Microbial Ecology

, Volume 55, Issue 4, pp 583–594

Assessment of Microzooplankton Grazing on Heterosigma akashiwo Using a Species- Specific Approach Combining Quantitative Real-Time PCR (QPCR) and Dilution Methods

Authors

  • Elif Demir
    • College of Marine and Earth StudiesUniversity of Delaware
  • Kathryn J. Coyne
    • College of Marine and Earth StudiesUniversity of Delaware
  • Martina A. Doblin
    • Institute of Water and Environmental Resource Management/Department of Environmental ScienceUniversity of Technology, Sydney
  • Sara M. Handy
    • College of Marine and Earth StudiesUniversity of Delaware
    • College of Marine and Earth StudiesUniversity of Delaware
    • Department of Biological SciencesUniversity of Southern California
Original Article

DOI: 10.1007/s00248-007-9263-9

Cite this article as:
Demir, E., Coyne, K.J., Doblin, M.A. et al. Microb Ecol (2008) 55: 583. doi:10.1007/s00248-007-9263-9

Abstract

Delaware’s Inland Bays (DIB) are subject to numerous mixed blooms of harmful raphidophytes each year, and Heterosigma akashiwo is one of the consistently occurring species. Often, Chattonella subsalsa, C. cf. verruculosa, and Fibrocapsajaponica co-occur with H. akashiwo, indicating a dynamic consortium of raphidophyte species. In this study, microzooplankton grazing pressure was assessed as a top–down control mechanism on H. akashiwo populations in mixed communities. Quantitative real-time polymerase chain reaction (QPCR) with species-specific primers and probes were used in conjunction with the dilution method to assess grazing pressure on H. akashiwo and other raphidophytes. As a comparison, we measured changes in chlorophyll a (chl a) to determine whole community growth and mortality caused by grazing. We detected grazing on H. akashiwo using QPCR in samples where chl a analyses indicated little or no grazing on the total phytoplankton community. Overall, specific microzooplankton grazing pressure on H. akashiwo ranged from 0.88 to 1.88 day−1 at various sites. Experiments conducted on larger sympatric raphidophytes (C. subsalsa, C. cf. verruculosa and F. japonica) demonstrated no significant microzooplankton grazing on these species. Grazing pressure on H. akashiwo may provide a competitive advantage to other raphidophytes such as Chattonella spp. that are too large to be consumed at high rates by microzooplankton and help to shape the dynamics of this harmful algal bloom consortium. Our results show that QPCR can be used in conjunction with the dilution method for evaluation of microzooplankton grazing pressure on specific phytoplankton species within a mixed community.

Introduction

Microzooplankton grazing accounts for a significant loss factor of phytoplankton and bacteria in marine systems [45]. Selective grazing by zooplankton that are <200 μm shapes phytoplankton community structure [70] and assists nutrient recycling and regeneration [26], thus establishing a link between lower and higher tropic levels [59]. Microzooplankton grazing can remove enough cells to prevent or diminish blooms [60], creating a top–down control (removal of organisms by consumers) on community dynamics that may determine the fate of harmful algal blooms. Several studies have demonstrated the direct negative effects of grazing on potentially harmful algae such as Chattonella antiqua [1, 69], Micromonas spp. [16], and Alexandrium minutum [7, 19]. Heterotrophic dinoflagellates have also been found to graze on laboratory cultures of the raphidophytes Fibrocapsa japonica [66] and Heterosigma akashiwo [9]. Conversely, grazing pressure on a mixed community may also indirectly favor the growth of harmful algal species, as is the case with Aureococcus anophagefferens [20].

Blooms of a mixed raphidophyte consortium composed of C. subsalsa, Heterosigma akashiwo, C. cf. verruculosa (this species is currently undergoing reclassification [6]), and F. japonica are common in Delaware Inland Bays (DIB) [24]. In 2000, C. cf. verruculosa reached a maximum density of 1 × 107 cells L−1 in the DIB and was linked to a major fish kill [3]. Since then, mixed blooms of raphidophytes have occurred each year from early May to late October [73]. Although no toxic events have been reported since 2000, understanding all the factors involved in bloom formation is important in assessing their potential to become harmful and their future management.

Interspecific competition may play an important role in determining the dynamics between raphidophyte species in mixed blooms. A series of laboratory experiments on local isolates of C. subsalsa and H. akashiwo indicated that H. akashiwo has a higher maximum growth rate (μmax) and lower half saturation constant (Ks) for phosphate, nitrate, and ammonium and can grow on urea, whereas C. subsalsa cannot. In fact, H. akashiwo appears to be a better competitor for nutrients under almost all conditions and can tolerate a wider salinity, light, and temperature range than C. subsalsa [74]. However, in natural bloom samples, C. subsalsa is often more abundant than H. akashiwo [24], suggesting that top–down control mechanisms like grazing could offset any competitive advantage that this latter species gains from bottom–up (growth limitation of organisms as a result of environmental parameters such as nutrients) control factors.

Landry and Hassett [40] introduced the dilution method to measure microzooplankton grazing rates in natural communities, and it has been used to measure grazing in a variety of aquatic systems, from estuaries ([48] and references therein) to the open ocean [41, 63]. The design is based on a series of dilutions, with each dilution decreasing the potential of grazer/prey encounter, so that the probability of prey consumption is proportional to the dilution factor [40, 41]. As is the case with all direct environmental grazing methods, the dilution method has limitations, and there is controversy over the assumption that it is applicable to all types of marine systems. Some evidence has been presented indicating adverse effects for coastal ciliate communities, which can starve in highly diluted experimental conditions (i.e., 10% of whole seawater dilutions) and potentially lead to overestimation of grazing pressure [14]. However, this limitation has not been fully tested, and in various regions, reasonable data have been collected using the dilution method [39]. The advantage of using this method is that it requires minimal handling and few manipulations of the natural phytoplankton and microzooplankton communities. The dilution method has recently been applied in harmful algae research (e.g., [7, 8, 19, 20]) and may be the most applicable method for evaluating grazing on fragile species such as raphidophytes that require minimum disturbance.

One of the difficulties associated with the dilution method is identification and quantitative enumeration of the taxon of interest. Several methods have been developed to investigate species-specific effects of microzooplankton grazing such as use of labeled prey for establishing a qualitative relationship between specific species of predators and prey [9]. Typically, species-specific grazing experiments are performed using cultures grown in the lab. Although these are valid in determining the potential for specific microzooplankton to consume specific algae, they cannot demonstrate selective grazing (or lack thereof) in natural communities. In environmental studies, taxon-specific pigment markers have been used to differentiate selective grazing pressure on general phytoplankton groups [18, 63, 71]. The pigment analysis method requires in-depth knowledge of the entire phytoplankton community and their pigment compositions, and the sample analyses can be time consuming. Although pigment analysis is a valuable method to evaluate grazing pressure on broad taxonomic groups of phytoplankton, it is seldom able to differentiate grazing at the species level because of the lack of unique pigment signatures. In general, it is more suited for open ocean phytoplankton communities where total community structure can be better assessed using pigments compared to estuarine communities.

Another species-specific method to determine grazing pressure is by microscopic cell counts. This approach requires a high level of expertise for experiments dealing with mixed communities. Microscope cell counts are also challenging for species such as raphidophytes, which cannot be identified after preservation longer than a few days [28]. Enumeration of fixed natural samples via microscopic cell counts is challenging because raphidophytes have no cell wall and often burst after fixation [28]. Quantitative real-time polymerase chain reaction (QPCR) is a practical and efficient alternative to microscope cell counts to accurately determine changes in cell abundances in mixed natural communities [4, 5, 10, 11, 47, 52]. For QPCR, DNA is extracted from the total community, and a target gene (such as 18S rDNA) is amplified using species-specific primers and a fluorescent probe. Accumulation of the PCR product is measured as a change in fluorescence and is directly proportional to the amount of starting template [27] or number of cells of the target species (e. g., [10, 24, 52]). The product accumulation is normalized to a reference standard to correct for differences in extraction and amplification efficiencies [10]. QPCR has recently been applied to investigate vertical migration of raphidophyte species (H. akashiwo and C. subsalsa) in mesocosm experiments containing natural phytoplankton communities [24] and germination of cysts produced by these algae in DIB [53].

In this study, we used the dilution method to examine grazing pressure on H. akashiwo in natural blooms occurring from June to October in 2004 and 2005. The grazing rates on H. akashiwo obtained with QPCR were compared to grazing rates for the whole community derived from chlorophyll a (chl a). We also performed a limited number of grazing experiments on natural populations of C. subsalsa, C. cf. verruculosa, and F. japonica. We found that H. akashiwo, which ranges in size from 12 to 20 μm, was subject to microzooplankton grazing pressure at various intensities, whereas the other larger raphidophyte species were not grazed. This could contribute to the dynamics of the consortium by giving other co-occurring raphidophytes, which may be too large (20–60 μm) to be consumed by microzooplankton, an advantage over H. akashiwo. In addition, our results demonstrate that the QPCR method can be integrated with the dilution method to evaluate microzooplankton grazing pressure on specific phytoplankton species in natural populations.

Materials and Methods

Dilution Experiments

We sampled blooms of H. akashiwo, other raphidophytes, and non-harmful algae in the DIB from June to November of 2004 and 2005. Water samples were collected from five different locations on 13 different dates at varying salinities (Table 1). Each location name identifies a different canal system linked to DIB. Sampling dates were within the period when H. akashiwo and other raphidophyte blooms are common in this area.
Table 1

Microzooplankton grazing results for the bloom seasons in 2004 and 2005 (numbered 1 through 13)

Expt

Date

Location

Salinity (psu)

Total number of species

Target prey species

Initial cell density (L−1)

Total chl a

QPCR

r2

k (day−1) 95% conf. int.

g (day−1) 95% conf. int.

r2

k (day−1) 95% conf. int.

g (day−1) 95% conf. int.

2004

 1

June 29

Williamson Creek

12

16

H. akashiwo

3.08 × 107

NS

ND

   

 2

July 22

Love Creek

15

NA

H. akashiwo

1.20 × 107

0.78

0.85 (0.64 to 1.06)

0.82 (1.12 to −0.5)

   

 3

July 31

Torquay Canal

24

10

H. akashiwo

1.00 × 104

0.42

1.52 (1.09 to 1.99)

0.80 (1.46 to 0.14)

   

 4

Sept 29

Russell’s Canal

21

8

H. akashiwo

3.30 × 106

0.51

0.39 (0.26 to 0.52)

0.28 (0.46 to 0.09)

0.47

1.61 (0.65 to 2.57)

1.88 (3.38 to 0.48)

 5

Sept 30

Russell’s Canal

21

8

H. akashiwo

4.09 × 107

0.35

0.21 (0.09 to 0.33)

0.19 (0.36 to 0.01)

0.65

0.18 (−0.51 to 0.88)

1.61 (2.90 to 0.92)

 6

Oct 5a

Derickson Creek

13

NA

Pseudopedinella sp.

9.00 × 106

0.97

1.11 (1.04 to 1.18)

0.77 (0.87 to 0.66)

   

 7

Oct 8

Derickson Creek

6

10

H. akashiwo

8.80 × 106

NS

ND

   

 8

Oct 11a

Russell’s Canal

22

7

H. akashiwo

1.14 × 107

NS

0.46 (0.32 to 0.60)

0.11 (0.31 to 0.08)

0.84

0.87 (0.63 to 1.10)

1.12 (1.48 to 0.75)

 9

Nov 17a

Indian River Bay Canal

20

NA

Euglena sp.

6.00 × 107

0.65

0.42 (0.22 to 0.63)

0.56 (0.85 to 0.27)

   

2005

 10

June 30

Indian River Bay Canal

22

8

C. subsalsa

8.78 × 105

NS

ND

NS

ND

C. cf. verruculosa

1.15 × 107

NS

ND

 11

July 14

Holt’s Landing

20

5

H. akashiwo

1.72 × 106

NS

ND

0.55

0.85 (0.47 to 1.24)

0.88 (1.44 to 0.32)

 12

July 26

Russell’s Canal

19

5

C. cf. verruculosa

4.00 × 105

0.38

0.34 (0.20 to 0.49)

0.24 (0.45 to 0.02)

NS

ND

C. subsalsa

5.00 × 104

NS

ND

 13

Sept 2

Russell’s Canal

24

NA

C. cf. verruculosa

8.50 × 105

0.73

0.33 (0.21 to 0.44)

0.39 (0.55 to 0.22)

NS

ND

F. japonica

1.38 × 105

NS

ND

Coefficients of grazing (g) and apparent growth rate (k), (n = 3), are given using chlorophyll a and QPCR results with upper and lower 95% confidence intervals.

NA Not available, NS insignificant r2 (used when no grazing was detected, included for comparison reasons), ND grazing not detected

aRaphidophyte preblooms/blooms were used for all experiments except Pseudopedinella sp. and Euglena sp.

Samples were collected in 20-L carboys using a bilge pump and were gently gravity filtered through a 202-μm size mesh to exclude meso- (zooplankton that are 0.2 to 20 mm in size) and macrograzers (zooplankton that are 2 to 20 cm in size). After collection and pre-filtration, the communities were evaluated via microscopy to ensure that the phytoplankton and microzooplankton communities were not harmed. Part of the 202-μm pre-filtered water was filtered through 0.2-μm filter cartridges using air pressure. This filtrate was then used as a diluent for the whole water. Dilution treatments were 25, 50, 75, and 100% of whole water, in triplicates of 1-L sterilized bottles. Each bottle was enriched with NO3, PO43−, trace metals, vitamins, and FeEDTA at f/2 phytoplankton medium concentrations [21]. No nutrient addition treatments used in open ocean dilution experiments [41] were not applied in this study because the study sites within DIB are highly eutrophic [54, 56]. Therefore, it was assumed that phytoplankton growth was not limited by nutrients in our experiments. In the absence of this treatment, values reported in this article may be considered upper limit estimates of actual microzooplankton grazing rates. Bottles were incubated at the ambient temperature of the collection site under a 12:12 light/dark cycle. Samples for extracted chl a and molecular analyses were collected before and after the 24-h incubation for each experiment.

Chlorophyll a was extracted in 90% acetone and measured fluorometrically [72]. Cell counts for raphidophytes and total community composition and density were performed on live samples using the droplet estimation method (45 μl) under light microscopy, which has a detection limit of 50,000–100,000 cells L−1 [2]. This method is valuable for counting total phytoplankton communities because the cells are still alive and swimming patterns can be observed, which may be essential in differentiating similarly structured algae. Microzooplankton species were identified using light microscopy from samples fixed in Lugol’s solution. Mortality because of microzooplankton grazing (g) and apparent growth rates (k) of phytoplankton were obtained by plotting apparent growth rates vs dilution factor, and this rate is reported as per day (day−1), according to Landry and Hasset [40]. GraphPad Prism (GraphPad Software, San Diego, CA, USA) was used for statistical comparisons of the slopes of regression analysis [22].

Molecular Methods

DNA from initial (T0) and final (T24) samples were collected by gently filtering (∼85 kPa) the water samples onto 3- or 5-μm polycarbonate filters. The filters were submersed in cetyltrimethylammonium bromide (CTAB) buffer (100 mM Tris–HCl [pH 8], 1.4 M NaCl, 2% [w/v] CTAB, 0.4% [v/v] β-mercaptoethanol, 1% [w/v] polyvinylpyrollidone, 20 mM EDTA [13]). The plasmid pGEM was included in the buffer at a concentration of 20 ng mL−1 as an internal standard to correct for differences between samples based on DNA extraction and amplification efficiencies [10]. Samples were then stored at −80°C, and DNA was extracted as described in Coyne et al. [11]. After spectrophotometric quantification, each DNA sample was diluted 1:200 with LoTE (3 mM Tris–HCl [pH 7.5], 0.2 mM EDTA) to obtain DNA concentrations of 1.84–62.12 ng μL−1 for QPCR.

We used QPCR to quantify raphidophyte species’ cell densities from initial (T0) and final (T24) samples for each experiment using an ABI Prism 7500 Real Time PCR detection system (Applied Biosystems). Microscope cell counts of raphidopytes in undiluted whole water samples were performed as described in Coyne et al. [10] and used as density calibrators for calculating cell densities from QPCR results [10]. Heterosigma akashiwo calibrators were prepared from samples that were confirmed (by PCR) to be free of a co-occurring newly described raphidophyte species [12] that is morphologically indistinguishable from Heterosigma by light microscopy [10, 12]. Thermal cycling parameters for the QPCR reactions are described in Coyne et al. [10]. Briefly, we amplified DNA in triplicate 25-μL reactions using 2.5 μL of 1:200 diluted template, 12.5 μL of Taqman Universal Master Mix (Applied Biosystems), 0.9 μM of a species-specific forward primer, 0.9 μM of a species-specific reverse primer, and 0.2 μM of a species-specific probe. The species-specific primers and probes (listed in Table 2) were designed to amplify a unique fragment of the 18S rRNA gene. All these primer/probe sets were previously reported in other studies [10, 23], with the exception of the F. japonica set. The specificity of the primers and probe were evaluated as described in Coyne et al. [10]. Primer/probe concentrations were optimized as described in Coyne et al. [10] and for all species were 0.9 μM and 0.2 μM, respectively, except for C. cf. verruculosa for which concentrations of 0.3 μM for primers and 0.25 μM for probe were used. The threshold cycle number (Ct) was determined for each sample and calibrator. Co-extracted pGEM plasmid DNA was also amplified for the diluted DNA samples in separate reactions using pGEM-specific primers and probe. The concentrations of raphidophytes were calculated using the comparative (ΔΔCt) method [10, 44] in which amplification of the target species in each sample is first normalized to the internal standard, pGEM, and then compared to the abundance of the target species in the calibrator samples. These concentrations were then used to calculate apparent phytoplankton growth and grazing mortality as described previously. Reaction efficiencies for the pGEM and H. akashiwo assays were 91 and 90%, respectively.
Table 2

List of primer and probe sequences used for QPCR analysis of the raphidophytes Chattonella subsalsa, C. cf. verruculosa, Fibrocapsa japonica, and Heterosigma akashiwo, and for amplification of the pGEM plasmid

DNA Target

Primer/Probe

Sequence (5′–3′)

References

Chattonella subsalsa 18S rDNA

Cs 1350F

CTAAATAGTGTGGGTAATGCTTAC

Coyne et al. [10]

Cs 1705R

GGCAAGTCACAATAAAGTTCCAA

Raph Probe

CAACGAGTACTTTCCTTGGCCGGAA

Chattonella cf. verruculosa 18S rDNA

Cv 1561F

ATGCATACAGCGAGTCTAGA

Handy et al. [23]

Cv 1780R

TCACTCCGAAAAGTGTCAAC

Cv Probe

CAAGAGTACCCAGGCCTCTCGACC

Fibrocapsa japonica 18S rDNA

Fc 1350F

TGCTTTAGTCATTGTGTGCAG

This study

Fc 1705R

ACCACAAACTAATGAGGAGGC

Fc Probe

CCCAGGCCTACCGGCCAAGGTTGTA

Heterosigma akashiwo 18S rDNA

Hs 1350F

CTAAATAGTGTCGGTAATGCTTCT

Coyne et al. [10]

Hs 1705R

GGCAAGTCACAATAAAGTTCCAT

Hs Probe

CAAGGAGTAACGACCTTTTGCCGGAA

Chattonella cf. verruculosa 18S rDNA

M13F

CCCAGTCACGACGTTGTAAAACG

Coyne et al. [10]

pGEM R

TGTGTGGAATTGTGAGCGGA

pGEM Probe

CACTATAGAATACTCAAGCTTGCATGCCTGCA

F Forward primer, R reverse primer

Results

Microzooplankton Grazing Pressure on Heterosigma akashiwo

Temperatures of the water collected ranged from 19–30°C, except for Expt 9 where the water was 6°C (Table 1). There was considerable variability in salinity (6–24 psu); however, neither of these parameters appeared to affect the grazing rates directly. Heterosigma akashiwo abundances were greater than 106 cells L−1 in all of these experiments except for Expt 3 where the density was 1 × 104 cells L−1 (Table 1). The highest abundance was observed in Expt 5 (40 × 106 cells L−1) followed by Expt 1 (30  × 106 cells L−1). In Expt 1, 4, 7, and 8, the phytoplankton community composition was evaluated by the droplet estimation method [2]. However, these estimates included only those species that could be identified by light microscopy and likely do not include very small flagellates that are below the level of detection. For these experiments, H. akashiwo abundances comprised 40, 99, 92, and 96%, respectively, of the phytoplankton community [73]. In Expt 1, where microzooplankton grazing was not detected, the phytoplankton community was highly diverse with 16 species compared to an average of 8 in our other experiments. Our results indicate that total chl a-derived microzooplankton-grazing rates are not directly related to the chl a levels in the environment or to H. akashiwo densities (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-007-9263-9/MediaObjects/248_2007_9263_Fig1_HTML.gif
Figure 1

Apparent growth rate versus fraction of whole water in Expt 4 (H. akashiwo) calculated by total community chl a (a) and QPCR (b) and, in Expt 5, calculated by total community chl a (c) and QPCR (d). Lines are fit by least squares linear regression.

Grazing rates on H. akashiwo and the total community for Expt 4, 5, 8, and 11 were determined by QPCR and chl a analyses, respectively. Expt 4 and 5 were set up within 24 h of each other, both with water collected during a bloom of H. akashiwo from Russell Canal, Little Assawoman Bay, DE. Heterosigma akashiwo density increased from 3 10−6 cells L−1 in Expt 4 to 40 × 106 cells L−1 in Expt 5 (Table 1). In Expt 4, grazing and growth of the total community (by chl a) were 0.28 and 0.39 day−1, respectively (Fig. 1a). Grazing rates on H. akashiwo obtained from QPCR data (Fig. 2b) were significantly higher (1.88 day−1, p = 0.019) than grazing rates on the total community (0.28 day−1, Fig. 1a). In Expt 5, as well, the grazing and growth rates for the whole community were significantly different (p = 0.001) than for H. akashiwo (Fig. 1c and d). Within the course of the consecutive experiments (Expts 4 and 5), grazing on the total community obtained by chl a values (Fig. 1a and c) did not change significantly and neither did species-specific grazing values on H. akashiwo obtained through QPCR (Fig. 1b and d). However, growth of the total community significantly decreased (p = 0.0011), and a greater decrease was observed for H. akashiwo growth rates based on QPCR results (p < 0.0001).
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-007-9263-9/MediaObjects/248_2007_9263_Fig2_HTML.gif
Figure 2

Apparent growth rate versus fraction of whole water in dilution Expt 8 (H. akashiwo) calculated by total community chl a (a) and QPCR (b) and, in Expt 11, calculated by total community chl a (c) and QPCR (d). Lines are fit by least squares linear regression.

In Expt 8, growth rates for the total community versus H. akashiwo were 0.46 and 0.87 day−1, respectively (Fig. 2a,b). Grazing rates using chl a concentrations (0.11 day−1, Fig. 2a) suggest that grazing on the total community was minimal, whereas species-specific results obtained by QPCR analysis of H. akashiwo indicate significantly higher grazing pressure on H. akashiwo (1.19 day−1, Fig. 2b) compared to the total community (p < 0.0001).). Similar results were obtained in Expt 11, where microzooplankton grazing on H. akashiwo yielded a significant slope (0.88 day−1, Fig. 2d), but was not detected on the total community (Fig. 2c).

Microzooplankton Grazing Pressure on Co-existing Raphidophytes

In Expt 10, 12, and 13, QPCR methods were used to investigate the microzooplankton grazing rates on other DIB raphidophytes (Table 1), including C. subsalsa (Expts 10 and 12), C. cf. verruculosa (Expts 10, 12, and 13), and F. japonica (Expt 13). The specificity of primers for F. japonica (reported in this study for the first time) in field samples from the DIB was confirmed by sequence analysis of positive PCR reactions. Application of the comparative Ct method was validated for this species with the pGEM internal standard as described in Coyne et al. [10]. In Expt 10, both Chattonella species dominated the community with a total of 1.24 × 107 cells L−1, whereas in Expts 12 and 13, total raphidopyte cell densities were ∼27 and 12 times lower, respectively, than Expt 10 (Table 1). Grazing on the total community, determined by chl a, was either minimal (Expts 12 and 13) or not detectable (Expt 10), whereas species-specific grazing rates calculated using QPCR for C. subsalsa, C. cf. verruculosa, and F. japonica were consistently undetectable.

Microzooplankton Grazing Pressure on Non-harmful Algae

We also measured microzooplankton grazing during two blooms of small (<20 μm) non-harmful algae (Pseudopedinella sp. and Euglena sp.) that occurred in Dirickson Creek (Expt 6) and Indian River Bay Canal (Expt 9; tributaries of the DIB). Grazing pressure on these species was assessed for comparative purposes because both are closer in size to H. akashiwo than other species included in this study. In Expt 9, grazing pressure on a small (∼10 μm) euglenoid species (Euglenophyceae, [65]) at a density of 60 × 106 L−1 was determined by dilution experiments (Fig. 3a and 3b). In this sample, we identified two large mixotrophic dinoflagellates, Oxyrrhis marina and Protoperidinium sp. with growth rates (0.8–1.4 day−1 [32, 51] and 0.21–0.33 day−1 [46]) that are within the range of growth rates (0.21–1.11 day−1) observed for total communities throughout this study. We evaluated microzooplankton grazing pressure by total chl a on unfractionated water (Fig. 3a) and for the <20-μm fraction (Fig. 3b) with no significant differences. Growth rates, however, were significantly greater for the <20-μm size fraction compared to whole community rates (p = 0.008).
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-007-9263-9/MediaObjects/248_2007_9263_Fig3_HTML.gif
Figure 3

Apparent growth rate versus fraction of whole water in dilution Expt 9 (Euglena sp.) calculated by total community chl a (a) and chl a for the <20-μm size fraction (b) and, in Expt 6 (Pseudopedinella sp.), calculated by total community chl a (c) and chl a for the <20-μm size fraction (d). Lines are fit by least squares regression to the points.

In Expt 6, we evaluated microzooplankton grazing on the dictyophyte Pseudopedinella sp. (<10 μm) that occurred at a density of 9 × 106 cells L−1 in Dirickson Creek. Grazing pressure measured by total chl a (0.77 day−1, Fig. 3c) was comparable to grazing on the total community (as determined by chl a) in Expts 2 and 3 (0.82, 0.80 day−1) when H. akashiwo was present. However, growth rates were higher for the phytoplankton community in Expt 6 (1.11 day−1) compared to growth rates of the total communities that included H. akashiwo in Expts 2, 4, 5, and 8 (Table 1). As expected, grazing rates were significantly higher for the <20-μm size fraction (1.09 day−1) of Expt 6 compared to grazing pressure on the unfractionated community (0.77 day−1, p = 0.0002, Fig. 3d), a result of the higher density of smaller phytoplankton in that particular location. Grazing pressures on the total phytoplankton communities in Expt 6 and Expt 9 were not significantly different; however, growth rates for Expt 6 were significantly higher than Expt 9 (p < 0.0001). When comparing the <20-μm size fraction, grazing rates were also significantly higher in Expt 6 than Expt 9 (p = 0.001).

Discussion

A consortium of at least four harmful raphidophyte species blooms annually in the DIB from mid May to the end of October. Our previous field and experimental data [74] suggested that top–down control mechanisms contribute to the dynamics of this group of algae, and we hypothesized that H. akashiwo (12–20 μm) can be removed by microzooplankton grazing, whereas larger raphidophyte species (C. subsalsa, C. cf. verruculosa, and Fibrocapsa japonica), are too large (20–50 μm) to be consumed by microzooplankton.

Although studies on laboratory cultures reveal important information on physiological characteristics and grazing rates of harmful algae [9, 55, 58, 62], they may have limited value in investigating predator/prey interactions within the natural environment. Grazing rates in natural estuarine samples, however, are much more difficult to interpret compared to laboratory cultures or offshore sites because of the complexity of environmental factors and phytoplankton communities involved [48]. For this reason, grazing pressure on natural populations of H. akashiwo has not been evaluated under natural environmental conditions before this study. In this article, we were able to measure variable rates of microzooplankton grazing on H. akashiwo in the DIB system using the dilution method in combination with species-specific molecular assays.

Grazing was measured on the total phytoplankton community by traditional chl a measurements. Although chl a is an efficient way of evaluating total community biomass changes over time, in estuarine settings such as the DIB, the complexity of the phytoplankton community structure makes it nearly impossible to detect grazing on a specific species. Even when the species of interest is numerically dominant, total community chl a can often be skewed toward smaller phytoplankton or can be disproportionately contributed by a few cells of much larger but rarer species.

Quantification of phytoplankton density using microscopic cell counts is another commonly used method that provides specific information. However, some of the problems associated are: raphidophytes, as a group, cannot be reliably quantified once fixed because of their fragile nature [10, 28] and H. akashiwo is a relatively small alga, which is difficult to distinguish from other species, particularly, when a phylogenetically distinct species that closely resembles H. akashiwo in morphology is present in collected samples [8, 12]. Finally, the error in cell counts is progressively greater, as the water is diluted to measure microzooplankton grazing. Because the grazing mortality and growth rates are calculated using linear regression analysis of the dilution treatments, the propagated error in cell counts could potentially distort the results.

QPCR has been established as a sensitive and accurate method for estimating changes in cell densities of harmful algal species in environmental samples [4, 10, 17, 24, 52]. In a recent study, the application of QPCR to the analysis of H. akashiwo demonstrated a linear response over eight orders of magnitude with a sensitivity of four copies of the target gene [8]. In consideration of the problems associated with microscopic enumeration for this species, linear regression analysis using QPCR data to evaluate grazing pressure on H. akashiwo may be less subject to various errors and more accurate than cell counts. In this article, we investigated the use of QPCR in conjunction with the dilution method to differentiate between microzooplankton grazing on H. akashiwo and grazing on the whole community (evaluated by chl a) in mixed natural estuarine communities.

Potential errors and limitations of QPCR are described in [10] and include inefficiencies in filtration or errors in sample volume measurement, inaccurate pipetting of the lysis buffer, and differences in lysis efficiencies. However, these problems can be avoided by consistent treatment of all samples and the use of an internal standard (pGEM) that reduces downstream errors. Cell counts for calibrator samples need to be performed carefully. Using ambient samples as calibrators may introduce some error because target cells in mixed communities may not be as easy to count as samples from unialgal cultures. Nonetheless, the advantages of this approach outweigh the potential limitations because calibrator samples collected from the same environment contain a similar complexity of DNA and potential PCR inhibitors as the unknown samples. Furthermore, because our objectives were to determine the relative changes in H. akashiwo abundance, the results of this study were not affected by errors in cell counts for calibrator samples.

The comparison of QPCR analyses with chl a measurements allowed us to detect species-specific and community level changes over a short time scale and at different cell densities. Our results indicate that microzooplankton grazing pressure on H. akashiwo can be greater than the grazing on the total phytoplankton community (Expts 4, 5, 8, and 11). These results contradict a previous study where two copepod species selected against H. akashiwo (in culture) and did not feed on this alga when it was provided in high concentrations [67]. However, considering their different feeding mechanisms and trophic interactions, differences between microzooplankton (e.g., ciliates) and mesozooplankton (e.g., copepods) grazing are to be expected. Heterosigma akashiwo is considered to be of low nutritional value [67], and toxic strains are known to deter grazers when provided as food [9]. These previous experiments were performed, however, using cultured H. akashiwo, and palatability to grazers will undoubtedly vary depending on strain and environmental conditions. Although the toxicity of H. akashiwo populations in the DIB is unknown, the relatively high grazing rates we observed and the lack of fish kills during blooms of H. akashiwo suggest they were not toxic during our study.

In September 2004, two experiments were conducted with water collected during the progression of a short-lived H. akashiwo bloom (Expts 4 and 5). Growth rates of the total phytoplankton community in both experiments were not significantly different (p = 0.438), and grazing on the total community did not change. Heterosigma akashiwo-specific grazing also remained relatively constant with no significant difference (1.88 vs 1.61 day−1, p = 0.934) between the two experiments. Although the species-specific grazing rate on H. akashiwo did not change, the growth rate of Heterosigma decreased from 1.61 to 0.18 day−1, suggesting that the bloom had reached stationary phase during Expt 5. We were able to verify this in field samples taken over the following 2 days when H. akashiwo cell densities fell to 3.75 × 105 cells L−1. The cause of this decline in cell densities is unknown. Although grazing by microzooplankton has been reported as the only controlling factor in a few studies [41, 43], others have found that it is not the sole controlling factor for some algae [48]. Calbet et al. [7], for example, reported that as much as 12% of the decline in biomass during the recession phase of an A. minutum bloom was in addition to the daily loss via microzooplankton grazing. There may be other top–down controls such as viruses [42, 49, 64] or algicidal/algistatic bacteria [15, 25, 29] leading to the decrease in density that was observed.

In addition to H. akashiwo, we also evaluated microzooplankton grazing pressure on naturally occurring blooms of other raphidophyte species common to the DIB: C. subsalsa (45–50 μm, Expt 10 and 12), C. cf. verruculosa (30–40 μm, Expt 10, 12, and 13), and F. japonica (20–24 μm, Expt 13). Microzooplankton grazing specifically on C. subsalsa and C. cf. verruculosa was not detected in our experiments. Grazing pressure on the total phytoplankton communities (determined by chl a) in these experiments was also either minimal (Expts 12 and 13) or undetectable (Expt. 10). This may be the result of dominance by these relatively large raphidophytes. Although Tillmann and Reckermann [66] reported that several mixotrophic dinoflagellates could readily ingest F. japonica and had positive growth rates, in our study, microzooplankton grazing on F. japonica was also not detected. However, we were limited to one experiment in observing microzooplankton grazing pressure on this species. The Chattonella and Fibrocapsa species are much larger than H. akashiwo and were possibly not consumed by micrograzers for this reason.

As a comparison, we also evaluated grazing pressure on other small, non-harmful algal species collected from DIB (Fig. 3). When non-harmful small phytoplankton (<20 μm) were dominant, microzooplankton grazing pressure for the <20-μm size fraction was equal to or higher than grazing on the total community. Comparison of grazing rates from these experiments (Expt 6 and 9) to grazing pressure observed on H. akashiwo supports our hypothesis that size selection by microzooplankton community may possibly be affecting raphidophyte dynamics in the DIB.

This study focused on changes in abundance of specific algae (esp. H. akashiwo) within the context of a natural population with unknown species of grazers. However, we observed that large dinoflagellates (>50 μm) generally had high densities at the end of the experiments, which may be due in part to the addition of nutrients. However, several of the large dinoflagellates common to the DIB, such as O. marina [66] and Gyrodinium spp. [38, 68], are capable of mixotrophy and grazing on the smaller phytoplankton including H. akashiwo [33, 37]. Other mixotrophic dinoflagellates of similar size (Noctiluca scintillans [9, 50], Cochlodinium polykrikoides [34], Gonyaulax polygramma [36], Prorocentrum spp., Heterocapsa triquetra, Scrippsiella trochoidea, Gymnodinium impudicum, Alexandrium tamarense, Akashiwo sanguinea, Gymnodinium catenatum, and Lingulodinium polyedrum [35]) and some small heterotrophic dinoflagellates [30, 31] have also been reported to feed on H. akashiwo. The presence of these mixotrophic dinoflagellates could potentially affect our comparison of species-specific QPCR-based grazing rates on H. akashiwo with chl a-based total community grazing rates because these organisms are known to sometimes retain ingested chl a for extended periods of time [33]. Other general groups of microzooplankton that were commonly seen in our experiments were identified as oligotrichous ciliates (15–60 μm), tintinnids (40–60 μm), Strombidium sp. (35–55 μm), rotifers (150–200 μm), barnacle nauplii (200 μm), and copepod nauplii (≥200 μm).

Our results support the hypothesis that removal of H. akashiwo by microzooplankton grazing may provide larger raphidophytes with a competitive advantage in the DIB. When Delaware isolates are compared, H. akashiwo is a better competitor for nutrients than C. subsalsa [74]. Our data suggest that the removal of H. akashiwo from the environment via microzooplankton grazing may be a key factor for the comparatively greater abundance of C. subsalsa in DIB, despite the latter species being competitively inferior with regard to most “bottom–up” control factors. Prey toxicity, however, can also help to determine the feeding preferences of microzooplankton and potentially decrease grazing pressure [61]. Further studies evaluating differences in interactions between toxic and non-toxic strains and higher trophic levels in the natural environment will be necessary to fully understand the bloom dynamics of this species. Other trophic interactions such as grazing on bacteria by H. akashiwo and C. subsalsa [57] can also contribute to the bloom dynamics and need to be considered when evaluating growth and grazing rates. As demonstrated in this article, integration of QPCR with the dilution method offers a valuable species-specific enumeration tool to carry out these investigations, and this approach can be used to study other interactions of harmful algal bloom species with grazers in aquatic ecosystems. In a broader scale, this species-specific method can also be utilized in investigating selected dominant phytoplankton species’ contribution to microbial loop carbon budgets during spring blooms.

Acknowledgments

Authors would like to thank K. Portune and Y. Zhang for laboratory assistance. Financial support was provided by Delaware Sea Grant 235445, EPA STAR ECOHAB R83-1041, EPA ECOHAB R83-3221, NOAA-MERHAB NA04NOS4780240, and the Center for the DIB. We thank D. Kirchman, C. Hare, R. Dale, C. Gobler, and P. Gaffney for their input to the manuscript and also the Delaware Volunteer Phytoplankton Monitoring Group (especially Coordinator Dr. E. Whereat and M. Farestad) for valuable field assistance and field data.

Copyright information

© Springer Science+Business Media, LLC 2007