sequencing, potentially hindering insights into overall species distribution ^{13, 14}. However, different species have highly divergent pathogenic characteristics, and knowledge of species distributions would improve our understanding of the consequences of human exposure and infectious risk in the environment.

Untreated (IN) and anaerobically digested (OUT) sewage sludge was sampled from five wastewater treatment plants geographically distributed throughout the continental U.S. resulting in a total of ten samples. Sampling was conducted during winter months, and all plants received residential wastewater, served populations ranging from 100,000 to 1,000,000 people, and produced U.S. EPA designated class B biosolids. DNA was extracted directly from sewage sludge using a Mobio Powersoil DNA kit (Mobio, California, USA), as previously described ⁸, and amplified with previously developed hexon gene targeted primers ^{15, 16} that were amended with a 5’ sample-specific barcode. Adenovirus hexon genes are known to contain variable regions that are useful for identifying species and types ^{1, 17}-¹⁹ and the primers used have previously been shown to be able to amplify and differentiate adenovirus species A-F ¹⁵. A table detailing species-specific primer mismatches is included in the supplementary information (Table S1), an analysis of primers and primer target genes for adenovirus species A, B, C, D, E, and F, demonstrated no statistical difference (p<0.05) in the number primer-target mismatches between any of the species. PCR amplicons were ~150bp in length. All 10 samples were PCR-positive for adenovirus. It has previously been estimated that approximately 0.1% of adenoviruses detected by molecular methods would be detected by cell culture methods ²⁰; however, it is not known what percentage is infectious in an environmental context. PCR products were gel-verified, purified using a Mobio PCR clean-up kit (Mobio, California, USA), and sequenced on an Ion Torrent 314 chip at GeneWiz (New Jersey, USA). Raw sequences are deposited in MG RAST ²¹ under accession number 4501936.3. Sequences underwent initial QC using DynamicTrim from the solexaQA toolkit ²² for a minimum phred score of 15 and were split using the pyrosequencing pipeline on the RDP website ²³ with a minimum read length of 100 bp. Trimmed reads underwent initial annotation through a BLASTn search of the NCBI viral genome database with a maximum E-value of 0.001. This E-value has been commonly used in metagenomic studies, and has previously been shown to reduce false negatives ⁷. A table summarizing read count data is included in the supplementary information (Table S2). The database contains representatives from HAdV-A to -F. Addition of adenovirus 52, the only member of species G, to the database did not result in any species G identifications. Approximately 99.8% of annotated reads were annotated as adenovirus. Comparison with the NCBI “nt” database was not practical due to the high number of adenoviral vectors in the database that also match these sequences. Adenovirus species annotations are summarized in Figure 1.

The most abundant adenoviral species was C in nine out of ten cases, with an average relative abundance of 78% ± 17% standard deviation, followed by adenovirus species B, with an average relative abundance of 20% ± 17%. Although most of the emphasis in prior environmental monitoring literature is placed on gastrointestinal adenovirus species F, the average relative abundance of adenovirus species F was only 0.7% ± 1.7%, with a maximum of 5.5%. However, it is important to note that due to varying infectious characteristics, higher abundance of certain species may not necessarily correlate with higher infectious risk from environmental exposure. There was no treatment effect on species distribution. For example, average abundance of species HAdV-C decreased from 84% ± 15% in untreated samples to 71% ±18% in anaerobically digested samples; however, these differences were not statistically significant.

It is interesting to observe that these findings in sewage sludge reflect previous findings of clinical adenoviral infection in the general public. Gray et al. reported the most abundant clinical adenovirus species in the US were HAdV-C (57.6%) and -B (34.6%) in the general population ²³. Madelboim et al. reported similar findings from clinical presentations in Israel, for adenovirus species C (57%) and B (14%) ²⁴. As sewage sludge is the concentrated waste from the general population and adenoviruses are excreted into this waste stream, the relative abundance of adenoviruses in sewage may mirror that in the general population. In addition to supporting our species distribution results in sewage sludge, these clinical observations highlight the importance of respiratory adenoviruses in human populations and suggest the potential for population disease monitoring through sewage.

As an alternative means of interpreting the data, and to obtain a phylogenetic, rather than taxonomic, assessment of adenovirus diversity, a subset of sequences were aligned using clustalX ²⁶, and an unrooted tree was produced and formatted in MEGA ²⁷ (Figure 2). None of the sequences included in this subset were annotated as HAdV-F. This independent analysis again highlights the abundance of HAdV-C, and also demonstrates that the read length produced provides the resolution necessary to show taxonomic and phylogenentic diversity of human adenoviruses.