Thiamine status of rainbow smelt (Osmerus mordax) eggs in the Great Lakes, USA

Abstract

During spring 2006–2009, eggs were collected for analysis of total thiamine from gravid rainbow smelt (Osmerus mordax) captured in each of the Great Lakes and two other waters as references for comparison. Mean standard length (mm ± standard error) of gravid females significantly differed between sample waters, with the Atlantic Ocean population being the longest (189 ± 12.3 mm) and Lake Michigan population the shortest (122 ± 0.3 mm). Mean thiamine concentrations (nmol/g ± standard error) for single-year samples for Lake Huron, Lake Michigan, and Little Clear Pond (New York) were 9.9 ± 0.8, 3.9 ± 0.7, and 8.1 ± 2.3 nmol/g, respectively. Thiamine concentrations for multiple-year samples ranged from 1.1 to 15.6 for Lake Ontario, from 2.6 to 3.3 for Lake Erie, from 5.0 to 9.9 for Lake Superior, and from 10.9 to 13.3 for the Atlantic Ocean (Fore River). Although highly variable within populations and across years, thiamine concentrations in most spawning adults appeared to be adequate in all the waters for the years sampled except for 2006 and 2009 in Lake Ontario and 2009 in Lake Erie.

Keywords:

• rainbow smelt
• thiamine
• thiaminase

Introduction

During the twentieth century, introductions of exotic species have altered native fish communities in the Great Lakes. While exotic species introductions threaten the integrity of ecosystems worldwide, freshwater systems are among the most susceptible to these invasions (Mack et al. 2000; Mooney and Hobbs 2000; Sala et al. 2000; Lodge 2001). The rainbow smelt (Osmerus mordax), alewife (Alosa pseudoharengus), ruffe (Gymnocephalus cernuus), bloody red shrimp, (Hemimysis anomala), and dreissenid mussels (Dreissena bugensis and D. polymorpha) are all introduced species associated with major disruptions and the extirpation of several native species (the amphipod Diporeia spp., the Atlantic salmon Salmo salar). However, following establishment of self-sustaining populations, the rainbow smelt has become a prominent food for many top piscivores in the Great Lakes, including lake trout (Salvelinus namaycush), Atlantic salmon, brown trout (Salmo trutta), Chinook salmon (Oncorhynchus tshawytscha), coho salmon (Oncorhynchus kisutch), burbot, (Lota lota), lake whitefish (Coregonus clupeaformis), and walleye (Sander vitreus) (Brandt 1986; Oldenburg et al. 2007; Stapanian et al. 2007; Madenjian et al. 2008a). Because the rainbow smelt has thiaminase, an enzyme which destroys thiamine (vitamin B₁), this may contribute to reduced salmonine recruitment by causing early mortality syndrome in their fry (Fitzsimons 1995; Riley and Evans 2008).

During the early 2000's, rainbow smelt populations in the Great Lakes experienced severe reductions in recruitment (Gorman et al. 2010; Kocovsky et al. 2010; Madenjian et al. 2010; Schaeffer et al. 2010; Walsh et al. 2010). In 2008, the lowest population of rainbow smelt in 30 years was observed in Lake Ontario, raising concerns whether the population would be able to rebound (Mills et al. 2006; O’Gorman et al. 2008; Walsh et al. 2010). Reasons for the declines of rainbow smelt were not known but these authors suggested a possible increase in predator abundance and greater competition for food by several introduced species. Chalupnicki et al. (2010) proposed that consumption of alewives may explain the low thiamine content in rainbow smelt eggs causing fry mortality in Owasco Lake, New York.

Our hypothesis was that adult rainbow smelt may be affected by thiamine deficiency associated with the consumption of thiaminase-containing prey (i.e., young smelt and alewife). The objective of this study was to measure the thiamine concentration in the eggs of gravid rainbow smelt from the Great Lakes and two additional water bodies (Atlantic Ocean and Little Clear Pond near Saranac Lake, New York).

Methods

Gravid rainbow smelt were collected during spring trawl surveys conducted by the United States Geological Survey (USGS) in the Great Lakes in 2006–2009 (Figure 1). Two locations were selected to represent Lake Superior (Cloud Bay and Keweenaw Peninsula, Wisconsin) and one location each for Lake Michigan (Waukegan, Illinois), Lake Huron (Carp River, Michigan), Lake Erie (Fairport Harbor, Ohio), and Lake Ontario (Rochester, New York). For comparison, smelt were also collected from two additional waters outside the Great Lakes drainage during the spring spawning runs- Little Clear Pond near Saranac Lake, New York (a freshwater lake without alewives) and the Fore River, Weymouth, Massachusetts (representing the Atlantic Ocean). Total length was measured (mm), and the eggs were extracted from females and immediately frozen (−78°C) until analyzed. We analyzed 2–10 lots of eggs per lake, each representing 1–20 females/lot per year to better represent the mean of lake populations. Total thiamine concentrations in all samples were determined by the method of Brown et al. (1998) using fluorescence measured with a Perkin Elmer model 204 fluorescence spectrophotometer (excitation set at 375 nm and emission monitored at 433 nm).

Figure 1. Sites encompassing the Great Lakes and two additional waters (Little Clear Pond (LCP) and Fore River (FR)), where gravid rainbow smelt were captured and their eggs collected for thiamine analysis.

In statistical analyses of our data, we considered p < 0.05 to be significant. Because the Shapiro–Wilks test showed that data were normally distributed, we assessed the significance of differences by one-way analyses of variance (ANOVA) followed by Tukey's HSD all-pairwise comparison test using Statistix 8.0 Analytical Software (Tallahassee, Florida). We also used a paired t-test to determine the significance of differences between inland and ocean thiamine values with the same analytical software.

Results

Mean total length of rainbow smelt collected differed significantly (p < 0.001), with the Fore River samples being the longest (189 ±  12.3 mm) and the Lake Michigan population the shortest (122 ± 0.3 mm; Table 1). Three significantly different sized groups were observed in the following order; Fore River > Lake Erie, Little Clear Pond, Lake Ontario, Lake Huron > Lake Superior, Lake Michigan.

Table 1. Mean concentrations (nmol/g) of total thiamine (± SE) in the eggs of rainbow smelt from the Great Lakes, USA and two comparison groups (Little Clear Pond and Fore River). Numbers in parentheses (N/n): N is the number of samples analyzed; n the number of individual females whose eggs are represented in each sample. Length is the mean total length (mm) of all fish sampled. Values not followed by a common superscript significantly differ (p < 0.05).

Lake (N/n)	Length	2006	2007	2008	2009
Lake Ontario	141 ± 3.7^A	1.4 ± 0.2^ab(10/1)	15.6 ± 3^c(10/1)	11.3 ± 2.6^cde(2/16)	1.1 ± 0.2^b(3/8)
Lake Erie	143 ± 2.3^A	–	–	3.3 ± 0.5^abe(6/20)	2.6 ± 0.1^abe(3/10)
Lake Michigan	122 ± 0.3^B	–	3.9 ± 0.7^abe(5/1)	–	–
Lake Huron	133 ± 3.1^A	–	–	–	9.9 ± 0.8^acde(3/10)
Lake Superior	128 ± 4.2^B	–	5.0 ± 0.8^abde(7/2)	–	9.9 ± 2.8^acde(3/9)
Little Clear Pond	141 ± 4.3^A	–	–	8.1 ± 2.3^abcde(3/20)	–
Fore River	189 ± 12.3^C	–	–	13.3 ± 4.4^cd(5/1)	10.9 ± 1.1^cde(5/1)

Mean total thiamine concentrations (±standard error) in eggs of rainbow smelt from Lake Ontario varied greatly (1.1–15.6 nmol/g) and significantly (p < 0.05) between years (2006 versus 2007 and 2008 versus 2009). Mean total thiamine concentrations in eggs from Lake Erie smelt were low (3.3 and 2.6 nmol/g, respectively) and not significantly different for years 2008 and 2009. Concentrations in eggs (5.0 and 9.9 nmol/g, respectively) from Lake Superior did not significantly differ during 2007 and 2009. The mean thiamine concentration in eggs from Lake Michigan rainbow smelt in 2007 (3.9 nmol/g) was low and not significantly different from the lowest values for Lake Ontario (1.1 and 1.4 nmol/g). The mean concentration of total thiamine for eggs from Little Clear Pond (in 2008) was intermediate (8.1 nmol/g) and not significantly different from any other values. The mean concentration of total thiamine for eggs from Lake Huron (in 2009) was intermediate (9.9 nmol/g) and not significantly different from any other values except from the lowest value (1.1 nmol/g) for Lake Ontario (Table 1). The mean concentrations of total thiamine for eggs from Lake Superior (2007 and 2009) were intermediate (5.0 and 9.9 nmol/g) and were not significantly different from other values.

The two-year mean of total thiamine concentration (12.1 ± 1.1 nmol/g) in rainbow smelt eggs from the Atlantic Ocean (Fore River) was significantly higher (p < 0.001, t  = −6.00, df = 9) than the overall mean (5.5 ± 0.6 nmol/g) for all inland lakes (omitting Lake Ontario due to high variability). Rainbow smelt eggs from Lake Ontario had both the highest and lowest numerical values observed.

Discussion

Although our data are limited, it appears that eggs of land-locked freshwater rainbow smelt varied more in thiamine content than those from the ocean suggesting that variable thiamine status is related to instability of rainbow smelt populations. For instance, rainbow smelt populations in the Great Lakes have historically oscillated in abundance and recruitment success since their introduction in the early 1900s (Van Oosten 1937; Smith 1985; O’Gorman et al. 1987; Henderson and Nepszey 1989). Current rainbow smelt populations are low, especially in Lake Ontario, which has experienced the lowest recorded abundance in the past 30 years. Differences in gravid rainbow smelt lengths can be attributed to predator–prey interactions and difference in the forage base across systems. Following their introduction, rainbow smelt and alewife have replaced native forage species such as cisco (Coregonus artedi) and bloater (Coregonus hoyi) as the primary food source for salmonids in the Great Lakes (Christie 1974). The upper Great Lakes (Superior and Michigan) have healthy salmon and trout populations that have depressed rainbow smelt populations where the reproductive year-class has become smaller and younger. Although the lower Great Lakes have healthy salmon populations, the lake trout is still recovering and requires less forage, which has permitted extra growth in reproductive adult smelt. In Little Clear Pond (≈162 ha), rainbow smelt is subjected to Atlantic salmon predation pressures and has no alewife to compete with for food. Although rainbow smelt has also existed longer in the Atlantic Ocean than in the Great Lakes, spawning runs are recently very low for reasons unknown (Chase 2006, 2009a, 2009b).

Differences in thiamine content for rainbow smelt within lakes across years indicate that there may be different forage and/or feeding patterns which produce these variations. Foltz and Norden (1977) and Janssen and Brandt (1980) both noted that within-lake thiamine variation is a common occurrence among rainbow smelt and alewife populations. Despite all-time low abundance records for the Lake Ontario population, minor increases have been observed since the low in 2008 (Walsh et al. 2010).

The diet of rainbow smelt from the Great Lakes has been reported to include fish as a major food source. Gordon (1961) reported up to 20% of the diet of smelt from Lake Huron contained young smelt and Notropis spp. Johnson et al. (2004) observed that up to 3.5% of the diet of smelt from Lake Superior was comprised of small fish. Foltz and Norden (1977), Stedman and Argyle (1985), and Davis et al. (1998) observed young alewife as the main food source (up to 60%) for adult rainbow smelt from Lake Michigan. Pothoven et al. (2009) described the diet of rainbow smelt from Lake Erie to include up to 48% emerald shiner (Notropis atherinoides). Walsh et al. (2008) and Brandt and Madon (1986) both observed up to 77% of rainbow smelt diet contents to include young rainbow smelt and slimy sculpin (Cottus cognatus). Fish become more common in the diet of larger and older rainbow smelt as they become more piscivorous and pelagic where benthic invertebrates are scarce and do not fulfill nutritional requirements.

Although thiamine deficiency has not been reported in rainbow smelt populations in the Great Lakes, Chalupnicki et al. (2010) reported that survival of fry hatched from eggs from rainbow smelt captured in Owasco Lake, New York was significantly increased by treatment with thiamine. Therefore, one possible cause of fluctuations of rainbow smelt population may be from consumption of thiaminase-rich forage fishes. Alewife, rainbow smelt, and Notropis spp. all contain moderate to high activity levels (2640 to 32,700 pmol/g min) of thiaminase (Tillitt et al. 2005), which has been linked to early mortality syndrome in fry of Great Lakes salmonines (Honeyfield et al. 2005; Madenjian et al. 2008a, 2008b). Wolf (1942) reported a thiamine deficiency in brook trout induced by feeding them emerald shiners, also high in thiaminase activity. Despite thiamine variations in rainbow smelt populations from the Great Lakes, most values appear to be adequate for successful reproduction except for some years in Lakes Ontario (years 2006 and 2009) and Erie (year 2009). Chalupnicki et al. (2010) demonstrated that survival of fry hatched from eggs from rainbow smelt captured in Owasco Lake having a mean thiamine content of 2.7 nmol/g egg (±SE = 0.3) was significantly increased by immersion in thiamine. Though not previously reported, the concentration of total thiamine in the rainbow smelt eggs, resulting in thiamine-responsive mortality in their fry, ranged from 2.1 to 3.0 nmol/g (Marc Chalupnicki et al. 2010, USGS unpublished data). Therefore values below 3.0 nmol/g thiamine in smelt eggs may be deficient. Based on the findings of Tillitt et al. (2005), which showed large annual fluctuations in thiamine content of whole bodies of rainbow smelt in Lake Michigan, we suspect that in some years the eggs will become thiamine-deficient, which will result in poor survival and recruitment.

Acknowledgments

We gratefully acknowledge Ed Roseman (USGS Great Lakes Science Center), Bill Edwards (USGS Lake Erie Biological Station), Kevin Kayle (Ohio Department of Natural Resources), Dawn Dittman (USGS Tunison Lab of Aquatic Science), Maureen Walsh (USGS Lake Ontario Biological Station), Gary Cholwek (USGS Lake Superior Biological Station), Chuck Madenjian (USGS Great Lakes Science Center), Ed Grant (New York Department of Environmental Conservation Adirondak Fish Hatchery), and Brad Chase and Matt Ayer (Massachusetts Marine Fisheries Division) who helped with sample collection. We also appreciate the reviewers who helped improve this article. This article is Contribution 1652 of the USGS Great Lakes Science Center.