Evaluating Bacteriophage P22 as a Tracer in a Complex Surface Water System: The Grand River, Michigan CHAOPENG SHEN, † MANTHA S. PHANIKUMAR,* ,† THENG T. FONG, ‡ IRFAN ASLAM, † SHAWN P. MCELMURRY, † STEPHANIE L. MOLLOY, ‡ AND JOAN B. ROSE* ,‡ Department of Civil and Environmental Engineering and Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan 48824 Received September 14, 2007. Revised manuscript received December 20, 2007. Accepted January 02, 2008. Viruses are important pathogens in both marine and fresh water environments. There is a strong interest in using bacteriophages as tracers because of their role as model viruses, since dissolved chemical tracers may not adequately describe the behavior of viruses that are suspended colloids. Despite a large number of studies that examined the transport of bacteriophages in the subsurface environment, few studies examined phage transport in large and complex surface water systems. In this paper we report the results of a dual tracer study on a 40 km reach of the Grand River, the longest river in Michigan, and we examine the performance of bacteriophage P22 relative to a chemical tracer (Rhodamine WT). Our analysis based on the transient storage (TS) model indicated that P22 can be successfully used as a tracer in complex surface water environments. Estimated P22 inactivation rates were found to be in the range 0.27-0.57 per day (0.12-0.25 log 10 per day). The highest inactivation rate was found in a reach with high suspended solids concentration, relatively low dissolved organic carbon content, and sediment with high clay content. Estimated TS model parameters for both tracers were found to be consistent with surficial geology and land use patterns. Maximum storage zone sizes for the two tracers were found in different river reaches, indicating that different processes contributed to TS within the same reach for the two tracers. This model can be used to examine the arrival times and concentrations of human viral pathogens released from untreated sewage at recreational areas. Introduction Bacteriophages are increasingly being used as tracers in hydrology because of their low detection limit, rapid and inexpensive enumeration methods (1), and their role as model viruses (2, 3). Bacteriophages are viruses that selectively invade specific bacterial cells and have no adverse effect on humans or animals. Viruses were implicated in 80% of the disease outbreaks for which etiological agents were identified in the past (4). Abundance tests (e.g., ref 5) have shown that viruses are ubiquitous in both the water column and sediments of marine and fresh waters. Of the major rivers sampled in Michigan, 33% tested positive for the presence of viable enteric viruses, and they are suspected to be the chief cause of swimming-associated diseases in recreational waters (6). Specifically, point sources of inadequately treated sewage discharge from urban areas can be one of the most significant virus pollution contributors (7). The majority of studies investigating virus fate have been performed under laboratory conditions (8) or in the subsurface environment (e.g., ref 9). In natural streams, statistical approaches were used in the past (e.g., refs 10 and 11). Earlier studies mostly focused on virus removal in constructed wetlands (12–14) and waste stabilization ponds (2). In a study conducted on the Areuse River in Switzerland (1), bacteriophage H40/1 was used with the chemical tracer uranine. The authors reported similar distribution patterns for the two tracers but a 20% higher recovery rate for the phage compared to uranine. To the best of our knowledge, the performance of bacte- riophages as surface water tracers (e.g., slug injection experiments) has not been evaluated in major rivers in the past. In the few studies reported in the literature, the focus was mainly on travel times and recovery rates, and the role of environmental factors contributing to phage inactivation/ attenuation has not been explored. Bacteriophages are known to undergo natural decay or inactivation as time progresses, adsorb readily to suspended particulate matter, and due to their colloidal nature, can aggregate into clumps large enough to settle out of the water. Factors that influence the inactivation of bacteriophages in the natural environment are complex and varied and have been described in several excellent review articles (8, 15, 16). Temperature and solar radiation are generally regarded as important contributing factors in surface waters (8). P22 is an icosahedral-shaped DNA bacteriophage, 52-60 nm in size, that belongs to the family podoviridae, contains dsDNA that is approximately 43 400 bp, and has a very short tail (17). Bacteriophage P22 infects smooth strains of Salmonella typhimurium (those that carry O-antigen surface polysac- charide). It was confused with PRD1 at one time, used extensively in groundwater tracer studies, and has recently been genetically characterized by our laboratory. The current strain used in our laboratory was shown to have inactivation rates of 0.02-0.05 log 10 per day in groundwater at temper- atures below 25 °C(15). Transport of solutes in rivers is controlled by dispersion as well as by transient storage (TS), which includes contribu- tions from both surface storage (e.g., due to in-stream vegetation) and hyporheic exchange (due to interaction with near-bed sediments). TS can significantly delay the down- stream transport of solute mass. In large streams, the trans- port of solutes is also influenced by catchment hydrology and watershed characteristics. Using Rhodamine WT (RWT) as a tracer, Gooseff at al. (18) recently found that increased geomorphic complexity from urban to natural settings in a Wyoming stream increased the potential for TS. Few studies, however, examined the impact of watershed-scale processes on TS in mixed land-use settings. In addition, the differences between responses of chemical and biological agents in these settings remain largely unknown. The aims of this paper were to (1) evaluate the perfor- mance of bacteriophage P22 in a major US river (the Grand River, Michigan) relative to RWT, (2) to evaluate the relative importance of different environmental factors that influence the inactivation (loss of virus per unit time) of bacteriophage * Addresscorrespondencetoeitherauthor.E-mail:phani@egr.msu.edu (M.S.P.); rosejo@msu.edu (J.B.R.). † Department of Civil and Environmental Engineering. ‡ Department of Fisheries and Wildlife. Environ. Sci. Technol. 2008, 42, 2426–2431 2426 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008 10.1021/es702317t CCC: $40.75  2008 American Chemical Society Published on Web 02/29/2008