Aquaculture 310 (2010) 130138

Contents lists available at ScienceDirect

Aquaculture
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a q u a - o n l i n e

Characterization of microbial communities in minimal-exchange, intensive aquaculture systems and the effects of suspended solids management
Andrew J. Ray a,, Gloria Seaborn b, John W. Lefer c, Susan B. Wilde d, Alisha Lawson a, Craig L. Browdy e
a

Waddell Mariculture Center, South Carolina Department of Natural Resources, 211 Sawmill Creek Rd., Bluffton, SC 29910, USA Center for Coastal Environmental Health and Biomolecular Research, NOAA, 219 Fort Johnson Rd., Charleston, SC 29412, USA Marine Resources Research Institute, South Carolina Department of Natural Resources, 217 Fort Johnson Rd., Charleston, SC 29412, USA d School of Forestry and Natural Resources, The University of Georgia, Athens, GA 30602, USA e Novus International, 5 Tomotley Ct., Charleston, SC 29407, USA
b c

a r t i c l e

i n f o

a b s t r a c t
Minimal-exchange, intensive culture systems require little, if any, water exchange and have high animal stocking densities. Intensive nutrient inputs lead to an abundant community of microorganisms. These microbes are partially contained within suspended biooc particles and contribute to water quality maintenance and provision of supplemental nutrition to the culture species. Optimal function of minimal-exchange, intensive systems is likely dependent on the structure of the microbial communities within them. This document offers a short review of microbial groups important for intensive marine aquaculture and descriptions of three methods for quantifying their abundance. The document also describes an experiment during which these methods were used to monitor the effects of partial biooc removal on microbe abundance. The rst method uses light microscopy, with the option of epiuorescence, along with a ranking system to enumerate the abundance of microbial taxa. The second method exclusively uses epiuorescence to illuminate chlorophyll and cyanobacteria pigments. Images are taken of each uorescing group of pigments and processed using image analysis software to quantify the respective abundance of the two pigment types. Using the third method, changes in bacterial abundance were determined by gas chromatographic measurement of bacteria-specic fatty acids in solvent extracted water column lipids. Using these techniques, it was determined that removing solids from the culture water signicantly (P 0.01) reduced the abundance of nematodes, rotifers, cyanobacteria, and bacteria. Understanding microbial composition and the effects that management protocols have on that composition may help system managers make better informed decisions. 2010 Elsevier B.V. All rights reserved.

Article history: Received 25 April 2010 Received in revised form 8 October 2010 Accepted 13 October 2010 Keywords: Biooc Shrimp Biomarkers Epiuorescence Intensive aquaculture Microbial ecology

1. Introduction 1.1. Minimal-exchange, intensive systems Minimal-exchange, intensive aquaculture systems offer an environmentally attractive means of shrimp and sh production, allowing for high density culture and little or no water exchange. With high animal density comes intensive nutrient input in the form of feed. When water is not exchanged with surrounding water bodies, expensive nutrients from feed are retained within the system. In minimal-exchange, intensive systems excess nutrients are assimilated and mineralized by a dense microbial community in the water column, thus alleviating potential toxicity (Avnimelech, 2006; Bratvold and Browdy, 1998; Ebeling et al., 2006; Ray et al., 2009). The microorganisms not only remove excess nutrients, but have been implicated in nutritional provision for animals, including shrimp and
Corresponding author. Gulf Coast Research Laboratory, 703 East Beach Dr., Ocean Springs, MS 39564, USA. Tel.: + 1 843 367 9407. E-mail address: AndrewJRay@gmail.com (A.J. Ray). 0044-8486/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.aquaculture.2010.10.019

tilapia, that can result in improved growth rate, feed conversion ratio (FCR), and weight gain (Azim and Little, 2008; Burford et al., 2004; Moss and Pruder, 1995; Wasielesky et al., 2006). 1.2. Microorganism groups Important microorganism groups in minimal-exchange, intensive aquaculture systems include algae, zooplankton, and bacteria. Algae utilize toxic total ammonia nitrogen (TAN), as well as less dangerous nitrate-nitrogen and phosphate compounds to construct cellular structures such as proteins and sugars. Various forms of algae, most notably diatoms, are nutritious and can benet shrimp production by contributing qualities such as essential amino acids and highly unsaturated fatty acids (Ju et al., 2009; Moss et al., 2001). However, potentially harmful algae can also be found in aquaculture systems. One group that has been problematic for shrimp culture is cyanobacteria, also known as blue-green algae. Some cyanobacteria are capable of producing toxins that may diminish shrimp growth or directly cause mortality (Alonso-Rodriguez and Paez-Osuna, 2003; Zimba et al., 2006).

A.J. Ray et al. / Aquaculture 310 (2010) 130138

131

Zooplankton consume algae and bacteria and can play an important role in the transfer of nutrients from primary producers to secondary consumers. Zooplankton such as rotifers can contribute signicantly to the protein and energy requirements of shrimp (Focken et al., 1998). However, zooplankton also consume oxygen and can cause a reduction in alkalinity through respiration. Two functional categories of bacteria are primarily responsible for water quality maintenance in minimal-exchange, intensive systems: heterotrophic ammonia-assimilative and chemoautotrophic nitrifying bacteria (Ebeling et al., 2006; Hargreaves, 2006). The heterotrophic group removes TAN from the water column to build cellular proteins. Nitrifying bacteria acquire energy through the oxidation reactions of TAN to nitrite-N and nitrite-N to nitrate-N, the latter of which is much less toxic than its predecessors in this sequence. Both the bacterial assimilation and nitrication processes consume oxygen and reduce alkalinity, thereby often requiring the supplementation of those two components. A portion of the microbial community in minimal-exchange, intensive culture systems is contained on or within particles, commonly called biooc (Fig. 1A). Biooc is composed of a variety of microorganisms, uneaten feed, feces, and detritus, and the particles are kept in suspension with water propulsion and aeration. Biooc offers numerous ecological advantages for microbes, including protection from predators, direct access to nutrients, and necessary substrate area (De Schryver et al., 2008). Biooc is thought to provide a packaging of microbial proteins and nutrients that is directly accessible to culture animals (Avnimelech, 2009; Burford et al., 2004). Biooc technology (BFT) can be considered a culture technique in which water quality is maintained and in situ animal feed is simultaneously produced in the form of biooc particles (Crab et al., 2007). Although biooc can be benecial, some level of control over the concentration of particles is likely required for optimal system performance. Ray et al. (2010) demonstrated that shrimp biomass production (kg m 3) was increased 41% when biooc concentration was managed through the use of external settling chambers. They also showed a 60% reduction in nitrate-nitrogen concentration and a 61% reduction in phosphate concentration when biooc concentration was managed. To optimize system function, a benecial microbial community should be developed and sustained. This requires knowledge of microbial community composition and the ability to monitor changes. By investigating microbial composition, inferences may be made of the underlying reasons for differences in community structure and the effects of management techniques, such as biooc removal, on that structure. Monitoring of microorganisms may help characterize opportunities for supplemental nutrition, and help managers recognize the occurrence of problematic organisms, thereby guiding informed management decisions. The purpose of this document is to describe three techniques that can be used to characterize microbial communities in minimalexchange, intensive aquaculture systems. To illustrate these techniques in a practical application they were used in an experiment that explored the effects of removing suspended solids (biooc) on microorganism abundance. 1.3. Visual microscopy abundance quantication

categories important and common to intensive marine aquaculture systems: chlorophytes (green algae), diatoms, dinoagellates, nematodes, rotifers, and cyanobacteria. The abundance of organisms belonging to these categories was assessed using light and uorescence microscopy with live samples. The categories used were selected based on the authors' experience with shrimp culture systems at the Waddell Mariculture Center (WMC) in Bluffton, South Carolina, USA. Other categories may be needed, depending on location and endemic biota. Broad classications such as these save time, minimize the amount of taxonomical training required by the analyst, and can be less susceptible to biases (DeVantier et al., 1998). This technique can be performed with an ordinary compound light microscope, although epiuorescence technology is advantageous. Many of the cyanobacteria and algal cells in intensive culture systems are small and become concealed by particles in the water, making visualizing and discerning between these two groups difcult. Epiuorescence can illuminate the chlorophyll-a pigment, contained by both eukaryotic algae and cyanobacteria, and can uoresce the cyanobacteria pigment phycocyanin exclusively. This facilitates ease of observation and differentiation between eukaryotic algae and cyanobacteria. Kastovska et al. (2007) described using epiuorescence in this manner to discern between the two groups. Epiuorescence functions by powering a broad-spectra mercury light bulb; lters are then used to select the specic wavelengths of light that cause chlorophyll pigments, cyanobacteria pigments, or stains such as 46-diamidino-2-phenylindole (DAPI), a DNA stain, to uoresce. An advantage that epiuorescence has over other pigment analyses is that many types of lters are available to discern a variety of biological components, not limited to pigments. An inverted compound microscope allows the observer to look up, through a larger sample than would be possible with an ordinary compound microscope. This technology allows the microorganisms in a relatively large sample (3 mL in this case) to settle down onto one plane. The three-dimensional biooc particles that are common in intensive culture systems are better visualized using an inverted scope because the sample is not compressed onto a at slide. 1.4. Epiuorescence with image analysis quantication This procedure utilizes the epiuorescence technology described above, except that rather than using live samples, images of uorescing chlorophyll (chlorophyll-a) and cyanobacteria pigment (phycocyanin) are captured using preserved samples on at slides. An image of uorescing chlorophyll is taken, and an image of cyanobacteria pigment is then taken in the same location on the slide. This enables a direct comparison between chlorophyll (including cyanobacteria) and cyanobacteria pigment exclusively. The details of wavelengths used for pigment excitation and uorescence emission are given in Section 2.1. To determine the relative abundance of uorescing pigments in each image, the pictures are processed using the computer program ImageJ. Researchers have used this technique in other applications. Selinummi et al. (2005) processed DAPI-stained images with image analysis software to quantify bacterial cells. Verity and Sieracki (1993) outlined the general procedures for measuring plankton biomass using this methodology. 1.5. Bacterial fatty acid assessment by gas chromatography

In this case, abundance quantication describes a system of ranking the relative abundance of organismal groups. Organisms are categorized and ranked according to a predetermined scale, similar to the previous work of researchers. Newall et al. (2006) used such a scale to rank the abundance of diatoms in a river system. DeVantier et al. (1998) described trained researchers who used a subjective scale to rank macro-invertebrate abundance. The current document explores the use of an abundance quantication system to describe the relative abundance of six microorganism

In this method, gas chromatography with mass spectrometry (GCMS) is used to separate, identify, and quantify fatty acids (FAs) derived from the lipid component obtained by solvent extraction of biooc. The sum of odd and branched chain fatty acids thus obtained serves as a measure of the relative abundance of bacteria. Odd and branched chain fatty acids, produced primarily by both aerobic and anaerobic bacteria, are widely accepted as biomarkers for bacteria and sums are frequently used to estimate bacterial contributions; examples

132

A.J. Ray et al. / Aquaculture 310 (2010) 130138

Fig. 1. Common components of minimal-exchange, intensive culture systems. (A) A biooc particle, which contains some of the microbial components in minimal-exchange, intensive culture systems; some evidence suggests that adult shrimp may be able to consume these particles. (B) A high abundance of chlorophytes, also called green algae, (C) three types of diatoms commonly observed at the Waddell Mariculture Center, (D) a heterotrophic dinoagellate has recently consumed green algae, (E) a nematode, (F) a rotifer. Images B through F are organisms used for the visual microscopy quantication method. The cyanobacteria quantied for that method were best visualized using epiuorescence (Fig. 2).

include Alfaro et al. (2006), Budge et al. (2001), Kastovska et al. (2007), Parrish et al. (2000), and Harvey and Macko (1997). Because these specic fatty acids represent only a portion of the total FAs found in the bacteria prole and the major FAs are those commonly found in other organisms, these sums are not necessarily related to absolute bacteria biomass. However, relative differences in the proportion of bacteria

from water samples can be readily assessed. Because the entire fatty acid prole for the biooc is obtained with this method, the potential availability of various fatty acids for animal nutrition may also be assessed. For instance, the concentration of eicosapentaenoic acid (EPA), an essential fatty acid, can be obtained simultaneously while evaluating bacterial abundance.

A.J. Ray et al. / Aquaculture 310 (2010) 130138

133

2. Materials and methods 2.1. Visual microscopy abundance quantication All microscopy was performed with an Olympus IX-51 Inverted Microscope (Olympus America Incorporated, Center Valley, Pennsylvania, USA) outtted with epiuorescence technology. It was equipped with a lter which enabled chlorophyll-a excitation at 480 nm and uorescence emission at 660 nm. The microscope also had a lter enabling excitation of the cyanobacterial pigment phycocyanin at 570 nm and uorescence emission at 620 nm. The chlorophyll and cyanobacteria pigment lters were used extensively to distinguish between eukaryotic algae and cyanobacteria, to illuminate minute cells, and to visualize cells that otherwise would be hidden by particles in the water, they were also used for the method described in Section 2.2 of this document. Three mL of thoroughly mixed sample water was placed in one 5 mL well of a two square well Chamber Slide (Lab TekNalge Nunc International, Naperville, Illinois, USA). Each sample was carefully examined by moving the stage of the microscope from the bottom of the well to the top along the y-axis, then horizontally across the x-axis and back down from the top of the well to the bottom along the y-axis. The well was then examined from side to side across the x-axis and one corner was inspected. The microscope stage was moved along the described course and randomly stopped at ten unique elds of view to screen for organisms in each sample. The observer helped to ensure randomness in choice of elds of view by not looking through the microscope oculars when moving the slide. During visual microscopy abundance quantication, six groups of organisms were recorded: chlorophytes, diatoms, dinoagellates, nematodes, rotifers, and cyanobacteria (Figs. 1B through E and 2B). The presence or absence of each group of organisms was noted and all groups were ranked numerically, relative to their abundance following the quantication system presented in Table 1. 2.2. Epiuorescence with image analysis quantication Epiuorescence and image analysis software were used to quantify the abundance of chlorophyll pigment and cyanobacteria pigment. Samples were preserved in 2% gluteraldehyde solution (nal concentration). The gluteraldehyde xed samples were stored in a 10 C refrigerator to be examined at a later date. A dilution was created by combining 2.0 mL preserved sample water with 3.0 mL of sterile seawater. The solution was then mixed using a vortex mixer (Fisher Scientic, Pittsburg, Pennsylvania, USA) for 30 s to

Table 1 Scoring system used for visual microscopy quantication. Each of the six categories of microorganisms was assigned a numerical score as a method of quantifying their observed abundance. Abundance Absent Rare Moderate Common Abundant Numerical score 0 1 2 3 4 Description No organism observed As few as one organism observed, as abundant as being present in 4 elds of view Present in between 5 and 9 elds of view Present in every eld of view, occupies up to 50% of each Occupies over 50% of every eld of view

break up biooc particles. The diluted sample was ltered using a 25 mm diameter, black 0.22 m pore-size, Poretics polycarbonate lter (GE Osmonics, Minnetonka, Minnesota, USA) placed atop a 0.7 m poresize GF/F glass microber lter (Whatman, Maidstone, Kent, United Kingdom). The GF/F lter acted as a cushion to protect the polycarbonate lter, upon which ltered material collected. The polycarbonate lter was transferred to a pre-cleaned 1.0 mm thick glass microscope slide (VWR International, West Chester, Pennsylvania, USA). Several drops of Type FF immersion oil (Cargille Laboratories, Cedar Grove, New Jersey, USA) were applied between the slide, lter, and a cover slip to prevent desiccation. Photo-micrographs were taken of the prepared slides using the Olympus IX51 inverted microscope with an attached four megapixel digital camera (Qimaging, Surrey, British Columbia, Canada) connected to a desktop computer (Dell Corporation, Round Rock, Texas, USA). A slide was placed on the microscope stage and the chlorophyll (chlorophyll-a) uorescence lter was engaged. Without looking at the image created, the stage was moved so that light penetrated the lter at approximately the bottom center point of ltered material and a photograph was taken (Fig. 2A). After taking an image with the chlorophyll lter engaged, the cyanobacteria (phycocyanin) lter was engaged without moving the microscope stage and another image was taken (Fig. 2B) so that two images, one of each uorescence type was taken in a single location. Both epiuorescence lters are described in more detail in Section 2.1 of this document. The stage was then moved without looking at the image so that the slide moved upward, along the y-axis, into a unique eld of view, and the same set of two different uorescence images was captured. This was performed ten times so that a total of ten chlorophyll and ten cyanopigment images were created per slide. Picture quality adjustments were made as needed such as focus,

Fig. 2. Two images taken in the same location on a slide using different epiuorescence wavelengths. Image A was taken using the chlorophyll wavelength, which causes chlorophyll-a, contained in both eukaryotic algae and cyanobacteria, to uoresce and image B was taken using the cyanobacteria wavelength, which causes only the cyanobacteria pigment phycocyanin to uoresce. These images were assessed using image analysis software that allowed quantication of the amount of uorescing material in each image, reported in uorescing pixels mL 1. In this particular example, there is a relatively high abundance of cyanobacteria, a sign of potentially poor culture conditions.

134

A.J. Ray et al. / Aquaculture 310 (2010) 130138

exposure time, and brightness so that the clearest possible image was obtained. Images were imported into ImageJ (Image Processing and Analysis in JAVA, National Institutes of Health, Bethesda, Maryland, USA) as .TIFF (Tagged Image File Format) les. Each picture was converted into an eight bit, grayscale image, and then inverted. Contrast was enhanced by normalizing the image to further improve image quality and maintain continuity between images. The Entropy Threshold tool was applied to help resolve the amount of material that was counted as uorescing. Finally, the option analyze particles was turned on, which generated the number of pixels in each image that were uorescing. The mean number of uorescing pixels in the ten images of each uorescence type on each slide was determined. Based on the number of pixels uorescing, the number of pixels available in each image, and the total volume of sample ltered, these numbers were converted to uorescing pixels mL 1. To determine the abundance of photosynthetic pigments excluding cyanobacteria (eukaryotic algae only), cyanobacteria uorescence was subtracted from the chlorophyll uorescence value. 2.3. Bacterial fatty acid assessment by GC Fifty mL of sample water was ltered through 47 mm diameter, 0.7 m pore-size, GF/F Whatman, borosilicate, glass microber lters (Sigma-Aldrich Incorporated, St. Louis, Missouri, USA). Filters were placed into a centrifuge tube containing 7 mL of chloroform and immediately stored in a 20 C freezer for subsequent analysis. An alternative to ltering that has proven successful is to centrifuge samples and collect the pellet. Lipids were extracted with a 2:1 ratio of chloroformmethanol (Folch et al., 1957) as described by Budge and Parrish (2003) with modication to accommodate sample size and the use of an internal standard (IS). Briey, just prior to analysis, samples were allowed to reach room temperature. After addition of 3.5 mL methanol and C23:0 methyl ester (ME) as an IS, the samples were vortexed until sample material appeared to be fully dispersed, then sonicated for 10 min. Aqueous sodium chloride (0.73%) was added with volume calculated to achieve a chloroformmethanolwater ratio of 8:4:3 as specied by Folch et al. (1957), and the contents were thoroughly mixed and centrifuged. The lower chloroform layer was removed, the aqueous layer washed with 3 2 mL of chloroform and chloroform extracts combined. The chloroform was dried over anhydrous sodium sulfate, ltered into a clean, dry tube and the chloroform removed under a gentle stream of dry N2. The residue was dissolved in 1 mL hexane and fatty acid methyl esters (FAMEs) prepared using BF3 (10% in methanol, Supelco) as described by Metcalfe et al. (1966). The hexane layer containing the FAMEs was concentrated for instrumental analysis. The FAMEs were analyzed by gas chromatography with splitless injection. Separation was achieved with oven temperature programming as follows: Initial temperature 50 C with a 2 min hold, ramped at 20 min 1 to 150 C followed by a 1.25min 1 ramp to 220 C. With the use of dual injection modules, samples were simultaneously analyzed on dual DB225ms columns (50%-cyanopropylphenyl-methylpolysiloxane 30 m 0.25 mm, J&W Scientic, Folsom, California, USA) with detection by ame ionization (FID) and mass spectrometry (MS). Scans from the MS detector were used in conjunction with comparison of retention times to those of known standards for peak identication. Correction factors (determined by analysis of quantitative standards, GLC 85 and GLC 411 (Nu-Chek Prep, Elysian, Minnesota, USA)), and verication against theoretical correction factors were applied to fatty acid peak areas from the FID. The amount of each fatty acid was calculated using IS methodology and reported as g analyte L 1 of water. The sum of the branched and odd chain fatty acids (iso 13:0, anteiso 13:0, C13:0, iso 14:0, anteiso 14:0, iso 15:0, anteiso 15:0, C15:0, C15:1, iso 16:0, iso 18:0, iso 18:0, and C19:1) was used as a measure of the levels of bacteria in the

samples. Because the C17 fatty acids (C17:0 and C17:1n-8) are commonly found in a variety of organisms they were omitted from the sum. 2.4. Practical method application 2.4.1. Experimental setting and design An experiment that assessed the effects of suspended solids management on microbial communities in minimal-exchange, superintensive shrimp culture systems illustrates the employment of the described microbial characterization techniques. The project was carried out using 32 outdoor, polyethylene, 3.5 m diameter tanks, lled with 6.3 m3 of ltered seawater at approximately 20 ppt salinity, receiving blown air delivered through ceramic air stones. The experimental system was the same as that described by Ray et al. (2010) except that more tanks were used. The settling chambers described by Ray et al. (2010) were used to remove suspended solids from half of the experimental tanks, the other half received no solids removal. Water was airlifted to the settling chambers, where turbulence slowed and solids were allowed to settle, water near the surface of the settling chambers owed back into shrimp culture tanks. The suspended solids removed are synonymous with biooc particles. Within each treatment (solids removal versus no solids removal), half of the tanks were stocked at approximately 140 shrimp m 3 and the other half were stocked at approximately 420 shrimp m 3. The shrimp in this experiment were Litopenaeus vannamei with a mean SE initial weight of 1.16 0.01 g. For the purpose of this document, data from both shrimp stocking densities were combined and only comparisons between treatments with and without solids removal are made. 2.4.2. Water quality and microbial analyses Temperature, dissolved oxygen, pH, and salinity were each measured twice per day at approximately 0830 and 1600 h in each tank using a YSI Model 556 (YSI Incorporated, Yellow Springs, Ohio, USA). Total ammonia nitrogen (TAN), nitrite nitrogen (NO2-N), nitrate nitrogen (NO3-N), orthophosphate (PO4), alkalinity (as CaCO3), total suspended solids (TSS), volatile suspended solids (VSS), and turbidity were each measured once per week in each tank. Aliquots for these eight water quality parameters and for the microbial analyses were taken from 500 mL samples collected just below the water surface in each tank. The three nitrogen species were measured using Hach Methods 8155, 8507, and 8039, respectively (Hach Company, 2003). Dilutions were made as necessary due to some high nitrogen concentrations and samples were read in a Hach DR/4000 V Spectrophotometer (Hach Company, Loveland, Colorado, USA). Absorbance was measured at 655 nm, 507 nm, and 500 nm for TAN, NO2-N, and NO3-N, respectively, and concentrations were determined using standard curves derived from the analysis of known standards. Because of the variability that can be associated with these tests, two replicate samples were assessed for TAN and NO2-N, and three replicates were used for NO3-N. Orthophosphate concentration was measured using the PhosVer 3 (ascorbic acid) method outlined in Hach Method 8048 (Hach Company, 2003). Absorbance was measured at 890 nm using the Hach DR/4000 V Spectrophotometer. Dilutions were made for orthophosphate analysis as necessary and two replicates were used for each test. Known standards were analyzed with each sample set during nitrogen and orthophosphate analyses to monitor accuracy. Alkalinity was measured following the Potentiometric Titration to Preselected pH procedure, outlined in section 2320 B by the APHA (1989). Sodium bicarbonate was added as needed to maintain alkalinity above 100 mg CaCO3 L 1 throughout the study. TSS and VSS were measured following ESS Method 340.2 (ESS, 1993). Turbidity was measured using a LaMotte model 2020e Turbidity Meter (LaMotte Company, Chestertown, Maryland, USA).

A.J. Ray et al. / Aquaculture 310 (2010) 130138

135

At week four of this 13 week study, settling chamber operation and sampling for microbial abundance data was initiated. Samples were collected and analyzed weekly, except the bacterial fatty acid assessment, which was conducted at 3-week intervals. 2.4.3. Statistical analyses The data collected were analyzed using SigmaPlot Version 11 (Systat Software, Incorporated, San Jose, California, USA). One-way Repeated Measures (RM) ANOVAs were used to analyze the microbial abundance data. However, because there were only four samples of bacterial indicators, the power of the RM ANOVA for this data set was weak. As a result, and because the data did not have equal variance, a MannWhitney Rank Sum Test was used to compare the bacterial data from only the nal sample date between tanks with and without solids removal. A Pearson Product Moment Correlation Test was used to test whether cyanobacteria abundance data collected using visual microscopy and those collected using epiuorescence were correlated. 3. Results and discussion By the end of this experiment, tanks with suspended solids removal had a mean total suspended solids (TSS) concentration 28% lower than tanks without suspended solids removal (567 mg L 1 versus 785 mg L 1, respectively). Overall water quality conditions during the experiment are presented in Table 2. 3.1. Visual microscopy abundance quantication The results of this technique indicate that systems with solids management had signicantly lower abundances of nematodes (P=0.010), rotifers (P=0.004), and cyanobacteria (P=0.004) compared to those without solids removal. By the last sample period of the study, nematode abundance had been reduced by 60.0%, rotifer abundance by 18.6%, and cyanobacteria abundance had been reduced by 22.7% through the use of settling chambers. The abundance of the other groups assessed with visual microscopy: chlorophytes, diatoms, and dinoagellates, were not signicantly affected by solids removal. The abundance of organisms over time is presented in Fig. 3. Table 3 gives the overall mean, standard error, and range of the abundance of each group throughout the study, with the exception of bacterial abundance for which only data from the last sample date is reported. Chlorophytes were highly abundant throughout the study (Fig. 3a), regardless of whether solids were managed, and most belonged to the genus Oocystis sp. Based on visual observations, these organisms appeared to be buoyant, perhaps preventing the cells from settling in the settling chambers. Moreno-Ostos et al. (2008) observed that Oocystis sp. were neutrally buoyant, which prevented settling. Diatoms were present in low abundance near the beginning of the study and became less common as the experiment progressed, regardless of whether solids were removed (Fig. 3b). Shrimp may have consumed a portion of the diatom community; previous authors have shown that these nutritious algae are an important food item for shrimp (Moss et al., 2001; Sullivan and Moncreiff, 1990). Some competition for resources may have existed between chlorophytes and diatoms, with chlorophytes being favored in these experimental systems. Dinoagellates had a generally low relative abundance throughout the study, regardless of whether solids were removed (Fig. 3c). These organisms are free-swimming which may have prevented them from settling in the settling chambers, potentially helping to explain why their abundance was not different between the treatments. The organisms that were reduced in abundance by removing solids appeared to be closely associated with the biooc particles. Nematodes (Fig. 3d) and rotifers (Fig. 3e) were seen almost exclusively grazing on and within the particles and cyanobacteria were also principally located within biooc particles. Both nematodes and rotifers consume oxygen and reducing their abundance may consequentially reduce biochemical

Table 2 Overall water quality throughout the shrimp culture period. Data are presented as mean standard error (range). Nitrate, orthophosphate, total suspended solids (TSS), volatile suspended solids (VSS), and turbidity were all generally lower, and alkalinity was generally higher in the solids removal treatment; Ray et al. (2010) showed that settling chambers can signicantly affect these parameters. Treatment Solids removal Temperature (C) AM 27.0 0.4 (22.230.5) PM 28.2 0.4 (23.531.7) Dissolved oxygen (mg L 1) AM 6.0 0.2 (3.89.2) PM 6.8 0.2 (4.611.6) pH AM 7.5 0.1 (6.38.2) PM 7.9 0.1 (6.48.6) Salinity (g L 1) AM 20.3 0.7 (13.925.4) PM 20.4 0.7 (13.925.8) 1 Ammonia (mg TAN L ) 0.4 0.2 (0.04.7) Nitrite (mg NO2-N L 1) 0.5 0.2 (0.04.8) 1 Nitrate (mg NO3-N L ) 38.6 9.4 (0.4266.9) Orthophosphate 23.5 2.4 (5.950.4) (mg PO4 L 1) 135.3 11.6 (40.0360.0) Alkalinity (mg CaCO3 L 1) 1 397.0 42.5 (60.01065.0) TSS (mg L ) VSS (mg L 1) 270.1 24.5 (43.3573.3) Turbidity (NTU) 39.9 4.7 (5.2104.1) No solids removal 26.8 0.4 (21.330.3) 28.1 0.4 (22.631.7) 6.1 0.2 (3.49.3) 6.6 0.2 (4.710.5) 7.4 0.1 (6.28.2) 7.7 0.1 (6.48.6) 19.7 0.7 (13.724.9) 19.7 0.7 (13.724.8) 0.3 0.1 (0.04.5) 0.3 0.2 (0.06.2) 53.0 12.6 (0.7293.3) 40.7 6.4 (10.2115.8) 127.5 11.1 (40.0250.0) 616.0 57.9 (51.71215.0) 427.7 44.9 (26.7880.0) 69.8 5.1 (6.7111.6)

oxygen demand, leaving greater oxygen resources available for the target crop. They also ingest algae and bacteria, suggesting that a change in the abundance of these primary consumers could have a top-down trophic effect. However, the two zooplankton groups are thought to be nutritious prey for shrimp. In terms of potential nutrition, a decrease in the abundance of these zooplankton may not be desirable. The fact that cyanobacteria (Fig. 3f) were removed with solids management is an encouraging reason to manage solids concentration. Cyanobacteria can hinder aquaculture efforts by producing lethal toxins and contributing to off-avors of animal esh (Alonso-Rodriguez and Paez-Osuna, 2003; Zimba et al., 2006). The visual microscopy abundance quantication method is a relatively simple technique. An analyst can be taught to distinguish between broad groupings of organisms, such as the ones described here, with minimal training. Visual microscopy may also be used to perform algal and zooplankton counts. This is achieved by counting the number of organisms in a given area of a specially designed microscope slide or hemacytometer, then extrapolating to determine the number of organisms per volumetric unit. However, in limited exchange intensive culture systems, microorganisms are more abundant than in typical, natural aquatic systems as a result of hypereutrophic conditions. If assessments of multiple culture units at a shrimp production facility or research center are needed, routinely counting microorganisms in these systems would likely be unreasonably time consuming. It is best to use live samples for visual microscopy quantication as most preservatives will damage organisms such as dinoagellates, potentially rendering them unrecognizable. Using live samples allows real-time assessment of the microbial community. Routinely performing this procedure can alert the analyst to shifts in community structure such as an algae crash or dinoagellate bloom, either of which may have negative impacts on shrimp production. Recognizing that such situations are the cause of diminished shrimp growth can save system managers from making the mistake of over-feeding shrimp to compensate for slow growth (Ray et al., 2009). Advantageous events can also be detected, such as a bloom of diatoms, during which time enhanced shrimp growth may be expected. Knowing the

136

g
Rotifer Relative Abundance Dinoflagellate Relative Abundance Chlorophyte Relative Abundance
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4

Cyanobacteria (fluorescence) pixels mL - 1 x 10 6

80

20

40

60

Week

10

12

h
Cyanobacteria (visual) Relative Abundance
0 1 2 3 4 0 1

b
Diatom Relative Abundance
2 3 4 0 1 2 3 4

g L - 1 Bacterial Indicator Fatty Acids

1000 1200

Nematode Relative Abundance

200

400

600

800

A.J. Ray et al. / Aquaculture 310 (2010) 130138

Solids Removal No Solids Removal

Week

10

12

A.J. Ray et al. / Aquaculture 310 (2010) 130138 Table 3 The overall relative abundance of each microbial group throughout the study. Data are presented as mean standard error (range). Different letters within a row indicate signicant differences (P 0.01) between the two treatments. Treatment Solids removal No solids removal

137

Chlorophytes (visual) 3.98 0.02 (3.004.00) 3.97 0.02 (3.004.00) Diatoms (visual) 0.14 0.04 (0.002.00) 0.21 0.04 (0.002.00) Dinoagellates (visual) 0.96 0.11 (0.004.00) 0.74 0.09 (0.004.00) Nematodes (visual) 0.24 0.04 (0.002.00)a 0.57 0.05 (0.002.00)b Rotifers (visual) 1.66 0.07 (0.004.00) 2.16 0.08 (0.004.00)b Cyanobacteria (visual) 2.58 0.87 (1.004.00)a 3.12 0.05 (1.004.00)b Cyanobacteria 50.612.57 (32.8698.60)b 39.101.73 (20.4165.83)a (pixels mL 1 106) Bacteria fatty acid 828.13 53.02 395.00 23.08 indicators (g L1) (240.001700.00)b (140.001000.00)a Fatty acid bacterial indicator concentrations were analyzed based on the values during the nal week of the experiment using a MannWhitney Rank Sum test. All other abundance data were assessed using a repeated measures ANOVA which evaluated differences in abundance throughout the experiment.

constraints. Images of live or immediately preserved samples can be used for real-time, or near real-time assessments with this technique. Preserved samples likely reduce errors associated with measuring mobile organisms more than once. Spectrophotometric analysis of pigments is another common means of assessing pigment concentrations in water samples. An advantage that epiuorescence has is that only active pigments are detected, whereas spectrophotometry can include the inactive pigments of dead cells. An alternative to using image analysis software with this technique would be to count the uorescing microorganisms and arrive at a concentration per volumetric unit. However, as with light microscopy counts, this becomes exceedingly labor intensive. Other authors have used epiuorescence to perform such cell counts (Kawahara et al., 2009) or simply to visualize microorganisms (Azim and Little, 2008). However, the technique presented here offers a means of quantication that is more resistant to human bias and is relatively expedient. 3.3. Bacterial fatty acid assessment by GC By the end of the experiment, fatty acid bacterial biomarkers were signicantly reduced (P b 0.001) by solids removal (Fig. 3h). In systems with solids management, the nal mean concentration of bacterial fatty acids was 60.3% lower than in those systems without solids management (411.3 44.1 g L 1 versus 1036.9 119.0 g L 1, respectively). Table 3 shows the overall mean concentrations of bacterial fatty acids throughout the study. By reducing the abundance of bacteria in shrimp culture systems respiration may be reduced, in turn reducing the total oxygen demand of the system and the need for expensive supplemental oxygen injection. Although these systems rely, in large part, on bacteria for water quality maintenance, excess bacteria may reduce the amount of oxygen available to shrimp. There was no indication of water quality deterioration during this experiment (Table 2), suggesting that an abundance of microbes sufcient to cycle harmful nutrients remained in the systems. Bacteria associated with biooc particles may provide a packaging of valuable nutrients (De Schryver et al., 2008). However, some groups of bacteria are pathogenic to shrimp. Therefore, further explorations of the effects of partial bacterial exclusion on shrimp production are likely needed. The patterns by which bacterial abundance changes over time and with respect to management protocol have important implications for system function; the described technique of bacterial fatty acid assessment is effective at documenting those changes. Ju et al. (2008) described using HPLC to assess muramic acid levels which served as a biomarker to infer bacterial abundance. The same method to assess muramic acid abundance can be used to determine amino acid abundances. Assessing either amino acid or fatty acid proles of biooc can also be useful for evaluating nutritional qualities of the material. In this respect, both methods offer dual purposes. In terms of cost, both amino acid and fatty acid characterizations require comparably expensive equipment. This likely makes both methods better suited to research interests; however, infrequent sampling by an outsourced laboratory is an option for commercial aquaculture operations as well. Both methods are likely equal in their effectiveness, but it is important to develop a variety of tools for microbial assessment. The ability to use multiple techniques allows more specic questions to be addressed, in this case nutritional as well as biological. In the future it may be desirable to describe entire fatty acid proles of microbial communities in aquaculture systems, allowing for a more complete characterization of those communities.

abundance of nutritionally benecial or potentially toxic organisms that can reduce growth may help system managers better calculate feed rations based on expected shrimp growth. A drawback to this technique is that it can be somewhat subjective; care must be taken to eliminate bias and monitor consistency among analysts. 3.2. Epiuorescence with image analysis quantication Cyanobacteria pigment abundance, as determined by epiuorescence with image analysis quantication, was signicantly reduced during the study (P b 0.001) through the practice of solids removal (Fig. 3g). By the nal week of the study, cyanobacteria pigments were reduced by 17.1% with solids removal. The abundance of eukaryotic algae pigments (chlorophyll uorescence minus cyanobacteria uorescence) was not signicantly affected by solids removal. The results of the epiuorescence with image analysis quantication method substantiated those of the visual microscopy assessment. The relative abundance of eukaryotic algae (in this case chlorophytes, diatoms, and photoautotrophic dinoagellates) did not appear to be affected by solids removal. Eukaryotic algal groups other than chlorophytes, diatoms, and photoautotrophic dinoagellates were rarely seen during the current experiment. The abundance of the eukaryotic algal groups that were monitored with visual microscopy seemed unaffected by solids removal, the uorescence of eukaryotic chlorophyll-a likewise appeared unchanged. A signicant reduction in cyanobacteria abundance with solids removal was indicated by both visual microscopy and by epiuorescence with image analysis. This correspondence helps to validate the methods; however, cyanobacteria abundance data collected with the two methods were not statistically correlated. Samples for the two methods were collected on different days of the week, perhaps hindering correlation. Also, the two methods measure different things. Visual microscopy measures the abundance of cells, while epiuorescence measures the abundance of uorescing pigments within those cells. Therefore, epiuorescence may comparatively underestimate cyanobacteria abundance, but still elucidate changes in abundance. This technique allows for direct comparison between eukaryotic algae and cyanobacteria. Unlike visual microscopy, samples can be preserved for later analysis, potentially helping to overcome time

Fig. 3. Graphs a through f depict the relative abundance of chlorophytes, diatoms, dinoagellates, nematodes, rotifers, and cyanobacteria, respectively, based on light microscopy abundance quantication; the scoring system for this method is described in Table 1. Graph g depicts the results of cyanobacteria quantication using epiuorescence and image analysis, shown as uorescing pixels per milliliter. Graph h represents the concentration of branched and odd chain fatty acids, used as bacterial indicators to assess the relative abundance of bacteria. Week four is the rst week that settling chambers were used to remove suspended solids and this began the period that changes in microbial abundance were explored.

138

A.J. Ray et al. / Aquaculture 310 (2010) 130138 Ebeling, J.M., Timmons, M.B., Bisogni, J.J., 2006. Engineering analysis of the stoichiometry of photoautotrophic, autotrophic, and heterotrophic removal of ammonia-nitrogen in aquaculture systems. Aquaculture 257, 346358. ESS (Environmental Sciences Section), Inorganic Chemistry Unit, Wisconsin State Lab of Hygiene, 1993. ESS Method 340.2: Total Suspended Solids, Mass Balance (Dried at 103105 C) Volatile Suspended Solids (Ignited at 550 C). Wisconsin State Lab of Hygiene, Madison, WI, USA. Focken, U., Groth, A., Coloso, R.M., Becker, K., 1998. Contribution of natural food and supplemental feed to the gut content of Penaeus monodon Fabricius in a semiintensive pond system in the Philippines. Aquaculture 164, 105116. Folch, J.M., Lees, M., Stanley, G.H.S., 1957. A simple method for the isolation and purication of total lipids from animal tissues. Journal of Biological Chemistry 226, 497509. Hach Company, 2003. Method 8155, DR/4000 Procedure, Salicylate Method (Ammonia), Method 8507, USEPA Diazotization Method (Nitrite), Method 8039, DR/4000, Procedure, Cadmium Reduction Method (Nitrate), Method 8048 Phosphorus Reactive (Orthophosphate). Hach Company, Loveland, Colorado, USA. Hargreaves, J.A., 2006. Photosynthetic suspended-growth systems in aquaculture. Aquacultural Engineering 34, 344363. Harvey, H.R., Macko, S.A., 1997. Catalysts or contributors? Tracking bacterial mediation of early diagenesis in the marine water column. Organic Geochemistry 26, 531544. Ju, Z.Y., Forster, I., Conquest, L., Dominy, W., Kuo, W.C., Horgen, F.D., 2008. Determination of microbial community structures of shrimp oc cultures by biomarkers and analysis of oc amino acid proles. Aquaculture Research 39, 118133. Ju, Z.Y., Forster, I.P., Dominy, W.G., 2009. Effects of supplementing two species of marine algae or their fractions to a formulated diet on growth, survival and composition of shrimp (Litopenaeus vannamei). Aquaculture 292, 237243. Kastovska, K., Stibal, M., Sabacka, M., Cerna, B., Santruckova, H., Kastovska, K., Stibal, M., Sabacka, M., Cerna, B., Santruckova, H., Elster, J., 2007. Microbial community structure and ecology of subglacial sediments in two polythermal Svalbard glaciers characterized by epiuorescence microscopy and PLFA. Polar Biology 30, 277287. Kawahara, N., Shigematsu, K., Miyadai, T., Kondo, R., 2009. Comparison of bacterial communities in sh farm sediments along an organic enrichment gradient. Aquaculture 287, 107113. Metcalfe, L.D., Schmitz, A.A., Pelka, J.R., 1966. Rapid preparation of fatty acid esters from lipids for gas chromatographic analysis. Analytical Chemistry 38, 514515. Moreno-Ostos, E., Cruz-Pizaro, L., Basanta, A., George, D.G., 2008. The spatial distribution of different phytoplankton functional groups in a Mediterranean reservoir. Aquatic Ecology 42, 115128. Moss, S.M., Pruder, G.D., 1995. Characterization of organic particles associated with rapid growth in juvenile white shrimp, Penaeus vannamei Boone, reared under intensive culture conditions. Journal of Experimental Marine Biology and Ecology 187, 175191. Moss, S.M., Divakaran, S., Kim, B.G., 2001. Stimulating effects of pond water on digestive enzyme activity in the Pacic white shrimp Litopenaeus vannamei (Boone). Aquaculture Research 32, 125131. Newall, P., Bate, N., Metzeling, L., 2006. A comparison of diatom and macroinvertebrate classication of sites in the Kiewa River system, Australia. Hydrobiologia 572, 131149. Parrish, C.C., Abrajano, T.A., Budge, S.M., Helleur, R.J., Hudson, E.D., Pulchan, K., Ramos, C., 2000. Lipid and phenolic biomarkers in marine ecosystems: analysis and applications. In: Wangersky, P. (Ed.), The Handbook of Environmental Chemistry, vol. 5. Springer-Verlag, Berlin Heidelberg, pp. 193223. Part D. Ray, A.J. 2008. The effects of simple management techniques on microbial community dynamics within biooc-based culture systems and the relationship to shrimp (Litopenaeus vannamei) production. Master's Thesis, The College of Charleston, Charleston, South Carolina, USA. Ray, A.J., Shuler, A.J., Lefer, J.W., Browdy, C.L., 2009. Microbial ecology and management of biooc systems. In: Browdy, C.L., Jory, D.E. (Eds.), The Rising Tide. Proceedings of the Special Session on Sustainable Shrimp Farming. World Aquaculture Society, Baton Rouge, Louisiana, USA, pp. 255266. Ray, A.J., Lewis, B.L., Browdy, C.L., Lefer, J.W., 2010. Suspended solids removal to improve shrimp (Litopenaeus vannamei) production and an evaluation of a plantbased feed in minimal-exchange, superintensive culture systems. Aquaculture 299, 8998. Selinummi, J., Seppl, J., Yli-Harja, O., Puhakka, J.A., 2005. Software for quantication of labeled bacteria from digital microscope images by automated image analysis. BioTechniques 39, 859863. Sullivan, M.J., Moncreiff, C.A., 1990. Edaphic algae are an important component of salt marsh food-webs: evidence from stable isotope analysis. Marine Ecology Progress Series 62, 149159. Verity, P.G., Sieracki, M.E., 1993. Use of color image analysis and epiuorescence microscopy to measure plankton biomass. In: Kemp, P.F., Sherr, B.F., Sherr, E.B., Cole, J.J. (Eds.), Handbook of Methods in Aquatic Microbial Ecology. CRC Press, Boca Raton, Florida, USA, pp. 327338. Wasielesky Jr., W., Atwood, H., Stokes, A., Browdy, C.L., 2006. Effect of natural production in a zero exchange suspended microbial oc based super-intensive culture system for white shrimp Litopenaeus vannamei. Aquaculture 258, 396403. Zimba, P.V., Camus, A., Allen, E.H., Burkholder, J.M., 2006. Co-occurrence of white shrimp, Litopenaeus vannamei, mortalities and microcystin toxin in a southeastern USA shrimp facility. Aquaculture 261, 10481055.

4. Conclusion Assessing the structure of the microbial communities upon which minimal-exchange, intensive culture systems are heavily reliant is an important step in understanding and predicting system function. This document offers strategies of evaluation useful for the microbial communities and conditions that occur in minimal-exchange, intensive culture systems. These techniques were used to successfully demonstrate the effects that suspended solids removal can have on the abundance of certain microorganisms. By removing suspended solids, the abundance of nematodes, rotifers, cyanobacteria, and bacteria were all signicantly reduced while chlorophyte, diatom, and dinoagellate abundances were not signicantly affected. These results indicate the effects that relatively simple management strategies can have on the microbial communities in intensive aquaculture systems. Acknowledgements The techniques described in this document were adapted and applied at the Waddell Mariculture Center in Bluffton, South Carolina, USA as part of a Master of Science Thesis (Ray, 2008) through the College of Charleston, Charleston, South Carolina, USA. Thank you Heidi Atwood, Kirsten Ayers, Maggie Broadwater, Amy Dickson, Sylvia Galloway, Traci Holstein, Martin Jones, Peter Kingsley-Smith, Beth Lewis, Brad McAbee, Mark McConnel, Stacy Ray, Andrew Shuler, Jesus Venero, Joe Wade, David White, and the staff of the Waddell Mariculture Center. Any mention of a trademark or product manufacturer is in no way an endorsement of that business or a suggestion that one product is superior to another. This research was supported by grants from the USDA Integrated Organic Program and the USDA US Marine Shrimp Farming Program. This is contribution number 657 from the South Carolina Department of Natural Resources Marine Resources Research Institute. References
Alfaro, A.C., Thomas, F., Sergent, L., Duxbury, M., 2006. Identication of trophic interactions within an estuarine food web (northern New Zealand) using fatty acid biomarkers and stable isotopes. Estuarine, Coastal, and Shelf Science 70, 271286. Alonso-Rodriguez, R., Paez-Osuna, F., 2003. Nutrients, phytoplankton and harmful algal blooms in shrimp ponds: a review with special reference to the situation in the Gulf of California. Aquaculture 219, 317336. APHA (American Public Health Association), American Water Works Association, and Water Pollution Control Association, 1989. Standard Methods for the Examination of Water and Wastewater, 17th edition. American Public Health Association, Washington, D.C., USA. Avnimelech, Y., 2006. Bio-lters: the need for a new comprehensive approach. Aquacultural Engineering 34, 172178. Avnimelech, Y., 2009. Biooc Technology: A Practical Guide Book. World Aquaculture Society, Baton Rouge, Louisiana, USA. 182 pp. Azim, M.E., Little, D.C., 2008. The biooc technology (BFT) in indoor tanks: water quality, biooc composition, and growth and welfare of Nile tilapia (Oreochromis niloticus). Aquaculture 283, 2935. Bratvold, D., Browdy, C.L., 1998. Simple electrometric methods for estimating microbial activity in aquaculture ponds. Aquacultural Engineering 19, 2939. Budge, S.M., Parrish, C.C., Mckenzie, C.H., 2001. Fatty acid composition of phytoplankton, settling particulate matter and sediments at a sheltered bivalve aquaculture site. Marine Chemistry 76, 285303. Budge, S.M., Parrish, C.C., 2003. FA determination in cold water marine samples. Lipids 38 (7), 781791. Burford, M.A., Thompson, P.J., McIntosh, R.P., Bauman, R.H., Pearson, D.C., 2004. The contribution of occulated material to shrimp (Litopenaeus vannamei) nutrition in a high-intensity, zero-exchange system. Aquaculture 232, 525537. Crab, R., Avnimelech, Y., Defoirdt, T., Bossier, P., Verstraete, W., 2007. Nitrogen removal techniques in aquaculture for a sustainable production. Aquaculture 270, 114. De Schryver, P., Crab, R., Defoirdt, T., Boon, N., Verstraete, W., 2008. The basics of bioocs technology: the added value for aquaculture. Aquaculture 277, 125137. DeVantier, L.M., De'ath, G., Done, T.J., Turak, E., 1998. Ecological assessment of a complex natural system: a case study from the Great Barrier Reef. Ecological Applications 8 (2), 480496.