Journal of Aquatic Animal Health Improvements Are Needed in Reporting of Accuracy Studies for Diagnostic Tests Used for Detectio

This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:12 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Improvements are Needed in Reporting of Accuracy Studies for Diagnostic Tests Used for Detection of Finfish Pathogens Ian A. Gardnera, Timothy Burnleyb & Charles Caraguelc a Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue, Charlottetown, Prince Edward Island C1A 4P3, Canada b Eastern Epidemiological Services, 23 Karen Drive, Cornwall, Prince Edward Island C0A 1H8, Canada c School of Animal and Veterinary Sciences, Roseworthy Campus, University of Adelaide, Roseworthy, South Australia 5371, Australia Published online: 17 Sep 2014.

To cite this article: Ian A. Gardner, Timothy Burnley & Charles Caraguel (2014) Improvements are Needed in Reporting of Accuracy Studies for Diagnostic Tests Used for Detection of Finfish Pathogens, Journal of Aquatic Animal Health, 26:4, 203-209, DOI: 10.1080/08997659.2014.938867 To link to this article: http://dx.doi.org/10.1080/08997659.2014.938867

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:203–209, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.938867

COMMUNICATION

Improvements are Needed in Reporting of Accuracy Studies for Diagnostic Tests Used for Detection of Finﬁsh Pathogens

Ian A. Gardner* Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue, Charlottetown, Prince Edward Island C1A 4P3, Canada Timothy Burnley Eastern Epidemiological Services, 23 Karen Drive, Cornwall, Prince Edward Island C0A 1H8, Canada Charles Caraguel School of Animal and Veterinary Sciences, Roseworthy Campus, University of Adelaide, Roseworthy, South Australia 5371, Australia

Accurate diagnostic tests are central to the detection, control, Abstract and prevention of pathogen spread in aquaculture. The World Indices of test accuracy, such as diagnostic sensitivity and speci- Organization for Animal Health (OIE) specifies that test accu- ficity, are important considerations in test selection for a defined racy in aquatic species should be estimated to assess a test’s purpose (e.g., screening or confirmation) and affect the interpretation of test results. Many biomedical journals recommend that au- “fitness for the intended purpose(s).” Such purposes include the thors clearly and transparently report test accuracy studies follow- demonstration or certification of freedom from disease, the con- ing the Standards for Reporting of Diagnostic Accuracy (STARD) firmation of clinically suspected cases, and the estimation of guidelines (www.stard-statement.org). This allows readers to eval- disease prevalence in a fish population, to list a few. Indices uate overall study validity and assess potential bias in diagnostic of test accuracy, such as diagnostic sensitivity and specificity, sensitivity and specificity estimates. The purpose of the present study was to evaluate the reporting quality of studies evaluating are important considerations when selecting a test for a defined test accuracy for finfish diseases using the 25 items in the STARD purpose and interpreting its results. Regardless of the intended checklist. Based on a database search, 11 studies that included purpose of the test, complete and transparent reporting of the estimates of diagnostic accuracy were identified for independent design elements and the results of a test accuracy study are evaluation by three reviewers. For each study, STARD checklist essential so that readers can appraise the validity of the test, items were scored as “yes,” “no,” or “not applicable.” Only 10 of the 25 items were consistently reported in most (≥80%) pa- including the potential for bias in diagnostic sensitivity and pers, and reporting of the other items was highly variable (mostly specificity estimates.

Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 between 30% and 60%). Three items (“number, training, and ex- The Standards for Reporting of Diagnostic Accuracy pertise of readers and testers”; “time interval between index tests (STARD) statement (www.stard-statement.org) was published and reference standard”; and “handling of indeterminate results, in 2003 to encourage improved reporting of the important fea- missing data, and outliers of the index tests”) were reported in less than 10% of papers. Two items (“time interval between index tests tures of test accuracy studies in human medicine (Bossuyt et al. and reference standard” and “adverse effects from testing”) were 2003). The statement was motivated by widespread evidence of considered minimally relevant to fish health because test samples bias in test accuracy studies and the finding that incomplete re- usually are collected postmortem. Modification of STARD to fit porting or a lack of reporting on items in the STARD checklist finfish studies should increase use by authors and thereby improve was associated with overly optimistic estimates of diagnostic the overall reporting quality regardless of how the study was designed. Furthermore, the use of STARD may lead to the improved sensitivity and specificity (Lijmer et al. 1999). The STARD design of future studies. does not prescribe directly how to design a test accuracy study. Rather, it focuses on clear and transparent reporting of its key

*Corresponding author: [email protected] Received September 24, 2013; accepted May 15, 2014 203 204 GARDNER ET AL.

TABLE 1. Reviewer scoring of STARD checklist items for 11 test accuracy studies in ﬁnﬁsh that were published between 2006 and 2012. The percentage of yes responses was calculated based on a denominator of the total number of yes and no responses. Abbreviations are as follows: NA = item considered not applicable to the reviewed study and NC = not calculated because of a zero denominator.

Reviewer 1 Reviewer 2 Reviewer 3a

Section, topic, Description Yes Yes Yes Median % and total Item of item (%) No NA (%) No NA (%) No NA yes scores

TITLE, ABSTRACT, 1 Identify the article as a study 7 (64) 4 0 7 (64) 4 0 8 (80) 2 0 64 KEYWORDS of diagnostic accuracy (recommend medical subject headings “sensitivity” and “specificity”). INTRODUCTION 2 State the research questions or 11 (100) 0 0 11 (100) 0 0 10 (100) 0 0 100 study aims, such as estimating diagnostic accuracy or comparing accuracy between tests or across participant groups. METHODS Participants 3 The study population: the 11 (100) 0 0 11 (100) 0 0 10 (100) 0 0 100 inclusion and exclusion criteria, setting, and locations where data were collected. 4 Participant recruitment: was 10 (91) 1 0 10 (91) 1 0 9 (90) 1 0 91 recruitment based on the presenting symptoms, the results from previous tests, or the fact that the participants had received the index tests or the reference standard? 5 Participant sampling: was the 9 (82) 2 0 9 (82) 2 0 9 (90) 1 0 82 study population a consecutive series of participants defined by the selection criteria in items 3 and 4? If not, specify how participants were further selected. 6 Data collection: was data 11 (100) 0 0 11 (100) 0 0 10 (100) 0 0 100 Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 collection planned before the index test and reference standard were performed (prospective study) or after (retrospective study)? Test methods 7 The reference standard and its 7 (100) 0 4 9 (82) 0 2 4 (40) 6 0 82 rationale. 8 Technical specifications of 10 (91) 1 0 10 (91) 1 0 10 (100) 0 0 91 materials and methods involved, including how and when measurements were taken, or cite references for index tests and reference standard. (Continued on next page) COMMUNICATION 205

TABLE 1. Continued.

Reviewer 1 Reviewer 2 Reviewer 3a

Section, topic, Description Yes Yes Yes Median % and total Item of item (%) No NA (%) No NA (%) No NA yes scores 9 Definition of and rationale for 6 (55) 5 0 6 (55) 5 0 5 (50) 5 0 55 the units, cutoffs, and categories of the results of the index tests and the reference standard. 10 The number, training, and 0 (0) 11 0 1 (10) 10 0 0 (0) 10 0 0 expertise of the persons executing and reading the index tests and the reference standard. 11 Whether or not the readers of 2 (22) 9 0 2 (22) 9 0 2 (20) 8 0 22 the index tests and reference standard were blind (masked) to the results of the other test and describe any other clinical information available to the readers. Statistical methods 12 Methods for calculating or 6 (55) 5 0 6 (55) 5 0 7 (70) 3 0 55 comparing measures of diagnostic accuracy and the statistical methods used to quantify uncertainty (e.g., 95% confidence intervals). 13 Methods for calculating test 2 (100) 0 9 2 (100) 0 9 2 (50) 2 6 100 reproducibility, if done. RESULTS Participants 14 When study was performed, 5 (56) 4 2 7 (70) 3 1 6 (60) 4 0 60 including beginning and end dates of recruitment. 15 Clinical and demographic 5 (45) 6 0 4 (36) 7 0 5 (50) 5 0 45 characteristics of the study population (at least information on age, gender, and spectrum of presenting symptoms). 16 The number of participants 3 (38) 8 0 3 (38) 8 0 4 (40) 6 0 38 Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 satisfying the criteria for inclusion who did or did not undergo the index tests or the reference standard; describe why participants failed to undergo either test (a flow diagram is strongly recommended). Test results 17 Time interval between the 0 (NC) 0 11 0 (0) 3 8 1 (10) 9 0 5 index tests and the reference standard and any treatment administered in between.

(Continued on next page) 206 GARDNER ET AL.

TABLE 1. Continued.

Reviewer 1 Reviewer 2 Reviewer 3a

Section, topic, Description Yes Yes Yes Median % and total Item of item (%) No NA (%) No NA (%) No NA yes scores

18 Distribution of severity of 1 (10) 10 0 2 (22) 9 0 4 (40) 6 0 22 disease (define criteria) in those with the target condition; other diagnoses in participants without the target condition. 19 A cross tabulation of the 7 (64) 4 0 7 (64) 4 0 8 (80) 2 0 64 results of the index tests (including indeterminate and missing results) by the results of the reference standard; for continuous results, the distribution of the test results by the results of the reference standard. 20 Any adverse events from 0 (NC) 0 11 0 (NC) 0 11 0 (NC) 0 10 NC performing the index tests or the reference standard. Estimates 21 Estimates of diagnostic 6 (55) 5 0 7 (64) 4 0 6 (60) 4 0 60 accuracy and measures of statistical uncertainty (e.g., 95% confidence intervals). 22 How indeterminate results, 1 (9) 10 0 1 (9) 10 0 1 (10) 9 0 9 missing data, and outliers of the index tests were handled. 23 Estimates of variability of 3 (38) 8 0 5 (45) 6 0 5 (50) 5 0 45 diagnostic accuracy between subgroups of participants, readers, or centers, if done. 24 Estimates of test 2 (100) 0 9 2 (100) 0 9 3 (100) 0 7 100 reproducibility, if done. DISCUSSION 25 Discuss the clinical 10 (91) 1 0 10 (91) 1 0 9 (90) 1 0 91 applicability of the study findings. TOTAL 135 94 46 143 92 40 138 89 23 52 Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 (49) (34) (17) (52) (33) (15) (55) (36) (9)

a Reviewer 3 only evaluated 10 studies; he did not review 1 of the studies (Caraguel et al. 2012) because he is an author.

elements, namely the 25 items in its checklist (see description of STARD specifically directed to ruminant paratuberculosis of items in Table 1). The STARD is endorsed by more than 200 represents the first attempt to make STARD more applicable biomedical journals, but to date endorsement by animal health to animal diseases (Gardner et al. 2011) and was based on the journals has been limited. Although STARD principles apply consensus of an expert group with experience in test accuracy broadly to test accuracy studies in all species, the guidelines studies and paratuberculosis. do not account for unique considerations in animal production, The purpose of the present study was to evaluate the reporting such as the use of experimental challenge studies, varied testing quality of test accuracy studies for finfish diseases based on the purposes, different epidemiological units and sampling designs, 25 items in the STARD checklist and to determine the frequency and latent class analysis methods to account for imperfect refer- of and potential reasons for disagreements among reviewers in ence tests (Branscum et al. 2005; Gardner 2010). A modification classification of author reporting. COMMUNICATION 207

METHODS was recorded if the item was considered not relevant to the We searched PubMed, Scopus, Web of Science, and EBSCO reviewed study. HOST for test accuracy studies published in English between The percentage of yes scores for the 25 items was calculated 2006 and 2012. We chose 2006 as the starting year because, for each reviewer and summarized as the median percentage theoretically, authors would have had the opportunity to read across reviewers. A scoring disagreement for an item was de- the STARD guidelines, which were published 3 years earlier, fined as occurring when at least one reviewer provided a different and follow its recommendations in their submission. The search score from the others. strategy included the use of the recommended medical subject headings of sensitivity and specificity for PubMed searches of RESULTS diagnostic accuracy studies (item 1 in the STARD checklist); the use of the term validation, which is often used in reference Description of Studies to the OIE pathway (OIE 2013); the classification of infectious Out of the 11 studies, 10 were for viral infections (including diseases of finfish as viral, bacterial, or parasitic; and a list of three listed by OIE: infectious salmon anemia, viral hemor- some specific diseases and tests used to diagnose them. The fol- rhagic septicemia, and infectious hematopoietic necrosis) and lowing search terms were used alone or in combination: sensitiv- 1 was for the myxosporean parasite Parvicapsula minibicor- ity, specificity, diagnostic, test agreement, validation, accuracy, nis. All studies involved the evaluation of PCR (sometimes reproducibility, repeatability, viral disease, bacterial disease, with other tests) with or without a reference test. In 7 stud- parasitic disease, infectious salmon anemia, viral hemorrhagic ies, tests were evaluated in Atlantic Salmon Salmo salar, and septicemia, infectious hematopoietic necrosis, infectious pan- 4 studies included other species, such as European Sea Bass creatic necrosis, heart and skeletal muscle inflammation, pan- Dicentrarchus labrax (also known as European Bass Morone creas disease, sleeping disease, Kudoa, PCR, RT-PCR (reverse labrax), Gilthead Seabream Sparus auratus, Rainbow Trout On- transcription PCR), ELISA (enzyme-linked immunosorbent as- corhynchus mykiss, Brown Trout Salmo trutta, Chinook Salmon say), IFAT (indirect fluorescent antibody test), virus isolation, Oncorhynchus tshawytscha, and Arctic Char Salvelinus alpinus. fish, and salmon. There were 8 studies that used field samples only, 2 that used Of the 66 papers identified in the initial screening, 54, 6, and experimental samples, and 1 that used both. All studies evalu- 6 were for viral, bacterial, and parasitic diseases, respectively. ated samples collected postmortem. Out of the 11 studies, 2 in- There were 49 papers that were excluded mostly because they cluded the evaluation of the reproducibility of test results among were descriptions of the development and optimization of tests, laboratories (item 13). the application of tests in observational studies, and the use of existing diagnostic tests in new ways (e.g., nonlethal sampling) Completeness of Reporting and Relevance of Items or with modified protocols (e.g., PCR). None of these studies The frequency of reporting items by reviewer and the median had estimates of test accuracy. Out of the 66 studies, 17 were percentage of yes scores are presented in Table 1. Out of the 25 deemed appropriate for further evaluation because they were items, 10 (items 2–8, 13, 24, and 25 in Table 1) were consistently test accuracy studies that used either experimentally challenged reported in most (≥80%) papers, but the reporting of the other fish or samples from either wild or cultured fish populations. Six items was variable and mostly in the 30–60% range. “Blinding studies were subsequently excluded because they were limited of test readers” (item 11) was only reported in 20% of the stud- to the estimation of analytical characteristics (e.g., repeatability, ies. Three items (item 10 [number, training, and expertise of limit of detection) without estimation of diagnostic sensitivity readers and testers], item 17 [time interval between index tests and specificity. and reference standard], and item 22 [handling of indetermi-

Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 The 11 papers that met the inclusion criteria (Knusel¨ et al. nate results, missing data, and outliers of the index tests]) were 2007; Gustafson et al. 2008; McClure et al. 2008; Nerette´ et al. reported in less than 10% of the papers. For item 22 (missing 2008; True et al. 2009; Abayneh et al. 2010; Jansen et al. 2010; data and outliers), the required information was occasionally Garver et al. 2011; Hodneland et al. 2011; Caraguel et al. 2012; in tables of results but was not mentioned explicitly in the text. Jonstrup et al. 2013) were evaluated by two of us (I.G. and T.B.). Reporting of item 20 (adverse effects from performing the index The third reviewer (C.C.) scored 10 out of 11 papers to avoid po- test or reference standard) was not considered to be applicable tential conﬂict of interest because of authorship (Caraguel et al. by reviewers because all testing was done on postmortem sam- 2012). All reviewers were familiar with STARD before doing ples, and two out of the three reviewers also considered that the evaluation. Papers were independently evaluated according item 17 was not applicable for the same reason. to the STARD checklist, and the 25 items (Table 1) were scored as “yes,” “no,” or “not applicable” (NA), depending on whether Variability in Reporting Frequency and Disagreements the authors had reported the pertinent information. For the three among Reviewers items on the checklist that have the qualiﬁer “if done” (“methods Reporting frequency (Table 1), which was calculated using for calculating test reproducibility,” “estimates of reproducibil- only yes and no scores, was similar among reviewers for most ity,” and “estimates of variability of diagnostic accuracy”), NA items (22 out of the 25 items) and ranges were mostly less than 208 GARDNER ET AL.

the unit of interest often is a group of fish rather than individual animals. Groups of fish are clustered within the housing unit (net-pen or tank) and within a farm or site when there are multiple housing units. Hence, the assessment of accuracy may be at the population level rather than the individual fish level, necessitating the inclusion of descriptions of the pen and site and the assessment of whether clinical signs, including elevated mortality, were evident at the population level or not. These population-level components have been added to item 14 for paratuberculosis in ruminants (Gardner et al. 2011), but other STARD items (e.g., items 17 and 20) need modification for finfish because noninvasive detection methods are rarely used and FIGURE 1. Frequency of disagreements among three reviewers in “yes,” “no,” most testing is based on postmortem samples. and “not applicable” classifications of the STARD checklist items (see Table Experimental challenge studies are not used for test accuracy 1) for 11 test accuracy studies in finfish that were published between 2006 and 2012. studies of infectious diseases in humans, but they may be the only source of samples for similar assessments in finfish. Inter- 20%. Ranges of ≥ 50% among reviewers were evident in two estingly, although challenge studies are not explicitly considered items: item 7 (reference standards and its rationale) had a range in STARD, the most relevant items related to experimental an- of 60% and item 13 (methods for calculating test reproducibility, imals (items 3–6) were well reported even in the absence of if done) had a range of 50%. This high variability was due to specific guidelines about critical information to be included in differences in the frequency of yes and NA scores. For example a manuscript. for item 7, one reviewer scored this item as NA in 4 out of the 11 Findings in the present study indicated highly variable report- studies when latent class analysis (Branscum et al. 2005) was ing (0–100%) of the 25 STARD items in the 11 eligible studies, used for estimation of diagnostic sensitivity and specificity when as evaluated by the median percentage of yes scores. Although the reference standard (e.g., virus isolation) was imperfect. The the three reviewers were familiar with STARD, their classifi- other two reviewers recorded this item as NA in 0 and 2 studies. cation of items was not always consistent. Disagreements were The ranking of disagreement percentages for individual items mostly between the no and NA categories and less frequently on a paper-by-paper basis yielded slightly different findings between yes and no categories. Items for which NA scores were (Figure 1), with the highest percentages of disagreements occur- recorded (items 7, 13, 14, 17, 20, and 24) likely require addi- ring for item 17 (time interval between index tests and reference tional explanation or modification. In the present study, we did standard) at 91% and item 7 (reference standard and its ratio- not attempt to reach a consensus of the reviewers using an it- nale) at 64%. For both items, most disagreements were between erative Delphi technique process or use a pseudo-gold standard no and NA scores, indicating confusion about item relevance. (e.g., two of three reviewers agreeing) as the definitive criterion, Across the 11 papers, three items (6, 20, and 25) had the com- and this approach may have yielded different findings. Dis- plete agreement of the three reviewers and another eight items agreements among reviewers not familiar with STARD may be (2–5, 8, 10, 11, and 24) had a single disagreement. greater because of the lack of awareness of the overall purpose of STARD and the reasons for each of the checklist items. Many of the identified reporting deficiencies in the 11 stud- DISCUSSION ies could be easily rectified without substantial increases in

Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 The OIE Manual describes a pathway, including some design manuscript length (e.g., item 1 [identifying the article as a study considerations, to guide the validation of tests for important in- of diagnostic accuracy], item 10 [number, training, and exper- fectious diseases of aquatic animals (OIE 2013). The estimation tise of readers and testers], item 11 [blinding of readers], and of diagnostic sensitivity and specificity is an important step in item 22 [how indeterminate results, missing data, and outliers the pathway. Additional guidance for quantitative PCR tests is of index tests were handled]). Other changes, such as recom- provided in Purcell et al. (2011). The focus of the present study mendations for reporting diagnostic sensitivity and specificity was the quality of the reporting of the test accuracy or valida- estimates from a latent class analysis, description of the unit of tion study rather than its design. Our inclusion criteria restricted interest, and description of the sampling design, may require evaluation to published studies in which test accuracy (diag- additional paragraphs. nostic sensitivity and specificity) was evaluated, irrespective of In summary, a modified version of STARD is needed to whether the OIE validation pathway was followed or not. improve the reporting quality of finfish test validation studies. The original focus of STARD was the reporting quality of Modified guidelines that are appropriate for finfish, and per- test accuracy studies in individual human patients presenting haps for mollusks and crustaceans, would be helpful for authors with clinical symptoms of disease and whose follow-up testing submitting papers and for reviewers and journal editors during depended on the preceding test results. In finfish aquaculture, the review process. Modifications are also needed to address COMMUNICATION 209

data from both experimental and field studies because the for- Gustafson, L., S. Ellis, D. Bouchard, T. Robinson, F. Marenghi, J. Warg, mer may be an important source of samples for the evaluation and C. Giray. 2008. Estimating diagnostic test accuracy for infectious of test accuracy. The methodological quality of the study, as salmon anaemia virus in Maine, USA. Journal of Fish Diseases 31:117– 125. evaluated using the quality assessment for diagnostic accuracy Hodneland, K., R. Garcia, J. A. Balbuena, C. Zarza, and B. Fouz. 2011. Real-time studies (QUADAS) tool guidelines (Whiting et al. 2011), was RT-PCR detection of betanodavirus in naturally and experimentally infected not assessed in the present study because of the small sample fish from Spain. Journal of Fish Diseases 34:189–202. size and because, to our knowledge, there are no published sys- Jansen, M. D., M. A. Wasmuth, A. B. Olsen, B. Gjerset, I. Modahl, O. Breck, R. tematic reviews and meta-analyses of test accuracy in aquatic N. Haldorsen, R. Hjelmeland, and T. Taksdal. 2010. Pancreas disease (PD) in sea-reared Atlantic Salmon, Salmo salar L., in Norway; a prospective, animals. Because reporting is considered to mirror the design of longitudinal study of disease development and agreement between diagnostic studies, the use of STARD for test accuracy studies of aquatic test results. Journal of Fish Diseases 33:723–736. diseases may indirectly lead to the improved design of future Jonstrup, S. P., S. Kahns, H. F. Skall, T. S. Boutrup, and N. J. Olesen. 2013. studies in this field. Development and validation of a novel Taqman-based real-time RT-PCR assay suitable for demonstrating freedom from viral haemorrhagic septicaemia virus. Journal of Fish Diseases 36:9–23. ACKNOWLEDGMENTS Knusel,¨ R., S. M. Bergmann, K. Elner-Jensen, J. Casey, H. Segner, and This research was undertaken, in part, thanks to funding from T. Wahli. 2007. Virus isolation vs RT-PCR: which method is more successful the Canada Excellence Research Chairs Program. in detecting VHSV and IHNV in fish tissue sampled under field conditions? Journal of Fish Diseases 30:559–568. Lijmer, J. G., B. W. Mol, S. Heisterkamp, G. J. Bonsel, M. H. Prins, J. H. P. van REFERENCES der Meulen, and P.M. M. Bossuyt. 1999. Empirical evidence of design-related Abayneh, T., N. Toft, A. B. Mikalsen, E. Brun, and M. Sandberg. 2010. Eval- bias in studies of diagnostic tests. Journal of American Medical Association uation of histopathology, real-time PCR and virus isolation for diagnosis of 282:1061–1066. infectious salmon anaemia in Norwegian salmon using latent class analysis. McClure, C., S. Saksida, G. Karreman, J. Constantine, J. Robinson, G. Traxler, Journal of Fish Diseases 33:529–532. and L. Hammell. 2008. Evaluation of a reverse transcriptase polymerase chain Bossuyt, P. M., J. B. Reitsma, D. E. Bruns, C. A. Gatsonis, P. P. Glasziou, L. reaction test and virus isolation on field samples collected for the diagnosis of M. Irwig, D. Moher, D. Rennie, H. C. deVet, and J. G. Lijmer. 2003. The infectious hematopoietic necrosis virus in cultured Atlantic Salmon in British STARD statement for reporting studies of diagnostic accuracy: explanation Columbia. Journal of Aquatic Animal Health 20:12–18. and elaboration. Annals of Internal Medicine 138:1–12. Nerette,´ P., H. Stryhn, I. Dohoo, and L. Hammell. 2008. Using pseudogold Branscum, A. J., I. A. Gardner, and W. O. Johnson. 2005. Estimation of standards and latent-class analysis in combination to evaluate the accuracy of diagnostic-test sensitivity and specificity through Bayesian modeling. Pre- three diagnostic tests. Preventive Veterinary Medicine 85:207–225. ventive Veterinary Medicine 68:145–163. OIE (World Organisation for Animal Health). 2013. Manual of Diagnostic Caraguel, C., H. Stryhn, N. Gagne,´ I. Dohoo, and L. Hammell. 2012. Use Tests for Aquatic Animals 2013. Chapter 1.1.2 - Principles and meth- of a third class in latent class modelling for the diagnostic evaluation of ods of validation of diagnostic assays for infectious diseases. Avail- five infectious salmon anaemia virus detection tests. Preventive Veterinary able at www.oie.int/international-standard-setting/aquatic-manual/access- Medicine 104:165–173. online/(accessed April 28, 2014). Fleiss, J. L., B. Levin, and M. C. Paik. 2003. Statistical methods for rates and Purcell, M. K., R. G. Getchell, C. A. McClure, and K. A. Garver. 2011. proportions. 3rd edition, Wiley, New York. Quantitative polymerase chain reaction (PCR) for detection of aquatic an- Gardner, I. A. 2010. Quality standards are needed for reporting of test accuracy imal pathogens in a diagnostic laboratory setting. Journal of Aquatic Animal studies for animal diseases. Preventive Veterinary Medicine 97:136–143. Health 23:148–161. Gardner, I. A., S. S. Nielsen, R. J. Whittington, M. T. Collins, D. Bakker, B. True, K., M. K. Purcell, and J. S. Foott. 2009. Development and validation of Harris, S. Sreevatsan, J. E. Lombard, R. Sweeney, D. R. Smith, J. Gavalchin, a quantitative PCR to detect Parvicapsula minibicornis and comparison to and S. Eda. 2011. Consensus-based reporting standards for diagnostic test histologically ranked infection of juvenile Chinook salmon, Oncorhynchus accuracy studies for paratuberculosis in ruminants. Preventive Veterinary tshawytscha (Walbaum), from the Klamath River, USA. Journal of Fish Dis- Medicine 101:18–34. eases 32:183–192. Garver, K. A., L. M. Hawley, C. A. McClure, T. Schroeder, S. Aldous, F. Whiting, P. F., A. W. S. Rutjes, M. E. Westwood, S. Mallett, J. J. Deeks, J. Downloaded by [Department Of Fisheries] at 00:12 17 November 2014 Doig, M. Snow, S. Edes, C. Baynes, and J. Richard. 2011. Development and B. Reitsma, M. M. G. Leeflang, J. A. C. Sterne, P. M. M. Bossuyt, and the validation of a reverse transcription quantitative PCR for universal detection QUADAS-2 Group. 2011. QUADAS-2: a revised tool for the quality assess- of viral hemorrhagic septicemia virus. Diseases of Aquatic Organisms 95: ment of diagnostic accuracy studies. Annals of Internal Medicine 155:529– 97–112. 536. This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:13 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Chronic Pathology and Longevity of Drepanocephalus spathans Infections in Juvenile Channel Catfish Matt J. Griffina, Lester H. Khooa, James M. Steadmana, Cynthia Warea, Sylvie M. Quinioub, Charles C. Mischkec, Terrence E. Greenwayc & David J. Wisec a Thad Cochran National Warmwater Aquaculture Center, Aquatic Research and Diagnostic Laboratory, College of Veterinary Medicine, Mississippi State University, Post Office Box 197, Stoneville, Mississippi 38776, USA b U.S. Department of Agriculture, Agricultural Research Service, Thad Cochran National Warmwater Aquaculture Center, Warmwater Aquaculture Research Unit, 127 Experiment Station Road, Post Office Box 38, Stoneville, Mississippi 38776, USA c Thad Cochran National Warmwater Aquaculture Center, Mississippi Agriculture and Forestry Experiment Station, Delta Research and Extension Center, 127 Experiment Station Road, Post Office Box 197, Stoneville, Mississippi 38776, USA Published online: 17 Sep 2014.

To cite this article: Matt J. Griffin, Lester H. Khoo, James M. Steadman, Cynthia Ware, Sylvie M. Quiniou, Charles C. Mischke, Terrence E. Greenway & David J. Wise (2014) Chronic Pathology and Longevity of Drepanocephalus spathans Infections in Juvenile Channel Catfish, Journal of Aquatic Animal Health, 26:4, 210-218, DOI: 10.1080/08997659.2014.938869 To link to this article: http://dx.doi.org/10.1080/08997659.2014.938869

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:210–218, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.938869

ARTICLE

Chronic Pathology and Longevity of Drepanocephalus spathans Infections in Juvenile Channel Catﬁsh

Matt J. Griffin,* Lester H. Khoo, James M. Steadman, and Cynthia Ware Thad Cochran National Warmwater Aquaculture Center, Aquatic Research and Diagnostic Laboratory, College of Veterinary Medicine, Mississippi State University, Post Office Box 197, Stoneville, Mississippi 38776, USA Sylvie M. Quiniou U.S. Department of Agriculture, Agricultural Research Service, Thad Cochran National Warmwater Aquaculture Center, Warmwater Aquaculture Research Unit, 127 Experiment Station Road, Post Office Box 38, Stoneville, Mississippi 38776, USA Charles C. Mischke, Terrence E. Greenway, and David J. Wise Thad Cochran National Warmwater Aquaculture Center, Mississippi Agriculture and Forestry Experiment Station, Delta Research and Extension Center, 127 Experiment Station Road, Post Office Box 197, Stoneville, Mississippi 38776, USA

Abstract Drepanocephalus spathans (Digenea: Echinostomatidae) is a common parasite of the double-crested cormorant Phalacrocorax auritus. The cercariae of D. spathans have been shown infective to juvenile Channel Catfish Ictalurus punctatus. The developing metacercariae concentrate in the cranial regions, often occluding blood vessels at the base of the branchial arch, occasionally resulting in death. The purpose of this study was to determine how long metacercariae of D. spathans persist in experimentally challenged Channel Catfish. Two separate infectivity trials were conducted. In both trials, metacercariae persisted at least 49 d postinfection, although prevalence and intensity of infection decreased over time. In the first trial, juvenile catfish (1–3 g) were exposed over three consecutive days to 100, 100, and 80 cercariae/fish/d, respectively. Fish were sampled 7 d after the final exposure, and metacercariae were observed in 83.3% (five of six) of challenged fish. At 21 d postexposure, metacercariae were present in only 50% of exposed fish (three of six). No metacercaria were observed in fish sampled at 35 d, however, metacercariae were present in one of six (16.7%) fish sampled 49 and 70 d postexposure, respectively. A second challenge consisted of a 24-h pooled exposure of 500 cercariae per fish. Again, metacercariae were present in most (six of seven; 85.7%) fish Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 at 7 d postexposure. At 21 d postexposure, metacercariae were only evident in one of seven (14.3%) sampled fish. No metacercariae were present in any fish at 35 d postchallenge, yet one of seven (14.3%) fish was positive at 49 d postchallenge. The second study was terminated at 63 d postchallenge, as all fish sampled (n = 14) were negative for metacercariae. These data suggest that cercariae of D. spathans are infective to juvenile Channel Catfish, although the infection appears short lived as metacercariae rarely persisted longer than 2 months.

Trematodes of the subclass Digenea are commonly associ- great egret Ardea alba, the great blue heron Ardea herodias, ated with catﬁsh aquaculture in the southeastern United States the American white pelican Pelecanus erythrorhynchos, and the (Overstreet and Curran 2004). The readily available source of double-crested cormorant Phalacrocorax auritus. In addition, ﬁsh attracts large numbers of piscivorous birds, namely the unless routine control measures are employed, most operations

*Corresponding author: grifﬁ[email protected] Received February 14, 2014; accepted May 30, 2014 210 DREPANOCEPHALUS SPATHANS IN CHANNEL CATFISH 211

have an abundance of potential snail intermediate hosts (Glahn used for this study. Snails shedding other trematode cercariae, and King 2004; Wise et al. 2004). As a result, digeneans are or harboring mixed infections, were excluded from the ex- commonly associated with catfish aquaculture and have been posure pool. Prior to challenge, all water was removed from linked to significant production losses since the discovery of each snail vial and replaced with fresh pond water. Snails Bolbophorus damnificus (Digenea: Bolbophoridae) in the late actively shedding cercariae of D. spathans were incubated 1990s (Venable et al. 2000; Levy et al. 2002; Overstreet et al. overnight at 27◦C. Following overnight incubation, all newly 2002; Terhune et al. 2002; Wise et al. 2008b). released cercaria were pooled and counted prior to challenge, The trematode Drepanocephalus spathans (Digenea: ensuring all cercariae used in fish exposures were less than Echinostomatididae) is a common intestinal parasite of the 24-h-old. double-crested cormorant (Flowers et al. 2004; Robinson et al. Longevity trial: 2012.—Fish were challenged in six 3.8-L 2008; Wagner et al. 2012). The parasite is highly prevalent containers (three challenge, three control) holding 3-L filtered within cormorant populations in the catfish farming region of pond water supplied with constant aeration. Twenty juvenile the Mississippi Delta, evident by a recent survey, which found D. Channel Catfish (1–3 g) were placed in each container. Fish spathans in 91% of birds sampled (O’Hear et al. 2014). More- were challenged with 100, 100, and 80 cercariae per fish over 3 over, recent research has shown the cercariae of D. spathans, d, respectively. Each day, cercariae were pooled, counted, and released by the marsh rams-horn snail Planorbella trivolvis, added to each challenge container. After 24 h, a complete water are infective to juvenile Channel Catfish Ictalurus punctatus. exchange was performed before adding new cercariae. On the In experimental challenges, metacercariae were observed in the fourth day, fish from each group were transferred to six differ- cranial region of infected fish, demonstrating an affinity for the ent aquaria and maintained in 20-L well water (27◦C), under branchial chamber, often found within, or adjacent to, blood flow-through conditions (1 L/min), with constant aeration. To vessels of the branchial arch. Metacercariae within or adjacent minimize growth, feeding consisted of a maintenance ration of to blood vessels may inhibit the fish’s ability to carry out nor- commercial feed delivered once weekly. Two fish were ran- mal physiological processes and in heavy infections can result domly sampled from each aquarium on 7, 21, 35, 49, and 70 d in death (Griffin et al. 2012). This would suggest D. spathans after the last challenge dose was administered. may have limiting effects on Channel Catfish production in the Longevity trial: 2013.—Fish were challenged in eight 3.8-L Mississippi Delta, similar to what has been observed for other containers (seven challenge, one control) holding 3-L filtered digeneans (Labrie et al. 2004; Overstreet and Curran 2004; Wise pond water with constant aeration. Twenty juvenile Channel et al. 2008b). Catfish (1–3 g) were placed in each container. Fish were chal- There is substantial variability in how long trematode metac- lenged with 500 cercariae per fish for 24 h. Prior to challenge, ercariae persist in fish hosts, ranging from months to years (Fis- cercariae were pooled, counted, and added to each challenge chthal 1949; Hoffman 1950, 1956, 1958; Hoffman and Putz container. After 24 h, fish from each group were transferred, 1965; Lemly and Esch 1983; Ferguson et al. 2010). In Channel respectively, to aquaria (seven challenge tanks, one control) and Catfish, metacercariae of Clinostomum marginatum were found maintained as above. One fish was randomly sampled from each 45 months postinfection, while those of B. damnificus have been aquarium at 7, 21, 35, and 49 d postchallenge. Two fish were observed for up to 30 months following experimental challenge sampled from each aquarium at 63 d postchallenge to determine (Mitchell et al. 2011). The purpose of this study was to determine if the trial needed to continue. how long metacercariae of D. spathans persist in experimentally Histological analysis.—All fish were initially fixed in a 10% challenged Channel Catfish and how long lesions are associated solution of neutral buffered formalin for at least 24 h. Tissues infections. were then decalcified in Cal-Ex solution (Fisher Scientific, Fair

Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 Lawn, New Jersey) until the calcified portions were pliable upon manipulation. Decalcified tissues were then rinsed in running METHODS water for a minimum of 3 h before being trimmed. Fish were Snail collection.—In 2012 and again in 2013, marsh rams trimmed in ∼5-mm cross sections starting rostrally and moving horn snails (n = 800) were collected from commercial catfish caudally, with small adjustments to thickness to ensure all major ponds with large numbers of snails. Snails were individually organs were sampled. Thinner sections were made for the head placed in glass scintillation vials with 10-mL filtered pond wa- region to ensure multiple sections through the branchial cham- ter and incubated overnight at 27◦C. All pond water used in ber, identified in previous experiments as a predilection site for this study was collected from an enclosed reservoir pond at the metacercariae. Tissues were processed routinely, embedded the Thad Cochran National Warmwater Aquaculture Facility in paraffin, sectioned at approximately 4 µm, stained with hema- in Stoneville, Mississippi, and passed through a 20-micron ny- toxylin and eosin (HE), and examined using light microscopy. lon mesh prior to use. The echinostomatid xiphidiocercariae of Tissue sections were examined with a BX-50 Olympus mi- Drepanocephalus spathans were identified based on the char- croscope (Olympus Optical, Tokyo, Japan), and representative acteristic morphology and motility described by Griffin et al. images were captured with an Olympus DP72 camera and DP2- (2012). Only snails shedding cercariae of D. spathans were TWAIN–cellSens software (Olympus Optical). 212 GRIFFIN ET AL.

Molecular identification.—Morphological identification was purified directly using the QIAquick columns (Qiagen). Purified confirmed by sequencing the cytochrome c oxidase subunit 1 PCR products were then sequenced directly, from both strands, (CO1) gene. Briefly, for each infectivity trial, genomic DNA using ABI BigDye chemistry (Applied Biosystems, Foster City, was isolated from three separate pools of five cercariae using California), alcohol precipitated, and run on an ABI Prism 3730 a commercial DNA isolation kit, following the manufacturer’s automated sequencer (Applied Biosystems). suggested protocol for animal tissues and eluted in 100 µLof Contiguous sequences were assembled using the correspond- elution buffer (DNEasy Blood and Tissue Kit; Qiagen, Valencia, ing chromatograms and the SeqMan utility of the Lasergene California). software package (DNAStar, Madison, Wisconsin). Obtained Initial amplification of the CO1 fragment was performed sequences were compared with published CO1 sequences of D. using the forward primer, JB3 (5-TTT TTT GGG CAT CCT spathans deposited in GenBank (JN993267; JN993266; Griffin GAG GTT TAT-3), and the reverse primer, JB13 (5-TCA TGA et al. 2012). AAA CAC CTT AAT ACC-3; Bowles et al. 1995; Morgan µ and Blair 1998). The 50- L amplification reactions consisted of RESULTS 25 µL EconoTaq PLUS GREEN 2 × master mix, 20 pmol each primer, 5 µL resuspended gDNA, and nuclease-free water to Longevity Trial: 2012 volume. Thermal cycling conditions consisted of an initial de- Minimal mortalities (10/60 fish; 16.7%) occurred sporadi- naturation of 95◦C for 5 min; 45 cycles of 95◦C for 30 s, 48◦C cally throughout the study, with most deaths occurring in the for 30 s, 72◦C for 45 s; and a final extension of 72◦Cfor5min. first 7 d postchallenge. It was assumed post mortem decompo- Aliquots from each amplification reaction were visualized af- sition and tankmate cannibalism would confound subsequent ter electrophoretic migration through a 1.5% agarose gel con- histopathological analysis. As such, dead fish were not exam- taining ethidium bromide (0.5 µg/mL) to confirm the presence ined. At the first sampling, 7 d postexposure, metacercariae of appropriate-sized bands (∼260 bp). The PCR products were were observed in five of six (83.3%) challenged fish. Encysted Downloaded by [Department Of Fisheries] at 00:13 17 November 2014

FIGURE 1. Bilateral metacercariae (arrows) surrounded by a mononuclear inflammatory infiltrate within the medial portion of the branchial chamber (7 d postchallenge). Note the inflammatory infiltrate here is poorly organized at this stage (bar ∼200 µm). [Color figure available online.] DREPANOCEPHALUS SPATHANS IN CHANNEL CATFISH 213

FIGURE 2. Metacercaria (arrow) at the base of the gill arch (7 d postchallenge; bar ∼200 µm). [Color ﬁgure available online.]

metacercariae were limited to the cranial region, infecting the tric rings of fibrocytes and flattened epithelioid macrophages, branchial chamber (medial surface at the level of the esophagus; and had a necrotic core with no metacercaria present. At 70 d Figure 1), the base of the gill arch (Figure 2), and the subdermal postchallenge, one fish had mature granulomas at the gill base periocular tissue. These encysted parasites were often accom- (Figure 5). These granulomas, like those described above, were panied by a mild mononuclear inflammatory infiltrate. well circumscribed with concentric rings of fibrocytes, epith- At 21 d postexposure, metacercariae were observed in three eloid macrophages, and other mononuclear inflammatory cells of six (50%) challenged fish. These encysted metacercariae were surrounding the metacercaria or necrotic debris. No signs of Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 in apposition to the external surface of the neurocranium (frontal parasitism were noted in any other challenged fish. At this point, bone) and just beneath the somatic musculature and dermis the study was terminated. Metacercariae were not observed in (Figure 3) as well as in the subdermal tissues just below the any control fish, and no control fish died during the study. operculum at the base of the cleithrum. They were also present at the base of the gill arch and the thin layer of musculature on the inner surface of the urohyal bone at the base of the Longevity Trial: 2013 branchial chamber. Again, these parasites were often accom- Again, mortalities (32/140; 22.9%) occurred sporadically panied by a minimal-to-mild mononuclear inflammatory infil- throughout the study, although most deaths occurred in the trate. No metacercariae were observed in any fish sampled 35 d first 7 d postchallenge. Similar to the earlier challenge, dead postchallenge. At 49 d postchallenge, one of six (16.7%) chal- fish were not examined due to poor post mortem condition. lenged fish had an encysted metacercaria in the suborbital bone At 7 d postchallenge metacercaria were present in six of at the dorsal surface close to the eye (Figure 4). One fish also had seven (85.7%) challenged fish. Developing metacercariae were a mature, well-organized granuloma attached to the outer surface mainly limited to the head (gill base, subdermal, or within of the bulbus arteriosus. This granuloma consisted of concen- bone [intraosseous], and also submucosal in the pharyngeal 214 GRIFFIN ET AL.

FIGURE 3. Metacercaria (arrow) in apposition to the external surface of the neurocranium (frontal bone) and just beneath the somatic musculature and dermis near the eye with minimal associated inﬂammatory response (21 d postchallenge). The adjacent empty cystic structure is indicative of a metacercaria that has resolved or was lost during processing (bar ∼200 µm). [Color ﬁgure available online.]

region). A mild-to-moderate mononuclear inflammatory to determine if the trial needed to proceed. No metacercariae infiltrate was associated with these developing metacercariae, or resolving granulomas were observed in any challenged fish which was sometimes more organized and included epithelioid (n = 14). At this point, the trial was terminated. As before, macrophages to form well-circumscribed granulomas. At 21 d metacercariae were not observed in any control fish, and no postchallenge, developing metacercariae were evident in only control fish died during the study. one (14.2%) experimentally challenged fish. Two metacercariae

Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 were observed in the cranial region together with a mild mononuclear inflammatory infiltrate. There were no metacercariae in any fish sampled 35 d Molecular Identification postchallenge. However, one fish had a granuloma within the The CO1 sequences obtained from cercariae used in these tissue adjacent to the branchial cavity, at the level of the heart. challenges matched (261/261 bp; 100%) the CO1 sequence from Concentric rings of inflammatory cells had filled in the central an adult specimen of D. spathans collected from a double- portion of the granuloma, which appeared to be resolving. There crested cormorant in Yazoo County, Mississippi (GenBank was no metacercaria within this granuloma. At 49 d postchal- JN993266; Griffin et al. 2012). These sequences were also lenge, one metacercaria was observed in one of seven (14.2%) highly similar (260/261 bp; 99.7%) to COI sequences from sampled fish. The metacercaria in this fish appeared to be intact cercaria released by marsh rams horn snails collected from a (nondisrupted tegument and internal structures) and within a commercial catfish operation in Humphrey’s County, Missis- mature granuloma, which was in the space between the olfac- sippi (GenBank JN993267; Griffin et al. 2012). The next most tory lobe and the cartilaginous portion of the skull, at the level similar sequence (221/258; 85.6%) available for comparison of the nares. Based on the observations in experimental trial 1, was from an unidentified Echinostoma sp. from Australia (Gen- two fish were sampled from each tank at 63 d postchallenge Bank AF025827; Morgan and Blair 1998). DREPANOCEPHALUS SPATHANS IN CHANNEL CATFISH 215

FIGURE 4. Metacercaria within the suborbital bone with minimal inﬂammatory response (49 d postchallenge; bar ∼200 µm). [Color ﬁgure available online.]

DISCUSSION persist in the catfish host as long as C. marginatum or B. damnifi- These data support previous findings indicating D. spathans cus, on rare instances the metacercariae were observed as late cercariae are infective to juvenile Channel Catfish, although this as 70 d postchallenge. During early stages of infection, encap- infection appears to be short lived. In both trials, multiple metac- sulated metacercariae and the associated inflammatory process ercariae were observed in challenged fish, confined mainly to were observed in many tissues; however, as the trials progressed, the subdermal connective tissues and muscles around the head, metacercaria were limited to the bones or cartilage of the neu- in close apposition or within the bones of the neurocranium, or rocranium, suggesting bone or cartilage may provide protection in tissues at the base of the gill arch. As the study continued, the against host defenses as immunoprivileged sites, with metacer- Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 number of positive fish, as well as the number of metacercariae, cariae developing in other tissues exposed to the hosts responses decreased. In both studies, parasite presence was limited to just and more readily broken down. 5 weeks postchallenge and by 9 weeks postchallenge was nearly Predilection to the fish host cranial region has been reported nonexistent. for several digenetic trematodes infecting fish. Encystment in The longevity of trematode metacercariae has been well stud- the luminal surfaces of the buccal cavity and the gill arch is not ied for several parasite species and their respective fish hosts limited to D. spathans (Anderson and Cribb 1994; Olsen and (Fischthal 1949; Hoffman 1950, 1956, 1958; Hoffman and Putz Pierce 1997; Velez-Hern´ andez´ et al. 1998; Mitchell et al. 2000), 1965; Lemly and Esch 1983; Ferguson et al. 2010; Boerlage and neither is encystment in the cranial cavity. Metacercaria et al. 2013). In general, metacercaria of digeneans persist for have been reported from the brian ventricles (Mueller 1972); extended periods of time, maximizing the chance for transmis- the interlobular infoldings, ventricular cavities, and external sion to the definitive host by extending the availability to op- meninges (Siegmund et al. 1997); the cerebellum, medulla portunistic predation. In Channel Catfish, the metacerceriae of oblongata, and optic lobes (Barber and Crompton 1997; Dezfuli C. marginatum and B. damnificus persist for several years (Ed- et al. 2007); the cranial fluid surrounding the brain (Johnston ney 1940; Mitchell et al. 2011). Although D. spathans does not and Halton 1981); the neural tissue within and on the surface of 216 GRIFFIN ET AL.

FIGURE 5. Two mature granulomas, close to the base of the gills, one of which contains a metacercaria (70 d postchallenge); this illustrates resolving infection (bar ∼500 µm). [Color ﬁgure available online.]

the brain (Heckman 1992); tissues of the optic tecta and valvula 2008b). This study demonstrates the pathology associated with cerebelli (Sandland and Goater 2001); and the adipose tissue D. spathans infection is primarily limited to the acute stages surrounding the brain (Zhokhov et al. 2010). However, no of infection. Consistent with previous findings (Griffin et al. reports list encystment in the neurocranial bone and cartilage, 2012), mortality occurred in fish exposed to D. spathans, while or gill arch base, as reported previously for D. spathans (Griffin no control fish died. However, most mortality occurred within et al. 2012) and observed here. Moreover, there are limited the first 7 d postexposure and the infection appears to start re- reports of intraosseus metacercariae development in fish. Taylor solving as early as 7-21 d postinfection. This would indicate the

Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 et al. (1994) reported associations between the metacercariae observed mortality is associated with acute stages of infection of Apophallus brevis and ectopic ossicles in Yellow Perch with D. spathans and removing the source of infection results Perca flavescens. Similarly, Kent et al. (2004) identified an in recovery of infected fish. This has been demonstrated for Apophallus sp. causing vertebral deformities in cyprinid other parasites associated with farm-raised catfish (Wise et al. fishes in the Willamette River, Oregon, noting the parasite 2008a, 2013). As a result, reducing or limiting the intermediate elicited proliferation of existing bone and actually penetrated snail host in production systems, similar to management prac- vertebrae. Although not unique, intraosseus development of the tices employed to control B. damnificus on farm-raised catfish, metacercariae of D. spathans and predilection to the bones of should also limit the impact of D. spathans on catfish aquacul- the neurocranium and the tissues at the base of the gill arch ture (Wise et al. 2004, 2006). appears to be uncommon among digeneans reported from fish. The Channel Catfish is an important aquaculture species in the southeastern United States, with significant production ACKNOWLEDGMENTS losses attributed to digeneans (Venable et al. 2000; Labrie et al. The authors would like to thank Holly Whitehead and 2004; Overstreet and Curran 2004; Wise et al. 2004; Wise et al. Kyle Christopher for their technical assistance. This work was DREPANOCEPHALUS SPATHANS IN CHANNEL CATFISH 217

supported by the U.S. Department of Agriculture Catfish Health Kent, M. L., V. G. Watral, C. M. Whipps, M. E. Cunningham, C. D. Criscione, Initiative (CRIS 6402-31320-002-02), the Mississippi State J. R. Heidel, L. R. Curtis, J. Spitsbergen, and D. F. Markle. 2004. A digenean University College of Veterinary Medicine, and the Mississippi metacercaria (Apophallus sp.) and a myxozoan (Myxobolus sp.) associated with vertebral deformities in cyprinid fishes from the Willamette River, Ore- Agriculture and Forestry Experiment Station (MAFES). This is gon. Journal of Aquatic Animal Health 16:116–129. MAFES publication J-12517. Reference to trade names does Labrie, L., C. Komar, J. Terhune, A. Camus, and D. Wise. 2004. Effect of not imply endorsement by the U.S. Government. sublethal exposure to the trematode Bolbophorus spp. on the severity of enteric septicemia of catfish in Channel Catfish fingerlings. Journal of Aquatic Animal Health 16:231–237. REFERENCES Lemly, D. A., and G. W. Esch. 1983. Differential survival of metacercariae Anderson, G. R., and T. H. Cribb. 1994. Five new didymozoid trematodes of Uvelifer ambloplitis (Huges 1927) in juvenile centrarchids. Journal of (Platyhelminthes, Digenea) from Australian platycephalid fishes. Zoologica Parasitology 69:746–749. Scripta 23:83–93. Levy, M. G., J. R. Flowers, M. F. Poore, J. E. Mullen, L. H. Khoo, L. M. Pote, Barber, I., and D. W. T. Crompton. 1997. The distribution of the metacercariae I. Paperna, R. Dzikowski, and R. W. Litaker. 2002. Morphologic, pathologic, of Diplostomum phoxini in the brain of minnows, Phoxinus phoxinus. Folia and genetic investigations of Bolbophorus species affecting cultured Channel Parasitologica 44:19–25. Catfish in the Mississippi Delta. Journal of Aquatic Animal Health 14:235– Boerlage, A. S., E. A. M. Graat, J. A. Verreth, and M. C. M. De Jong. 2013. 246. Survival of heterophyid metacercaria in common carp (Cyprinus carpio). Mitchell, A. J., M. J. Salmon, D. G. Huffman, A. E. Goodwin, and T. M. Brandt. Parasitology Research 112:2759–2762. 2000. Prevalence and pathogenecity of a heterophyid trematode infecting the Bowles, J., D. Blair, and D. P. McManus. 1995. A molecular phylogeny of the gills of an endangered fish, the Fountain Darter, in two central Texas spring- human schistosomes. Molecular Phylogenetics and Evolution 4:103–109. fed rivers. Journal of Aquatic Animal Health 12:283–289. Dezfuli, B. S., S. Capuano, E. Simoni, L. Giari, and A. P. Shinn. 2007. Mitchell, A., M. Yost, L. Pote, B. Farmer, and C. Panuska. 2011. Longevity Histopathological and ultrastructural observations of metacercarial infections of Bolbophorus damnificus infectious in Channel Catfish. Journal of Aquatic of Diplostomum phoxini (Digenea) in the brain of minnows Phoxinus phoxi- Animal Health 23:103–109. nus. Diseases of Aquatic Organisms 29:51–59. Morgan J. A. T., and D. Blair. 1998. Relative merits of nuclear ribosomal internal Edney, J. M. 1940. The longevity of Clinostomum marginatum metacercariae transcribed spacers and mitochondrial CO1 and ND1 genes for distinguishing in Channel Catfish. Journal of the Tennessee Academy of Science 15:380. among Echinostoma species (Trematoda). Parasitology 116:289–297. Ferguson, J. A., C. B. Schreck, R. Chitwood, and M. L. Kent. 2010. Per- Mueller, J. 1972. Cranial deformation in a fish resulting from intracra- sistence of infection by metacercariae of Apophallus sp., Neascus sp., and nial parasitism by strideid metacercariae. Journal of Parasitology 58: Nanophyetus salmincola plus two myxozoans (Myxobolus insidiosus and 183–184. Myxobolus fryeri) in Coho Salmon Oncorhynchus kisutch. Journal of Para- O’Hear, M., L. Pote, M. Yost, C. Doffitt, T. King, and C. Panuska. 2014. sitology 96:340–347. Mophologic and molecular identifications of digenetic trematodes in double- Fischthal, J. H. 1949. The over-wintering of black grubs and yellow grubs in crested cormorants (Phalocrocorax auritus) from the Mississippi Delta, USA. fish. Journal of Parasitology 29:350–353. Journal of Wildlife Diseases 50:42–49. Flowers, J. R., M. F. Poore, J. E. Mullen, and M. G. Levy. 2004. Digeneans Olsen, R. E., and J. R. Pierce. 1997. A trematode metacercaria causing gill collected from piscivorous birds in North Carolina, USA. Comparative Para- cartilage proliferation in steelhead trout from Oregon. Journal of Wildlife sitology 71:243–244. Diseases 33:886–890. Glahn, J. F., and D. T. King. 2004. Bird depredation. Pages 503–529 in C. S. Overstreet, R. M., and S. S. Curran. 2004. Defeating diplostomoid dangers in Tucker and J. A. Hargreaves, editors. Biology and culture of the Channel USA catfish aquaculture. Folia Parasitologica 51:153–165. Catfish. Elsevier, Amsterdam. Overstreet, R. M., S. S. Curran, L. M. Pote, D. T. King, C. K. Blend, and W. Griffin, M. J., L. H. Khoo, S. M. Quiniou, M. M. O’Hear, L. M. Pote, T. E. D. Grater. 2002. Bolbophorus damnificus n. sp. (Digenea: Bolbophoridae) Greenway, and D. J. Wise. 2012. Genetic sequence data identifies the cercaria from the Channel Catfish Ictalurus punctatus and American white pelican of Drepanocephalus spathans (Digenea: Echinostomatidae), a parasite of the Pelecanus erythrorhynchos in the USA based on life-cycle and molecular double-crested cormorant (Phalocrocorax auritus), with notes on its pathol- data. Systematic Parasitology 52:81–96. ogy in juvenile Channel Catfish (Ictalurus punctatus). Journal of Parasitology Robinson, S. A., M. R. Forbes, C. E. Hebert, and J. D. McLaughlin. 2008. Male- 98:967–972. biased parasitism by common helminthes is not explained by sex differences Heckman, R. A. 1992. Host records and tissue locations for Diplostomum mor- in body size or spleen mass of breeding cormorants Phalacrocorax auritus. Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 dax (Metacercariae) inhabiting the cranial cavity of fishes from Lake Titicaca, Journal of Avian Biology 39:272–276. Peru. Journal of Parasitology 78:541–543. Sandland, G. J., and C. P. Goater. 2001. Parasite-induced variation in host Hoffman, G. L. 1950. Studies on the development and biology of two trematodes, morphology: brain-encysting trematodes in Fathead Minnows. Journal of Bunodera eucaliae and Posthodiplostomum minimum. Doctoral dissertation. Parasitology 87:267–272. State University of Iowa, Iowa City. Siegmund, I., R. Franjola, and P. Torres. 1997. Diplostomatid metacercariae Hoffman, G. L. 1956. The life cycle of Crassiphiala bulboglossa (Trematoda: in the brain of Silversides from Lake Rinihue,˜ Chile. Journal of Wildlife Strigeida). Development of the metacercaria and cyst, and effect on the fish Diseases 33:362–364. hosts. Journal of Parasitology 42:435–444. Taylor, L. H., B. K. Hall, T. Miyake, and D. K. Cone. 1994. Ectopic ossicles Hoffman, G. L. 1958. Experimental studies on the cercaria and metacercaria of associated with metacercariae of Apophallus brevis (Trematoda) in Yellow a strigeoid trematode, Posthodiplostomum minimum. Experimental Parasitol- Perch, Perca flavescens (Teleostei): development and identification of bone ogy 7:23–50. and chondroid bone. Anatomy and Embryology 190:29–46. Hoffman, G. L., and R. E. Putz. 1965. The black-spot (Uvulifer ambloplitis: Terhune, J. S., D. J. Wise, and L. H. Khoo. 2002. Bolbophorus confuses in- Trematoda: Strigeoidea) of centrarchid fishes. Transactions of the American fections in Channel Catfish in northwestern Mississippi and effects of water Fisheries Society 94:143–151. temperature on emergence of cercariae from infected snails. North American Johnston, B. R., and D. W. Halton. 1981. Occurrence of Bucephaloides Journal of Aquaculture 64:70–74. gracilescens metacercariae in three species of gadoid fish. Journal of Fish Velez-Hern´ andez,´ E. M., F. Constantino-Casas, L. J. Garc´ıa-Marquez,´ and D. Biology 18:685–691. Osorio-Sarabia. 1998. Gill lesions in Common Carp, Cyprinus carpio L., in 218 GRIFFIN ET AL.

Mexico due to the metacercariae of Centrocestus formosanus. Journal of Fish Wise, D. J., T. R. Hanson, and C. S. Tucker. 2008b. Farm-level impacts of Bol- Diseases 21:229–232. bophorus infections of Channel Catfish. North American Journal of Aqua- Venable, D. L., A. P. Gaude´ III, and P. L. Klerks. 2000. Control of the trematode culture 70:382–387. Bolbophorus confuses in Channel Catfish Ictalurus punctatus ponds using Wise, D. J., M. H. Li, M. J. Griffin, E. H. Robinson, L. H. Khoo, T. E. Greenway, salinity manipulation and polyculture with Black Carp Mylopharyngodon T. S. Byars, J. R. Walker and C. C. Mischke. 2013. Impacts of Bolbophorus piceus. Journal of the World Aquaculture Society 31:158–166. damnificus (Digenea: Bolbophoridae) on production characteristics of Chan- Wagner, B. A., E. P. Hoberg, C. M. Somers, C. Soos, H. Fenton, and E. J. Jenk- nel Catfish, Ictalurus punctatus, raised in experimental ponds. Journal of the ins. 2012. Gastrointestinal helminth parasites of double-crested cormorants World Aquaculture Society 44:557–564. (Phalacrocorax auritus) at four sites in Saskatchewan, Canada. Comparative Wise, D. J., C. C. Mischke, T. E. Greenway, T. S. Byars, and A. J. Mitchell. Parasitology 79:275–282. 2006. Uniform application of copper sulfate as a potential treatment for Wise, D. J., A. Camus, T. Schwedler, and J. Terhune. 2004. Health management. controlling snail populations in Channel Catfish production ponds. North Pages 482–488 in C. S. Tucker and J. A. Hargreaves, editors. Biology and American Journal of Aquaculture 68:364–368. culture of the Channel Catfish. Elsevier, Amsterdam. Zhokhov, A. E., D. A. Morozova, and A. Tessema. 2010. Trematode metac- Wise, D. J., M. J. Griffin, J. S. Terhune, L. M. Pote, and L. H. Khoo. 2008a. ercariae from the cranial cavity of African catfish Clarias gariepinus Induction and evaluation of proliferative gill disease in Channel Catfish fin- (Burchell 1822) from Lake Tana, Ethiopia. Inland Water Biology 3: gerlings. Journal of Aquatic Animal Health 20:236–244. 160–164. Downloaded by [Department Of Fisheries] at 00:13 17 November 2014 This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:14 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Observations on Side-Swimming Rainbow Trout in Water Recirculation Aquaculture Systems Christopher Gooda, John Davidsona, Christin Kinmana, P. Brett Kenneyb, Grete Bæverfjordc & Steven Summerfelta a The Conservation Fund's Freshwater Institute, 1098 Turner Road, Shepherdstown, West Virginia 25443, USA b Division of Animal and Nutritional Sciences, West Virginia University, Morgantown, West Virginia 26506-6108, USA c Nofima Marin, Sjølseng N-6600, Sunndalsøra, Norway Published online: 17 Sep 2014.

To cite this article: Christopher Good, John Davidson, Christin Kinman, P. Brett Kenney, Grete Bæverfjord & Steven Summerfelt (2014) Observations on Side-Swimming Rainbow Trout in Water Recirculation Aquaculture Systems, Journal of Aquatic Animal Health, 26:4, 219-224, DOI: 10.1080/08997659.2014.938870 To link to this article: http://dx.doi.org/10.1080/08997659.2014.938870

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Versions of published Taylor & Francis and Routledge Open articles and Taylor & Francis and Routledge Open Select articles posted to institutional or subject repositories or any other third-party website are without warranty from Taylor & Francis of any kind, either expressed or implied, including, but not limited to, warranties of merchantability, fitness for a particular purpose, or non-infringement. Any opinions and views expressed in this article are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor & Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

It is essential that you check the license status of any given Open and Open Select article to confirm conditions of access and use. Journal of Aquatic Animal Health 26:219–224, 2014 Published with license by American Fisheries Society ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.938870

COMMUNICATION

Observations on Side-Swimming Rainbow Trout in Water Recirculation Aquaculture Systems

Christopher Good,* John Davidson, and Christin Kinman The Conservation Fund’s Freshwater Institute, 1098 Turner Road, Shepherdstown, West Virginia 25443, USA P. Brett Kenney Division of Animal and Nutritional Sciences, West Virginia University, Morgantown, West Virginia 26506-6108, USA Grete Bæverfjord Noﬁma Marin, Sjølseng N-6600, Sunndalsøra, Norway Steven Summerfelt The Conservation Fund’s Freshwater Institute, 1098 Turner Road, Shepherdstown, West Virginia 25443, USA

Side swimming is an understudied anomaly occasionally ex- Abstract hibited by various fish species under intensive culture condi- During a controlled 6-month study using six replicated water tions (Branson and Turnbull 2008). Side swimmers swim in a recirculation aquaculture systems (WRASs), it was observed that coordinated manner but are oriented abnormally in the water Rainbow Trout Oncorhynchus mykiss in all WRASs exhibited a higher-than-normal prevalence of side swimming (i.e., controlled, column (Figure 1), such that their dorsal and ventral aspects forward swimming but with misaligned orientation such that the are approximately perpendicular to the direction of gravity. De- fish’s sagittal axis is approximately parallel to the horizontal plane). spite the change in orientation, side swimmers appear to con- To further our understanding of this abnormality, a substudy was sume feed well and grow similarly to normally swimming fish. conducted wherein side swimmers and normally swimming fish The authors have observed side-swimming fish in a variety of were selectively sampled from each WRAS and growth performance (length, weight), processing attributes (fillet yield, visceral aquaculture settings, including flow-through raceways, circu- index, ventrum [i.e., thickness of the ventral “belly flap”] index), lar tanks in water recirculation aquaculture systems (WRASs), blood gas and chemistry parameters, and swim bladder morphol- and freshwater net-pens. The authors are familiar with several ogy and positioning were compared. Side swimmers were found commonly held beliefs concerning the reason(s) for side swim- to be significantly smaller in length and weight and had less fillet ming, including swim bladder malformation, increased intra- yield but higher ventrum indices. Whole-blood analyses demon- coelomic adipose tissue, increased adipose tissue deposition in Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 strated that, among other things, side swimmers had significantly lower whole-blood pH and higher PCO2. Side swimmers typically the ventral connective tissue (i.e., the “belly-flap”), and pro- exhibited swim bladder malformations, although the positive pre- longed exposure to rotational water currents in circular tanks; dictive value of this subjective assessment was only 73%. Overall, however, very little scientific research has been carried out to this study found several anatomical and physiological differences investigate these theories or to characterize the basic differ- between side-swimming and normally swimming Rainbow Trout. Given the reduced weight and fillet yield of market-age side swim- ences between side-swimming and normally swimming fish. In mers, producers would benefit from additional research to reduce one of the few studies documenting side swimming, Davidson side-swimming prevalence in their fish stocks. et al. (2011a) found that side swimming was exacerbated in Rainbow Trout Oncorhynchus mykiss cultured in WRASs with

C The Conservation Fund This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The moral rights of the named author(s) have been asserted. *Corresponding author: [email protected] Received March 20, 2014; accepted May 20, 2014 219 220 GOOD ET AL.

METHODS

Water recirculation aquaculture systems.—The experimental WRASs used during the present study have been previously described by Davidson et al. (2009), and the specific operation of these systems for the initial controlled experiment was described in detail by Davidson et al. (2011b). Briefly, the study used six identical 9.5-m3 WRASs with 5.3-m3 circular dual-drain tanks, fluidized sand biofilters, and low-head oxygenators that provided rearing-unit dissolved oxygen at approximately saturation. Three WRASs were operated at low exchange (0.26% makeup water relative to a total recirculation flow of 380 L/min) and received water ozonation via generators (model G22; Pa- cific Ozone Technology, Benecia, California) that converted a portion of pure-oxygen feed gas to ozone, which was then transferred to water within the systems’ low-head oxygenators. An SC100 Universal Controller (Hach) provided proportional- integral-derivative control of generator output to maintain ozone oxidation–reduction potential at a set point of 250 mV, which ensured that ozone dosage was approximately 20–25 g ozone/kg FIGURE 1. View from a circular tank side window illustrating both side- feed throughout the study. The remaining three WRASs were swimming (top) and normally swimming (bottom) Rainbow Trout. [Color figure operated at high exchange (2.6% makeup water relative to the available online.] total recirculation flow) and did not receive water ozonation.

low water exchange rates and identified accumulating concen- Rainbow Trout.—The Rainbow Trout used in this experiment trations of nitrate nitrogen (NO3-N) and dissolved potassium as were acquired as eyed eggs from a commercial producer and potential causal factors; further controlled research (Davidson raised on site, initially as fry in small (0.5-m3) flow-through cir- et al. 2014) found a significant association between Rainbow cular tanks. Fish were stocked into the replicated WRASs (1,000 Trout side swimming and elevated NO3-N. fish/system) at a mean ± SE size of 151 ± 3 g and were main- An important consideration for producers is whether side tained for 6 months at a density of 30–80 kg/m3. A constant swimming is actually related to reduced fish performance, i.e., 24-h photoperiod was provided, and feed was administered ev- whether reducing side swimming in fish populations would lead ery alternate hour (with two feed events occurring within that to increased growth, survival, fillet yield, or overall farm prof- hour) using automated feeders (T-drum 2000CE; Arvo-Tec, Fin- itability. Additionally, in the authors’ experience side swimmers land). Feeding rates were based on standardized feeding charts are inherently conspicuous and are routinely commented upon but were modified occasionally according to feeding activity by members of the public touring aquaculture facilities; there- and observable wasted feed. A slow-sinking trout feed (42:16 fore, side swimming might contribute to negative perceptions protein-to-fat ratio; Zeigler Brothers, Gardners, Pennsylvania) regarding the health, welfare, and product quality of farmed fish. was used throughout the study. A controlled 6-month study using six replicated WRASs

Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 was carried out examining Rainbow Trout performance in low- Fish sampling.—At the end of the study (370 d posthatch, over- water-exchange, ozonated systems versus high-water-exchange, all average fish size = 1,364 g), three side swimmers and three nonozonated systems (Davidson et al. 2011b). Approximately normally swimming Rainbow Trout from each WRAS (36 fish midway through this experiment, it was noted that a higher- in total) were selectively sampled and sequentially euthanized than-normal prevalence of Rainbow Trout were exhibiting side with an overdose (200 mg/L) of tricaine methanesulfonate (MS- swimming in all WRASs, with the prevalence of side swimming 222; Western Chemical, Ferndale, Washington). Fish were indi- being significantly higher in low-exchange WRASs (Davidson vidually brought to a team of researchers who were blinded as et al. 2011a). In light of this unexpected finding, a substudy was to whether each fish was a side swimmer or a normal swimmer. planned to investigate side swimmers at the end of the original Fish were first bled via caudal venipuncture using a 21-guage, experiment when the fish were at, or slightly larger than, market 38-mm needle and 3-mL syringe. Whole-blood samples were size. The objective of this study was to characterize anatomical immediately analyzed using an i-Stat 1 portable analyzer (Ab- and physiological differences between side-swimming and nor- bott Laboratories, Abbott Park, Illinois) with CG4 + (measur- mally swimming fish, in particular to determine whether side ing PCO2 [partial pressure of the gas], PO2,HCO3, total CO2, swimming was associated with reduced performance or product O2 saturation, and lactate) and CHEM8 + (measuring sodium, yield. potassium, chloride, calcium, glucose, creatinine, hematocrit, COMMUNICATION 221

and hemoglobin) cartridges. Fish were measured for length and TABLE 1. Summary of swim bladder (n = 32) assessments by three blinded weight, then the viscera was removed through a ventral inci- raters compared with the actual status of fish as side swimmers or normal swimmers. The letters in parentheses (a–d) identify the four status and rating sion, while taking care not to disturb the position or integrity combinations and correspond to formulas for positive and negative predictive of the swim bladder. With the viscera removed, photographs values, sensitivity, and specificity, as listed in the Methods section. were taken of each swim bladder in situ (32 swim bladder photographs were obtained overall). Fish were then butterfly Actual status filleted, the fillets were weighed, and measurements were taken Swim bladder for fillet thickness and ventral “belly-flap” thickness using dig- rating and total Side swimmer Normal swimmer Total ital microcalipers. These measurements were taken at three locations along the length of the carcass, at approximately the Abnormal 47 (a) 17 (b) 64 5th (A), 50th (B), and 95th (C) percentile of the overall fil- Normal 1 (c) 31 (d) 32 let or belly-flap length (cranial to caudal). Lastly, at a later Total 48 48 96 date swim bladder photographs were independently assessed by three on-site researchers with advanced experience in fish carcass processing and were categorized as normal or abnormal based on the swim bladder position, level of inflation, or general A two-way mixed-effects model, type absolute agreement, morphology. single-measures intraclass correlation (Norman and Streiner 2008) was calculated to determine the scoring consistency Calculations and statistical analyses.—Total fillet yield for each between raters. fish was calculated by dividing the weight of both right and left fillets by fish weight. Percentage viscera was likewise calcu- RESULTS lated as the visceral weight divided by the fish weight. Ventrum Swim bladders that were visually assessed ranged from nor- indices were calculated at points A, B, and C by dividing the mal (typical morphology, positioning, and inflation level) to belly-flap width by fish weight, and fillet thickness indices were having a variety of malformations, some of which are illus- calculated at the three points by dividing the thickness of the trated in Figure 2. In determining the reliability of the visual right fillet by the sum of the right and left fillet thicknesses. swim bladder assessment, the PPV of the test was 73%, with Linear regression models were run on each of the fish- a NPV of 97%, and the test sensitivity and specificity were processing metrics listed above and on the individual blood 98% and 65%, respectively. The intraclass correlation was de- chemistry parameters using SPSS (IBM, Armonk, New York). termined to be 0.772, and there were no significant differences For each regression model, swimming status served as the inde- between the raters in their responses determining normal and ab- pendent variable with “tank” added as a covariate to control for normal swim bladders. Visual swim bladder assessment results the effects of both rearing-unit clustering and treatment effects are summarized in Table 1. (i.e., low exchange, ozone versus high exchange, no ozone) Linear regression modeling of blood chemistry parameters from the original controlled study. Individual blood chem- revealed significant (P < 0.05) differences in potassium, glu- istry parameters and processing metrics served as dependent cose, pH, lactate, and PCO2 between side swimmers and nor- variables. mally swimming fish (Table 2). In side swimmers, whole-blood To assess whether swim bladder categorization (i.e., normal potassium, pH, and lactate were significantly lower, while glu- or abnormal) was associated with side-swimming fish, positive cose and PCO2 were significantly higher. and negative predictive values (PPVs and NPVs, respectively) Regression analyses also revealed significant differences in Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 (Szklo and Nieto 2004; Duffy et al. 2005) were calculated based performance and processing metrics between side swimmers on the results from the three raters. Given the cell identifiers (a and normal swimmers (Table 3). Overall, side swimmers were through d) listed in Table 1, which identify the four status (side significantly smaller in both length and weight and had lower swimmer or normal swimmer) and swim bladder rating (normal fillet yields than normally swimming fish. As well, side swim- or abnormal) combinations, the PPV and NPV were calculated mers had significantly thicker ventrum indices at points B and as follows: C, and their right fillet index at point A was also significantly PPV = a/(a + b) lower. NPV = d/(c + d) DISCUSSION The sensitivity and specificity (Altman and Bland 1994a) of this The results of this study suggest that, despite appearing visual assessment was calculated by the following: otherwise healthy and feeding well, side-swimming Rainbow Trout do not perform as well, in terms of growth and fillet Sensitivity = a/(a + c) yield, as their normally swimming conspecifics. These differ- Specificity = d/(b + d) ences appear to be subtle, albeit statistically significant, as side 222 GOOD ET AL. Downloaded by [Department Of Fisheries] at 00:14 17 November 2014

FIGURE 2. Photographs of eviscerated Rainbow Trout (cranial = bottom of photo), demonstrating qualitative swim bladder abnormalities observed during sampling. The panels are as follows: top left = normal inflation and position; top right = normal inflation, lateral deviation; bottom left = under inflation, s-shaped deviation; and bottom right = under inflation, tortuous malformation, and cranial retroflexion. [Color figure available online.]

swimmers in general are mostly indistinguishable from normal than the majority of fish in a rearing unit, as is evidenced by swimmers once they are removed from the water and their be- differences in whole-blood parameters between the two types havior can no longer be observed. It is likely that the small of swimmers. Increased blood carbon dioxide (PCO2)inside but significant differences in size and fillet yield of side swim- swimmers implies a relatively higher level of muscular ex- mers can be attributed to less efficient swimming and an inher- ertion, which corresponds with observations that side swim- ent stress associated with swimming in a different orientation mers tend to have an increased frequency of body movement COMMUNICATION 223

TABLE 2. Whole-blood gas and chemistry analysis results (mean ± SE) despite moving forward in the water column at approximately for side-swimming and normally swimming Rainbow Trout. Parameter values the same speed as normally swimming fish. The higher PCO2, that are significantly different (P < 0.05) between side swimmers and normal swimmers are indicated in bold italics. in turn, appears to have affected other blood parameters in side- swimming fish, such as the lower pH due to CO2-associated Parameter Side swimmers Normal swimmers acidosis and lower PO2 and O2 saturation due to the resultant ± ± Bohr and Root effects of low blood pH (Wedemeyer 1996). Sodium (mmol/L) 147.0 0.5 148.3 0.7 Blood glucose was also relatively elevated in side-swimming 2.772 ± 0.060 3.106 ± 0.116 Potassium (mmol/L) fish, indicating either higher exertion in swimming, stress, or Chloride (mmol/L) 134.1 ± 0.4 134.9 ± 0.5 ± ± both. However, a corresponding increase in blood lactate, as Calcium (mmol/L) 1.454 0.019 1.483 0.015 would be expected in fish with relatively increased exercise or Glucose (mg/dL) 61.78 ± 1.18 58.06 ± 1.50 ± ± higher levels of stress, was not noted in side-swimming fish; Urea nitrogen (mg/dL) 9.056 1.116 10.83 1.33 in fact, side swimmers had statistically lower levels of blood Hematocrit (% PCVa) 28.56 ± 1.01 25.94 ± 0.10 ± ± lactate than normally swimming fish. This apparent inconsis- Hemoglobin (g/dL) 9.711 0.345 8.828 0.340 tency requires further research for adequate explanation, in- pH 7.080 ± 0.012 7.135 ± 0.013 ± ± cluding assessments of blood cortisol and time-to-exhaustion Total CO2 (mmol/L) 11.11 0.36 10.65 0.38 swimming exercise studies. At present, it is clear from the data PCO2 (mm Hg) 34.27 ± 1.79 29.18 ± 1.18 ± ± obtained in this study that alterations in blood chemistry and PO2 (mm Hg) 21.72 1.12 25.82 2.84 gas physiology are associated with side swimming, and it is O2 saturation (%) 22.61 ± 2.20 33.00 ± 5.19 ± ± likely that these changes, in the long term, have affected growth HCO3 (mmol/L) 10.02 0.30 9.771 0.335 and muscle yield. Further research is necessary to determine 0.8822 ± 0.0845 1.215 ± 0.100 Lactate (mmol/L) the economic impact of increased side swimming in farmed fish a Percent packed-cell volume. populations. Anecdotal information from producers has suggested that increased adipose deposition in the ventral connective tissue (i.e., the belly flap) can lead to disruption in normal buoyancy and TABLE 3. Summary of processing attributes (mean ± SE) for side- subsequent abnormalities in swimming, including side swim- swimming and normally swimming Rainbow Trout. Measurements for the fillet ming. Ventral indices calculated in this study support the notion thickness index and the ventrum index (ventral belly-flap thickness) were taken at three locations along the length of the carcass, at approximately the 5th (A), that side swimmers have relatively greater belly flap thicknesses, 50th (B), and 95th (C) percentile of the overall fillet or belly-flap length. Param- and therefore it is possible that buoyancy is disrupted in side eter values that are significantly different (P < 0.05) between side swimmers swimmers due to morphological differences in the ventral body and normal swimmers are indicated in bold italics. wall. However, despite being statistically different, the empirical values of the ventrum indices were not overly different Parameter Side swimmers Normal swimmers between the two swimming types, and therefore it is likely that Length (mm) 389.7 ± 5.2 412.7 ± 3.7 increased belly flap thickness is, at best, only a contributing Weight (g) 1,139 ± 45 1,394 ± 34 factor in the development of abnormal swimming. This latter Ventrum 0.00898 ± 0.00026 0.00859 ± 0.00027 notion is supported by the more compelling evidence obtained index (A) in this study that swim bladder malformation has a relatively Ventrum 0.01081 ± 0.00042 0.00973 ± 0.00033 strong association with side swimming. index (B) Swim bladder malformation most likely occurs in the early

Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 Ventrum 0.01090 ± 0.00040 0.00934 ± 0.00026 stages of anatomical development (Poppe et al. 1997); because index (C) the Rainbow Trout used in the present study were raised as a Right fillet 0.4773 ± 0.0032 0.4880 ± 0.0020 single cohort prior to study commencement, any environmental thickness conditions promoting swim bladder malformation were com- index (A) mon for all individuals regardless of their subsequent exper- Right fillet 0.5041 ± 0.0019 0.5060 ± 0.0018 imental treatment assignment. We report a strong correlation thickness between malformed swim bladders (subjectively determined index (B) to be abnormal in morphology or position) and side swim- Right fillet 0.5163 ± 0.0032 0.5094 ± 0.0029 ming, based on the strong PPV of the subjective assessment. thickness Predictive values are normally used to judge the usefulness index (C) of a diagnostic test (Altman and Bland 1994b) but are lim- Total fillet 48.50 ± 0.42 50.67 ± 0.46 ited in that they are, in part, determined by the prevalence of yield (%) disease in a population, which can change over time. In the Viscera (%) 15.52 ± 0.56 14.91 ± 0.48 present study, pretest probability was set at 50% (i.e., an equal number of side swimmers and normally swimming fish were 224 GOOD ET AL.

selectively sampled), and therefore the effect of side-swimming ACKNOWLEDGMENTS prevalence change in a population on the predictive values of This research was supported by the U.S. Department of Agri- the visual test was not assessed. However, as the visual assess- culture, Agricultural Research Service under Agreement Num- ment of swim bladders is, by nature, a lethal test, predictive ber 59-1930-5-510. The experimental protocols described are value calculations carried out in this study were not intended to in compliance with the Animal Welfare Act (9CFR) and were determine the usefulness of this test for “diagnostic” purposes. approved by our Institutional Animal Care and Use Committee. Instead, these values were calculated to determine the associa- The use of trade names does not imply endorsement by the U.S. tion of swim bladder malformation with side swimming in light Government. of the fact that swim bladder assessment is an inherently subjective test. It is interesting to note that almost all the assessments of true side swimmers indicated abnormal swim bladders, while REFERENCES a sizeable proportion of normal swimmers also had swim blad- Altman, D. G., and J. M. Bland. 1994a. Diagnostic tests 1: sensitivity and ders assessed as abnormal. The very high NPV indicates that specificity. British Medical Journal 308:1552. swim bladder malformation may be a necessary component of Altman, D. G., and J. M. Bland. 1994b. Diagnostic tests 2: predictive values. British Medical Journal 309:102. the side-swimming risk factor constellation; however, from the Branson, E. J., and T. Turnbull. 2008. Welfare and deformities in fish. Pages 202– lower PPV it is clear that certain fish are still able to swim nor- 216 in E. J. Branson, editor. Fish welfare. Blackwell Scientific Publications, mally despite possessing a malformed swim bladder and, hence, Ames, Iowa. are not influenced by additional risk factors that would be suf- Davidson, J., C. Good, C. Welsh, B. Brazil, and S. T. Summerfelt. 2009. ficient to induce side swimming. From these results, it appears Heavy metal and waste metabolite accumulation and their effect on Rainbow Trout performance in a replicated water reuse system oper- that side swimming is a complex, multifactorial abnormality, ated at low or high system flushing rates. Aquacultural Engineering 41: and further research to understand the factors associated with 136–145. this condition, particularly prospective study in which causation Davidson, J., C. Good, C. Welsh, and S. T. Summerfelt. 2011a. Abnormal swim- can be more appropriately assessed, is required. ming behavior and increased deformities in Rainbow Trout Oncorhynchus Overall, the results of this study indicate that, due to the mykiss cultured in low exchange water recirculating aquaculture systems. Aquacultural Engineering 45:109–117. reduced performance of side swimmers relative to normally Davidson, J., C. Good, C. Welsh, and S. T. Summerfelt. 2011b. The effects swimming fish, successful efforts to reduce side swimming in of ozone and water exchange rates on water quality and Rainbow Trout farmed fish populations could lead to increased farm profitabil- Oncorhynchus mykiss performance in replicated water recirculating systems. ity. Further research is required to understand this phenomenon Aquacultural Engineering 44:80–96. and its genesis in Rainbow Trout and in other important farmed Davidson, J., C. Good, C. Welsh, and S. T. Summerfelt. 2014. Com- paring the effects of high vs. low nitrate on the health, performance, fish species, as the causes of side swimming are still unknown and welfare of juvenile Rainbow Trout Oncorhynchus mykiss within and most likely vary from one species to another. Given the asso- water recirculating aquaculture systems. Aquacultural Engineering 59: ciation between swim bladder malformation and side swimming 30–40. determined in this study, it is important to fully understand the Duffy, M., B. Hazard Munro, and B. Jacobsen. 2005. Key principles of statistical mechanisms and timing of swim bladder formation and inflation inference. Pages 89–91 in B. H. Munro, editor. Statistical methods for health care research. Lippincott Williams and Wilkins, Philadelphia. in Rainbow Trout to reduce malformation prevalence or to aid Norman, G., and D. Streiner. 2008. Biostatistics: the bare essentials. People’s in the separation of fish capable of optimal growth from those Medical Publishing House, Shelton, Connecticut. without functional swim bladders. Better and earlier diagnostic Poppe, T. T., H. Hellberg, D. Griffiths, and H. Meldal. 1997. Swimbladder testing needs to be available so that fish with uninflated swim abnormality in farmed Atlantic Salmon (Salmo salar). Diseases of Aquatic bladders can be identified and removed. Additional research fo- Organisms 30:73–76. Szklo, M., and F. J. Nieto. 2004. Epidemiology: beyond the basics. Jones and Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 cusing on other risk factors for side swimming would also be Bartlett, Sudbury, Massachusetts. very useful, given the < 100% PPV for swim bladders assessed Wedemeyer, G. A. 1996. Physiology of fish in intensive culture systems. Chap- to be abnormal in morphology or position. man and Hall, New York. This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:14 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Diversity of Bacteria Cultured from the Blood of Lesser Electric Rays Caught in the Northern Gulf of Mexico Zhen Taoab, Stephen A. Bullardc & Cova R. Ariasa a Aquatic Microbiology Laboratory, School of Fisheries, Aquaculture, and Aquatic Sciences, Auburn University, 559 Devall Drive, Auburn, Alabama 36832, USA b School of Marine Sciences, Ningbo University, Ningbo 315211, China c Aquatic Parasitology Laboratory, School of Fisheries, Aquaculture, and Aquatic Sciences, Auburn University, 352 Upchurch Hall, Auburn, Alabama 36849, USA Published online: 16 Oct 2014.

To cite this article: Zhen Tao, Stephen A. Bullard & Cova R. Arias (2014) Diversity of Bacteria Cultured from the Blood of Lesser Electric Rays Caught in the Northern Gulf of Mexico, Journal of Aquatic Animal Health, 26:4, 225-232, DOI: 10.1080/08997659.2014.922513 To link to this article: http://dx.doi.org/10.1080/08997659.2014.922513

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:225–232, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.922513

ARTICLE

Diversity of Bacteria Cultured from the Blood of Lesser Electric Rays Caught in the Northern Gulf of Mexico

Zhen Tao Aquatic Microbiology Laboratory, School of Fisheries, Aquaculture, and Aquatic Sciences, Auburn University, 559 Devall Drive, Auburn, Alabama 36832, USA; and School of Marine Sciences, Ningbo University, Ningbo 315211, China Stephen A. Bullard Aquatic Parasitology Laboratory, School of Fisheries, Aquaculture, and Aquatic Sciences, Auburn University, 352 Upchurch Hall, Auburn, Alabama 36849, USA Cova R. Arias* Aquatic Microbiology Laboratory, School of Fisheries, Aquaculture, and Aquatic Sciences, Auburn University, 559 Devall Drive, Auburn, Alabama 36832, USA

Abstract The prevalence and taxonomic diversity of bacteria cultured from the blood of apparently healthy Lesser Electric Rays Narcine bancroftii captured from open beach habitat in the north-central Gulf of Mexico are reported herein. The blood of 9 out of 10 Lesser Electric Rays was positive for bacteria, and bacterial isolates (n = 83) were identified by 16S rRNA gene sequencing. The majority of the isolates belonged to the phylum Proteobacteria (91.5%). Vibrio spp. comprised 53% of all isolates and were recovered from all Lesser Electric Rays with culture-positive blood. Among them, V. harveyi (n = 14) and V. campbellii (n = 11) were most common, followed by a group of unidentified Vibrio sp. (n = 10) related to V. nigripulchritudo. Isolates representing other species of Proteobacteria included Pseudoal- teromonas (n = 13), Shewanella (n = 5), Amphritea (n = 3), Nautella (n = 3), and Arenibacter (n = 1). Higher bacterial diversity was observed in blood cultured on marine agar relative to blood agar, but gram-positive bacteria were isolated from the latter only. The 16S rRNA gene sequences of bacterial isolates were compared phylogenetically to those from related type strains. Most isolates were identified to the level of species, but some clustered independently from reference strains, likely representing new species of Vibrio, Amphritea, Shewanella,andTenacibaculum. The present study is the first record of any bacterium from this ray species and reveals a taxonomically and phylogenetically diverse microbiota associated with its blood. Moreover, these data document that the presence of bacteria in elasmobranch blood is not coincident with clinical signs of disease, thereby rejecting the paradigm of septicemia Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 indicating a disease condition in aquatic vertebrates.

Little published information exists on the biodiversity, preva- blood of apparently healthy elasmobranchs (Grimes et al. lence, and physiological effects of bacteria that infect the 1993; Borucinska and Frascas 2002; Mylniczenko et al. 2007) blood and other tissues of cartilaginous ﬁshes (class Chon- Based on this foundational taxonomic work with blood-borne drichthyes: sharks, skates, rays, and chimaeras) (Borucinska bacteria in sharks, seemingly, the classical assumption that and Frascas 2002; Mylniczenko et al. 2007). However, sev- bacterial presence in blood indicates disease is no longer eral studies have reported the isolation of bacteria from the robust. Yet, other bacteria, including Vibrio spp., recovered

*Corresponding author: [email protected] Received April 30, 2013; accepted April 16, 2014 225 226 TAO ET AL.

from sharks are indeed considered opportunistic pathogens Immediately before necropsy, each ray was euthanized with an (Grimes et al. 1984a). Hence, whether the taxonomic spec- overdose (300 mg/L) of tricaine methanesulfonate (MS-222). trum of these elasmobranch-associated bacteria comprise oppor- All animal protocols were approved by the Auburn University tunistic or obligate pathogens, benign commensals, or bona fide Institutional Animal Care and Use Committee (IACUC number tissue-dwelling symbionts that serve a critical role in elasmo- 2012–2098). Immediately after euthanasia, the area of skin cir- branch physiology is indeterminate. Nevertheless, documenting cumscribed by the gill slits, mouth, and pectoral girdle was dried microbial taxonomic diversity in other elasmobranch lineages with a clean paper towel, disinfected with 70% ethanol, and cut is a good first step towards deciphering the nature of these asso- away to expose the pericardial chamber. The exposed surfaces ciations between elasmobranchs and bacteria. Detailed studies of the heart, including ventricle and conus arteriosus, were dis- based on materials sampled from nonshark elasmobranchs, i.e., infected with 70% ethanol before a blood sample was taken by skates, rays, or chimaeras, are largely missing from the literature inserting a sterile syringe into the lumen of the heart. Each blood (Mylniczenko et al. 2007). sample from each ray was immediately spread onto blood agar The Lesser Electric Ray Narcine bancroftii (order Torpedini- (BA) (MOLTOX, Boone, North Carolina) and marine agar (MA) formes, family Narcinidae; also known as Narcine brasiliensis) (Difco, Sparks, Maryland) using aseptic methods. Agar plates ranges in shallow waters of tropical and subtropical continen- were incubated at 28◦C for 48 h under aerobic conditions. A rep- tal shelves to 37 m deep, including the Gulf of Mexico, the resentative of each colony type on the primary isolation plate Caribbean Sea, and the islands of the West Indies (Robins and was restreaked on MA to obtain pure cultures for identification. Ray 1986). Three other species of Narcinidae have geographic A total of 86 single isolates were preserved as glycerol stocks ranges that overlap with the Lesser Electric Ray, but the Lesser (marine broth supplemented with 20% glycerol) at −80◦C Electric Ray is the only narcinid that reportedly ranges in the until subsequent analysis. Individual blood samples were la- north-central Gulf of Mexico (the focus area for the present beled NB-01 through NB-09. The isolates were designated as study) (McEachran and de Carvalho 2002). This ray is a slow FMR (Fort Morgan Ray) followed by the colony number. swimming fish that can be seasonally aggregated on sandbars Bacterial identification.—Bacterial isolates were identified and surf zones along open beaches and barrier islands. It can be by partially sequencing the 16S rRNA gene. The DNA tem- regionally abundant in the summer months, during which time plate was prepared using a rapid-boiling method as follows. pregnant females birth viviparous offspring, but then moves to Five colonies from a pure isolate were selected from a 24-h offshore deep waters in the winter (Rudloe 1989). During fall culture on MA and resuspended in a centrifuge tube with (August–October) in the northern Gulf of Mexico, Lesser Elec- 100-µL sterile distilled H2O. Proteinase K was added to the tric Rays can be observed commonly by snorkeling in waters of cell suspension to a final concentration of 30 unit/µL. After 0.2−3.0 m; the spiracles of the nearly completely buried rays a 20-min digestion at 55◦C, the lysate was heated to 100◦C appear as characteristic holes in the sand (S.A.B., personal ob- for 15 min and spun down at 15,000 g for 5 min. The su- servations). Perhaps because this ray species is seldom caught pernatant was transferred to a new tube and used as tem- by commercial fishermen, has no recreational or commercial plate DNA. The nearly complete 16S rRNA gene of each iso- value, and is typically hidden, nearly completely buried in the late was amplified using the following primers: 63V (forward) sand, it is rarely included in faunal surveys of beach habitat 5-CAGGCCTAACACATGCAAGTC-3 and 1387R (reverse) in the Gulf of Mexico. As a result, there is little substantive 5-GGGCGGWGTGTACAAGGC-3 (Marchesi et al. 1998). information on the abundance and population structure of this The PCR conditions and reagents have been described else- species throughout its range or in the northern Gulf of Mexico. where (Arias et al. 2006). Sequencing was conducted on the Concomitantly, we know little of its general biology, including PCR-amplified products, using 63V as sequencing primer, in an

Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 its parasites, pathogens, and symbionts. The objective of this ABI 3730xl sequencer at Lucigen (Madison, Wisconsin). study was to determine if bacteria could be isolated from the Sequence analysis.—Sequence trace files were edited with blood of apparently healthy Lesser Electric Rays and, if so, to BioEdit version 7.1.9 (Hall 1999) to remove noise and untrusted characterize the bacterial diversity present in their blood. ends. Sequences (n = 3) having < 500 bp or > 3 ambiguous positions were excluded from the analysis. The resulting 83 sequences were assigned to taxonomic units by (1) the Ribo- METHODS somal Database Project (RDP) Na¨ıve Bayesian Classifier (Cole Sample collection.—The studied Lesser Electric Rays et al. 2007), (2) the GreenGenes web classification tool (De- were hand-netted off Fort Morgan, Alabama (30◦1345N, Santis et al. 2006), and (3) the Basic Local Alignment Search 87◦547W), maintained alive in enclosed plastic transport con- Tool (BLAST; Altschul et al. 1997). Query sequences that had tainers filled with water from the collection site and fitted with a 98–100% and 95–98% identity match to those present in the water pumps and aerators powered by a car battery, and trans- databases were considered identified at the species and genus ported alive to Auburn University (within 5 h after collec- level, respectively (Tindall et al. 2006). tion). A total of 10 Lesser Electric Rays (24–47 cm in total Phylogenetic analysis.—Partial 16S rRNA gene sequences length; 7 females and 3 males) were examined in the study. were aligned using Clustal X2 (Larkin et al. 2007). Multiple BACTERIA FROM BLOOD OF ELECTRIC RAYS 227

sequence alignment (MSA) was conducted by trimming the sequences to cover the entire alignment and subsequent re- alignment. The trimmed MSA spanned the hypervariable V2, V3, and V4 regions corresponding to the Escherichia coli 16S rRNA gene base pair positions (Van de Peer et al. 1996). Se- quences of the type strains identiﬁed as nearest to the Lesser Electric Ray isolates by RDP and BLAST were incorporated into the phylogenetic trees as reference. Phylogenetic analysis was conducted in MEGA 5.0 software (Tamura et al. 2011). Trees were constructed using the neighbor-joining method (Saitou and Nei 1987) with the Jukes–Cantor correction (Jukes and Cantor 1969). The partial 16S rRNA gene sequences of bacterial isolates recovered from Lesser Electric Rays were submitted to the GENBANK nucleotide sequence database (accession numbers KC439161 to KC439244).

RESULTS FIGURE 2. Distribution of isolates from the blood of Lesser Electric Rays Isolate Identification cultured in marine agar (MA) and blood agar (BA). [Color figure available All blood samples but one (NB-09) were culture positive, al- online.] though the number of colony types (from approximately 3–16) varied among specimens (Figure 1). A total of 86 pure isolates them to Arenibacter sp. Five isolates were ascribed to Vibrio by were recovered from the blood samples. Isolates were recov- GreenGenes, but RDP ascribed them to Vibrionaceae. We re- ered on both MA (47 colonies) and BA (39 colonies) culture solved the divergence by assigning the sequence to the lowest media. Three isolates yielded poor 16S rRNA gene sequence taxonomic level. quality and were removed from the study. The remaining 83 Among the Proteobacteria, 79 isolates were from the class sequences were ascribed to specific taxa using three databases. Gammaproteobacteria while only 3 isolates were identified as Overall, results from RDP, GreenGene, and GENBANK were in class Alphaproteobacteria. Within the Gammaproteobacteria, agreement and isolates were identified unambiguously to genus Vibrio was the predominant genus with 45 isolates. Other iso- (sequence similarity of 95% or higher). Isolates were classified lates representing genera of Gammaproteobacteria comprised into 14 genera, 11 families, 6 orders, 4 classes, and 3 phyla. The Pseudoalteromonas (n = 13), Shewanella (n = 5), Ferrimonas majority of the isolates (91.5%) were ascribed to the phylum (n = 2), Amphritea (n = 3), Photobacterium (n = 2), Thalas- Proteobacteria, followed by the phylum Bacteroidetes (6.0%) somonas (n = 2), Aestuariibacter (n = 1) and Pseudomonas and the phylum Actinobacteria (2.4%). In a few cases, there (n = 1). All Alphaproteobacteria were assigned to Nautella was a disagreement between the results obtained from differ- (n = 3). Figure 1 shows the distribution of the predominant gen- ent databases. For example, GreenGenes could not place four era in each individual fish. Vibrio was the only genus recovered Flavobacteriaceae isolates below family; whereas, RDP ascribed from all fish with culture-positive blood. In fact, it was the most common genus in all fish except in NB-07 from which only four isolates were recovered and all of them belonged to differ- Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 ent genera. Pseudoalteromonas and Shewanella were recovered from five and four Lesser Electric Rays, respectively. The MA not only yielded more isolates but also provided a higher diversity of genera than did the BA (Figure 2), but a few genera (gram-positive bacteria) were recovered on BA only.

Phylogenetic Analysis The majority of Vibrio isolates (35 out of 45) comprised three clades (Figure 3). Clade I included V. harveyi (n = 14) plus the types species, Clade II included 11 isolates of V. campbellii or V. sagamiensis (the partial 16S rRNA gene sequence used did FIGURE 1. Bacteria recovered from the blood of Lesser Electric Rays (indi- not allow for differentiation between these two species), and cated as NB followed by their ID number). [Color ﬁgure available online.] Clade III included 11 isolates not ascribed to any reference or 228 TAO ET AL. Downloaded by [Department Of Fisheries] at 00:14 17 November 2014

FIGURE 3. Phylogeny (partial 16S rRNA gene sequences) of bacterial isolates from the blood of Lesser Electric Rays and ascribed to species in the genus Vibrio. The isolate number is followed by the GENBANK accession number. Sequences from type strains, or the closest match, were used for comparison. The tree topology was obtained by the neighbor-joining method (Jukes–Cantor correction). The three main clades are indicated. The numbers at the nodes indicate bootstrap values (1,000 replicates). Scale bar = 0.5% sequence divergence. BACTERIA FROM BLOOD OF ELECTRIC RAYS 229

type strain sequence but with V. nigripulchritudo as the closest cluding V. harveyi (17% of all isolates) and V. campbellii (13%). relative. These Vibrio spp. have been previously reported as part of the Analysis of the non-Vibrio Gammaproteobacteria isolates re- normal flora in sharks (Grimes et al. 1985). Conversely, we failed sulted in nine principal groups (Figure 4). A Pseudoalteromonas to recover any isolate ascribed to V. alginolyticus, a species also clade had 13 Lesser Electric Ray isolates ascribed to P. pheno- common in sharks (Grimes et al. 1993). Similarly, the common lica (n = 7), P. prydzensis (n = 1), and P. spongiae (n = 4). marine bacterium Photobacterium damselae, which has been Two of the ray isolates could not be ascribed to named species. isolated from the internal organs of healthy fish (Gomez-Gil In two instances, a few ray isolates shared identical sequences et al. 2006; Mylniczenko et al. 2007), was not isolated during but did not cluster with any reference strain. Five isolates within our study. These discrepancies could be a factor of host speci- the Shewanella clade had a nonculturable bacterium as its clos- ficity, culture medium used for isolation, culture conditions, and est neighbor and could not be assigned to any named species of habitat characteristics comprising the geographic locality where Shewanella. Similarly, within the Amphritea clade, three isolates the Lesser Electric Rays were captured (Austin 2006). Species clustered separately from all known species of Amphritea. of Pseudoalteromonas and Shewanella were common in the The only two Actinobacteria recovered clustered along with blood of the Lesser Electric Rays studied herein, and isolates Micrococcus luteus or M. yunnanensis and Microbacterium ho- representing these genera frequently have been reported from minis (Figure 5). Two isolates of Bacteroidetes clustered with fish; however, only in a few instances have they been isolated Arenibacter nanhaiticus, and another was the sister taxon to from viscera. Some of the lesser-known bacteria isolated in the Zhouia amylolytica. The remaining two isolates were most sim- present study include Amphritea atlantica, Arenibacter nan- ilar to Tenacibaculum sp. but could not be ascribed to a named haiticus, and Zhouia amylolytica. These species or their closest species. phylogenetic species were first discovered in marine sediments (Liu et al. 2006; Gartner et al. 2008; Miyazaki et al. 2008; Sun et al. 2010). Species of Vibrio, Pseudoalteromonas, and DISCUSSION Shewanella have also been recovered from marine sediments It is well known that the blood of elasmobranchs, as in other (Urakawa et al. 2000; Holmstrom et al. 2006; Huang et al. vertebrates, harbors a diverse assemblage of parasites, including 2010). Dean and Motta (2004) theorized that the suction feed- flagellates, amoebas, apicomplexans, microsporidians, and cili- ing behavior of the Lesser Electric Ray facilitates the ingestion ates (see Goertz 2004 and references therein), as well as meta- of sediment, and we think it is plausible that many bacteria zoan parasites (Bullard and Dippenaar 2003; Benz and Bullard would also be ingested during this feeding activity; however, 2004; Patella and Bullard 2013). Bacteria, while markedly less we lack adequate behavior observations and microbial data to studied in elasmobranchs than the aforementioned symbionts, accept or reject this notion. The interstitial and benthic habitat of have been detected in various elasmobranch tissues, including the Lesser Electric Ray could drive the taxonomic composition in the blood, liver, muscle, and epithelium (Knight et al. 1987; of the microbiota. Regardless, how bacteria enter the blood is Terrell 2004; Mylniczenko et al. 2007). Some of these bacteria unknown and also seemingly exceedingly difficult to test in an have been identified as pathogens, e.g., Vibrio harveyi (as V. open, natural system. A comparison of the present results with carchariae) (Grimes et al. 1984a, 1984b; Grimes et al. 1993; those from a pelagic ray that is phylogenetically related to the Pedersen et al. 1998), Aeromonas salmonicida (Briones et al. Lesser Electric Ray may be informative along these lines. 1998), and Flavobacterium sp. (Terrell 2004), but many oth- The culture techniques used in this study likely underesti- ers are described as opportunistic pathogens or have not been mated the bacterial diversity of the tested samples since only associated with disease previously (Mylniczenko et al. 2007). 1–10% of all bacteria can be cultured under laboratory condi-

Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 However, one should not assume an elasmobranch is diseased tions (Amann et al. 1995). We chose a culture-based strategy be- if its blood is infected with parasites or bacteria nor should one cause culture methods are still the “gold standard” in fish disease assume a link between the presence or absence of parasites and diagnostics laboratories (AFS–FHS 2014) and because the low that of bacteria in the blood of elasmobranchs. Doubtless, ample cost associated with this approach makes it seemingly more ac- plausible scenarios exist for how the blood of elasmobranchs cessible to a broader spectrum of researchers. The type of culture can be exposed to bacteria, and there is insufficient data to medium, even general media such as MA and BA, inadvertently conclude whether these bacteria are commensals or symbionts can select for specific bacterial groups, and we observed this of the elasmobranchs or simply reflect transient bacteremias in the present study: species of Pseudoalteromonas, Amphritea, that course asymptomatically. In any case, we still lack a firm Nautella, Arenibacter, Tenacibaculum, and Zhouia were recov- understanding of the taxonomic and phylogenetic diversity of ered on MA only; whereas, species of Ferrimonas, Microbac- bacteria that live in the blood of sharks and rays. terium, Micrococcus, Photobacterium, and Pseudomonas were Most isolates recovered from Lesser Electric Rays belonged recovered on BA only. to the phylum Proteobacteria (91.5%), which is a result that is In summary, the present study reported a high prevalence of in agreement with previous reports (Horsley 1977; Grimes et al. bacteria in the blood of wild, apparently healthy Lesser Elec- 1985). As expected, several species of Vibrio were isolated, in- tric Rays. Vibrio spp. were found in all but one individual and 230 TAO ET AL. Downloaded by [Department Of Fisheries] at 00:14 17 November 2014

FIGURE 4. Phylogeny (partial 16S rRNA gene sequences) of bacterial isolates from the blood of Lesser Electric Rays and assigned as non-Vibrio Gammapro- teobacteria species. The isolate number is followed by the GENBANK accession number. Sequences from type strains, or the closest match, were used for comparison. The tree topology was obtained by neighbor-joining methods (Jukes–Cantor correction). Each genus clade is indicated. The numbers at the nodes indicate bootstrap values (1,000 replicates). Scale bar = 5% sequence divergence. BACTERIA FROM BLOOD OF ELECTRIC RAYS 231

FIGURE 5. Phylogeny (partial 16S rRNA gene sequences) of bacterial isolates from the blood of Lesser Electric Rays and assigned as non-Gammaproteobacteria. The isolate number is followed by the GENBANK accession number. Sequences from type strains, or the closest match, were used for comparison. The tree topology was obtained by neighbor-joining methods (Jukes–Cantor correction). Each genus clade is indicated. The numbers at the nodes indicate bootstrap values (1,000 replicates). The scale bar represents 5% sequence divergence.

included opportunistic fish pathogens (V. harveyi) and poten- research fellowship funded by the Chinese Scholarship Council tially unnamed species (Figure 3, Clade III). Non-Vibrio Pro- and Ocean University of China. teobacteria were also common and contained putative unnamed species within the Shewanella and Amphritea clades. These putative new species require further corroboration by full-length REFERENCES sequence of their 16S rRNA gene and additional taxonomic AFS–FHS (American Fisheries Society–Fish Health Section). 2014. FHS blue markers. Taken together, these insights on the elasmobranch book: suggested procedures for the detection and identification of cer- microbiota are relevant to the fundamental ecology and evolu- tain finfish and shellfish pathogens, 2014 edition. AFS–FHS, Bethesda, Maryland. tionary biology of aquatic symbioses. They are also vital to hus- Altschul, S. F., T. L. Madden, A. A. Schaffer,¨ J. Zhang, Z. Zhang, W. Miller, bandry and veterinary staffers who are employed by the aquar- Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 and D. J. Lipman. 1997. Gapped blast and psi-blast: a new generation ium industry and tasked with keeping exhibited sharks and rays of protein database search programs. Nucleic Acids Research 25:3389– healthy, oftentimes following protocols that use blood culture 3402. as a means of assessing the overall health status of the exhibited Amann, R. I., W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic identification and in situ detection of individual microbial cells without cultivation. elasmobranch. Microbiological Reviews 59:143–169. Arias, C. R., J. W. Abernathy, and Z. Liu. 2006. Combined use of 16s ribosomal DNA and automated ribosomal intergenic spacer analysis (ARISA) to study ACKNOWLEDGMENTS the bacterial community in catfish ponds. Letters in Applied Microbiology 43:287–292. This work was funded by the National Oceanic and Atmo- Austin, B. 2006. The bacterial microflora in fish, revised. ScientificWorld Jour- spheric Administration (NOAA-NA08NMF4720545), Marine nal 6:931–945. Environmental Science Consortium, Gulf of Mexico Research Benz, G. W., and S. A. Bullard. 2004. Metazoan parasites and associates of Institute, and National Science Foundation (NSF DEB num- chondrichthyans with emphasis on taxa harmful to captive hosts. Pages 325– 416 in M. F. L. Smith, D. A. Thoney, and R. E. Hueter, editors. The husbandry bers 1112729, 1051106, and 1048523) and is a contribution of of elasmobrach fishes, volume 16. Ohio Biological Survey, Columbus. the Southeastern Cooperative Fish Parasite and Disease Project Borucinska, J. D., and S. Frascas. 2002. Naturally occuring lesions and micro- (Auburn University). Zhen Tao is the recipient of a graduate organisms in two species of free-living sharks: the Spiny Dogfish, Squalus 232 TAO ET AL.

acanthias L., and the Smooth Dogfish, Mustelus canis (Mitchill), from the Larkin, M., G. Blackshields, N. Brown, R. Chenna, P. McGettigan, H. north-western Atlantic. Journal of Fish Diseases 25:287–298. McWilliam, F. Valentin, I. Wallace, A. Wilm, and R. Lopez. 2007. Clustal w Briones, V., A. Fernandez, M. Blanco, F. Ramiro, J. Garcia, J. L. Mendez, and clustal x version 2.0. Bioinformatics 23:2947–2948. J. Goyache, and M. L. de Vicente. 1998. Hemorrhagic septicemia by Liu, Z.-P.,B.-J. Wang, X. Dai, X.-Y.Liu, and S.-J. Liu. 2006. Zhouia amylolytica Aeromonas salmonicida salmonicida in a Black-tip Reef Shark (Carcharhinus gen. nov., sp. nov., a novel member of the family flavobacteriaceae isolated melanopterus). Journal of Veterinary Medicine 45:443–445. from sediment of the South China Sea. International Journal of Systematic Bullard, S. A., and S. M. Dippenaar. 2003. Branchotenthes robinoverstreeti and Evolutionary Microbiology 56:2825–2829. N. gen. and N. sp. (monogenea: Hezabothriidae) from gill filaments on the Marchesi, J. R., T. Sato, A. J. Weightman, T. A. Martin, J. C. Fry, S. J. Hiom, D. Bowmouth Guitarfish, Rhina ancyclostoma (Rhynchobatidae), in the Indian Dymock, and W. G. Wade. 1998. Design and evaluation of useful bacterium- Ocean. Journal of Parasitology 89:595–601. specific PCR primers that amplify genes coding for bacterial 16s rRNA. Cole, J. R., B. Chai, R. Farris, Q. Wang, A. Kulam-Syed-Mohideen, D. McGar- Applied Environmental Microbiology 64:795–799. rell, A. Bandela, E. Cardenas, G. Garrity, and J. Tiedje. 2007. The ribosomal McEachran, J. D., and M. de Carvalho. 2002. Narcinidae. Food and Agriculture database project (rdp-ii): introducing myrdp space and quality controlled Organization of the United Nations, Rome. public data. Nucleic Acids Research 35:D169–D172. Miyazaki, M., Y. Nogi, Y. Fujiwara, M. Kawato, T. Nagahama, K. Kubokawa, Dean, N. M., and P. J. Motta. 2004. Feeding behavior and kinematics of the and K. Horikoshi. 2008. Amphritea japonica sp. nov. and Amphritea bale- Lesser Electric Ray, Narcine brasiliensis (elasmobranchii: Batoidea). Zool- nae sp. nov., isolated from the sediment adjacent to sperm whale carcasses ogy 107:171–189. off Kagashima, Japan. International Journal of Systematic and Evolutionary DeSantis, T., P. Hugenholtz, N. Larsen, M. Rojas, E. Brodie, K. Keller, T. Huber, Microbiology 58:2815–2820. D. Dalevi, P. Hu, and G. G. Andersen. 2006. Greengenes, a chimera-checked Mylniczenko, N. D., B. Harris, and R. E. Wilborn. 2007. Blood culture results 16s rrna gene database and workbench compatible with ARB. Applied and from healthy captive and free-ranging elasmobranchs. Journal of Aquatic Environmental Microbiology 72:5069–5072. Animal Health 19:159–167. Gartner, A., J. Wisese, and J. Imhoff. 2008. Amphritea atlantica gen. nov., sp. Patella, R., and S. A. Bullard. 2013. Hexabothriids of Devil Rays (Mobulidae): nov., a gammaproteobacterium from the logatchev hydrothermal vent field. new genus and species from gill of Mobula hypostoma in the Gulf of Mexico International Journal of Systematic and Evolutionary Microbiology 58:34– and redescription of a congener from Mobula rochebrunei in the eastern 39. Atlantic Ocean. Journal of Parasitology 99:856–867. Goertz, C. 2004. Protozoan diseases of elasmobranchs. Pages 417–426 in Pedersen, K., L. Verdonck, B. Austin, D. A. Austin, A. R. Blanch, P. A. D. M. F. L. Smith, D. A. Thoney, and R. E. Hueter, editors. The husbandry Grimont, J. Jofre, S. Koblavi, J. L. Larsen, T. Tiainen, M. Vigneulle, and J. of elasmobrach fishes, volume 16. Ohio Biological Survey, Columbus. Swings. 1998. Taxonomic evidence that Vibrio carchariae Grimes et al. 1985 Gomez-Gil, B., E. Fajer-Avila, and F. Garcia-Vargas. 2006. Vibrios of the Spot- is a junior synonym of Vibrio harveyi (Johnson and Shunk 1936) Baumann ted Rose Snapper lutjanus guttatus Steindachner, 1969 from northwestern et al. 1981. International Journal of Systematic Bacteriology 48:749–758. Mexico. Journal of Applied Microbiology 102:1518–1526. Robins, C. R., and G. C. Ray. 1986. A field guide to Atlantic coast fishes of Grimes, D. J., P. R. Brayton, R. R. Colwell, and S. Gruber. 1985. Vibrios as North America. Houghton Mifflin, Boston. autochthonous flora of neritic sharks. Systematic and Applied Microbiology Rudloe, A. 1989. Habitat preferences, movement, size frequency patterns and 6:221–226. reproductive seasonality of the Lesser Electric Ray. Northeast Gulf Science Grimes, D. J., R. R. Colwell, J. Stemmler, H. Hada, D. Maneval, F. M. Hetrick, 10:103–112. E. B. May, R. T. Jones, and M. Stoskopf. 1984a. Vibrio species as agents of Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for elasmobranch disease. Helgolander Meeresunters 37:309–315. recostructing phylogenetic trees. Molecular Biology and Evolution 4:406– Grimes, D. J., D. Jacobs, D. G. Swartz, P. R. Brayton, and R. R. Colwell. 425. 1993. Numerical taxonomy of gram-negative, oxidase-positive rods from car- Sun, F., B. Wang, Y. Du, X. Liu, Q. Lai, G. Li, J. Luo, and Z. Shao. 2010. charhinid sharks. International Journal of Systematic Bacteriology 1993:88– Arenibacter nanhaiticus sp. nov., isolated from marine sediment of the South 98. China Sea. International Journal of Systematic and Evolutionary Microbiol- Grimes, D. J., J. Stemmler, H. Hada, D. Maneval, F. M. Hetrick, E. B. May, R. ogy 60:78–83. T. Jones, M. K. Stoskopf, and R. R. Colwell. 1984b. Vibrio species associated Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and S. Kumar. 2011. with mortality of sharks held in captivity. Microbial Ecology 10:271–282. Mega5: molecular evolutionary genetics analysis using maximum likelihood, Hall, T. 1999. BioEdit, version 7.0.0. Available: http://www.mbio.ncsu. evolutionary distance, and maximum parsimony methods. Molecular Biology edu/bioedit/biodoc.pdf. (October 2014). and Evolution 28:2731–2739. Holmstrom, C., S. Egan, A. Franks, S. McCloy, and S. Kjelleberg. 2006. Terrell, S. P. 2004. An introduction to viral, bacterial, and fungal diseases of Downloaded by [Department Of Fisheries] at 00:14 17 November 2014 Antifouling activities expressed by marine surface associated Pseudoal- elasmobranchs. Pages 427–432 in M. F. L. Smith, D. A. Thoney, and R. teromonas species. FEMS Microbiology Ecology 41:47–58. E. Hueter, editors. The husbandry of elasmobrach fishes, volume 16. Ohio Horsley, R. W. 1977. A review on the bacterial flora of teleosts and elamobrachs, Biological Survey, Columbus. includings methods for its analysis. Journal of Fish Biology 10:529–553. Tindall, B. J., P. Kampfer, J. Euzeby, and A. Oren. 2006. Valid publication of Huang, J. W., B. Sun, and X. Zhang. 2010. Shewanella xiamenensis sp. nov., names of prokaryotes according to the rules of nomenclature: past history isolated from coastal sea sediment. International Journal of Systematic and and current practice. International Journal of Systematic and Evolutionary Evolutionary Microbiology 60:1585–1589. Microbiology 56:2715–2720. Jukes, T. H., and C. R. Cantor. 1969. Evolution of protein molecules. Pages Urakawa, H., T. Yoshida, M. Nishimura, and K. Ohwada. 2000. Characterization 21–132 in H. N. Munro, editor. Mamalian protein metabolism, volume 3. of depth-related population variation in microbial communities of a coastal Academic Press, New York. marine sediment using 16s rDNA-based approaches and quinone profiling. Knight, I. I., D. J. Grimes, and R. R. Colwell. 1987. Bacterial hydrolysis of Environmental Microbiology 5:542–554. urea in the tissues of carcharhinid sharks. Journal of Fisheries and Aquatic Van de Peer, Y., S. Nicolai, P. De Rijk, and R. Wachter. 1996. Database on the Sciences 45:357–360. structure of small ribosomal subunit DNA. Nucleic Acids Research 24:86–91. This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:15 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Occurrence of the Parasitic Copepod Ergasilus labracis on Threespine Sticklebacks from the South Coast of Newfoundland Alexandra A. Eavesa, Keng Pee Angb & Harry M. Murrayc a Stantec Consulting, Ltd., 141 Kelsey Drive, St. John's, Newfoundland A1B 0L2, Canada b Cooke Aquaculture, Inc., 1 Fundy Bay Drive, St. George, New Brunswick E5C 3E2, Canada c Fisheries and Oceans Canada, Aquatic Resources Division, Science Branch, Aquaculture, Biotechnology, and Aquatic Animal Health Section, 80 East White Hills Road, Post Office Box 5667, St. John's, Newfoundland A1C 5X1, Canada Published online: 16 Oct 2014.

To cite this article: Alexandra A. Eaves, Keng Pee Ang & Harry M. Murray (2014) Occurrence of the Parasitic Copepod Ergasilus labracis on Threespine Sticklebacks from the South Coast of Newfoundland, Journal of Aquatic Animal Health, 26:4, 233-242, DOI: 10.1080/08997659.2014.938871 To link to this article: http://dx.doi.org/10.1080/08997659.2014.938871

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:233–242, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.938871

ARTICLE

Occurrence of the Parasitic Copepod Ergasilus labracis on Threespine Sticklebacks from the South Coast of Newfoundland

Alexandra A. Eaves Stantec Consulting, Ltd., 141 Kelsey Drive, St. John’s, Newfoundland A1B 0L2, Canada Keng Pee Ang Cooke Aquaculture, Inc., 1 Fundy Bay Drive, St. George, New Brunswick E5C 3E2, Canada Harry M. Murray* Fisheries and Oceans Canada, Aquatic Resources Division, Science Branch, Aquaculture, Biotechnology, and Aquatic Animal Health Section, 80 East White Hills Road, Post Ofﬁce Box 5667, St. John’s, Newfoundland A1C 5X1, Canada

Abstract A study conducted from August to October 2013 surveyed Threespine Sticklebacks Gasterosteus aculeatus (n = 822) for the presence of parasitic copepods in the vicinity of large sea-cage salmonid farms in Bay d’Espoir, Newfoundland. The majority of parasitic copepods surveyed were Ergasilus labracis (n = 4,684). Other parasitic copepods observed on Threespine Sticklebacks during the survey included chalimus-stage Lepeophtheirus spp. (n = 3), adult Argulus alosae (n = 2), and a single Thersitina gasterostei. This represents a new host record for E. labracis. The copepods were present on fish collected in a broad range of temperatures (6.9–17.7◦C) and salinities (10.2–30.2 [Practical Salinity Scale]). The parasitic copepods were most commonly found on larger hosts estimated to be age 1 or older. Surprisingly, the highest infestations (approximately 65%) were found on regions of the hosts outside of the gills (behind the pectoral fins and pelvic spines); in some cases, the copepods had inflicted significant damage to the skin of their hosts. Among host fish with evidence of an additional infection, such as microsporidian tumors (xenomas) or hemorrhagic-like symptoms (dark red abdomens and bloody mucus), the prevalence of E. labracis was significantly higher (43.4%) than among healthy fish (28.9%) despite there being no significant difference in size between the two fish health groups. In contrast, intensity (mean number of individual parasites per host) was significantly higher among healthy hosts (23.6) than among unhealthy ones (7.63). Although this parasite has been listed as present in Newfoundland previously, it has a broad host range and has been reported to be pathogenic to

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 farmed salmonids. Therefore, its potential impact on wild and farmed ﬁsh populations around Newfoundland should not be underestimated.

Parasitic copepods are a persistent global problem affect- or commercial interest, particularly those affecting aquaculture ing wild and farmed fish alike (Johnson et al. 2004; Costello operations (reviewed by Pike and Wadsworth 1999; Johnson 2009; Saksida et al. 2011). There are several species of parasitic et al. 2004). A broader knowledge of the dynamics of parasitic copepods, many with broad host specificity (Costello 2009), but copepods infecting both wild and cultured fishes would improve the majority of studies have focused on species of ecological our understanding of how epidemics develop and persist in the

*Corresponding author: [email protected] Received March 28, 2014; accepted June 16, 2014 233 234 EAVES ET AL.

marine environment and assist with the development of strate- vironment (Kabata 1988; McDonald and Margolis 1995) and gies for reducing their negative impacts (Costello 2009; Bellay no information detailing the effects of infestation on the host. et al. 2013). Moreover, due to the expansion of finfish aquaculture on the Ergasilid copepods are an important group of parasites that south coast of Newfoundland in the last two decades, it is im- occur globally in marine, brackish, and freshwater habitats (Wil- perative to understand the parasite ecology of this region (Khan son 1911; Bere 1930; Hudson et al. 1994; Paperna 1996; Kil- 2009), but there are no published reports of parasitic copepods ian and Avenant-Oldewage 2013). So far as is known, ergasilid on farmed salmonids for this area. In response to this knowl- copepods remain free-swimming for the entire lifecycle of males edge gap, we examined resident nonsalmonid finfish (primarily and all developmental stages of females (Wilson 1911; Kabata Threespine Sticklebacks) to identify potential sentinel species 1988; Hogans 1991). Only the adult females are parasitic, and for monitoring the presence and effects of parasitic copepods on they have disproportionately large second antennae which func- host fish in the vicinity of aquaculture operations off southern tion as claws for grasping the host, most commonly by the gill Newfoundland and possibly other areas in the North Atlantic. lamellae (Wilson 1911; Kabata 1988). The parasitic females use their oral appendages for grazing on the epithelium, mucosal glands, and blood of the gills (Wilson 1911; Einzporn 1965). Prolific ergasilids are pathogenic to their hosts, causing destruc- METHODS tion of gill tissue, emaciation, and sometimes death (reviewed Field sampling.—Sampling was conducted in the Bay by Paperna and Zwerner 1982). Many common food fishes are d’Espoir region of the south coast of Newfoundland (Figure 1). suitable hosts on which parasitic adult females can be found in Three sampling trips were made between August and October abundance (Wilson 1911; Paperna and Zwerner 1976a; Hogans 2013: August 22–24, October 1–2, and October 29–30. For each 1991). Thus, due to their low host specificity and tolerance of a trip, 2 d of sampling were conducted; on day 1 beach seining was broad range of environmental conditions, ergasilids pose a po- done with a 1-m × 6-m pole seine (Wildlife Supply Company), tential threat to both wild and farmed fish stocks (Johnson and and on day 2 trawls were made using a 1.5-m2 small pelagic Rogers 1973; Hogans 1989; Johnson et al. 2004; Dezfuli et al. seine (Filmar, Inc.). Seining was conducted according to beach 2011). accessibility as determined by weather and tides (beach 2 was Recently, Threespine Sticklebacks Gasterosteus aculeatus in only sampled on trip 3), and pelagic trawls were made adjacent the Northeast Pacific were found to harbor higher abundances to two salmonid marine production sites. After collection, fish of developing parasitic copepods (Lepeophtheirus and Caligus were bagged collectively at each location on the first trip and spp.) than were found on wild salmon Oncorhynchus spp. in the individually on trips 2 and 3. All samples were frozen at −20◦C same area (Jones et al. 2006; Jones and Prosperi-Porta 2011). until examination in the laboratory. Temperature and salinity Jones and Prosperi-Porta (2011) hypothesized that these fish were recorded for each location with a YSI Castaway (CTD) serve as a sentinel species for the abundance and diversity of par- device. Salinities were recorded in terms of the Practical Salin- asitic copepods in coastal ecosystems. Threespine Sticklebacks ity Scale, which defines salinity according to a dimensionless have a nearly circumpolar distribution (Williams and Delbeek conductivity ratio instead of by weight (pounds of salt per thou- 1989), and most populations are anadromous: adults migrate sand pounds of seawater [‰]); the numerical values are similar into freshwater to breed in the spring, and surviving adults and between the two, however (UNESCO 1981). young of the year return to the marine environment in the late Parasitic copepod and fish health survey.—In the laboratory summer (Williams and Delbeek 1989). Threespine Sticklebacks fish were thawed, then total body length and weight were mea- are planktivorous predators (Peltonen et al. 2004) and are not sured. Fish were examined with the aid of a dissecting micro-

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 regarded to be highly active swimmers (Williams and Delbeek scope (Leica Wild MZ8) at 10 × magnification and illuminated 1989). In addition, they are host to several species of parasites with a fiber optic light source (Cole Palmer 9745-00). Fish were throughout their range (Hanek and Threlfall 1970a; McDonald placed into one of two categories according to their relative and Margolis 1995). health status: (1) healthy, i.e., having no external evidence of The fish parasites endemic to Newfoundland and Labrador deformities, lesions, or disease (this category included healthy are relatively poorly known (Hanek and Threlfall 1970a). Fresh- reproductive adults) or (2) unhealthy, i.e., having external ev- water species of ergasilids, such as Ergasilus auritus, E. lu- idence of deformity or disease such as microsporidian tumors ciopercarum, and another unidentified species, are known to (xenomas) or hemorrhagic-like symptoms (bloody mucus and parasitize landlocked populations of Rainbow Smelt Osmerus inflamed, dark red abdomens) (Figure 2). Fish were also sur- mordax, Brook Trout Salvelinus fontinalis, Atlantic Salmon veyed for the presence and number of parasitic copepods as Salmo salar, Brown Trout Salmo trutta, and Threespine Stick- well as their locations on the host. Only copepods with grasping leback (Sandeman and Pippy 1967; Hanek and Threlfall 1970b; claws firmly attached to host tissue were counted. All fish and Threlfall 1981; Cone and Ryan 1984; McDonald and Margolis copepods were inventoried and stored in 95% ethanol. Identi- 1995). However, there have been scant reports of this family fication of parasitic copepods was made according to Kabata of parasitic copepods occurring in Newfoundland’s marine en- (1988) and Hogans (1991). OCCURRENCE OF ERGASILUS LABRACIS IN NEWFOUNDLAND 235

FIGURE 1. Map of the Bay d’Espoir region on the south coast of Newfoundland. The sites at which beach seining was conducted (circles) are as follows: (1) Hardy Cove, (2) Raymond Point, (3) Day Cove, (4) Brimball Harbour, and (5) Furbey Cove. Pelagic trawls were conducted at Atlantic Salmon farm sites (squares [6–7]). Digital imaging.—Images of parasitic copepods on hosts RESULTS were recorded with a Leica MZ6 dissecting microscope Among the Threespine Sticklebacks surveyed (n = 822), equipped with an Olympus Q Color 5 camera. The morpholog- the predominant species of infesting parasitic copepods (n = ical features of the parasitic copepods were documented with a 4,684) was Ergasilus labracis (Krøyer 1864). Other parasitic Zeiss AX10 compound microscope with a Nikon Coolpix 4500 copepods observed during the survey included the chalimus camera. Image plates were assembled using Adobe Photoshop stage Lepeophtheirus spp. (n = 3), adult Argulus alosae (n =

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 CS 5 version 12.1. 2), and a single Thersitina gasterostei. Statistical analyses.—Due to limited sample sizes and the inconsistent presence of fish at sites (i.e., n = 8 for all pelagic Characterization of Ergasilus labracis trawls adjacent to finfish farming sites), all fish were pooled per Specimens of E. labracis collected from host tissues were ap- collection trip for analysis. Prevalence and mean intensity (mean proximately 1 mm in length (excluding egg sacs) with a trans- number of individual parasites per host) as defined by Bush et al. parent carapace and the antennary area of the cephalothorax (1997) were used in the analysis. Mean prevalence and intensity protruding anteriorly (Figure 3a). The carapace contained an and their 95% confidence intervals were calculated according opaque white to pale pink-colored yolk-like material with dark to Rozsa et al. (2000) using QPweb (Reiczigel et al. 2013). indigo coloration under the carapace on the ventral half (Figure Comparisons of host size and the intensity of parasitic copepod 3b). The endopod of the first leg was three-segmented (Figure infections were made using the nonparametric Kruskal–Wallis 3c), and the second segment of the endopod in legs 2–4 had two test and the pairwise multiple comparison procedure (Dunn’s setae along the medial margin (Figure 3d). The second antenna method); Fisher’s exact test was employed for comparison of had an inflated membrane between the first and second segment < prevalences. Differences for which there were P-values 0.05 that did not extend to the distal end of the second segment (Fig- were regarded as statistically significant. ure 3a, e). Also on the second antenna, there was no pronounced 236 EAVES ET AL.

bers, including the inner edge of the operculum and the bran- chiostegal membranes (Figure 5a) and the lamellae (Figure 5b); (2) the pectoral fin (underside of the fin and axil; Figure 5c); and (3) the pelvic fin, axil, and belly (Figure 5d). The copepods were relatively cryptic on the skin of the host and were only evident by examination with a dissecting microscope. Around the gill region, E. labracis were oriented parallel to the anterior– posterior axis of the host and parallel to the gill filaments on the lamellae (Figure 5b). In the pectoral fin region they were also oriented parallel to the anterior–posterior axis of the host with their ventral surfaces being adjacent to the soft tissue of the fish (Figure 5e). In the pelvic fin region they were also posi- tioned parallel to the anterior–posterior axis in the vicinity of the fin axil, though they were oriented toward the ventral midline along the belly of the fish under the pelvic fin (Figure 5d). The attached E. labracis ranged from newly settled females with an elongate carapace and dark ventral pigment visible through the dorsal surface of the animal to large, ovigerous specimens with broader carapaces and dark ventral pigment obscured by the opaque contents of their cephalothorax (Figure 5d). Egg sacs were observed, confirming that the parasite was reproductive on Threespine Sticklebacks, but they varied in appearance from densely packed pale eggs to dissociating transparent strands with blue-pigmented nauplia. Beneath the sites of attachment minor lesions were observed in the flesh of the host. Most commonly (less than 30 copepods per host), the lesions were small pits in the silver skin of the host that were approximately half the diameter of the cephalothorax of the parasite (compare Figure 5e and 5f). However, in crowded, high-intensity conditions (more than 30 E. labracis per host), the cephalic regions of the parasites were closely associated on the flesh of the fish such that the posterior ends of the parasites (and egg strings when present) were at an oblique angle to the surface of the host. In instances of highest infection intensity (more than 100 per host), the margins of the grazing areas of the parasites converged, creating large, deep lesions through the flesh of the fish and were entirely occupied by parasitic copepods FIGURE 2. Health conditions observed among Threespine Sticklebacks par- (Figure 5g). asitized by the parasitic copepod Ergasilus labracis. The asterisk denotes the There was a slight variation in the distribution of the E.

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 region of a fish displaying hemorrhagic-like symptoms, and the arrowheads labracis on hosts among sampling trips (Figure 6). For all trips indicate the locations of microsporidian xenomas. Scale bar = 1 cm (0.5 cm for the majority of the parasitic copepods (66.8, 69.9, and 61.4%, fish with xenomas or hemorrhagic-like symptoms and xenomas). [Color figure respectively) were observed on regions of the host other than available online.] the gills. The predominant areas of infection were the pectoral fin (trip 1) and pelvic fin (trips 2 and 3). curvature along the inner margin of the second segment and the subchela had a protuberance on the inner margin of the shaft Prevalence of Ergasilus labracis in Bay d’Espoir (Figure 3e). The claw of the second antenna had two outgrowths Among all Threespine Sticklebacks surveyed from the Bay close together on the proximal part of the inner margin (Figure d’Espoir region (pooling all sites and collection dates), the aver- 3f). age ± SE length and weight of fish with E. labracis were 45.6 ± 1.0 mm and 1.166 ± 0.070 g; the corresponding values for Locations of Ergasilus labracis fish with no parasitic copepods were 29.0 ± 0.3 mm and 0.244 Ergasilus labracis are small and hard to see on their host. ± 0.016 g. The average size, prevalence, and intensity for each They were attached primarily in three unarmored and mucous- collection trip are summarized in Table 1. The mean length and secreting regions of the fish’s body (Figure 4): (1) the gill cham- weight of fish was significantly different between trips 1 and OCCURRENCE OF ERGASILUS LABRACIS IN NEWFOUNDLAND 237

FIGURE 3. Morphological characteristics of adult female Ergasilus labracis collected from Threespine Sticklebacks in Bay d’Espoir: (a) dorsal view of the cephalothorax and inflated membrane (arrowhead) on the first segment of the second antenna; (b) ventral view of the pigment pattern; (c) endopod of the first leg three-segmented (arrowheads); (d) paired setae (arrows) on the inner margin of the endopod of the fourth swimming leg; (e) lateral view of the second antenna with an inflated membrane (arrowhead) and protuberance on a subchela (arrow); and (f) close-up of a claw of the second antenna shown in panel (e) with two outgrowths (arrows) close together on the proximal part of the inner margin and a protuberance on the subchela (arrowhead) as shown in (e). Abbreviations are as follows: sa = second antenna, ct = cephalothorax, gs = genital segment, fa = first antenna, sl = swimming legs, te = telson, and ma = margin of second segment. Scale bars: (a)–(b) = 0.2 mm, (c) = 0.04 mm, (d) = 0.02 mm, (e) = 0.1 mm, and (f) = 0.025 mm. [Color figure available online.]

2, as was the intensity of parasitic copepods on fish between DISCUSSION trips 1 and 2, and 1 and 3, but not between trips 2 and 3. The prevalence values were significantly different between all trips. Geographic Range Overall Threespine Stickleback health was also assessed. Although Ergasilus labracis is endemic to the eastern The gross pathology of unhealthy fish included the presence of seaboard of North America (summarized below), the only pre- microsporidian tumors (xenomas) among 17.9, 3.4, and 17.4% vious record of this parasite as far north as Newfoundland are of fish surveyed in trips 1, 2, and 3, respectively and inflamed in a taxonomic key (Kabata 1988) and an inventory of fish para- dark red abdomens and bloody mucous among 28.1% of the sites in Canada (McDonald and Margolis 1995), but no location, fish surveyed in trip 1 (only). The prevalence and intensity of date, or host information is included. Therefore, this is the first parasitic copepods could not be analyzed for these conditions comprehensive account of this parasite in Newfoundland. First individually because the sample sizes were insufficient to test described on Striped Bass Morone saxatilis in Baltimore, Mary- for significance. Instead, they were analyzed in collective groups land (Krøyer 1864), E. labracis were subsequently found on the of “unhealthy” and “healthy.” The size of the host was not sig- same host species in Woods Hole, Massachusetts (Wilson 1932); nificantly different between groups (Table 2). The prevalence of the Potomac River, Maryland (Paperna and Zwerner 1976a, E. labracis was significantly greater on unhealthy fish. Intensity 1976b); the Kouchibouguac River region, New Brunswick and the upper Bay of Fundy, Nova Scotia (Hogans 1985); and the

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 was signiﬁcantly greater on healthy hosts. Hudson River, New York, Rock Hall, Maryland, the Tabusintac River, New Brunswick, and the Annapolis River, Nova Scotia (Hogans 1991).

Host Record In this paper we report on a new host record for E. labracis: the Threespine Stickleback. Since its original description on Striped Bass, E. labracis have been found on other anadromous fish, including on the gills of returning sea-run Rainbow Smelt in the Saguenay River, Quebec (Frechet´ and Dodson 1983; accord- FIGURE 4. Diagram of a Threespine Stickleback illustrating the soft-tissue ing to Hogans 1989 and McDonald and Margolis 1995, these areas where Ergasilus labracis were predominantly attached, including the gill specimens were later reexamined and identified as E. labracis chamber, behind the base of the pectoral fin, and the dorsal surface of the pelvic by Kabata 1986 [see also Kabata 1988]); Brook Trout from spine and belly. the Tabusintac River (Frimeth 1987); Rainbow Smelt from the 238 EAVES ET AL.

FIGURE 5. Adult female Ergasilus labracis on Threespine Stickleback hosts collected from Bay d’Espoir: (a) E. labracis with egg strings protruding from the gill opercular region (arrows) on the right side of the host; (b) ventral anterior view of E. labracis among gill lamellae, with an arrow indicating the louse body, and arrowheads the paired egg strings; (c) E. labracis protruding from beneath a pectoral fin (arrow) on the right side of the host; (d) lateral view of mature female E. labracis (arrowheads) and a newly settled adult female (arrow) attached in the region of the pelvic fin; (e) lateral view of E. labracis (arrowheads) attached to a pectoral fin axil (the fin is folded anteriorly); (f) same host specimen as in (e) on a lower plane of focus with E. labracis pulled away revealing dark grazing pits (ellipses) in the silver flesh of the host; and (g) lateral view of a high-intensity infection with numerous E. labracis (arrowheads) beneath the pectoral and pelvic fins (the arrows indicate the margin of the lesion where parasitic copepods grazed through the flesh of the host). Abbreviations are as follows: go = gill operculum, pfr = pectoral fin rays, spf = spine of the pelvic fin, and pf = pectoral fin. Scale bars: (a)–(d) = 1 mm, (e)–(f) = 0.5 mm, and (f) = 2 mm. [Color figure available online.]

Miramichi River, New Brunswick (Hogans 1991); and White Perch Morone americana from Tolchester, Maryland (Hogans 1991). Marques and Hogans (1996) reported the same species of parasitic copepods on juvenile Atlantic Tomcod Microgadus tomcod collected from the Miramachi River. They also suggest that an E. centrarchidarum—a freshwater ergasilid—recorded by Bere (1930) on the same host species in the Passamaquoddy region of New Brunswick was most likely E. labracis (Marques and Hogans 1996). In addition, repeated outbreaks of this parasite have occurred on parr-stage farmed Atlantic Salmon in the lower Saint John River, New Brunswick, and in one instance resulted in mass mortalities prior to treatment (Hogans 1989;

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 O’Halloran et al. 1992).

Locations on Host This is the first report of an ergasilid being predominantly located on the outer dermis of its hosts. The majority of ergasilids infest the gills of their hosts and are rarely found on the skin or other tissues (Wilson 1911; Rogers and Hawke 1978; Oldewage and van As 1987); a notable exception is E. rhinos found in the nasal fossae of three species of sunfish (Flier Centrarchus macropterus, Pumpkinseed Lepomis gibbosus, and Redbreast Sunfish Lepomis auritus; Burris and Miller 1972). Threlfall (1981) report three specimens of E. auritus attached to the cornea of a landlocked Rainbow Smelt in a freshwater FIGURE 6. Distributions of Ergasilus labracis locations on Threespine Stick- lebacks from Bay d’Espoir, by collection trip. “Other” includes the dorsal spines lake in Newfoundland (Black River Pond). In stark contrast to and anal fins of the host. these earlier reports, the present study found that the majority OCCURRENCE OF ERGASILUS LABRACIS IN NEWFOUNDLAND 239

TABLE 1. Prevalence, intensity, and abundance of Ergasilus labracis on Threespine Sticklebacks in Bay d’Espoir, along with host size and environmental data (temperature and salinity ranges in the upper 2 m of the water column) for the locations at which ﬁsh were collected. Within columns, means with signiﬁcant differences (P < 0.05) are indicated by different lowercase letters.

Ergasilids (means and 95% Fish (means ± SEs) conﬁdence intervals) Trip Temperature (◦C) Salinitya n Length (mm) Weight (g) Prevalence (%) Intensity 1 16.4–17.7 22.0–29.0 668 30.8 ± 0.35 y 0.318 ± 0.018 y 26.7 (23.6–30.5) x 10.0 (7.34–14.8) x 2 13.7–15.8 12.2–29.1 115 52.5 ± 1.67 z 1.711 ± 0.124 z 56.5 (47.0–65.7) y 31.4 26.6–38.9) z 3 6.9–8.8 10.2–30.2 39 46.9 ± 2.45 z 1.195 ± 0.180 z 79.5 (63.5–90.7) z 27.1 (17.5–42.9) y

aPractical Salinity Scale.

of E. labracis surveyed (∼65%) were attached outside of the Among the Threespine Sticklebacks surveyed in the present gill region of their hosts. Other reports of ergasilids parasitiz- study, those with a high level of infestation (>100 lice/fish) (Pa- ing the skin of their hosts are uncommon. Cope (1959) found perna and Zwerner 1982) also had large lesions on the skin, but freshwater E. turgidus on the gills and fins of Threespine Stick- whether or not this was a moribund condition for the host could lebacks in Alaska, but the majority were on the gills and none not be determined postmortem. Regardless, the potential con- were reported on the skin. Arai (1969) reported E. auritus on sequences of skin lesions associated with the site of ergasilid the gills and body surface of Threespine Sticklebacks in the attachment may be similar to those described for other para- Northeast Pacific, but no further details were provided. Kabata sitic copepods, such as Lepeophtheirus salmonis, whose dermal (1988) also reported E. auritus on the gills, fins, and skin of grazing habits compromise the osmoregulation and swimming Threespine Sticklebacks, but again, no further details were pro- ability of host Atlantic Salmon (Wagner et al. 2003) as well as vided. Rogers and Hawke (1978) reported E. clupeidarum on increasing secondary infections, stress, and loss of growth (Pike the skin of freshwater Gizzard Shad Dorosoma cepedianum in and Wadsworth 1999; Johnson et al. 2004). Tennessee and Alabama. Subsequently, Yamashita (1980) re- On Atlantic Salmon infested by L. salmonis, the parasitic ported ergasilids on the skin of Red Seabream Pagrus major copepods are most commonly found on regions of the host with in a hatchery in Japan but later corrected the identification to thin skin that are devoid of scales, such as the head and the caligid copepodids (Yamashita 1981). regions between the dorsal and adipose fins and the perianal region (Wootten et al. 1982; Jonsd´ ottir´ et al. 1992; summarized Damage Inflicted to Host by Mustafa et al. 2000). These regions of the host are typified The typical pathology of ergasilids is hyperplasia of the gills as having thin, soft skin and no scales, and parasitic copepods (Einzporn 1965; Paperna and Zwerner 1976a, 1982; Frechet´ were most commonly found attached to the same tissue type and Dodson 1983; Hogans 1985; Kilian and Avenant-Oldewage on Threespine Sticklebacks in the present study (i.e., from the 2013). In addition to our predominantly finding E. labracis out- soft mucous-secreting regions at the base of the fins and on side of the gill region, this is only the second report of significant the abdomen under the base of the pelvic girdle). This raises skin lesions being associated with the location of ergasilid at- the question why this pattern of Ergasilus distribution has not tachment to the host and the first report of it occurring in a marine been more commonly reported on host fish. Are hosts with Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 species. Previously, Rogers and Hawke (1978) attributed a die- suitable outer tissues limited to a single freshwater and marine off of freshwater Gizzard Shad in Tennessee and Alabama to parasite host combination (Rogers and Hawke 1978; present “large numbers” of E. clupeidarum in skin lesions, but no details study)? Another possibility is sampling bias; previous studies of parasite abundance or distribution on the host were provided. of Ergasilus spp. largely report dissecting out the gills without

TABLE 2. Prevalence, intensity, and abundance of Ergasilus labracis on Threespine Sticklebacks in Bay d’Espoir according to the health appearance of hosts. Within columns, means with signiﬁcant differences (P < 0.05) are indicated different lowercase letters.

Ergasilids (means and 95% Fish (means ± SEs) conﬁdence intervals) Host appearance n Length (mm) Weight (g) Prevalence (%) Intensity Healthy 557 35.3 ± 0.6 z 0.640 ± 0.041 z 28.9 (25.2–32.8) y 23.6 (19.7–28.9) z Unhealthy 265 32.9 ± 0.6 z 0.368 ± 0.030 z 43.4 (37.5–49.4) z 7.63 4.78–12.5) y 240 EAVES ET AL.

examining the rest of the host fish’s body (i.e., Paperna and ther investigation, as do the dynamics of co-occurring epizootic Zwerner 1976b, 1982; Hogans 1989, 1991; Dezfuli et al. 2011; events among fish populations. Kilian and Avenant-Oldewage 2013). Occurrence of Ergasilus labracis with Regard to Temperature and Salinity Prevalence of Ergasilus labracis by Host Size and Health Hogans (1991) suggested that E. labracis are an estuarine Among the Threespine Sticklebacks we surveyed there was or freshwater parasite which is unlikely to survive the transi- a significant positive correlation between host size and the in- tion to a marine environment on an anadromous host. How- tensity of parasitic copepods. The average size of fish infected ever, fish infested with E. labracis have been collected from a with gill lice was significantly larger than that of uninfected broad range of environmental conditions, ranging in temperature fish (45.6 ± 1.0 mm versus 29.0 ± 0.3 mm). Cope (1959) from 0–24◦C and salinities ranging from 0–32‰ (Paperna and also found that Threespine Sticklebacks without ergasilid in- Zwerner 1976b; Frimeth 1987; Hogans 1989). Paperna and festation in a freshwater stream in Alaska were relatively small Zwerner (1976b, 1982) describe these parasitic copepods to (average size, 35.8 mm). This disparity in average size likely be as euryhaline as their anadromous hosts. The present study represents different age-classes of fish; Reimchen (1990) and supports the conclusion that E. labracis are indeed tolerant of Barber (2003) classify Threespine Sticklebacks less than 40 mm a broad range of environmental conditions, having found them as young of the year and those greater than 40 mm as age 1 among fish collected in water temperatures ranging from 6.9◦C or older. Generally, larger fish harbor more parasitic copepods to 17.7◦C and salinities of 10.2–30.2 (Practical Salinity Scale). (Pike and Wadsworth 1999), and with respect to E. labracis on Consistent with these observations, in a study of fish assem- wild fish in particular, Paperna and Zwerner (1976b) reported blages in a Newfoundland estuary Methven et al. (2001) reported the prevalence of infection on all Striped Bass over 1 year old no correlation between salinity and the abundance of Three- (n = 192) and from all collecting localities to have exceeded spine Sticklebacks. However, considering that the majority of 90%. To date, the highest reported infection intensity of E. the female ergasilid developmental life cycle and the entirety of labracis on a wild fish involved an age-1 White Perch that had the males’ is spent free swimming (Wilson 1911; Kabata 1988), 2,757 individuals parasitizing its gills (Paperna and Zwerner other oceanographic factors such as current pattern, speed, wind 1976a). However, during an outbreak of E. labracis that killed influence, and site flushing (Revie et al. 2003) should be exam- several thousand parr-sized farmed Atlantic Salmon in only a ined as potential parameters affecting the distribution of this few days, Hogans (1989) reported a prevalence of 100% and species of parasite. an intensity of 84.6 E. labracis per host; the highest number of Lastly, the comparatively small sample size (n = 39) on trip parasitic copepods infesting a single fish was 2,229. Not only 3 corresponded to a marked decline in the temperature range can the immediate pathological effects of parasitic copepod in- in the near-surface sampling environment. Threespine Stickle- festation be detrimental to hosts, heavily infected hosts are less backs drop down to slightly deeper water in winter (Jones and likely to withstand environmental challenges such as oxygen John 1978); despite this, the prevalence of parasitic copepods depletion, algal blooms, and pollution (Paperna and Zwerner among the fish sampled remained high (79.5%). As described 1976b; Lafferty and Kuris 1999). above, E. labracis are tolerant of colder temperatures and can Host health also had a significant effect on parasitic cope- overwinter on hosts and even continue to produce egg strings pod infestation levels. When the infestation levels of healthy (albeit at a slower rate) during the coldest months of the year and unhealthy host groups were compared, we observed that the (Paperna and Zwerner 1976a, 1976b, 1982). prevalence of parasitic copepod infestation was greater among

Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 unhealthy fish. In contrast, the intensity of infection was signifi- Conclusions cantly lower among unhealthy fish despite there being no signif- The parasitic copepod E. labracis appears to be widespread icant difference in host size. Parasitic copepod infestation may among the Threespine Sticklebacks surveyed in the south coast render hosts more susceptible to secondary infection (Pike and region of Newfoundland. This raises the specter that this species Wadsworth 1999; Mustafa et al. 2000; Wagner et al. 2003) or be might serve as a reservoir for E. labracis in the western North At- a vector for the transmission of microsporidian, viral, and bac- lantic, similar to their role in the Northeastern Pacific (Jones and terial pathogens (Weissenberg 1968; Nylund et al. 1994; Barker Prosperi-Porta 2011). Considerable research will be required et al. 2009; Jakob et al. 2011; Oelckers et al. 2014). However, to understand the dynamics of these parasitic copepods and simultaneous infections have a significant effect on the available Threespine Stickleback ecology (Jones et al. 2006) and host resources of the host (Threlfall 1968; Ward et al. 2005). So even health (presented here). In addition, E. labracis are known to though lice infestation could increase the likelihood of a fish’s infest farmed salmonids in another region of the North Atlantic becoming infected by another pathogen, parasitic copepods may (Hogans 1989; O’Halloran et al. 1992). Both the presence of this have greater infestation success (i.e., achieve higher infestation parasite on farmed fish in Newfoundland and the potential for intensities) on healthy hosts. The influence of simultaneous host transmission between wild and farmed fish populations needs health afflictions on parasitic copepod infestation warrants fur- to be investigated. Ergasilus labracis are not as rare or host OCCURRENCE OF ERGASILUS LABRACIS IN NEWFOUNDLAND 241

specific as previously thought (Hogans 1989). Moreover, they of Quebec and their usefulness as biological indicators.] Canadian Journal of have the capacity to exploit more regions of the host and inflict Zoology 61:621–626. far greater damage to fish than has been previously recognized. Frimeth, J. P.1987. A survey of the parasites of nonanadromous and anadromous Brook Charr (Salvelinus fontinalis) in the Tabusintac River, New Brunswick, Given this and their tolerance of a broad range of aquatic en- Canada. Canadian Journal of Zoology 65:1354–1362. vironmental conditions, the potential impact of this parasite on Hanek, G., and W. Threlfall. 1970a. Parasites of the Threespine Stickleback wild and farmed fish populations should not be underestimated. (Gasterosteus aculeatus) in Newfoundland and Labrador. Journal of the Fish- eries Research Board of Canada 27:901–907. Hanek, G., and W.Threlfall. 1970b. Ergasilus auritus Markewitsch, 1940 (Cope- poda: Ergasilidae) from Gasterosteus aculeatus Linnaeus, 1758 in Newfound- ACKNOWLEDGMENTS land. Canadian Journal of Zoology 48:185–187. The authors are grateful to Dwight Drover, Curtis Pennell, Hogans, W. E. 1985. Northern range extension record for Ergasilus labracis and Sharon Kenny for technical assistance with field sampling; (Copepoda, Ergasilidae) parasitic on the Striped Bass (Morone saxatilis). to the farm operators and boat crew of Cooke Aquaculture, Crustaceana 49:97–98. Hogans, W. E. 1989. Mortality of cultured Atlantic Salmon, Salmo salar L., parr Inc. (Cold Ocean Salmon Ltd.), for conducting the trawls; caused by an infection of Ergasilus labracis (Copepoda: Poecilostomatoida) to Sebastien Donnet for processing CTD data; and to Pierre in the lower Saint John River, New Brunswick, Canada. Journal of Fish Goulet for preparing the map images used in Figure 1. This Diseases 12:529–531. work was funded by the Department of Fisheries and Oceans Hogans, W. E. 1991. Redescription of Ergasilus labracis (Krøyer, 1864) (Cope- (DFO), Aquaculture Collaborative Research and Development poda: Poecilostomatoida), a parasite of anadromous fishes from the East Coast of North America. Canadian Journal of Zoology 69:651–654. Program, and Cooke Aquaculture, Inc. (Cold Ocean Salmon Hudson, P. L., C. A. Bowen II, and R. M. Stedman. 1994. New records of Ltd.). An earlier version of this manuscript benefitted from com- Ergasilus (Copepoda: Ergasilidae) in the Laurentian Great Lakes, including ments by John Brattey at DFO and two anonymous reviewers. a lakewide review of records and host associations. Canadian Journal of Zoology 72:1002–1009. Jakob, E., D. E. Barker, and K. A. Garver. 2011. Vector potential of the salmon louse Lepeophtheirus salmonis in the transmission of infectious hematopoi- REFERENCES etic necrosis virus (IHNV). Diseases of Aquatic Organisms 97:155–165. Arai, H. P. 1969. Preliminary report on the parasites of certain marine fishes Johnson, S. K., and W. A. Rogers. 1973. Distribution of the genus Ergasilus in of British Columbia. Journal of the Fisheries Research Board of Canada several Gulf of Mexico drainage basins. Alabama Agricultural Experiment 26:2319–2337. Station Bulletin (Auburn University) 445. Barber, I. 2003. Parasites and size-assortative schooling in Three-Spined Stick- Johnson, S. C., J. W. Treasurer, S. Bravo, K. Nagasawa, and Z. Kabata. 2004. A lebacks. Oikos 101:331–337. review of the impact of parasitic copepods on marine aquaculture. Zoological Barker, D. E., L. M. Braden, M. P. Coombs, and B. Boyce. 2009. Preliminary Studies 43:229–243. studies on the isolation of bacteria from sea lice, Lepeophtheirus salmonis, Jones, D. H., and A. W. G. John. 1978. A Three-Spined Stickleback, Gas- infecting farmed salmon in British Columbia, Canada. Parasitology Research terosteus aculeatus L., from the North Atlantic. Journal of Fish Biology 105:1173–1177. 13:231–236. Bellay, S., E. F. de Oliveira, M. Almeida-Neto, D. P. Lima Junior, R. M. Take- Jones, S. R. M., and G. Prosperi-Porta. 2011. The diversity of sea lice (Copepoda: moto, and J. L. Luque. 2013. Developmental stage of parasites influences the Caligidae) parasitic on Three-Spine Stickleback (Gasterosteus aculeatus)in structure of fish–parasite networks. PLoS (Public Library of Science) ONE coastal British Columbia. Journal of Parasitology 97:399–405. [online serial] 8(e75710). Jones, S. R. M., G. Prosperi-Porta, E. Kim, P. Callow, and N. B. Hargreaves. Bere, R. 1930. The parasitic copepods of the fish of the Passamaquoddy region. 2006. The occurrence of Lepeophtheirus salmonis and Caligus clemensi Contributions to Canadian Biology and Fisheries 5:423–430. (Copepoda: Caligidae) on Three-Spine Stickleback Gasterosteus aculeatus Bush, A. O., K. D. Lafferty, J. M. Lotz, and A. W. Shostak. 1997. Parasitol- in coastal British Columbia. Journal of Parasitology 92:473–480. ogy meets ecology on its own terms: Margolis et al. revisited. Journal of Jonsd´ ottir,´ H., J. E. Bron, R. Wootten, and J. F. Turnbull. 1992. The Parasitology 83:575–583. histopathology associated with the preadult stages of Lepeophtheirus salmo- Burris, K. W., and G. C. Miller. 1972. Ergasilus rhinos sp. n. (Copepoda: Cy- nis on the Atlantic Salmon, Salmo salar L. Journal of Fish Diseases 13: Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 clopoida) from the nasal fossae of three centrarchid fishes in North Carolina. 303–310. Journal of Parasitology 58:600–604. Kabata, Z. 1986. Type specimens of Ergasilus funduli Krøyer, 1863 (Crustacea: Cone, D. K., and P. M. Ryan. 1984. Population sizes of metazoan parasites of Copepoda) re-examined. Steenstrupia 12:153–156. Brook Trout (Salvelinus fontinalis) and Atlantic Salmon (Salmo salar)ina Kabata, Z. 1988. Copepoda and Brachiura. Canadian Special Publication in small Newfoundland lake. Canadian Journal of Zoology 62:130–133. Fisheries and Aquatic Sciences 101. Cope, O. B. 1959. New parasite records from stickleback in an Alaska stream. Khan, R. A. 2009. Parasites causing disease in wild and cultured fish in New- Transactions of the American Microscopical Society 78:157–162. foundland. Icelandic Agricultural Sciences 22:29–35. Costello, M. J. 2009. The global economic cost of sea lice to the salmonid Kilian, E., and A. Avenant-Oldewage. 2013. Infestation and pathological alter- farming industry. Journal of Fish Diseases 32:115–118. ations of Ergasilus sarsi (Copepoda) on the Tanganyika Killifish from Africa. Dezfuli, B. S., L. Giari, A. Lui, M. Lorenzoni, and E. J. Noga. 2011. Mast cell Journal of Aquatic Animal Health 25:237–242. responses to Ergasilus (Copepoda), a gill ectoparasite of Sea Bream. Fish and Krøyer, H. 1864. Bidrag til kundskab om snyltekrebsene (slutning). [Contribu- Shellfish Immunology 30:1087–1094. tion to the knowledge of parasitic crustaceans (containment).] Naturhistorisk Einzporn, T. 1965. Nutrition of Ergasilus sieboldi Nordmann, II. The uptake of Tidsskrift 3:321–426. food and the food material. Acta Parasitologica Polonica 13:373–380. Lafferty, K. D., and A. M. Kuris. 1999. How environmental stress affects the Frechet,´ A., and J. J. Dodson. 1983. Les parasites de l‘eperlan´ d’Amerique´ impacts of parasites. Limnology and Oceanography 44:925–931. (Osmerus mordax) anadrome du Quebec´ et leur utilite´ comme etiquettes´ MacKinnon, B. M. 1998. Host factors important to sea lice infection. ICES biologiques. [Parasites of the anadromous Rainbow Smelt (Osmerus mordax) Journal of Marine Science 55:188–192. 242 EAVES ET AL.

Marques, F., and W. E. Hogans. 1996. Ergasilus labracis (Krøyer, 1864) (Cope- Revie, C. W., G. Gettinby, J. W. Treasurer, and C. Wallace. 2003. Identifying poda, Ergasilidae) on Microgadus tomcod (Osteichthyes, Gadidae) in estuar- factors affecting sea lice Lepeophtheirus salmonis abundance on Scottish ine waters off New Brunswick, Canada. Crustaceana 69:532–534. salmon farms using general linear models. Diseases of Aquatic Organisms McDonald, T. E., and L. Margolis. 1995. Synopsis of the parasites of fishes of 57:85–95. Canada: supplement (1978–1993). Canadian Special Publication of Fisheries Rogers, W. A., and J. P. Hawke. 1978. The parasitic copepod Ergasilus from and Aquatic Sciences 122. the skin of the Gizzard Shad Dorosoma cepedianum. Transactions of the Methven, D. A., R. L. Haedrich, and G. A. Rose. 2001. The fish assemblage American Microscopical Society 97:244. of a Newfoundland estuary: diel, monthly, and annual variation. Estuarine, Rozsa L., J. Reiczigel, and G. Majoros. 2000. Quantifying parasites in samples Coastal and Shelf Science 52:669–687. of hosts. Journal of Parasitology 86:228–232. Mustafa, A., D. J. Speare, J. Daley, G. A. Conboy, and J. F. Burka. 2000. Saksida, S. M., D. Morrison, M. Sheppard, and I. Keith. 2011. Sea lice man- Enhanced susceptibility of seawater cultured Rainbow Trout, Oncorhynchus agement on salmon farms in British Columbia, Canada. Pages 235–278 in mykiss (Walbaum), to the microsporidian Loma salmonae during a primary S. Jones and R. Beamish, editors. Sea lice: an integrated approach to under- infection with the sea louse, Lepeophtheirus salmonis. Journal of Fish Dis- standing parasite abundance and distribution. Wiley, Toronto. eases 23:337–341. Sandeman, I. M., and J. H. C. Pippy. 1967. Parasites of freshwater fishes Nylund, A., T. Hovland, K. Hodneland, F. Nilsen, and P. Løvik. 1994. Mech- (Salmonidae and Coregonidae) of insular Newfoundland. Journal of the Fish- anisms for transmission of infectious salmon anemia (ISA). Diseases of eries Research Board of Canada 24:1911–1943. Aquatic Organisms 19:95–100. Threlfall, W. 1968. A mass die-off of Three-Spined Sticklebacks (Gasterosteus Oelckers, K., S. Vike, H. Duesund, J. Gonzalez, S. Wadsworth, and A. Ny- aculeatus L.) caused by parasites. Canadian Journal of Zoology 46:105–106. lund. 2014. Caligus rogercresseyi as a potential vector for transmission of Threlfall, W. 1981. Metazoan parasites of two populations of Rainbow Smelt in infectious salmon anemia (ISA) virus in Chile. Aquaculture 420/421:126– Newfoundland, Canada. Journal of Parasitology 67:129–131. 132. UNESCO (United Nations Educational, Scientific, and Cultural Organization). O’Halloran, J., J. Carpenter, D. Ogden, W. E. Hogans, and M. Jansen. 1992. 1981. Tenth report of the Joint Panel on Oceanographic Tables and Standards. Ergasilus labracis on Atlantic Salmon. Canadian Veterinary Journal 33:75. Technical Papers in Marine Science 36. Oldewage, W. H., and J. G. van As. 1987. Observations on the attachment of a Wagner, G. N., R. S. McKinley, P. A. Bjørn, and B. Finstad. 2003. Physiological piscine gill parasitic ergasilid (Crustacea: Copepoda). South African Journal impact of sea lice on swimming performance of Atlantic Salmon. Journal of of Zoology 22:313–317. Fish Biology 62:1000–1009. Paperna, I. 1996. Parasites, infections, and diseases of fishes in Africa: an update. Ward, A. J. W., A. J. Duff, J. Krause, and I. Barber. 2005. Shoaling behaviour CIFA (Committee for Inland Fisheries of Africa) Technical Paper 31. of sticklebacks infected with the microsporidian parasite, Glugea anomala. Paperna, I., and D. E. Zwerner. 1976a. Parasites and diseases of Striped Bass, Environmental Biology of Fishes 72:155–160. Morone saxatilis (Walbum), from the lower Chesapeake Bay. Journal of Fish Weissenberg, R. 1968. Intracellular development of the microsporidian Glugea Biology 9:267–287. anomala Moniez in hypertrophying migratory cells of the fish Gasterosteus Paperna, I., and D. E. Zwerner. 1976b. Studies on Ergasilus labracis Krøyer aculeatus L., an example of the formation of “xenoma” tumors. Journal of (Cyclopidea: Ergasilidae) on Striped Bass, Morone saxatilis, from the lower Protozoology 15:44–57. Chesapeake Bay, I. Distribution, life cycle, and seasonal abundance. Canadian Williams, D. D., and J. C. Delbeek. 1989. Biology of the Threespine Stickleback, Journal of Zoology 54:449–462. Gasterosteus aculeatus, and the Blackspotted Stickleback, G. wheatlandi, Paperna, I., and D. E. Zwerner. 1982. Host–parasite relationship of Ergasilus during their marine pelagic phase in the Bay of Fundy, Canada. Environmental labracis Krøyer (Cyclopidea, Ergasilidae) and the Striped Bass, Morone sax- Biology of Fishes 24:33–41. atilis (Walbaum), from the lower Chesapeake Bay. Annales de Parasitologie Wilson, C. B. 1911. North American parasitic copepods belonging to the family (Paris) 57:393–405. Ergasilidae. Proceedings of the United States National Museum 39:263–400. Peltonen, H., M. Vinni, A. Lappalainen, and J. Ponni.¨ 2004. Spatial feeding Wilson, C. B., 1932. The copepods of the Woods Hole region, Massachusetts. patterns of herring (Clupea harengus L.), Sprat (Sprattus sprattus L.), and Bulletin of the United States National Museum 158. the Three-Spined Stickleback (Gasterosteus aculeatus L.) in the Gulf of Wootten, R., J. W. Smith, and E. A. Needham. 1982. Aspects of biology of Finland, Baltic Sea. ICES Journal of Marine Science 61:966–971. the parasitic copepods Lepeophtheirus salmonis and Caligus clemensi on Pike, A. W., and S. L. Wadsworth. 1999. Sea lice on salmonids: their biology farmed salmonids, and their treatment. Proceedings of the Royal Society of and control. Advances in Parasitology 44:233–337. Edinburgh, Section B 81:185–197. Reiczigel, J., L. Rozsa, and A. Reiczigel. 2013. Quantitative parasitology (QP- Yamashita, K. 1980. Parasitic copepod (Ergasilus sp.) found on the epidermis web). Available: http://www2.univet.hu/qpweb. (March 2014). of the larval fish of Red Seabream (Pagrus major). Fish Pathology 15:91–94. Downloaded by [Department Of Fisheries] at 00:15 17 November 2014 Reimchen, T. E. 1990. Size-structured mortality in a Threespine Stickleback Yamashita, K. 1981. Correction of the report entitled “The parasites found on (Gasterosteus aculeatus)–Cutthroat Trout (Oncorhynchus clarki) community. the epidermis of the larval fish of Red Seabream (Pagrus major)” in Fish Canadian Journal of Fisheries and Aquatic Sciences 47:1194–1205. Pathology 15:91–94. Fish Pathology 16:55. This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:16 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Florfenicol Residues in Rainbow Trout after Oral Dosing in Recirculating and Flow-through Culture Systems J. R. Meinertza, K. R. Hessa, J. A. Bernardya, M. P. Gaikowskia, M. Whitselb & R. G. Endrisc a U.S. Geological Survey, Upper Midwest Environmental Sciences Center2630 Fanta Reed Road, La Crosse, Wisconsin 54603, USA b MPI Research, Inc., 3058 Research Drive, State College, Pennsylvania 16801, USA c Merck Animal Health, 556 Morris Avenue, Summit, New Jersey 07901, USA Published online: 16 Oct 2014.

To cite this article: J. R. Meinertz, K. R. Hess, J. A. Bernardy, M. P. Gaikowski, M. Whitsel & R. G. Endris (2014) Florfenicol Residues in Rainbow Trout after Oral Dosing in Recirculating and Flow-through Culture Systems, Journal of Aquatic Animal Health, 26:4, 243-250, DOI: 10.1080/08997659.2014.945046 To link to this article: http://dx.doi.org/10.1080/08997659.2014.945046

PLEASE SCROLL DOWN FOR ARTICLE

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:243–250, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.945046

ARTICLE

Florfenicol Residues in Rainbow Trout after Oral Dosing in Recirculating and Flow-through Culture Systems

J. R. Meinertz,* K. R. Hess, J. A. Bernardy, and M. P. Gaikowski U.S. Geological Survey, Upper Midwest Environmental Sciences Center, 2630 Fanta Reed Road, La Crosse, Wisconsin 54603, USA M. Whitsel MPI Research, Inc., 3058 Research Drive, State College, Pennsylvania 16801, USA R. G. Endris Merck Animal Health, 556 Morris Avenue, Summit, New Jersey 07901, USA

Abstract Aquaflor is a feed premix for fish containing the broad spectrum antibacterial agent florfenicol (FFC) incorporated at a ratio of 50% (w/w). To enhance the effectiveness of FFC for salmonids infected with certain isolates of Flavobac- terium psychrophilum causing coldwater disease, the FFC dose must be increased from the standard 10 mg·kg−1 body weight (BW)·d−1 for 10 consecutive days. A residue depletion study was conducted to determine whether FFC residues remaining in the fillet tissue after treating fish at an increased dose would be safe for human consumption. Groups of Rainbow Trout Oncorhynchus mykiss (total n = 144; weight range, 126–617 g) were treated with FFC at 20 mg·kg−1 BW·d−1 for 10 d in a flow-through system (FTS) and a recirculating aquaculture system (RAS) each with a water temperature of ∼13◦C. The two-tank RAS included a nontreated tank containing 77 fish. Fish were taken from each tank (treated tank, n = 16; nontreated tank, n = 8) at 6, 12, 24, 48, 72, 120, 240, 360, and 480 h posttreatment. Florfenicol amine (FFA) concentrations (the FFC marker residue) in skin-on fillets from treated fish were greatest at 12 h posttreatment (11.58 µg/g) in the RAS and were greatest at 6 h posttreatment (11.09 µg/g) in the FTS. The half-lives for FFA in skin-on fillets from the RAS and FTS were 20.3 and 19.7 h, respectively. Assimilation of FFC residues in the fillets of nontreated fish sharing the RAS with FFC-treated fish was minimal. Florfenicol water concentrations peaked in the RAS-treated tank and nontreated tanks at 10 h (453 µg/L) and 11 h (442 µg/L) posttreatment, respectively. Monitoring of nitrite concentrations throughout the study indicated the nitrogen oxidation efficiency of the RAS biofilter was minimally impacted by the FFC treatment. Downloaded by [Department Of Fisheries] at 00:16 17 November 2014

Aquaflor, also marketed as Aquafen or Florocol, is a feed pre- terium psychrophilum at a dosage of 10 mg·kg−1 body weight mix for fish containing the broad spectrum antibacterial agent (BW)·d−1 for 10 consecutive days (USFDA CVM 2007: NADA florfenicol {FFC; 2,2-dichloro-N-[1-(fluoromethyl)-2-hydroxy- 141-246). Subsequent field experience suggested a higher dose 2-[4-(methylsulfonyl) phenyl]ethyl]acetamide} and is regis- rate might prove more efficacious, and as a result work was un- tered for use in countries throughout the world to control various dertaken to support development of a higher dose rate for FFC susceptible pathogens in a variety of commercially important in freshwater-reared salmonids. freshwater and marine fishes. Florfenicol (Aquaflor) was ap- With the potential need to increase the dose rate, animal proved for use in 2007 in freshwater-reared salmonids to control safety and human food safety issues had to be considered. With mortality due to coldwater disease associated with Flavobac- respect to animal safety, FFC was found to have no impact on

*Corresponding author: [email protected] Received January 8, 2014; accepted July 7, 2014 243 244 MEINERTZ ET AL.

fish pathology or feeding when fish were treated at the ele- 2,000 mg/kg to provide a nominal dose of 20 mg·kg−1 BW·d−1 − vated doses. Data indicated that FFC doses up to 34.9 mg·kg 1 when fish were fed at a rate of 1% BW/d. Feeds were stored − BW·d 1 for 20 consecutive days in Channel Catfish Ictalurus separately at about −20◦C until used. − − punctatus,upto75mg·kg 1 BW·d 1 for 20 d in hybrid Striped Nonmedicated and medicated feed samples were collected Bass (female White Bass Morone chrysops × male Striped Bass before the start and after the end of the treatment phase. The − − M. saxatilis), or up to 100 mg·kg 1 BW·d 1 for 10 d in Atlantic FFC concentrations in the feed were determined by Eurofins Salmon Salmo salar did not result in pathology (Inglis et al. Lancaster Laboratories (Portage, Michigan) according to meth- 1991; Gaikowski et al. 2003, 2012; Straus et al. 2012). With ods described in Hayes (2005). The mean FFC concentration in respect to human food safety, FFC distribution, metabolism, the medicated feed was 1,870 mg/kg before the treatment phase and depletion in a variety of fishes had been well characterized and 1,860 mg/kg after the treatment phase. Florfenicol was not (Martinsen et al. 1993; Horsberg et al. 1994, 1996; Pinault et al. detected in nonmedicated feed (reporting limit = 0.1 mg/kg). 1997; Samuelsen et al. 2003; Yanong et al. 2005; Wrzesinski Test fish.—The species, the number of test fish, animal wel- et al. 2006; Feng et al. 2008; Bowser et al. 2009; Kosoff et al. fare, and euthanasia techniques were approved by the U.S. Ge- 2009; Gaikowski et al. 2010). However, despite the studies pre- ological Survey, Upper Midwest Environmental Sciences Cen- viously conducted, data have not been developed to determine ter’s (UMESC; La Crosse, Wisconsin) Institutional Animal Care − whether after treatment with a dose greater than 10 mg· kg 1 and Use Committee. Rainbow Trout were reared at UMESC in − BW·d 1 the skin-on fillet from a salmonid would be safe for flowing well water at a nominal temperature of ∼13◦C. Fish human consumption (the maximum residue concentration con- were fed Silver Cup nonmedicated fish feed (Nelson and Sons, sidered safe for human consumption is 1 mg/kg). Therefore, Murray, Utah) throughout rearing. Fish were 2 years 3 months to fill the data gap, a marker-residue depletion study was con- old (weight range, 126–617 g) when the study was initiated and − − ducted by treating a salmonid at 20 mg FFC·kg 1 BW·d 1 for were used without regard to gender. 10 d to characterize the depletion of florfenicol residues from Test systems.—The commercial RAS consisted of two the skin-on fillet tissue. polyethylene tanks (volume, 1,900 L; height, 76 cm; diameter, In areas with limited water availability, recirculating aquacul- 89 cm), mechanical filters, and a biological filter (Fish FarmTM ture systems (RAS) are becoming more prominent. Therefore, II, Aquatic Eco-Systems, Apopka, Florida). Supplemental aer- fish were treated in RAS as well as a flow-through system (FTS) ation was provided through air stones immersed in each tank. to allow for a direct comparison of FFC residue depletion from The RAS was located in an environmentally controlled room the fillet tissue of fish treated in each system. The study was also in which the air temperature was adjusted to maintain a water designed to determine whether fish in a nontreated tank of an temperature of ∼13◦C, and incandescent lighting provided a RAS sharing water with fish in a treated tank would assimilate photoperiod similar to the natural photoperiod at the time of the FFC residues in their fillet tissue. study (∼12 h daylight). An RAS stabilization phase was initiated by filling the sys- METHODS tem with well water and inoculating the biofilter with a bacterial Test article and analytical standards.—The test article was culture (Fritz-Zyme Turbo Start 700, Fritz Industries, Mesquite, Aquaflor Type A Medicated Premix. Aquaflor Type A Medi- Texas). Proline ammonium chloride (Aquatic Eco-Systems) was cated Premix contains FFC (Chemical Abstract Service number added to provide an initial food source for the bacteria be- 73231-34-2; molecular weight, 358.22 atomic mass units; 50% fore adding fish. Sodium bicarbonate (Thermo Fisher Scientific, w/w; source, Merck Animal Health, Omaha, Nebraska) and was Waltham, Massachusetts) was added periodically throughout the incorporated into fish feed to prepare the FFC-medicated feed. study to maintain a suitable pH. After about 12 weeks, the water

Downloaded by [Department Of Fisheries] at 00:16 17 November 2014 Florfenicol (purity 99.7%) and florfenicol amine [FFA; chemistry in the system was considered stable, i.e., the alkalinity, (1R,2S)-2-amino-3-fluoro-1- (4-methylsulfonylphenyl) propan- pH, and concentrations of ammonia nitrogen (NH3-N), nitrite 1-ol; purity 97.3%] were supplied by Merck and Company (NO2), and nitrate (NO3) were within acceptable limits (Tim- (Whitehouse Station, New Jersey). The FFC was used as an mons and Ebeling 2007:39–77). Thereafter, 149 Rainbow Trout analytical reference standard during feed and water analyses were stocked into the tank assigned to receive FFC-medicated and the FFA was used as an analytical reference standard during feed and 77 Rainbow Trout were stocked into the tank assigned tissue analyses. to receive nonmedicated feed. Feed preparation and analyses.—Nonmedicated feed (ex- Daily RAS maintenance included measuring water flow truded, 4-mm, floating pellets) was prepared at Rangen (Buhl, through the system, dissolved oxygen concentrations (Yellow Idaho) according to standard aquaculture feed production pro- Springs Instruments model 55/12FT dissolved oxygen meter, cedures. Medicated feed with FFC (extruded, 4-mm, floating YSI, Yellow Springs, Ohio), pH (Beckman model theta 255 pH pellets) was also prepared at Rangen according to standard meter, Beckman Coulter, Fullerton, California), and the water procedures by mixing FFC into feed mash before pelleting. temperature (PT100 RTD thermometer, OMEGA Engineering, Medicated feed was prepared at a target FFC concentration of Stamford, Connecticut) in each tank. Total ammonia (ammonia FLORFENICOL AMINE DEPLETION FROM RAINBOW TROUT 245

nitrogen; test code 3642-SC), nitrate (nitrate-nitrogen; test code through each phase in all tanks was vigorous. Therefore, feed 3689-SC), and nitrite (nitrite-nitrogen; test code 3694-SC) were consumption was estimated at 100%. also measured in each tank using a LaMotte SMART2 colorime- Sampling fish and water.—During the pretreatment phase, ter (LaMotte, Chestertown, Maryland). Measurements were al- control fillet tissue was acquired by removing five fish from ways taken in the morning before the tanks were cleaned. Water each tank, euthanizing the fish by electrocution, weighing the samples were diluted with deionized water if the expected con- fish, removing skin-on fillets, rinsing the fillets with deionized centration was outside of the range of a test, and final readings water, individually sealing fillets in freezer bags, and storing were adjusted accordingly. Un-ionized ammonia was calculated the bags at colder than −70◦C. Fish were not removed from according to the procedures described in Timmons and Ebeling the tanks during the treatment phase. During the posttreatment (2007). After rinsing all filters and removing solids from the clar- phase, 16 fish were removed from the treated tanks and eight ifier, 20% (636 L) of the total RAS volume was removed using were removed from the nontreated tank at the following times a submersible pump. Thereafter, 636 L of well water, measured after the last portion of medicated feed was offered: 6, 12, 24, with a Daniel L. Jerman water meter (Hackensack, New Jersey), 48, 72, 120, 240, 360, and 480 h. Fish were processed by using were added to the RAS. Weekly maintenance included measur- the procedures previously described. ing alkalinity and hardness, as well as light intensity (Milwau- Each day during the treatment phase before maintenance ac- kee model SM7000 light meter, Milwaukee Instruments, Rocky tivities were performed, a water sample was taken from each Mount, North Carolina). tank. After administering the last portion of medicated feed, The FTS was also set up in an environmentally controlled water was sampled hourly from all tanks for 12 h. Thereafter, room. The temperature (∼13◦C) of water flowing through water was sampled from all tanks when fish were sampled from the tank (23 L/min) was controlled with an electronic mix- tanks, i.e., at 24, 48, 72, 120, 240, 360, and 480 h. Samples ing valve. The fiberglass tank (model RT-430, Frigid Units, were filtered through 0.45-µm (pore size) syringe filters (HPF Toledo, Ohio) had a diameter of 119 cm, a stand pipe height Millex-HV Durapore, Millipore, Billerica, Massachusetts) into of 63 cm, and a water volume of 700 L. Supplemental aer- 4-mL screw-cap vials (Cryo.s, Greiner Bio-one GmbH, Frick- ation was provided through air stones immersed in the tank. enhausen, Germany), and stored at about −70◦C. The FTS tank was stocked with 149 Rainbow Trout on the Determining FFA concentrations in fillet tissue.—Frozen fil- same day the RAS was stocked. Daily maintenance included lets were shipped overnight on dry ice to MPI Research (State cleaning and measuring flow through the system, dissolved College, Pennsylvania). Concentrations of FFA (the marker oxygen concentrations, pH, and the water temperature. Weekly residue for FFC) in the fillet tissue were determined using a maintenance included measuring alkalinity, hardness, and light method approved by the U.S. Food and Drug Administration intensity. (FDA) (M. Mushtaq, Merck Animal Health, personal commu- Administration of nonmedicated and medicated feed.— nication). The method is based on converting all FFC residues During the 7-d pretreatment phase, all fish were offered a daily to FFA by acid-catalyzed hydrolysis. A summarized description ration of nonmedicated feed at a mass equal to ∼1.5% of the of the method follows. A sample of ground fillet tissue (2 ± total body weight in the tank. During the 10-d treatment phase, 0.2 g) was hydrolyzed by adding 8 mL of 6 N hydrochloric acid fish in the nontreated recirculating tank were offered a daily (HCl) and heating, with agitation, in a water bath at 95–100◦C ration of nonmedicated feed at a mass equal to ∼1% of the total for approximately 2 h. The tissue hydrolysate was extracted with body weight in the tank. Fish in the treated recirculating tank 20 mL of ethyl acetate and centrifuged for 5 min using a relative and fish in the FTS tank were offered a daily ration of FFC- centrifugal force of 1,400 × g. The ethyl acetate was discarded medicated feed at a mass to achieve a FFC dose of 20 mg·kg−1 and the pH of the aqueous hydrolysate was adjusted to ≥12.5 −1 Downloaded by [Department Of Fisheries] at 00:16 17 November 2014 BW·d . The daily mass (443 g; ∼1% of the total fish mass) was by adding 8 mL of a 30% (w/w) sodium hydroxide solution. based on the FFC concentration in the feed (1,870 mg/kg), the The pH-adjusted solution was poured onto a Varian Chem Elut estimated mean weight of fish at the beginning of the treatment CE1020 sorbent column (Varian, Palo Alto, California) and al- phase (288 g), and the total number of fish in the tank (144; five lowed to adsorb for 45–60 min. The column was eluted with fish were removed before treatment for control tissue). During three 20-mL aliquots of methylene chloride. Methylene chlo- the 20-d posttreatment phase, all fish were offered a daily ration ride was evaporated from samples using a nitrogen stream and of nonmedicated feed at a mass equal to 1.7–2% of the estimated a temperature of 45–50◦C. The dried residue was dissolved in total body weight in the tank. The daily feed ration during the 2 mL of 10 mM potassium phosphate buffer, pH 4.0, containing posttreatment phase was recalculated periodically to account for 1% (v/v) acetonitrile. The resulting solution was filtered through fish growth and fish loss through sampling. a0.2-µm filter to yield the final extract, which was analyzed by Daily feed rations were divided into two equal portions and HPLC using ultraviolet (UV) detection at 220 nm. The method offered twice each day within an 8-h span with a minimum reporting limit was 0.05 µg/g. of 4 h between offerings. Feed was offered to fish more than A control fillet sample and four quality control samples (con- 1 h after completing general tank maintenance and sampling trol fillet tissue fortified with FFA at 0.5 and 2.0 µg/g, two for water chemistry. Feeding activity during each feed offering samples for each concentration) were processed each day from 246 MEINERTZ ET AL.

treated fish. Each day the performance acceptance criteria were of elimination kinetics (Gibaldi and Perrier 1982). Data were −βt met, i.e., an FFA concentration of ≤0.05 µg/g in the control fit to the model Yt = A , where Yt is the concentration in tissue extract and recoveries of FFA ranged from 80% to 110% the fillet tissue at time t. Model parameters were estimated in three of the four quality control samples. using nonlinear least-squares regression (SAS version 9.1.3). Determining FFC concentrations in water.—Water samples The elimination rate constant was used to calculate a half- were analyzed on an Agilent 1260 liquid chromatography (LC) life (t1/2) for FFA in fillet tissue using the following formula: system with an Agilent G6460A triple quadrupole mass spec- t1/2 = 0.693/β, where β = the rate constant (slope) of the model trometer (MS) (Agilent Technologies, Santa Clara, California) (Gibaldi and Perrier 1982). using a 3-µL injection volume, a Kinetex XB-C18 analytical × µ column (2.1 50 mm, 1.7 m, 100 Å; Phenomenex, Tor- RESULTS rance, California), an isocratic mobile phase of 88% (by volume) water : methanol : formic acid (898:100:2) and 12% (by vol- Water Quality ume) methanol, a flow rate of 500 µL/min, and a run time of A summary of the RAS water quality data are presented 2 min. The retention time was 1.20 min. The negative source pain Table 1. Ammonia, nitrite, and nitrate concentrations were rameters, drying gas (300◦C and 7 mL/min), nebulizer pressure within acceptable levels with the exception of a few days at (35 lb/in2 gauge [psig]), sheath gas (300◦C and 11 L/min), capil- the beginning of the treatment phase when nitrite concentra- lary voltage (3,500 V), and nozzle voltage (500 V) of the Agilent tions peaked at 2.24 mg/L. The nitrite levels did not affect fish Jet-Spray electro-spray interface were optimized for multiple re- feeding activity. Water quality measurements in the FTS tank action monitoring of the FFC. The FFC precursor ion (356 m/z) were consistent: dissolved oxygen was >6.3 mg/L, pH was and product ion (185 m/z) were set to unit resolution, and the ∼7.5, temperature was ∼13.5◦C, alkalinity was ∼130 mg/L as dwell (200 ms), fragmentor (110 V), collision energy (12 V), CaCO3, and hardness was ∼170 mg/L as CaCO3. Light intensity and cell accelerator (7 V) were optimized for negative polarity. over both systems ranged from 310 to 540 lx. MassHunter Workstation Data Acquisition (version B.04.01) and MassHunter Workstation Quantitative (version B.04.00) FFA Residues in the Fillet Tissue Analysis software (Agilent Technologies) were used to control Florfenicol amine was not detected in the fillet tissue from the LC-MS system and acquire and process data. The standards any fish sampled during the pretreatment phase. Florfenicol bracketed the expected sample concentrations and were injected amine concentrations in the fillet tissue from treated fish in the at the beginning, end, and after every 10 samples during the se- RAS and FTS are presented in Table 2. Results from the simple quential runs. The FFC peaks were identified by precursor to linear regression analysis of the combined data set indicated product ion generation and retention time comparisons with the that the system type did not affect the FFA concentrations in = = = standards. Peak areas of the FFC standard ions were used to the fillet tissue (F 0.39, df 1, P 0.53). For each system, construct the standard curves and the concentration of the FFC sample time did affect mean FFA concentrations in the fillet = = ≤ = sample ions were determined by comparison with the standard tissue (RAS: F 386.09, df 7, P 0.01; FTS: F 514.70, = < curves. The reporting limit was 4 µg/L. df 7, P 0.01). The elimination rate constants for FFA in − Data analysis.—Data from one fish taken from the FTS fillet tissue from fish in the RAS and FTS were 0.0342 and − tank at 24 h posttreatment was excluded from all data analy- 0.0352, respectively. The t1/2 for FFA in the fillet tissue from ses. The reported FFA concentration in that fish (0.05 µg/g) was markedly different from the FFA concentrations in any other fish TABLE 1. Water quality data for the recirculating aquaculture system used to study florfenicol amine depletion from Rainbow Trout. in the group (range, 2.90–13.95 µg/g) indicating the fish was

Downloaded by [Department Of Fisheries] at 00:16 17 November 2014 physically compromised and did not consume medicated feed. Concentrations of FFA in fillet tissue (uncorrected for method Water quality variable Mean Minimum Maximum accuracy) were modeled by log-linear regression. Individual data sets for each system and the combined data set were fit Dissolved oxygen (mg/L) 9.5 7.0 10.9 = + pH 7.8 7.5 8.2 to the log-linear equation, lnY mX b, where lnY was ◦ the natural log of the FFA concentration at time X, m was the Temperature ( C) 13.3 13.0 13.5 slope of the line, and b was the y-intercept. Data assumptions Flow rate (L/min) 13.6 9.5 16.1 of equal variance (Bartlett’s test for equal variances; Snedecor Total ammonia (mg/L as 0.64 0.18 1.64 and Cochran 1989) and normal distribution (Shapiro and Wilk ammonia nitrogen) 1965) were evaluated before analysis, and least-squares regres- Un-ionized ammonia (mg/L) 0.010 0.004 0.020 sion was used to estimate model parameters (GLM procedure: Nitrite (mg/L) 0.80 0.23 2.24 SAS version 9.1.3, SAS Institute, Cary, North Carolina). Nitrate (mg/L) 98 23 330 Concentrations of FFA were also fit to a monoexponen- Alkalinity (mg/L as CaCO3) 187 139 210 tial regression model that followed a one-compartment model Hardness (mg/L as CaCO3) 187 182 194 FLORFENICOL AMINE DEPLETION FROM RAINBOW TROUT 247

TABLE 2. Florfenicol amine concentrations in fillet tissue from Rainbow Trout treated with florfenicol-medicated feed at 20 mg·kg−1 BW·d−1 for 10 d in recirculating and flow-through systems with water temperatures of ∼13◦C. CV = 100·SD/mean, FFA range = minimum–maximum, ND = not detected.

Recirculating system Flow-through system CV (%) (FFA range, CV (%) (FFA range, Time after treatment (h) Mean FFA (µg/g) µg/g) Mean FFA (µg/g) µg/g) 6 9.82 31 (5.33–15.75) 11.09 23 (7.76–16.89) 12 11.58 26 (5.39–15.24) 11.02 24 (7.40–15.05) 24 7.07 33 (3.73–10.61) 7.70 41 (2.90–13.95) 48 2.50 32 (0.66–3.64) 2.54 46 (1.35–5.40) 72 1.00 34 (0.66–1.85) 1.21 32 (0.65–2.03) 120 0.43 31 (0.17–0.64) 0.49 57 (0.16–1.32) 240 0.25 130 (ND–1.42) 0.26 33 (0.15–0.43) 360 0.16 34 (0.05–0.31) 0.16 29 (0.07–0.26) 480 0.11 35 (ND–0.19) 0.11 25 (0.06–0.17)

fish in the RAS and FTS were 20.3 and 19.7 h, respectively. The current study at the same sample time, i.e., 7.1 µg/g in the RAS FFA elimination model for the RAS was lnY =−0.00938X + and 7.7 µg/g in the FTS. These data confirmed that increasing 1.530 (Figure 1a) and for the FTS was lnY =−0.00959X + the dose would increase the FFC residue burden in the fillet 1.595 (Figure 1b). tissue. The assimilation of FFC residues by fish in the nontreated The second conservative approach element was to treat the RAS tank was minimal. Only three fish sampled at 6 h post- Rainbow Trout at a water temperature considered to be relatively treatment had measurable FFA concentrations (0.06, 0.32, and cool for Rainbow Trout culture. The relatively low water tem- 0.72 µg/g). Thereafter, FFA was not detected in any other non- perature was intended to slow depletion or increase the half-life treated fish. of the FFC residues in the fillet tissue. The approach was based on Kosoff et al. (2009) who treated Nile Tilapia Oreochromis FFC Concentrations in Water niloticus, Walleye Sander vitreus, and hybrid Striped Bass with − − A summary of the FFC concentrations in the RAS tanks FFC at 10 mg·kg 1 BW·d 1 for 10 d, with each species be- ◦ throughout the study is presented in Table 3. The FFC concen- ing treated at two water temperatures that were 5 C different tration in the treated RAS tank peaked at 10 h posttreatment from each other. Their data indicated that the half-life of FFA in (453 µg/L) and decreased to 9 µg/L by 480 h. The FFC concen- the fillet tissue from each species was shorter in the warmwa- tration in the nontreated RAS tank peaked at 11 h posttreatment ter treatments, confirming a second, conservative, experimental (442 µg/L) and decreased to 9 µg/L by 480 h. The FFC concen- design element. tration in the FTS tank was always ≤7 µg/L. The third conservative approach element was to treat fish in an RAS where replacement of the system’s water with freshwater was limited. Limited water replacement in an RAS was DISCUSSION intended to minimize the loss of FFC residues from the system

Downloaded by [Department Of Fisheries] at 00:16 17 November 2014 Three principal elements in the experimental design were thereby exposing fish to waterborne FFC residues indefinitely. intended to provide a conservative approach to estimate the de- Assuming fish would assimilate waterborne FFC residues, depletion of FFC residues from the fillet tissue of fish treated at pletion of FFC residues from the fillet tissue would be prolonged. 15 mg·kg−1 BW·d−1 for 10 consecutive days (treatment regi- Implementation of this element was inconsequential to the out- men for the proposed label expansion). The first element was come of the study. Nontreated fish essentially did not assimilate to treat fish at 20 mg·kg−1 BW·d−1 for 10 consecutive days, a FFC residues; therefore, the waterborne residues did not affect rate 33% greater than the rate that was being pursued for label the depletion of FFC residues. In addition, the half-lives and expansion. The 20-mg·kg−1 BW·d−1 treatment was intended to FFA concentrations in fillet tissue from fish treated in the RAS provide a greater burden of FFC residues in the fillet tissue than and FTS were nearly identical. there would be if fish were treated at 15 mg·kg−1 BW·d−1.Previ- There was some concern over the relatively wide range (126– ously reported data supported the intention. Pinault et al. (1997) 617 g) of individual fish mass used in the present study. This was treated Rainbow Trout at half the dose (10 mg·kg−1 BW·d−1 for based on Bowser et al. (2009) who treated three sizes (100, 250, 10 d) used in the current study. They reported FFA fillet tissue and 500 g) of Nile Tilapia with FFC at 15 mg·kg−1 BW·d−1 for concentrations of 3.4 µg/g at 1 d after treatment, which was 10 d and found that there were trends toward more rapid half- approximately one-half the FFA concentrations reported in the lives and elimination times as fish decreased in size. Additional 248 MEINERTZ ET AL. Downloaded by [Department Of Fisheries] at 00:16 17 November 2014

FIGURE 1. Depletion of FFA from the skin-on fillet tissue of Rainbow Trout treated with florfenicol-medicated feed at 20 mg·kg−1 BW·d−1 in a (a) recirculating aquaculture system and (b) flow-through system for 10 consecutive days. The solid lines are the linear model developed from the natural log-transformed data. The dashed lines are 95% confidence limits. [Color figure available online.] FLORFENICOL AMINE DEPLETION FROM RAINBOW TROUT 249

TABLE 3. Florfenicol concentrations (µg/L) in water samples taken from for 10 d when the withdrawal period is calculated according a two-tank recirculating aquaculture system in which Rainbow Trout in one to FDA methods reported in Guidance for Industry Number 3 tank were treated with florfenicol-medicated feed at 20 mg·kg−1 BW·d−1 for 10 d and Rainbow Trout in the other tank were not treated. LOQ = limit of (USFDA CVM 2006; T. Katz, Merck Animal Health, personal quantitation. communication). Data from the current study also indicated that the implemen- Nontreated tation of a FFC treatment regimen in an RAS at ∼13◦Cmin- Phase Sample time Treated tank tank imally disrupts the performance of the system’s biofilter. This finding was similar to Gaikowski et al. (2012) where tilapia Pretreatment Day 5

Posttreatment 0 h 384 372 ACKNOWLEDGMENTS 1 h 391 376 The authors thank Doug Ramsey of the Rangen Aquaculture 2 h 392 377 Research Center (Hagerman, Idaho) for his expertise in prepar- 3 h 407 353 ing the FFC-medicated feed. The authors thank the FDA Minor 4 h 402 384 Use/Minor Species Grant Program for providing funds to con- 5 h 415 392 duct this study and Merck Animal Health for providing the test 6 h 417 392 article and funds to conduct the tissue analyses. Use of trade, 7 h 418 398 firm, or product names is for descriptive purposes only and does 8 h 432 395 not imply endorsement by the U.S. Government. 9 h 443 389 10 h 453 424 REFERENCES 11 h 451 442 Bowser, P. R., R. E. Kosoff, C.-Y. Chen, G. A. Wooster, R. G. Getchell, 12 h 443 394 J. L. Craig, P. Lim, S. E. Wetzlich, A. L. Craigmill, and L. A. Tell. 2009. Florfenicol residues in Nile Tilapia after 10-d oral dosing in feed: effect of 24 h 312 366 fish size. Journal of Aquatic Animal Health 21:14–17. 48 h 360 352 Feng, J. B., X. P. Jia, and L. I. Li. 2008. Tissue distribution of florfenicol in 72 h 286 293 tilapia (Oreochromis niloticus × O. aureus) after a single oral administration ◦ 120 h 194 196 in freshwater and seawater at 28 C. Aquaculture 276:29–35. 240 h 68 68 Gaikowski, M. P., M. Mushtaq, P. Cassidy, J. R. Meinertz, S. M. Schleis, D. Sweeney, and R. G. Endris. 2010. Depletion of florfenicol amine, marker 360 h 25 24 residue of florfenicol, from the edible fillet of tilapia (Oreochromis niloticus Downloaded by [Department Of Fisheries] at 00:16 17 November 2014 480 h 9 9 × O. niloticus and O. niloticus × O. aureus) following florfenicol administration in feed. Aquaculture 301:1–6. Gaikowski, M. P., J. C. Wolf, R. G. Endris, and W. H. Gingerich. 2003. Safety of Aquaflor (florfenicol, 50% type A medicated article), administered in statistical analyses grouping fish from the present study into feed to Channel Catfish, Ictalurus punctatus. Toxicologic Pathology 31:689– various mass categories indicated that fish mass had no effect 697. Gaikowski, M. P., J. C. Wolf, S. M. Schleis, D. Tuomari, and R. G. Endris. on the FFA concentration in the fillet tissue. 2012. Safety of florfenicol administered in feed to tilapia (Oreochromis sp.). Presently, the withdrawal period in the United States for Toxicologic Pathology 31:689–697. all freshwater-reared finfishes (other than warmwater-reared Gibaldi, M., and D. Perrier. 1982. Pharmacokinetics, 3rd edition. Marcel Dekker, finfish) treated at 10 mg·kg−1 BW·d−1 for10dis15d. New York. Despite the implementation of elements to provide a con- Hayes, J. M. 2005. Determination of florfenicol in fish feed by liquid chromatography. Journal of the Association of Official Analytical Chemists 88:1777– servative approach to estimate FFC residue depletion, the 1783. depletion profiles indicated that the present, approved with- Horsberg, T. E., K. A. Hoff, and R. Nordmo. 1996. Pharmacokinetics of flor- drawal period of 15 d (360 h) is sufficient for freshwater- fenicol and its metabolite florfenicol amine in Atlantic Salmon. Journal of reared salmonids dosed at rates up to 20 mg·kg−1 BW·d−1 Aquatic Animal Health 8:292–301. 250 MEINERTZ ET AL.

Horsberg, T. E., B. Martinsen, and K. J. Varma. 1994. The disposition of Straus, D. L., J. D. Bowker, M. P. Bowman, D. Carty, A. J. Mitchell, and B. 14C-florfenicol in Atlantic Salmon (Salmo salar). Aquaculture 122:97–106. D. Farmer. 2012. Safety of Aquaflor-medicated feed to sunshine bass. North Inglis,V.,R.H.Richards,K.J.Varma,I.H.Sutherland,andE.S.Bokken. American Journal of Aquaculture 74:1–7. 1991. Florfenicol in Atlantic Salmon, Salmo salar L., parr: tolerance and Timmons, M. B., and J. M. Ebeling. 2007. Recirculating aquaculture. assessment of efficacy against furunculosis. Journal of Fish Diseases 14: Northeastern Regional Aquaculture Center, Publication 01-007, Ithaca, 343–351. New York. Kosoff, R. E., C.-Y. Chen, G. A. Wooster, R. G. Getchell, P. R. Bowser, A. USFDA CVM (U.S. Food and Drug Administration, Center for Veterinary Clifford, J. L. Craig, P. Lim, S. E. Wetzlich, A. L. Craigmill, and L. A. Tell. Medicine). 2006. Guidance for industry 3—general principles for evaluat- 2009. Florfenicol residues in three species of fish after 10-d oral dosing in ing the safety of compounds used in food-producing animals. U.S. Depart- feed. Journal of Aquatic Animal Health 21:8–13. ment of Health and Human Services. Available: www.fda.gov/cvm/Guidance/ Martinsen, B., T. E. Horsberg, K. J. Varma, and R. Sams. 1993. Single dose published.htm. (April 2008). pharmacokinetic study of florfenicol in Atlantic Salmon (Salmo salar)in USFDA CVM (U.S. Food and Drug Administration, Center for Veteri- seawater at 11◦C. Aquaculture 112:1–11. nary Medicine). 2007. Supplemental new animal drug application, NADA Pinault, L. P., L. K. Millot, and P. J. Sanders. 1997. Absolute oral bioavail- 141-246 (AQUAFLOR type A medicated article (florfenicol), an an- ability and residues of florfenicol in the Rainbow Trout (Oncorhynchus tibiotic. U.S. Department of Health and Human Services. Available: mykiss). Journal of Veterinary Pharmacology and Therapeutics 20(Supple- www.fda.gov/cvm/drugsuseaqua.htm. (April 2008). ment 1):297–298. Wrzesinski, C., L. Crouch, P. Gaunt, D. Holifield, N. Bertrand, and R. Endris. Samuelsen, O. B, Ø. Bergh, and A. Ervik. 2003. Pharmacokinetics of florfeni- 2006. Florfenicol residue depletion in Channel Catfish, Ictalurus punctatus col in cod Gadus morhua and in vitro antibacterial activity against Vibrio (Rafinesque). Aquaculture 253:309–316. anguillarum. Diseases of Aquatic Organisms 56:127–133. Yanong, R. P., E. W. Curtis, R. Simmons, V. A. Bhattaram, M. Gopalakrishnan, Shapiro, S. S., and M. B. Wilk. 1965. An analysis of variance test for normality N. Ketabi, N. V. Nagaraja, and H. Derendorf. 2005. Pharmacokinetic studies (complete samples). Biometrika 52:591–611. of florfenicol in Koi carp and Threespot Gourami Trichogaster trichopterus Snedecor, G. W., and W. G. Cochran. 1989. Statistical methods, 8th edition. after oral and intramuscular treatment. Journal of Aquatic Animal Health Iowa State University Press, Ames. 17:129–137. Downloaded by [Department Of Fisheries] at 00:16 17 November 2014 This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:17 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Increased Survival of Juvenile Turbot Scophthalmus maximus by Using Bacteria Associated with Cultured Oysters Rosa M. Farto Seguína, María Bobo Bermúdeza, Leticia Riveraa & Teresa P. Nietoa a Universidad de Vigo, Vigo 36310, Spain Published online: 31 Oct 2014.

To cite this article: Rosa M. Farto Seguín, María Bobo Bermúdez, Leticia Rivera & Teresa P. Nieto (2014) Increased Survival of Juvenile Turbot Scophthalmus maximus by Using Bacteria Associated with Cultured Oysters, Journal of Aquatic Animal Health, 26:4, 251-262, DOI: 10.1080/08997659.2014.920734 To link to this article: http://dx.doi.org/10.1080/08997659.2014.920734

PLEASE SCROLL DOWN FOR ARTICLE

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:251–262, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.920734

ARTICLE

Increased Survival of Juvenile Turbot Scophthalmus maximus by Using Bacteria Associated with Cultured Oysters

Rosa M. Farto Segu´ın,* Mar´ıa Bobo Bermudez,´ Leticia Rivera, and Teresa P. Nieto Universidad de Vigo, Vigo 36310, Spain

Abstract Preventing vibriosis in juvenile cultured Turbot Scophthalmus maximus caused by Vibrio anguillarum frequently requires the use of feed supplemented with antibiotics in addition to vaccines. Whether the use of probiotics instead of antibiotics in juvenile Turbot is a safer strategy requires more study. The antibacterial potential of 148 Vibrio spp. strains (mostly isolated from cultures of healthy oysters, clams, and Turbot) was analyzed in vitro against V. anguillarum and other pathogens by means of an agar diffusion assay. A wide spectrum of inhibitory activity was shown by 9 strains. Based on their easy phenotypic differentiation from V. anguillarum, we selected two strains (S1 and S2, both isolated from the European flat oyster Ostrea edulis) for testing in juvenile Turbot (3 g). None of the strains were virulent by intraperitoneal or bath challenges, and all were susceptible to the antibiotics most frequently used in aquaculture. Three different stocks of Turbot, which were assayed separately, were significantly protected from infection with V. anguillarum. The final survival rates of fish treated in mixed challenges with S1 or S2 and V. anguillarum were 44% and 66%, respectively, whereas only 17% of the fish treated with only the pathogenic strain survived. The application of probiotic strains also increased the survival time of juvenile Turbot after infection with V. anguillarum. Both strains persisted in the epidermal mucus layer of the fish for 30 d, and they were not displaced by the pathogen. These data prove the efficacy of using bacteria well adapted to the dynamics of culture production as a way to provide juvenile Turbot immediate protection against infection by V. anguillarum. Moreover, the epidermal mucus sampling was useful for investigating the persistence of both probiotic strains when exposed to the pathogen.

The global production of Turbot Scophthalmus maximus in obtained by using a combination of bacteria from this genus 2011 was 72,413 metric tons, and it was valued at US$600 and Lactobacillus, although the beneficial effects have not been million. The largest producer by far is Spain, whose annual pro- studied in challenge trials against pathogens (Daga´ et al. 2013). Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 duction has nearly doubled since 1998 and represents more than A few studies have also addressed the use of bacteria iso- 75.9% of the global total (Serrano 2014). Protecting juvenile lated from Turbot aquaculture. In particular, two strains of Turbot during their development is necessary to avoid infection. Vibrio have been selected, but the authors believed that they The use of bacteria as probiotics in Turbot larval aquaculture would provide a weak defense against highly virulent pathogens has previously been examined for bacteria isolated from ma- such as V. anguillarum (Gatesoupe 1997; Ringo and Vadstein rine environments and other sources. Low survival rates were 1998). Some more promising candidates were bacteria from the found in larvae treated with Streptococcus thermophilus, Lac- Roseobacter clade, which appear to be universal colonizers of tobacillus helveticus, and Lactobacillus plantarum (Gatesoupe marine larval rearing units and displayed antibacterial activity 1991). Larvae treated with Pediococcus acidilactici showed no against V. anguillarum (Porsby et al. 2008). In fact, treatment significant differences in survival rate from the control (Vil- with a Roseobacter strain (currently identified as Phaeobac- lamil et al. 2010). On the contrary, a high survival rate was ter; Pintado et al. 2010) led to increased survival of larvae

*Corresponding author: [email protected] Received October 11, 2013; accepted March 10, 2014 251 252 FARTO SEGUIN´ ET AL.

during experimental infection with that pathogen. However, the and that they were positive for all tests except for being able to quantity and frequency of the additions were not optimized grow at 0% NaCl. We also analyzed Vibrio spp. strains that (Planas et al. 2006), nor was the effect on juvenile Turbot stud- had previously been isolated and phenotypically characterized ied. Although ingestion of V. proteolyticus by juvenile Turbot (Montes et al. 1999) from the skin mucus of Turbot (17) and the tended to produce positive effects on its feed efficiency, its water (2) at the aquaculture experiment system at the Instituto nutritional significance and potential as a prophylactic agent Oceanografico´ de Vigo in which they had been held. against vibriosis were not examined (Schrijver and Ollevier Reference strains pathogenic to fish or shellfish were used for 2000). Pseudomonas spp. and Vibrio spp. were also recently the agar diffusion assays. These strains were previously iden- administered to juvenile Turbot, which enhanced serum bacte- tified and provided by national or international bacteria col- ricidal activity against Edwardsiella tarda and Vibrio anguil- lections (see www.cect.org/english,http://bccm.belspo.be/db/ larum. However, both strains were also able to disseminate into lmg search form.php, and http://www.ncimb.com/). Pathogenic internal organs, and the immediate protection of the host was V. lentus strains (P52, P54, and P58) previously isolated and not evaluated (Wang et al. 2013).The probiotic effect in juvenile identified by us were also used (Farto et al. 2003; Table 1). Turbot would have to be studied in detail to produce Turbot that Characterization of strains.—The strains with strong in- were disease resistant until they reached market or broodstock hibitory activity (described below) and the three V. anguillarum size. Such study could provide important economic benefits for strains (CECT 7199, NCIMB 571, and NCIMB 6) were char- aquaculture. acterized by 90 physiological, morphological, and biochemical In this study, 148 Vibrio spp. strains previously isolated by us tests as previously described (Guisande et al. 2004). Cultures (mostly from healthy oysters, clams, and Turbot) were analyzed grown at 22◦C for 24 h on tryptic soy agar (Cultimed) sup- in vitro and in vivo in order to identify probiotic strains that plemented with up to 2% (weight : volume) NaCl (TSA-2) might be useful in preventing vibriosis caused by Vibrio anguil- were used as inocula. Additionally, both of the strains used in larum in juvenile Turbot. The epidermal mucus of the Turbot the virulence assays (S1 and S2) were characterized by their was sampled to evaluate the persistence of the probiotics in this susceptibility to the antibiotics most frequently used in aqua- host, along with their capacity to resist being excluded by the culture (ciprofloxacin [5 µg], florfenicol [30 µg], tetracycline pathogen. [10 and 30 µg], and sulfamethoxazole/trimetropin [25 µg]) by using the disk diffusion susceptibility method on TSA-2 plates ◦ METHODS as described by Huys et al. (2005). After incubation at 22 C for 24 h, the diameter of the inhibited growth zone surround- Bacterial Strains and Characterization ing the antibiotic disk was measured and the presence of any We used 129 Vibrio spp. strains previously isolated by us inhibition area was taken as indicating susceptibility to the (Guisande et al. 2004) from healthy oysters and clams at differ- antibiotic. ent stages of culture at several facilities in northwestern Spain The strains S1 and S2 were also characterized by sequenc- between 2008 and 2010. A total of 110 strains were isolated from ing the 16S rRNA gene according to the method described the European flat oyster Ostrea edulis (82 from broodstock, 24 by Guisande et al. (2008). Multiple alignment of sequences from larvae, and 4 from spat), 6 from the Manilla littleneck was created by ClustalW in Genious Editor version 7.0.6 clam Tapes japonica (larvae), 8 from the pullet carpet shell (Biomatters; http://www.geneious.com/). This included 1.138 Venerupis pullastra (broodstock), and 5 from the grooved car- positions after the removal of ambiguous ones. A phylogenetic pet shell Ruditapes decussatus (broodstock). Briefly, broodstock tree was constructed using Molecular Evolutionary Genetics (50 g of flesh and intervalval liquid) and larvae or spat (50 units Analysis (MEGA) version 6 (Tamura et al. 2013). This was per-

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 each) were homogenized in 100 mL of sterile marine phosphate- formed using the neighbor-joining method and Tamura–Nei dis- buffered saline (pH 7.1) and adequate serial dilutions were tance model, with the calculation of cluster stability by bootstrap plated onto marine agar (MA; Cultimed) or thiosulfate–citrate– analysis with 1,000 replicates. The partial 16S rRNA sequences bile–sucrose agar (TCBS; Cultimed). Colonies selected from the of the S1 and S2 strains have been deposited in the GenBank TCBS plates were spread on MA to obtain pure cultures, which (Mountain View, Maryland), EMBL (Heidelberg, Germany), were inoculated on nutritive broth (Cultimed) supplemented and DDBJ (Mishima, Japan) nucleotide sequence databases un- with 2% (weight : volume) NaCl (Panreac) and 15% (volume : der the accession numbers KJ364527 and KJ364528, respec- volume) of glycerol (Panreac) for preservation at −80◦C. Thaw- tively. ing was avoided each time they were needed in order to preserve Differentiation of probiotic strains from the pathogenic strain their original features. The strains were phenotypically charac- and other naturally occurring bacteria.—To distinguish the terized as Vibrio by using the classic tests: gram staining, ox- pathogenic strain (CECT 7199) from the potentially probiotic idase, oxidative or fermentative glucose metabolism, motility, strains (S1 and S2) used in the virulence assays, appropriate nitrate reduction, growth ability at 0% NaCl, and susceptibil- tests were conducted after biochemical and physiological pro- ity to O129 vibriostatic agent (2,4-diamino-6,7-di-isopropyl- ﬁling. The S1 and S2 strains were positive for β-hemolysis, with pteridine phosphate; 150 µg/disc). These assays indicated that susceptibility to ampicillin (TSA-2 supplemented with ampi- all of the strains were facultative, anaerobic, gram-negative rods cillin [25 µg/mL]) and able to grow at 7% NaCl. By contrast, INCREASED SURVIVAL OF JUVENILE TURBOT 253

TABLE 1. Inhibitory activity in an agar diffusion assay of potentially probiotic strains (109 CFU/mL) against various pathogenic strains (109). Values are the mean ± SD diameters of the zones of inhibited growth from four different experiments; blank cells indicate the absence of inhibitory activity.

Strains with inhibitory activity against pathogenic strains

Pathogenic straina S1b S2b S3c S4d S5e S6c S7c S8c S9c S10b S11c S12c

V. alginolyticus CECT 586 1.7 ± 1.1 V. anguillarum CECT 7199 2.5 ± 1.6 2.6 ± 2.0 3.3 ± 1.1 2.9 ± 0.8 4.1 ± 2.7 6.1 ± 1.0 8.1 ± 2.8 NCIMB 571 4.7 ± 2.4 5.5 ± 0.6 6.5 ± 1.1 6.3 ± 1.0 9.0 ± 0.0 5.0 ± 1.0 7.5 ± 0.6 7.2 ± 1.4 3.8 ± 0.5 9.0 ± 2.5 NCIMB 6 5.7 ± 4.5 2.1 ± 21.4± 2.1 2.3 ± 0.9 4.1 ± 2.3 V. harveyi LMG 4044 T 3.2 ± 1.7 1.4 ± 1.0 3.3 ± 2.8 2.1 ± 2.9 V. lentus CECT 5293 1.0 ± 01.0± 01.0± 03.0± 01.0± 05.0± 0 P52 2.8 ± 1.3 4.3 ± 0.9 3.0 ± 1.3 3.8 ± 0.5 1.8 ± 0.8 P54 2.5 ± 2.1 2.5 ± 0.7 4.0 ± 0.0 2.0 ± 0.0 4.5 ± 2.1 P58 3.5 ± 0.0 1.5 ± 1.2 V. neptunius LMG 20615 4.2 ± 2.6 9.8 ± 2.6 2.3 ± 1.9 1.3 ± 0.6 3.0 ± 4.0 5.3 ± 3.7 2.8 ± 3.4 8.7 ± 5.1 1.5 ± 1.0 6.3 ± 4.1 V. parahaemolyticus CECT 5306 3.5 ± 2.7 1.9 ± 1.2 3.0 ± 2.0 5.3 ± 2.0 5.1 ± 2.2 3.0 ± 1.1 3.2 ± 1.2 V. splendidus LMG 16751 1.5 ± 1.4 2.7 ± 1.2 1.5 ± 1.6 2.3 ± 1.6 2.6 ± 1.8 1.3 ± 1.4 1.2 ± 1.5 4.7 ± 1.8 V. tapetis CECT 4600T 3.7 ± 1.6 4.7 ± 0.9 3.4 ± 1.0 5.0 ± 1.3 4.6 ± 0.8 4.0 ± 0.8 2.9 ± 1.2 V. tasmaniensis LMG 20012T 2.0 ± 1.2

aT = type strain; abbreviations are as follows: CECT = Spanish Type Culture Collection, LMG = Belgian Coordinated Collections of Microorganisms (University of Ghent), NCIMB = Scottish National Collections of Industrial and Marine Bacteria. bOstrea edulis larvae. cOstrea edulis broodstock. dVenerupis pullastra broodstock. eTapes japonica larvae.

the pathogenic strain was positive for α-hemolysis, resistant to using the Replicator Mast Multipointer Inoculator (MAST) and ampicillin, and unable to growth at 7% NaCl. Naturally oc- incubated at 22◦C for 120 h. curring bacteria that were susceptible to ampicillin and posi- The diameters of the inhibited-growth zones surrounding the tive for α-hemolysis or resistant to ampicillin and positive for potentially probiotic strains were measured, and zones at least β-hemolysis could thus be distinguished from the probiotic and 1 mm in diameter were considered indicative of antibacterial pathogenic strains. activity. The inhibitory capacity of each tested strain was clas- Additionally, to differentiate our probiotic strains from other sified according to the total number of pathogenic strains that closely related Vibrio species (V. artabrorum, V. celticus, V. it inhibited (strong, medium, or weak depending on whether chagasii, V gigantis, V. hemicentroti, V. pomeroyi, and V. splen- the strain inhibited >5, 3–5, or 1–2 pathogenic strains, respec- didus), the following phenotypical tests from the literature (see tively), independently of the diameters of the inhibited zones. Table S.1 in the supplement that appears in the online version of Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 this article) were used: ability to grow at 6% NaCl; response to In Vivo Antibacterial Potential of Vibrio spp. Strains Thornley’s arginine dihydrolase and the Voges–Proskauer test; Experimental fish.—Healthy, unvaccinated Turbot (3 ± ability to produce acid from L-arabinose, D-lactose, D-melibiose, 0.5 g [mean ± SE]) were used for all challenges. The fish L-rhamnose, and sucrose; and degradation of esculin and gelatin. were the offspring of Turbot reared at an experimental fish farm at the University of Vigo (Marine Sciences Station at In Vitro Antibacterial Potential of Vibrio spp. Strains Toralla) with no history of disease. Prior to challenge, 10 fish The 148 Vibrio spp. strains were tested against 14 strains randomly selected from the stock were screened for bacterial pathogenic for fish and shellfish (Table 1) on MA plates. Briefly, pathogens using routine diagnostic procedures, as previously the strains tested were cultured in M9 minimal media supple- described (Lago et al. 2012). For each bacterial strain and mented with up to 1.5% NaCl and incubated at 22◦C for 24 h experiment (intraperitoneal or bath challenge), fish were held with shaking. The concentration of the bacterial cultures was for 2 weeks in separate experimental tanks containing 50 L of ◦ adjusted (OD590 = 0.5), and the MA plates were covered with sterile, 18 C aerated (with an air stone) seawater with a salin- an agar layer containing 109 or 1010 CFU per plate that was ity of 33‰ and were fed twice a day. These conditions were prepared for each pathogenic strain. Following that, 109 CFU of maintained during the challenges, except that tanks with non- each potentially probiotic strain were inoculated on these plates circulating seawater were used. The environmental parameters 254 FARTO SEGUIN´ ET AL.

+ were recorded daily in each tank; the values of NH4 and NH3 lum of the potentially probiotic strains was prepared and added were kept below 3.3 mg/L and 0.1 mg/L, respectively, at pH for 12 h as described above, with a final concentration of 105 7.7–8 and an oxygen concentration of 6 mg/L. Water was re- CFU/mL. After this, the seawater was changed and the bacte- newed after feeding when required. All animal experiments were rial inoculum of the pathogenic strain was also added for 12 h performed according to practices approved by the Spanish Eth- to a final concentration of 106 CFU/mL (stocks 1 and 2) or 107 ical Committee (http://www.boe.es/boe/dias/2013/02/08/). CFU/mL (stock 3). Then the seawater was changed. The starting Preliminary challenges to confirm the absence of virulence in bacterial numbers were confirmed by determining the number potentially probiotic strains.—For both the intraperitoneal and of CFU per milliliter of each tank at time 0. Each fish group bath challenges, we used S1 and S2 isolated from Ostrea edulis was monitored for 30 d and sampled as described above. Fish larvae. These were selected since they inhibited the growth of exposed only to the pathogenic strain were used as a positive pathogenic strain V. anguillarum (CECT 7199) and were easily control (Pc tank). On the contrary, fish exposed only to poten- differentiated from V. anguillarum by phenotypical tests. tially probiotic strains (S1c and S2c tanks) and those exposed For the intraperitoneal challenge, all strains were grown in only to M9 broth (M9B tank) were used as a negative control in shaken M9 broth supplemented with up to 1.5% NaCl (M9B; each stock. All fish were maintained under the same conditions 22◦C, 24 h); cells were harvested and resuspended in sterilized as in the mixed-challenge assay. 9 seawater to an OD590 of 0.4 (10 CFU/mL). The bacterial sus- All treatments, (M9B, S1c, S2c, Pc, S1mP, and S2mP) were pension (0.1 mL) was inoculated in Turbot (n = 30) by injection administered to three different stocks at different times to in- into the abdominal cavity, with doses of 106 or 107 CFU per fish, vestigate whether the protection of fish was reproducible or not. which were monitored for 7 d. The surviving fish were eutha- One tank per treatment was used for each stock, except for nized by an overdose of tricaine methanesulfonate (MS-222; assays in which probiotic and pathogenic strains were mixed, Sigma). For all moribund, dead, and euthanized fish, the liver where two tanks were used. The total numbers of fish used per and kidney tissues were sampled for bacteria by spreading them treatment and stock depended on availability (Table 5). on blood agar (Oxoid) and TSA-2 supplemented with ampicillin Additionally, the external surfaces of fish were examined (25 µg/mL; TSA-2A). After incubation at 22◦C for 48 h, iden- during the experiments in order to determine the persistence tification of suspicious re-isolated colonies was carried out as of the bacteria tested and the degree to which they were ex- described above. Fish intraperitoneally injected (0.1 mL) with cluded by the pathogen. At various times postchallenge (0, 12, M9 broth were used as negative controls and sampled as de- 24, 36, 60, 84, and 156 h as well as 14, 16, and 30 d), the scribed above. Virulent strains were defined as those showing epidermal mucus layer close to the pectoral fin (1 cm2)was any mortality after challenge at doses of 106 CFU per fish. removed aseptically from three randomly selected fish. These For the bath challenge, the bacterial inoculum of the poten- samples were swabbed with a sterile inoculating cotton bud tially probiotic or pathogenic strains was prepared by growth in for direct plating of the bacteria on blood agar, TSA-2 sup- shaken M9B supplemented with up to 1.5% NaCl (22◦C, 24 h). plemented with ampicillin (25 µg/mL), and TSA-2. The plates ◦ The inoculum was added to each tank, which contained 30 fish, were incubated at 22 C for 72 h, and one swab per fish was for 12 h (to a final concentration of 105 or 107 CFU/mL, adjusted analyzed. The presence of probiotic and pathogenic strains or as above). After that, the seawater was changed. The starting naturally occurring bacteria was determined by their differential bacterial numbers were confirmed by determining the number of responses to blood agar (BA) and TSA-2 containing ampicillin CFU per milliliter of each tank at 0 h. Each fish group was mon- (TSA-2-A), as described above. Predominance and morpholog- itored for 30 d and sampled as described above. Fish exposed ical differences in colonies were confirmed on those plates and to M9 broth and maintained under the same conditions were on TSA-2. Each colony from the BA, TSA-2-A, and TSA plates

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 used as negative controls. Virulent strains were defined as those was compared with those previously cultured with probiotic or showing any mortality after challenge at doses of 106 CFU/mL. pathogenic strains. When the result was unclear, the colonies Preliminary challenges to confirm the virulence of were isolated and cultured to purity on TSA-2. Then they were pathogenic strains.—Our model pathogen was V. anguillarum, tested for their ability to grow at 7% NaCl, and their responses which is a well-established pathogen of Turbot and considered to the different tests were noted as described above for other one of the leading fish pathogens, with a broad host range among Vibrio species closely related to our probiotic strains. The pres- farmed and wild fish (reviewed by Austin and Austin 2007). The ence of probiotic and pathogenic strains or naturally occurring strains CECT 7199, NCIMB 571, and NCIMB 6 were tested in bacteria was considered confirmed when a colony of each one order to select one that would be adequate. Intraperitoneal and was identified in at least one of the three sampled fish at each bath challenges and sampling were performed as above. time postchallenge. Virulence assays mixing potentially probiotic and pathogenic After 30 d the surviving fish from the three stocks previously strains.—Based on the results of the intraperitoneal and bath exposed only to S1 (S1c tank) or S2 (S2c tank) were divided in challenges, the pathogenic strain CECT 7199 was selected to two groups containing similar numbers of fish. One was bath- perform these assays. Potentially probiotic (S1 and S2) and challenged again with V. anguillarum (5 × 106 CFU/mL)to pathogenic strains were mixed in the same tank for a bath chal- determine whether fish retained their protection against vib- lenge (S1mP and S2mP tanks) as follows: the bacterial inocu- riosis. Another, which was used as a negative control, was INCREASED SURVIVAL OF JUVENILE TURBOT 255

not exposed again (S1cp or S2cp tanks). Fish exposed only Table 2. Furthermore, both strains were susceptible to all of the to M9 broth (M9B tank) or pathogenic strain (Pc tank) were antibiotics most frequently used in aquaculture (Table 2). also used as negative and positive controls, respectively, in The sequencing of 16S rRNA showed that the two probiotic each stock. Variable numbers of fish were used, depending on strains have 100% identity in gene sequences. They were also availability (see Results). Each fish group was monitored for 30 closely related to the V. splendidus group (having 99.6–100% d and sampled as described above. similarity with eight species) and showed the highest sequence similarity (100%) with V. crassostreae, V. celticus, V. gigantis, Statistical Analysis and V. pomeroyi (Figure 1). Kaplan–Meier survival curves for the three stocks analyzed The phenotypic profile of all the V. anguillarum strains ob- jointly were generated, and mean survival times were calculated. tained in this study confirmed their previous identification. Comparisons of the survival curves and the final survival of fish were made by means of the log-rank (Mantel–Cox) test. In Vivo Antibacterial Potential of Vibrio Strains Each curve was constructed from data for three stocks, which Preliminary challenges to confirm the absence of virulence included data from two replicates in assays in which probiotic in potentially probiotic strains.—We injected the strains S1 and and pathogenic strains were mixed. Additionally, Kaplan–Meier S2 into the abdominal cavities of the Turbot, and neither of survival curves were generated for each stock independently and them was virulent at 106 or 107 CFU per fish. Similar results mean survival was calculated. These analyses were calculated were obtained after bath challenges with 105 or 107 CFU/mL. using SPSS version 15.0. Bacteria were never recovered from the livers of euthanized fish Statistical differences in the persistence of probiotic or injected with probiotic strains or from uninfected control fish. pathogenic strains in epidermal mucus among the three stocks No clinical signs were seen. were also evaluated using paired-sample t-tests and one-way Preliminary challenges to confirm the virulence of analysis of variance (ANOVA; Tukey’s post hoc test). These pathogenic strains.—The NCIMB 571 and CECT 7199 strains tests were done in Sigma Plot version 12. In all cases, a signifi- of V. anguillarum caused 100% mortality after intraperitoneal cance level of 0.05 was used. challenge with 106 or 107 CFU per fish. Both were recovered in pure culture from all fish. These two strains were also tested by bath challenge; only CECT 7199 was virulent at 106 CFU/mL RESULTS and was recovered from fish in pure culture. The lethal effect In Vitro Antibacterial Potential of Vibrio Strains was detected within 3 and 8 d after the intraperitoneal and bath Of the 148 strains tested, 12 (S1–S12) showed inhibitory challenges, respectively. Clinical signs included anorexia, ab- activity against different pathogens, such as V. anguillarum. normal swimming behavior, hemorrhagic lesions on the liver Interestingly, 9 of them (most of which were isolated from and intestine, and pale liver. NCIMB 6 was avirulent according the broodstock of Ostrea edulis) had a strong inhibitory ac- to both virulence assays. Bacteria were never recovered from tivity, with the ability to inhibit at least 6 of the 14 pathogenic the liver of euthanized fish infected with the avirulent strain or strains tested. The other strains had weak inhibitory activity. from uninfected control fish. The probiotic strains showed different inhibitory activity against Virulence assays mixing potentially probiotic and pathogenic each pathogenic strain (Table 1). All of the potentially probiotic strains.—The potentially probiotic and pathogenic (CECT strains retained their inhibitory ability even with a 10-fold in- 7199) strains were mixed in the same tank to analyze the antibac- crease in the number of cells (1010 CFU/mL) of V. anguillarum terial potential of the former against vibriosis in Turbot. Kaplan– (data not shown). Meier survival curves for the three stocks analyzed jointly are

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 shown in Figure 2; mean survival times are given in Table 3. The Characterization of Potentially Probiotic Strains comparisons of the survival curves by the log-rank test revealed and Confirmation of V. anguillarum that there were significant differences between all of the Only strains S1 and S2 could be easily differentiated phe- negative controls (S1c, S2c, and M9B) and fish treated with the notypically from all of the pathogenic strains of V. anguil- pathogen alone (Pc) or with probiotic S1 (S1mP) or S2 (S2mP) larum tested, since both showed β degradation of sheep erythro- in the mixed challenges. The analyses also showed significantly cytes, susceptibility to ampicillin (25 µg/mL), and the ability better survival among fish treated with S1 or S2 in mixed chal- to grow at 7% NaCl. Conversely, the pathogenic strains showed lenges than among those exposed only to the pathogen (Table 4). α-hemolysis, were resistant to ampicillin, and were unable to tol- Whereas the final survival rates of fish treated with S1 and S2 in erate 7% NaCl. The two potentially probiotic strains produced the mixed challenges were 44% and 66%, respectively, that of identical results in 83 tests. They consisted of motile short rods; fish exposed to the pathogen alone was 17% (Figure 2; Table 3). were gram-negative, oxidase- and catalase-positive, and oxida- Significant differences between fish treated with S1 and S2 tive and fermentative in O/F medium with glucose; and grew in were found as well (Table 4). Furthermore, the application of yellow on TCBS medium and in crystal violet (0.001%). They probiotic strains significantly increased survival time (Figure 2). were able to tolerate from 3.5% to 7% NaCl or pH 7.5–10 but Additionally, Kaplan–Meier survival curves were generated not 10% NaCl or pH 4.5. Other characteristics are shown in for each stock independently (data not shown) in order to 256 FARTO SEGUIN´ ET AL.

TABLE 2. Phenotypic characteristics of the two potentially probiotic strains selected for analysis. A positive sign indicates that the strain tested positive for that characteristic, a negative sign that it tested negative.

Test S1 S2 Growth at: 4, 10, 22, 28◦C ++ 37◦C – + 44◦C – + Thornley’s arginine dihydrolase ++ Moeller’s lysine decarboxylase, ornithine decarboxylase, and histidine decarboxylase – – Methyl red test ++ Voges–Proskauer test –– Nitrate reduction ++ Indole production ++ Utilization of D-lactose or production of H2S on Kligler medium – – Hemolysis-α –– Acid from: arbutin, D-cellobiose, D-fructose, D-galactose, D-mannitol, D-mannose, ribose, sucrose, D-trehalose ++ amygdaline, L-arabinose, ethanol, inositol, D-lactose, D-raffinose, L-rhamnose, and D-sorbitol – – glycerol – + D-melibiose – + salicin – + Hydrolysis of: casein, chitin, DNA, esculin, gelatin, lecithin, starch ++ urea, Tween 20, chondroitin –– celullose + – Use as sole carbon source: aspartate, β-alanine, DL-alanine, alginate, L-arginine, glycine, L-lysine, malonate, L-phenylalanine, L-proline ++ propanol, pyruvate, L-serine, succinate, L-tartrate, L-tryptophan, uracil acetate, citrate –– inulin – + Susceptibility to: O129 (150 µg/disc) ++ ciprofloxacin (5 µg/disc) ++ florfenicol (30 µg/disc) ++ tetracycline (10 µg/disc) ++ tetracycline (30 µg/disc) ++ sulphamethoxazole/trimetropin (25 µg/disc) ++

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 derive stock-specific final survival rates and mean survival times. grown from M9 broth tanks showed higher diversity (>3differ- The different treatments produced different values for each ent colonies) in naturally occurring bacteria. By contrast, S1 or stock (Table 5). All dead fish exhibited the disease symptoms S2 was predominant on plates obtained from fish exposed only described above, and the pathogenic strain was recovered in to probiotics or to mixed challenges. Interestingly, in mixed pure culture from most fish. The behavior of the positive and and control challenges with strain S1, naturally occurring bac- negative controls was as expected. teria were significantly absent from the skin of several sampled The persistence of bacteria in the joint analysis of the three fish, at least for 36 h postchallenge (Figure 3). This was not stocks is shown in Figures 3 and 4. In challenges with only the case with S2, however (Figure 4). The pathogenic strain probiotics as well as in mixed challenges, both probiotic strains was detected in all experiments from 0 h after infection to 14 d were able to persist in the epidermal mucus layer of fish for 30 postchallenge and was able to significantly displace the naturally d. In the mixed challenges, however, the cumulative presence occurring bacteria for at least 60 h (Figures 3, 4). By contrast, S1 of S1 was significantly higher than that of S2 after 0, 12, 336, and S2 were not displaced by the pathogen, since they showed and 384 h. In comparing the diversity of bacteria on plates de- similar persistence when used alone and when mixed with the rived from fish subjected to those challenges with that of fish pathogenic strain in any of the stocks. These results were also subjected to M9 broth, clear differences were seen. Cultures confirmed by the mixed challenges involving S1 and S2. In fact, INCREASED SURVIVAL OF JUVENILE TURBOT 257

FIGURE 1. Phylogenetic tree based on partial 16S rRNA gene sequences obtained by the neighbor-joining method. Vibrio cholerae was used as an out group. Bootstrap percentages (≥50) after 1,000 simulations are shown at the nodes. The bar labelled 0.005 indicates the nucleotide substitution per position. GenBank sequence accession numbers are given in parentheses.

the cumulative presence of S1 was significantly higher than that control challenges still had the strains in their epidermal mucus of the pathogenic strain, though no significant differences be- layer (Figures 3, 4). However, when they were bath-challenged tween the persistence of S2 and that of the pathogenic strain again with V. anguillarum, the resulting number of dead fish were found. There were no significant differences in the rest of was similar to that among fish exposed only to the pathogen the comparisons. (Table S.2). The lethal effect was detected within 8 d. All of Thirty days postchallenge, the surviving fish from the three dead fish showed the disease symptoms described above, and the stocks that had been exposed to the probiotic strains in the pathogenic strain was recovered in pure culture from most fish.

TABLE 3. Case summaries, ﬁnal survival rates, and mean survival times for all Turbot stocks by the Kaplan–Meier method.

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 Mean for survival time 95% confidence limits Total no. No. of dead No. of Final survival Treatmenta of fishb fish surviving fish rate (%) Estimate SE Lower Upper M9B 49 0 49 100 30 0 30 30 S1c 49 0 49 100 30 0 30 30 S2c 49 0 49 100 30 0 30 30 Pc 58 48 10 17.20 7.93 1.35 5.29 10.57 S1mP 95 53 42 44.20 16.63 1.24 14.19 19.07 S2mP 73 25 48 65.80 22.04 1.30 19.48 24.60

a M9B = control with M9 broth, S1c = control with S1, S2c = control with S2, Pc = control with pathogenic bacteria, S1mP = S1 mixed with pathogenic bacteria, and S2mP = S2 mixed with pathogenic bacteria. b One replicate per treatment was used per stock except for treatments S1mP and S2mP, for which there were two replicates. c Estimate limited to the longest survival time that was recorded (30 d). 258 FARTO SEGUIN´ ET AL.

FIGURE 3. Cumulative presence of bacteria in the skin of Turbot after treatment with probiotic S1. The data are the means of three stocks, each with three fish randomly sampled at each sampling time postchallenge (error bars = SDs). Abbreviations are as follows: O (M9B) = naturally occurring bacteria (“other bacteria”) in the M9B tank; O (Pc) = other bacteria in the pathogen control tank; O(S1c)= other bacteria in the S1 control tank; O (mS1) = other bacteria in the mixed S1–pathogen tank; Pc = pathogenic bacteria in the pathogen control FIGURE 2. Kaplan–Meier survival curves of juvenile Turbot treated with the tank; P (mS1) = pathogenic bacteria in the mixed S1–pathogen tank; S1c = 5 potentially probiotic strains S1 and S2 (10 CFU/mL) to prevent infection with S1 in the S1 control tank; and S1 (mP) = S1 in the mixed S1–pathogen tank. 6 7 V. anguillarum (10 CFU/mL for stocks 1 and 2, 10 CFU/mL for stock 3). Each On the y-axis, 1 means that a colony of bacteria was found in three of the three curve was constructed from data for three stocks, which included data from two sampled fish at time 1 postchallenge and similarly for the other sampling times. replicates in assays for which probiotic and pathogenic strains were mixed. The number of fish per stock or treatment varied, depending on availability (Table 5). According to the log-rank analysis (Table 4), infected Turbot with no previous against V. anguillarum in vitro. Similar results have previously treatment with probiotic strains died sooner. been found (Prado et al. 2009). Thus, this method has proved itself to be useful for reducing the number of potential probiotic DISCUSSION strains to be tested in vivo and provides a means of quickly This study demonstrates that bacteria present in marine rear- differentiating between them. ing environments enhance the survival of juvenile Turbot ex- It is generally accepted that laboratory cultures do not survive posed to a pathogenic strain of V. anguillarum, thus qualifying if they are reintroduced into the natural environment because as probiotics (Verschuere et al. 2000). We chose bacteria iso- naturally occurring bacteria outcompete or antagonize them (related from healthy cultures of oysters and clams, the epidermal viewed by Austin and Austin 2007). Thus, the assays in vivo mucus of Turbot, and the ambient water at culture facilities. were crucial. Based on an easy phenotypical differentiation from Thus, all strains are well adapted to the dynamics of Turbot cul- V. anguillarum, we selected the potentially probiotic strains S1 ture. Interestingly, we found that 9 of 148 strains (most of them and S2. The 16S rRNA sequencing data revealed a close phylo- Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 isolated from Ostrea edulis) exerted strong antibacterial activity genetic relationship (above the level proposed as the threshold

TABLE 4. Pairwise comparisons of survival curves by the log-rank test for Turbot stocks analyzed jointly by the Kaplan–Meier method. Blank cells indicate no signiﬁcant differences (P > 0.05); df = 5. See Table 3 for treatments.

Treatmenta M9B S1c S2c Pc S1mP S2mP χ2 P χ2 P χ2 P χ2 P χ2 P χ2 P M9B 75.64 0.0 39.65 0.0 20.41 0.0 S1c 75.64 0.0 39.65 0.0 20.41 0.0 S2c 75.64 0.0 39.65 0.0 20.41 0.0 Pc 75.64 0.0 75.64 0.0 75.64 0.0 32.39 0.0 57.02 0.0 S1mP 39.65 0.0 39.65 0.0 39.65 0.0 32.39 0.0 8.39 0.004 S2mP 20.41 0.0 20.41 0.0 20.41 0.0 57.02 0.0 8.39 0.004 INCREASED SURVIVAL OF JUVENILE TURBOT 259

range of marine animals (reviewed by Lago et al. 2009; Kim et al. 2013), it is important to confirm the lack of virulence of S1 and S2 for Turbot. After being isolated our strains were kept under optimal conditions to preserve their original features, and they were used directly for in vivo virulence assays. They were avirulent by intraperitoneal challenge, which produced a faster development of disease and associated mortality than in the bath challenge (Lago et al. 2012). Similarly, they were avirulent by bath challenge (performed in a large number of treatments) and with different fish stocks (the immunological responses of the different stocks varied somewhat as a result of different rearing conditions). Furthermore, both strains were unable to invade fish tissues even when intraperitoneal assays were performed. Thus, it is unlikely that the strains will show virulence. All these data provide strong evidence that these strains, which were isolated from healthy Ostrea edulis larvae, are also safe for the stage FIGURE 4. Cumulative presence of bacteria in the skin of Turbot after treat- of Turbot tested. Even so, the strains could be pathogenic to ment with the probiotic S2. See Figure 3 for a key to the abbreviations and other other stages and/or species different from those studied. Infec- information. tion can be avoided if (1) they are used in closed hatcheries and (2) biosecurity standards are correctly applied. One effective for intraspecies variability [≥98%]) between our strains and tactic would be to segregate the fish groups by age in accor- most of the species in the Vibrio splendidus group, confirm- dance with good practice on fish farms. Another would be the ing the previously shown inability of such data to differentiate disinfection of water before it leaves the facility to diminish the among these species (reviewed by Beaz-Hidalgo et al. 2010). total number of bacteria in it (Yanong 2012). Both strains were However, our strains can be differentiated from V. celticus (us- also susceptible to all of the antibiotics most frequently used in ing conventional phenotypical traits) by the Voges–Proskauer aquaculture and thus cannot transfer resistance genes to other test and the production of acid from D-lactose, sucrose, and bacteria. L-rhamnose—which are four of the five traits used to distin- The in vivo assays revealed that fish can be protected by using guish this species from V. crassostreae (Beaz-Hidalgo et al. one order of magnitude fewer probiotic cells than pathogenic 2010). They can be also differentiated from V. gigantis using ones (i.e., 105 probiotic cells versus 106 pathogenic cells). How- their growth in 6% NaCl and the production of acid from su- ever, higher concentrations of probiotic were needed to ob- crose (Le Roux et al. 2005). Similarly, the Voges–Proskauertest, tain antibacterial activity against Vibrio species in different or- acid production from L-arabinose, and the degradation of gelatin ganisms (Gram et al. 1999; Vaseeharan and Ramasamy 2003; were useful for distinguishing our strains from V. artabrorum Kesarcodi-Watson et al. 2012). Although inhibitory activity was (Dieguez´ et al. 2011). All of these, in addition to ADH, were tested with 109 CFU of probiotic strains/mL in our in vitro as- unique discriminatory traits between this species and V. cras- says, protection of Turbot was also found with a dose of 105 sostreae. Our strains differ from V. pomeroyi and V. chagasii CFU/mL. As low doses would be preferable in aquaculture, the with respect to the degradation of esculin and, in the case of the dosage that achieves the highest efficacy should be determined latter, with respect to the production of acid from D-mannose (Nikoskelainen et al. 2001).

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 and sucrose (Kim et al. 2013; Lasa et al. 2013). Likewise, the Promising probiotic bacteria isolated from marine environ- ability to produce acid from L-arabinose or sucrose was useful ments have previously been selected for use in the larval aqua- in differentiating our strains from V. hemicentroti (Kim et al. culture of this host (Gatesoupe 1994; Huys et al. 2001; Planas 2013). Finally, our strains were distinguished from V. splen- et al. 2006); despite the high commercial value of Turbot, how- didus by means of three of the four traits used to distinguish ever, the use of probiotics with juveniles has not yet been de- this species from V. crassostreae (ADH, growth in 6% NaCl, veloped. The bacteria that we identified protected Turbot of and production of acid from L-rhamnose; Table S.1). All these three different stocks (which were assayed at different times) results and the source of isolation suggest that our probiotic from infection with V. anguillarum. Although there were dif- strains could be identified as V. crassostreae. Further molecular ferences in the final survival percentages, this is not surprising, studies are needed to confirm the identity of the species. as variation among individual fish is to be expected (Svendsen Our potentially probiotic strains were included in the V.splen- and Bøgwald 1997; Defoirdt et al. 2005; Ogut and Reno 2005; didus clade as expected. In fact, this group has been found to Croxatto et al. 2007). Variations in the immune responses of the be the predominant one in the genus Vibrio in coastal marine different stocks, along with the variable number of fish groups, environments (Lambert et al. 1998; Sobecky et al. 1998). Since could explain our results. Kaplan–Meier analysis is a powerful several species in this clade have caused mortality in a wide tool because it estimates the probability of individual survival 260 FARTO SEGUIN´ ET AL.

TABLE 5. Case summaries, ﬁnal survival rates, and mean survival times for individual Turbot stocks by the Kaplan–Meier method. See Table 3 for additional information.

Mean survival time 95% confidence Average for limits all stocks Total no. No. of No. of Final Treatment Stock of fish dead fish surviving fish survival (%) Estimate SE Lower Upper Mean SE M9B 1 21 0 21 100 30 0 30 30 30 0 2 11 0 11 100 30 0 30 30 3 17 0 17 100 30 0 30 30

S1c 1 21 0 21 100 30 0 30 30 30 0 2 11 0 11 100 30 0 30 30 3 17 0 17 100 30 0 30 30

S2c 1 21 0 21 100 30 0 30 30 30 0 2 11 0 11 100 30 0 30 30 3 17 0 17 100 30 0 30 30

Pc 1 32 31 1 3.1 3.4 0.9 1.7 5.1 10.6 2.7 2 10 5 5 50.0 17.1 4.3 8.7 25.5 3 16 12 4 25.0 11.3 2.8 5.8 16.8

S1mP 1 43 37 6 14.0 10.0 1.3 7.5 12.5 17.6 2.1 2 20 8 12 60.0 18.7 3.2 12.4 24.9 3 32 8 24 75.0 24.3 1.8 20.8 27.8

S2mP 1 21 13 8 38.1 15.1 2.6 9.9 20.2 21.3 2.2 2 20 7 13 65.0 22.4 2.4 17.7 27.1 3 32 5 27 84.4 26.4 1.5 23.4 29.3

independently of source (stock), and pooling data is recom- preferable to antibiotics. Interestingly, our probiotics showed a mended (Srinivasan and Zhou 1993). In fact, the accuracy of wide spectrum of antibacterial activity, which could be use- our results was improved by analyzing all three stocks together. ful against other pathogens that harm Turbot and other marine All of the data revealed that there was significantly higher final organisms. Further studies will confirm whether the probiotic survival when juvenile fish were treated with probiotics. Hence, effect can be achieved in other Turbot growth stages as well as

Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 they confirm the promise of using these strains as probiotics in their practical value against other pathogens. aquaculture. Both probiotic strains persisted in the epidermal mucus of Vaccines to prevent vibriosis in juvenile Turbot produce sur- Turbot for 30 d. Moreover, S1 caused the exclusion of natu- vival rates ≥60% (www.hipra.com/english; Santos et al. 1991). ral occurring bacteria there, at least during the early stages of Achieving an acceptable level of protection requires that the exposure. Competitive exclusion is widely thought to be an in- vaccine be administered at least 1 month prior to exposure to terference mechanism used by probiotic (Jin et al. 1996; Lee a pathogen (Santos et al. 1991; Wang et al. 2013). It is not et al. 2003; Chabrillon et al. 2005; Capkin and Altinok 2009; surprising that vaccination protocols are sometimes inadequate reviewed by Verschuere et al. 2000) and pathogenic bacteria due to poor form management (e.g., insufficient bath exposure, (Vine et al. 2004; reviewed by Austin and Austin 2007). Vibrio incorrect fish sizes, stressful handling, etc.). As a result, it is a anguillarum was previously shown to use the epidermal mucus common practice to use both vaccines and feed supplemented as an access point to fish (reviewed by Austin and Austin 2007). with antibiotics. The use of our probiotic strains in juvenile Tur- Interestingly, we found exclusion of V. anguillarum only during bot could be a way to achieve protection both immediately and the late stages of infection (16–30 d). However, V. anguillarum over the course of the fish’ development which would make them was unable to displace any of the probiotic strains. It was also INCREASED SURVIVAL OF JUVENILE TURBOT 261

unable to overcome the protective effect of the probiotics S1 REFERENCES and S2 in mixed challenges. Although our data are qualitative Austin, B., and D. A. Austin. 2007. Bacterial fish pathogens: diseases of farmed and the potential mechanism of action was not investigated, they and wild fish, 4th edition. Praxis Publishing, Chichester, UK. provide strong evidence that the presence of probiotic bacteria Beaz-Hidalgo, R., A. L. Dieguez,´ I. Cleenwerck, S. Balboa, A. Doce, P.deVos, and J. L. Romalde. 2010. Vibrio celticus sp. nov., a new Vibrio species belong- impedes the steps necessary for V. anguillarum to cause disease ing to the Splendidus clade with pathogenic potential for clams. Systematic in fish. In fact, the probiotic strains were applied 12 h before the and Applied Microbiology 33:311–315. pathogenic strain and are well adapted to the marine medium. Capkin, E., and I. Altinok. 2009. Effects of dietary probiotic supplementations This allows them to establish immediate protection. Therefore, on prevention/treatment of yersiniosis disease. Journal of Applied Microbi- our strains are good candidates for probiotics to be used with ology 106:1147–1153. Chabrillon, M., R. M. Rico, S. Arijo, P. Diaz-Rosales, M. C. Balebona, and M. juvenile Turbot. A. Morinigo.˜ 2005. Interactions of microorganisms isolated from Gilthead In the mixed challenges, significant differences were found Seabream, Sparus aurata L., on Vibrio harveyi, a pathogen of farmed Sene- between the two probiotic strains with respect to their persis- galese Sole, Solea senegalensis (Kaup). Journal of Fish Diseases 28:531–537. tence in the epidermal mucus layer at specific times during the Croxatto, A., J. Lauritz, C. Chen, and D. L. Milton. 2007. Vibrio anguillarum experiment. Furthermore, although the cumulative presence of colonization of Rainbow Trout integument requires a DNA locus involved in exopolysaccharide transport and biosynthesis. Environmental Microbiology S1 was significantly higher than that of the pathogenic strain, 9:370–382. it was similar for S2 and the pathogenic strain. Naturally oc- Daga,´ P., G. Feijoo, M. T. Moreira, D. Costas, A. G. Villanueva, and J. M. Lema. curring bacteria were also significantly absent from the skin of 2013. Bioencapsulated probiotics increased survival, growth and improved sampled fish after challenge with S1 but were never excluded by gut flora of Turbot (Psetta maxima) larvae. Aquaculture International 21:337– S2. Fish previously treated with this strain also showed a signifi- 345. Defoirdt, T., P. Bossier, P. Sorgeloos, and W. Verstraete. 2005. The impact of cantly higher final survival rate than those treated with S1. These mutations in the quorum sensing systems of Aeromonas hydrophila, Vibrio results suggest that the mechanisms by which the pathogen is anguillarum,andVibrio harveyi on their virulence towards gnotobiotically blocked are different for the two strains. Pilot plant–scale ex- cultured Artemia franciscana. Environmental Microbiology 7:1239–1247. periments could be useful in determining the best strain (or Dieguez,´ A. L., R. Beaz-Hidalgo, I. Cleenwerck, S. Balboa, P. de Vos, and J. combination of them) (Daga´ et al. 2013). L. Romalde. 2011. Vibrio atlanticus sp. nov. and Vibrio artabrorum sp. nov. isolated from the clams Ruditapes philippinarum and Ruditapes decussatus. No protection against vibriosis was found in any stock more International Journal of Systematic and Evolutionary Microbiology 61:2406– than 30 d postchallenge with either probiotic strain. Thus, al- 2411. though the strains have the ability to persist in the epidermal Farto,R.,S.P.Armada,M.Montes,J.A.Guisande,M.J.Perez,´ and T. P. Nieto. mucus layer, the data suggest that the presence of S1 or S2 at 2003. Vibrio lentus associated with diseased wild octopus (Octopus vulgaris). that time is insufficient to protect fish against a new exposure to Journal of Invertebrate Pathology 83:149–156. Gatesoupe, F. J. 1991. The effect of three strains of lactic bacteria on the V. anguillarum. Since we tested only a single application of pro- production rate of rotifers, Brachionus plicatilis, and their dietary value for biotics, additional treatments before the end of the 30-d period larval Turbot, Scophthalmus maximus. Aquaculture 96:335–342. could be assayed to determine whether they prevent infection Gatesoupe, F. J. 1994. Lactic acid bacteria increase the resistance of Turbot by V. anguillarum. Although oral administration of live probi- larvae, Scophthalmus maximus, against pathogenic Vibrio. Aquatic Living otic cells could be more effective than adding such cells to the Resources 7:277–282. Gatesoupe, F. J. 1997. Siderophore production and probiotic effect of Vibrio rearing water (Taoka et al. 2006), immediate protection is not sp. associated with Turbot larvae, Scophthalmus maximus. Aquatic Living expected. Despite this, incorporation of our strains into fish feed Resources 10:239–246. would also be essential to ensure the colonization of fish guts Gram, L., J. Melchiorsen, B. Spanggaard, I. Huber, and T. F. Nielsen. 1999. and achieve the highest protection. Inhibition of Vibrio anguillarum by Pseudomonas fluorescens AH2, a possible Our data prove the efficacy of using bacteria present in oyster- probiotic treatment of fish. Applied Environmental Microbiology 65:969– 973. Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 rearing environments to establish the immediate protection of Guisande, J. A., E. P. Lago, S. Prado, T. P. Nieto, and R. Farto. 2008. Genotypic juvenile Turbot against infection by V. anguillarum. Moreover, diversity of culturable Vibrio species associated with the culture of oysters epidermal mucus sampling was useful in investigating the per- and clams in Galicia and screening of their pathogenic potential. Journal of sistence of both probiotic strains and their resistance to being Shellfish Research 27:801–809. excluded by the pathogen. Guisande,J.A.,M.Montes,R.Farto,S.P.Armada,M.J.Perez,´ and T. P. Nieto. 2004. A set of tests for the phenotypic identification of culturable bacteria associated with Galician bivalve mollusc production. Journal of Shellfish Research 23:599–609. ACKNOWLEDGMENTS Huys, G., M. Cnockaert, K. Bartie, D. T. Oanh, N. T. Phuong, T. Somsiri, S. This study was supported by grants 10MMA312017PR from Chinabut, F. M. Yussoff, M. Shariff, M. Giacomini, S. Bertone, J. Swings, the Xunta de Galicia (Regional Government of Galicia) and and A. Teale. 2005. Intra- and interlaboratory performance of antibiotic 09VIA11 from the Universidad de Vigo. We thank D. Costas disc-diffusion-susceptibility testing of bacterial control strains of relevance and A. G. Villanueva for valuable comments, and the staff of for monitoring aquaculture environments. Diseases of Aquatic Organisms 66:197–204. ECIMAT for their helpful technical support. The authors ac- Huys, L., P. Dhert, R. Robles, F. Ollevier, P. Sorgeloos, and J. Swings. 2001. knowledge M. Reigosa for valuable comments on the statistical Search for beneficial bacterial strains for Turbot (Scophthalmus maximus L.). analysis. Aquaculture 193:25–37. 262 FARTO SEGUIN´ ET AL.

Jin, L. Z., Y. W. Ho, M. A. Ali, N. Abdullah, and S. Jalaludin. 1996. Effect Porsby, C. H., K. F. Nielsen, and L. Gram. 2008. Phaeobacter and Ruegeria of adherent Lactobacillus spp. on in vitro adherence of Salmonellae to the species of the Roseobacter clade colonize separate niches in a Danish Turbot intestinal epithelial cells of chicken. Journal of Applied Bacteriology 81:201– (Scophthalmus maximus) rearing farm and antagonize Vibrio anguillarum 206. under different growth conditions. Applied and Environmental Microbiology Kesarcodi-Watson, A., P. Miner, J. L. Nicolas, and R. Robert. 2012. Protective 74:7356–7364. effect of four potential probiotics against pathogen challenge of the larvae of Prado, S., J. Montes, J. L. Romalde, and J. L. Barja. 2009. Inhibitory activity three bivalves: Pacific oyster (Crassostrea gigas), flat oyster (Ostrea edulis), of Phaeobacter strains against aquaculture pathogenic bacteria. International and scallop (Pecten maximus). Aquaculture 344-349:29–34. Microbiology 12:107–114. Kim, D., K. S. Baik, Y. S. Hwang, J. S. Choi, J. Kwon, and C. N. Seong. Ringo, E., and O. Vadstein. 1998. Colonization of Vibrio pelagius and 2013. Vibrio hemicentroti sp. nov., an alginate lyase-producing bacterium, Aeromonas caviae in early developing Turbot (Scopthalmus maximus L.) isolated from the gut microflora of sea urchin (Hemicentrotus pulcherrimus). larvae. Journal of Applied Microbiology 84:227–233. International Journal of Systematic and Evolutionary Microbiology 63:3697– Santos, Y., I. Bandin, S. Nunez,˜ K. Gavningen, and A. E. Toranzo. 1991. 3703. Protection of Turbot, Scophthalmus maximus (L.), and Rainbow Trout, On- Lago, E. P., T. P. Nieto, and R. Farto. 2009. Fast detection of Vibrio species corhynchus mykiss (Richardson), against vibriosis using two different vac- potentially pathogenic for mollusc. Veterinary Microbiology 139:339–346. cines. Journal of Fish Diseases 14:407–4011. Lago, E. P., T. P. Nieto, and R. Farto. 2012. A study of virulence factors Schrijver, De R., and F. Ollevier. 2000. Protein digestion in juvenile Turbot of Aeromonas salmonicida subsp. salmonicida strains associated with in- Scophthalmus maximus and effects of dietary administration of Vibrio prote- fections in Turbot (Psetta maxima L.). Diseases of Aquatic Organisms 99: olyticus. Aquaculture 186:107–116. 145–151. Serrano, J. L. 2014. National aquaculture sector overview. Food and Agriculture Lambert, C., J. L. Nicolas, V. Cilia, and S. Corre. 1998. Vibrio pectenicida sp. Organization of the United Nations, FAO 2005-2014, Rome. nov., a pathogen of scallop (Pecten maximus) larvae. International Journal of Sobecky, P. A., T. J. Mincer, M. C. Chang, A. Toukdarian, and D. R. Helinski Systematic and Evolutionary Microbiology 48:481–487. 1998. Isolation of broad-host-range replicons from marine sediment bacteria. Lasa, A., A. L. Dieguez,´ and J. L. Romalde. 2013. Vibrio toranzoniae sp. nov., Applied and Environmental Microbiology 64:2822–2830. a new member of the Splendidus clade in the genus Vibrio. Systematic and Srinivasan, C., and M. Zhou 1993. A note on pooling Kaplan–Meier estimators. Applied Microbiology 36:96–100. Biometrics 49:861–864. Le Roux, F., A. Goubet, F. L. Thompson, N. Faury, M. Gay, J. Swings, and Svendsen, Y. S., and J. Bøgwald. 1997. Influence of artificial wound and nonin- D. Saulnier. 2005. Vibrio gigantis sp. nov. isolated from the haemolymph of tact mucus layer on mortality of Atlantic Salmon (Salmo salar L.) following cultured oysters (Crassostrea gigas). International Journal of Systematic and a bath challenge with Vibrio anguillarum and Aeromonas salmonicida.Fish Evolutionary Microbiology 55:2251–2255. and Shellfish Immunology 7:317–325. Lee, Y. K., K. Y. Puong, A. C. Ouwehand, and S. Salminen. 2003. Displacement Tamura, K., G. Stecher, D. Peterson, A. Filipski, and S. Kumar. 2013. MEGA6: of bacterial pathogens from mucus and Caco-2 cell surface by lactobacilli. Molecular Evolutionary Genetics Analysis, version 6.0. Molecular Biology Journal of Medical Microbiology 52:925–930. and Evolution 30:2725–2729. Montes, M., R. Farto, M. J. Perez,´ T. P. Nieto, J. L. Larsen, and H. Chris- Taoka, Y.,H. Maeda, J. Y.Jo, S. M. Kim, S. I. Park, T. Yoshikawa, and T. Sakata. tensen. 2003. Characterization of Vibrio strains isolated from Turbot (Scoph- 2006. Use of live and dead probiotic cells in Tilapia Oreochromis niloticus. thalmus maximus) culture by phenotypic analysis, ribotyping, and 16S Fisheries Science 72:755–766. rRNA gene sequence comparison. Journal of Applied Microbiology 95:693– Vaseeharan, B., and P. Ramasamy. 2003. Control of pathogenic Vibrio spp. by 703. Bacillus subtilis BT23, a possible probiotic treatment for black tiger shrimp Montes, M., M. J. Perez,´ and T. P. Nieto. 1999. Numerical taxonomy of gram- Penaeus monodon. Letters in Applied Microbiology 36:83–87. negative, facultative anaerobic bacteria isolated from skin of Turbot (Scoph- Verschuere, L., G. Rombaut, P. Sorgeloos, and W. Verstraete. 2000. Probiotic thalmus maximus) and surrounding water. Systematic and Applied Microbi- bacteria as biological control agents in aquaculture. Microbiology and Molec- ology 22:604–618. ular Biology Reviews 64:655–671. Nikoskelainen, S., A. Ouwehand, S. S. Salminen, and G. Bylund. 2001. Pro- Villamil, L., A. Figueras, M. Planas, and B. Novoa. 2010. Pediococcus acidilac- tection of Rainbow Trout Oncorhynchus mykiss from furunculosis by Lacto- tici in the culture of Turbot (Psetta maxima) larvae: administration pathways. bacillus rhamnosus. Aquaculture 198:229–236. Aquaculture 307:83–88. Ogut, H., and P. W. Reno. 2005. Evaluation of an experimental Aeromonas Vine, N. G., W. D. Leukes, H. Kaiser, S. Daya, J. Baxter, and T. Hecht. salmonicida epidemic in Chinook Salmon, Oncorhynchus tshawytscha (Wal- 2004. Competition for attachment of aquaculture-candidate probiotic and baum). Journal of Fish Diseases 28:263–269. pathogenic bacteria on fish intestinal mucus. Journal of Fish Diseases 27:319– Downloaded by [Department Of Fisheries] at 00:17 17 November 2014 Pintado, J., M. Perez-Lorenzo,´ A. Luna-Gonzalez,´ C. Sotelo, M. Prol, 326. and M. Planas. 2010. Monitoring of the bioencapsulation of a probiotic Wang, C., Y.-H. Hu, B.-G. Sun, H. Chi, J. Li, and L. Sun. 2013. Environmen- Phaeobacter strain in the rotifer Brachionus plicatilis using denaturing gra- tal isolates P1SW and V3SW as a bivalent vaccine induce effective cross- dient gel electrophoresis. Aquaculture 302:182–194. protection against Edwardsiella tarda and Vibrio anguillarum. Diseases of Planas, M., M. Perez-Lorenzo,´ M. Hjelm, L. Gram, I. U. Fiksdal, Ø. Bergh, and Aquatic Organisms 103:45–53. J. Pintado. 2006. Probiotic effect in vivo of Roseobacter strain 27-4 against Yanong, R. P. E. 2012. Biosecurity in aquaculture, part 2. Recirculating aqua- Vibrio (Listonella) anguillarum infections in Turbot (Scophthalmus maximus culture systems. Southern Regional Aquaculture Center, Publication 4708, L.) larvae. Aquaculture 255:323–333. Stoneville, Mississippi. This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:20 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Phenotypic and Genotypic Heterogeneity among Streptococcus iniae Isolates Recovered from Cultured and Wild Fish in North America, Central America and the Caribbean islands Lucy Chouab, Matt J. Griffinc, Trellor Fraitesb, Cynthia Warec, Hugh Fergusond, Natalie Keirsteadbe, John Brakeb, Judy Wilesf, John P. Hawkef, Michael T. Kearneyf, Rodman G. Getchellg, Patricia Gauntc & Esteban Sotoab a Center for Conservation Medicine and Ecosystem, Ross University, Main Island Road. West Farm, St. Kitts, West Indies. b Department of Biomedical Sciences, School of Veterinary Medicine, Ross University, Main Island Road, West Farm, St. Kitts, West Indies c Department of Pathobiology and Population Medicine, College of Veterinary Medicine, Mississippi State University, Post Office Box 197, Stoneville, Mississippi 38776, USA d School of Veterinary Medicine, St. George's University, Post Office Box 7, St. George's, Grenada e Drug Safety and Metabolism Unit, AstraZeneca Research and Development Waltham, Massachusetts 02451, USA f Department of Pathobiological Sciences, School of Veterinary Medicine, Louisiana State University, Skip Bertman Drive, Baton Rouge, Louisiana 70803, USA g Aquatic Animal Health Program, Department of Microbiology and Immunology, College of Veterinary Medicine, Cornell University, Upper Tower Road, Ithaca, New York 14853, USA Published online: 31 Oct 2014.

To cite this article: Lucy Chou, Matt J. Griffin, Trellor Fraites, Cynthia Ware, Hugh Ferguson, Natalie Keirstead, John Brake, Judy Wiles, John P. Hawke, Michael T. Kearney, Rodman G. Getchell, Patricia Gaunt & Esteban Soto (2014) Phenotypic and Genotypic Heterogeneity among Streptococcus iniae Isolates Recovered from Cultured and Wild Fish in North America, Central America and the Caribbean islands, Journal of Aquatic Animal Health, 26:4, 263-271, DOI: 10.1080/08997659.2014.945048 To link to this article: http://dx.doi.org/10.1080/08997659.2014.945048

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 Journal of Aquatic Animal Health 26:263–271, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.945048

ARTICLE

Phenotypic and Genotypic Heterogeneity among Streptococcus iniae Isolates Recovered from Cultured and Wild Fish in North America, Central America and the Caribbean islands

Lucy Chou Center for Conservation Medicine and Ecosystem and Department of Biomedical Sciences, School of Veterinary Medicine, Ross University, Main Island Road, West Farm, St. Kitts, West Indies Matt J. Griffin Department of Pathobiology and Population Medicine, College of Veterinary Medicine, Mississippi State University, Post Office Box 197, Stoneville, Mississippi 38776, USA Trellor Fraites Department of Biomedical Sciences, School of Veterinary Medicine, Ross University, Main Island Road, West Farm, St. Kitts, West Indies Cynthia Ware Department of Pathobiology and Population Medicine, College of Veterinary Medicine, Mississippi State University, Post Office Box 197, Stoneville, Mississippi 38776, USA Hugh Ferguson School of Veterinary Medicine, St. George’s University, Post Office Box 7, St. George’s, Grenada Natalie Keirstead Department of Biomedical Sciences, School of Veterinary Medicine, Ross University, Main Island Road, West Farm, St. Kitts, West Indies; and Drug Safety and Metabolism Unit, AstraZeneca Research and Development, Waltham, Massachusetts 02451, USA John Brake Department of Biomedical Sciences, School of Veterinary Medicine, Ross University, Main Island Road, West Farm, St. Kitts, West Indies

Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 Judy Wiles, John P. Hawke, and Michael T. Kearney Department of Pathobiological Sciences, School of Veterinary Medicine, Louisiana State University, Skip Bertman Drive, Baton Rouge, Louisiana 70803, USA Rodman G. Getchell Aquatic Animal Health Program, Department of Microbiology and Immunology, College of Veterinary Medicine, Cornell University, Upper Tower Road, Ithaca, New York 14853, USA Patricia Gaunt Department of Pathobiology and Population Medicine, College of Veterinary Medicine, Mississippi State University, Post Ofﬁce Box 197, Stoneville, Mississippi 38776, USA

*Corresponding author: [email protected] Received April 20, 2014; accepted July 7, 2014 263 264 CHOU ET AL.

Esteban Soto* Center for Conservation Medicine and Ecosystem and Department of Biomedical Sciences, School of Veterinary Medicine, Ross University, Main Island Road, West Farm, St. Kitts, West Indies

Abstract Streptococcus iniae, the etiological agent of streptococcosis in fish, is an important pathogen of cultured and wild fish worldwide. During the last decade outbreaks of streptococcosis have occurred in a wide range of cultured and wild fish in the Americas and Caribbean islands. To gain a better understanding of the epizootiology of S. iniae in the western hemisphere, over 30 S. iniae isolates recovered from different fish species and geographic locations were characterized phenotypically and genetically. Species identities were determined biochemically and confirmed by amplification and sequencing of the 16S rRNA gene. Repetitive-element palindromic PCR fingerprinting as well as biochemical and antimicrobial susceptibility profiles suggest that a single strain of S. iniae was responsible for two different disease outbreaks among reef fishes in the Caribbean, one in 1999 and another in 2008. Interestingly, a majority of the isolates recovered from cultured fish in the Americas were genetically distinct from the Caribbean isolates and exhibited a trend toward higher minimal inhibitory concentration with respect to several antibiotics as well as greater genetic variability. The biological significance of this genetic variability is unclear, but it could have implications for future vaccine development and treatment.

Streptococcus iniae is a gram-positive coccus that can infect serotypic variation among S. iniae strains antibodies aimed at a wide range of freshwater and marine animals (Keirstead et al. one strain often do not confer protection against other strains 2014). First described from a captive freshwater Amazon River (Aviles et al. 2013). With serious health risks to both humans dolphin Inia geoffrensis in 1976, there have since been several and aquatic animals, proper identification and treatment with confirmed reports of S. iniae outbreaks in finfish from warmwa- appropriate antibiotics should be a priority when dealing with ter regions across the globe (Agnew and Barnes 2007; Baiano S. iniae infections. and Barnes 2009). Often associated with farmed marine and The limited studies of S. iniae have focused on the bio- freshwater finfish, significant economic losses in global aqua- chemical, genetic, and antimicrobial diversity of these bacteria culture industries have been attributed to this pathogen (Agnew (Dodson et al. 1999; Kvitt and Colorni 2004; Nawawi et al. and Barnes 2007; Baiano and Barnes 2009). Equally important 2008; Erfanmanesh et al. 2012). To our knowledge, no study is the mortality that S. iniae can cause in wild fish populations has compared isolates from cultured and wild fish outbreaks (Ferguson et al. 2000). In certain regions of the world, such in the western hemisphere. To gain a better understanding of as the Caribbean islands, outbreaks in wild fish can be devas- the epizootiology of S. iniae in the western hemisphere, we tating for the ecosystem, let alone the people who depend on investigated the genotypic and phenotypic characteristics of S. reef species for economic sustenance and food (Stevens 1999; iniae isolates recovered from wild and cultured fish populations Keirstead et al. 2014). Bacteria shed by sick animals not only during outbreaks in the Caribbean and the Americas.

Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 serve as a source of bacteria to other aquatic animals but also increase the zoonotic potential in at-risk human populations, METHODS including biologists, divers, fishermen, veterinarians, and work- Bacteria.—A total of 31 archived isolates of S. iniae were ers in the aquaculture industry (Stevens 1999; Koh et al. 2009; used in this study (Table 1). Isolates recovered from cultured Keirstead et al. 2014). The increasing number and spread of fish in the Americas were kindly donated by Dr. Rodman cases supports the view that S. iniae is a reemerging pathogen in Getchell (Cornell University), Dr. Andy Goodwin (University of certain geographic locations and potentially associated with the Arkansas–Pine Bluff), Dr. Patricia Gaunt (Mississippi State Uni- emergence of new invasive serotypes and antibiotic resistance versity) and Dr. John Hawke (Louisiana State University). Iso- (Miller and Neely 2005). lates from wild reef fish were previously described as causative Though studies investigating cross-protective vaccines to agents of outbreaks in the Caribbean in 1999–2000 and again in combat this pathogen have been promising, there are unresolved 2006–2007 (Ferguson et al. 2000; Keirstead et al. 2014; Table 1). issues regarding the ability of antibodies to confer protection An Escherichia coli (ATCC accession number 25922) and an against S. iniae (Aviles et al. 2013). Most of these issues are S. agalactiae isolate (previously isolated from an outbreak of attributed to insufficient antigenic stimulation and minimal in- streptococcosis in cultured tilapia Oreochromis sp.) were also duction of the humoral immune response. In addition, owing to included. HETEROGENEITY AMONG STREPTOCOCCUS INIAE ISOLATES 265

TABLE 1. Isolates used in this study. Abbreviations are as follows: SECB = southeastern Caribbean basin, NECB = northeastern Caribbean basin, HSB = hybrid Striped Bass, HT = hybrid Tilapia, and ARS = albino Rainbow Shark.

Year of Hemolysis API 20 API 20 GTG5 Isolate Location Fish isolation pattern 4-h code 24-h code group Reference LSU 94-034 Massachusetts HSB 1994 α 4160010 4162111 A This paper LSU 94-036 Illinois Tilapia 1994 α 4160010 4162115 A This paper LSU 96-325 Iowa Tilapia 1996 α 4160010 4162111 A This paper LSU 97-003 Iowa Tilapia 1997 α 4162110 4163111 A This paper LSU 98-061 Minnesota Tilapia 1998 α 4160010 4160100 A This paper CAN LN 2848 2 Texas Tilapia 1994 β 4562114 4563117 A This paper LSU 01-105 Minnesota Tilapia 2001 β 4562114 4563117 B This paper LSU 07-278A Costa Rica Tilapia 2007 β 4460014 4462115 B This paper LSU 10-070 Florida Tilapia 2010 β 4562114 4563115 B This paper LSU 12-007A Minnesota Tilapia 2012 β 4562114 4563117 B This paper CAN LN 2848 3 Texas HT 1994 γ 4162110 4163111 B This paper ARK PB 03-62A Florida ARS 2003 γ 4462110 4463113 B This paper ARK PB 03-62B Florida ARS 2003 γ 4462110 4463113 B This paper MSU 107-7 Louisiana Tilapia 2000 γ 4562110 4563113 C This paper MSU 109-3 Louisiana Tilapia 2000 γ 4562110 4563113 C This paper MSU 110-4 Louisiana Tilapia 2000 γ 4162110 4163113 C This paper B1 SECB Reef fish 1999 β 4563014 4562017 C Ferguson et al. (2000) B2 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) B3 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) B4 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) B5 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) B6 SECB Reef fish 1999 β 4563014 4562017 C Ferguson et al. (2000) B7 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) B8 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) B9 SECB Reef fish 1999 β 4563014 4563017 C Ferguson et al. (2000) K08-213 NECB Reef fish 2008 β 4563014 4563017 C Keirstead et al. (2014) K08-302 NECB Reef fish 2008 β 4563014 4562017 C Keirstead et al. (2014) K08-408 NECB Reef fish 2008 β 4563014 4563117 C Keirstead et al. (2014) K08-409H NECB Reef fish 2008 β 4562014 4562017 C Keirstead et al. (2014) K08-409K NECB Reef fish 2008 β 4560014 4562017 C Keirstead et al. (2014) K09-302 NECB Reef fish 2008 β 4563014 4563017 C Keirstead et al. (2014)

Biochemical identiﬁcation.—Isolates were inoculated on of the different antibiotics with respect to isolates recovered

Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 trypticase soy agar plates supplemented with 5% sheep blood from the Caribbean and the Americas utilizing the SAS ver- at 37◦C for 48 h. Once pure colony growth was observed, a sion 9.3. All comparisons were considered significant at P ≤ 4.0 McFarland standard solution in 1 × phosphate-buffered 0.05. saline (PBS) was prepared for biochemical identification fol- DNA extraction.—All isolates recovered were used for molec- lowing the manufacturer’s instructions in the API 20 Strep test ular analysis. Initially, genomic DNA was isolated from pure (bioMerieux,´ Durham, North Carolina). The results were read colonies of each isolate. A loop of the bacterium was suspended and recorded and the numerical profile was entered on the man- in 500 µL of sterile 1 × PBS and subjected to DNA extrac- ufacturer’s API Web site for genus and species identification. tion using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, Antimicrobial susceptibility.—The minimal inhibitory con- California) following the manufacturer’s suggested protocol for centration (MIC) of 19 antimicrobial agents was tested against gram-positive bacteria. Isolated genomic DNA was stored at the S. iniae isolates on CMV3AGPF–Sensititre Gram Positive −20◦C until use. Plate Format (Trek Diagnostic System, Cleveland, Ohio) or Molecular identification and genotyping.—Molecular con- on in-house-made plates following the manufacturer’s instruc- firmation of identity was determined by amplification and tions and Clinical and Laboratory Standards Institute guidelines sequencing of the 16S ribosomal RNA gene performed follow- (CLSI 2008). A two-sample t-test was used to analyze the MIC ing previously established protocols (Dorsch and Stackebrandt 266 CHOU ET AL.

1992). Amplicons for sequencing were purified with the Qi- and LSU 12-007A) had a 24-h API fermentation code and β- aQuick PCR Cleanup Kit (Qiagen) according to the manufac- hemolysis pattern identical to those of the Caribbean isolate turer’s suggested protocol and sequenced commercially (Davis K08-408 (Table 1). Sequencing, Davis, California). The generated sequences were All isolates, regardless of their geographic origin, exhibited compared with 16S rRNA sequences deposited in the National a very low MIC for penicillin, streptomycin, kanamycin, orme- Center for Biotechnology Information’s GenBank following toprim/sulphadimethoxine, and gentamicin (Table 2). However, a BLASTn search of the nonredundant nucleotide database the MICs for tigecycline, tetracycline, oxytetracycline, florfeni- (Altschul et al. 1990). col, and chloramphenicol were significantly higher in the North Genomic fingerprinting.—Genomic fingerprinting by and Central American isolates than in the Caribbean isolates repetitive-sequence-mediated PCR was performed using the (P < 0.05; Table 2). In addition, the North and Central Amer- GTG5 primer and BOX primer sets, with slight modifications to ican isolates were variable in their antimicrobial susceptibility published protocols (Versalovic et al. 1991, 1994; Griffin et al. profiles, while the Caribbean isolates were relatively uniform in 2011, 2013). The Escherichia coli and S. agalactiae isolates their susceptibility growth patterns. mentioned previously were included as outliers for genetic µ analysis. Briefly, the analysis consisted of 50- L reactions com- Genotype Comparison prised of 25 µL of IQ Supermix (BioRad, Hercules, California), Molecular diagnosis provided by amplifying and sequencing 40 pmol of primer (GTG5:5-GTGGTGGTGGTGGTG-3 ; the 16S rRNA gene confirmed that the isolates were S. iniae.The ∼ BOX: 5 -CTACGGCAAGGCGACGCTGACG-3 ), 25 ng BOX primers generated relatively simple banding patterns. Re- of DNA template, and nuclease-free water to volume. Am- gardless of origin, largely identical banding patterns (consisting plifications were performed on a PTC-200 gradient cycler of four visibly distinct bands that were generally homogenous (MJ Research, Waltham, Massachusetts) using two different in size and intensity) were produced for a majority of the iso- temperature profiles. The temperature profile for the GTG5 ∼ ◦ lates (Figure 1). The lone exception was an 1,000-bp band that primer consisted of 1 cycle at 95 C for 10 min; 40 cycles of was markedly more intense in three isolates (LSU 97-003, LSU 95◦Cfor1min,46◦C for 1 min, and 72◦C for 3 min; and final ◦ 96-325, and CAN LN 2848 3), which clustered separately from extension at 72 C for 10 min. The temperature profile for the the rest of the group. However, based on banding patterns and BOX primer consisted of 1 cycle of 95◦C for 10 min; 5 cycles of ◦ ◦ ◦ UPGMA analysis, the isolates were largely homogenous with 95 Cfor1min,40 C for 1 min, and 72 C for 5 min; 35 cycles respect to the fingerprints generated from the BOX primers, of 95◦Cfor1min,55◦C for 1 min, and 72◦C for 1 min; and final ◦ forming one large, monophyletic group. By contrast, the GTG5 extension at 72 C for 5 min. Aliquots of each amplification primer generated much more complex fingerprints, with pat- µ reaction (10 L each) were electrophoresed through a 2.0% terns consisting of 10–14 bands of varying degrees of intensity (weight : volume) agarose gel supplemented with ethidium (Figure 1). The S. iniae isolates formed three distinct clusters µ bromide (0.25 g/mL) and visualized under ultraviolet light. (A, B, and C), with some geographic and hemolytic correla- Band sizes were determined by direct comparison with con- tion (Table 1; Figure 2). With the exception of three isolates currently run DNA standards (Hyperladder II, Bioline USA, (MSU 107-7, MSU 109-3, and MSU 110-4), the isolates from Taunton, Massachusetts). Visibly distinct bands were manually the Americas clustered together, forming two distinct subgroups annotated and genetic fingerprints were analyzed using the (A and B) within one large cluster. All of the α-hemolytic iso- Quantity One software version 4.6.5 (Bio-Rad Laboratories) lates from the United States fell into genogroup A. Similarly, to calculate the Dice coefficients and generate a dendrogram a majority of the β-hemolytic isolates from the Americas fell based on the unweighted pair-group method using arithmetic into genogroup B. The Caribbean isolates also grouped together, Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 averages (UPGMA). forming a single distinct cluster (C), which included the three MSU isolates from Louisiana. All genogroup C isolates from the Caribbean were β-hemolytic, while the three C-group isolates RESULTS from the Americas were γ-hemolytic. Phenotype Comparison Isolates demonstrated moderate growth on most blood agar DISCUSSION plates following 24 h of incubation at 37◦C, with most isolates Given the increased demand for wild and cultured fish stocks producing pinpoint whitish colonies. All of the Caribbean iso- as a source of dietary protein, it is important to develop a thor- lates showed β-hemolysis, whereas the isolates from North and ough understanding of the epizootic tendencies of important fish Central America produced α-, β-, and γ-hemolysis (Table 1). pathogens, especially those with zoonotic potential (Fulde and In addition, the North and Central American isolates possessed Valentin-Weigand 2013). In addition, efforts to create reliable different biochemical profiles as determined by the API 20 Strep diagnostic tests and establish effective preventative measures (Table 1). Interestingly, some isolates recovered from cultured and treatment options are critical, as devastating losses in fin- tilapia from the United States (CAN LN 2848 2, LSU 01-105, fish culture in Japan, Taiwan, Israel, and the United States have Downloaded by [Department Of Fisheries] at 00:20 17 November 2014

TABLE 2. Minimal inhibitory concentrations of 19 different antibiotics with respect to the Streptococcus iniae isolates used in this study as determined by broth microdilution. Antibiotics are as follows: tigecycline (TGC), erythromycin (ERY), tetracycline (TET), ciproﬂoxacin (CIP), chloramphenicol (CHL), penicillin (PEN), daptomycin (DAP), vancomycin (VAN), streptomycin (STR), nitrofurantoin (NTR), tylosin tartrate (TYLT), gentamicin (GEN), quinupristin/dalfopristin (SYN), lincomycin (LIN), linezolid (LZD), kanamycin (KAN), ﬂorfenicol (FFN), oxytetracycline (OXY), Sulfadimethoxine/ormetoprim (PRI).

Isolate TGC ERY TET CIP CHL PEN DAP VAN STR NIT TYLT GEN SYN LIN LZD KAN FFN OXY PRI Caribbean isolates B1 0.06 0.25 1 1 4 0.25 0.3 1 512 2 0.25 128 0.5 1 2 128 4 1 4.75/0.25 B2 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B3 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B4 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B5 0.06 0.25 1 1 4 0.25 0.3 1 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B6 0.06 0.25 1 1 4 0.25 0.3 1 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B7 0.06 0.25 1 1 4 0.25 0.3 1 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B8 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 B9 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 K08-213 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 K08-302 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 K08-408 0.03 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 K08-409H 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 1 4.75/0.25 K08-409K 0.03 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 0.5 4.75/0.25 K09-302 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 4.75/0.25 Non-Caribbean isolates ARK PB 03-62A 0.03 0.25 1 0.5 4 0.25 0.3 1 512 16 0.25 128 0.5 1 1 128 2 1 4.75/0.25 ARK PB 03-62B 0.03 0.25 1 0.5 4 0.25 0.3 0.5 512 32 0.25 128 0.5 1 2 128 2 1 4.75/0.25 CAN LN 2848 2 0.25 0.25 16 0.5 8 0.25 0.3 1 512 4 0.25 128 0.5 1 2 128 8 32 4.75/0.25 CAN LN 2848 3 0.03 0.25 1 4 4 0.25 1 0.5 512 2 0.25 128 0.5 1 2 128 2 1 4.75/0.25 LSU 01-105 0.25 0.25 16 0.5 8 0.25 0.3 1 512 4 0.25 128 0.5 1 2 128 4 16 4.75/0.25 LSU 07-278A 0.12 0.25 8 0.5 4 0.25 0.3 1 512 2 0.25 128 0.5 1 2 128 4 16 4.75/0.25 LSU 10-070 0.25 0.25 16 0.5 8 0.25 0.5 1 512 4 0.25 128 0.5 1 4 128 4 16 4.75/0.25 LSU 12-007A 0.25 0.25 16 0.5 8 0.25 0.3 1 512 4 0.25 128 0.5 1 2 128 4 16 4.75/0.25 LSU 94-034 0.12 0.25 4 0.5 4 0.25 0.3 0.5 512 4 0.25 128 0.5 1 2 128 4 16 4.75/0.25 LSU 94-036 0.03 0.25 1 0.3 4 0.25 0.3 0.25 512 4 0.25 128 0.5 1 2 128 4 1 4.75/0.25 LSU 96-325 0.06 0.25 1 0.5 4 0.25 0.3 0.5 512 4 0.25 128 0.5 1 2 128 4 1 4.75/0.25 LSU 97-003 0.03 0.25 1 0.5 4 0.25 0.3 0.5 512 4 0.25 128 0.5 1 2 128 4 1 4.75/0.25 LSU 98-061 0.06 0.25 4 0.5 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 4 8 4.75/0.25 MSU 107-7 0.06 0.25 1 1 4 0.25 0.3 1 512 2 0.25 128 0.5 1 2 128 2 2 9.5/0.5 MSU 109-3 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 9.5/0.5 MSU 110-4 0.06 0.25 1 1 4 0.25 0.3 0.5 512 2 0.25 128 0.5 1 2 128 2 2 9.5/0.5 267 268 CHOU ET AL.

FIGURE 1. Repetitive-sequence-mediated PCR ampliﬁcation of DNA from Streptococcus iniae (lanes 1–31), S. agalactiae (lane 32) and E. coli (lane 33) isolates. Genetic proﬁles were generated using BOX (A) and GTG5 (B) primers. Lane designations are as follows for all gels: L = hyperladder II; 1 = LSU 07-278A; 2 = LSU 01-105; 3 = LSU 12-007A; 4 = LSU 10-070; 5 = CAN LN 2848 2; 6 = ARK PB 03-62A; 7 = ARK PB 03-62B; 8 = CAN LN 2848 3; 9 = LSU 96-325; 10 = LSU 98-061; 11 = LSU 94-034; 12 = LSU 94-036; 13 = LSU 97-003; 14 = MSU 110-4; 15 = MSU 109-3; 16 = MSU 107-7; 17 = B1; 18 = B2; 19 = B3; 20 = B4; 21 = B5; 22 = B6; 23 = B7; 24 = B8; 25 = B9; 26 = K08-213; 27 = K08-302; 28 = K08-408; 29 = K08-409H; 30 = K08-409K; 31 = K09-302; 32 = S. agalactiae;33= E. coli; and 34 = negative control.

already been attributed to S. iniae (Lau et al. 2003). Initially, the in aquaculture (oxytetracycline dehydrate, Terramycin type A commercially available API 20 Strep test was used to compare medicated article for fish, Phibro Animal Health, Fairfield, New the biochemical profiles of the S. iniae isolates. The numeric Jersey; florfenicol, Aquaflor type A medicated article, Merck profile analysis showed a high degree of variation in the iso- Animal Health, Summit, New Jersey). These five antimicrobials lates from the Americas, indicating greater heterogeneity than act as protein synthesis inhibitors, and tetracycline, oxytetracy- among Caribbean API profiles, which were largely homologous line, and tigecycline actually share structural and mechanistic (Table 1). The Caribbean isolates were mostly similar in their similarities (Prescott 2004). The differences in MICs between 24-h API 20 profile code. However, it should be recognized the isolates from the Americas and the Caribbean could be re- that biochemical identification of S. iniae using commercial kits lated to the source of isolation, as all of the North and Central such as the API 20 Strep test is largely inaccurate since a major- American isolates were recovered from cultured fish while all

Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 ity of these kits do not test for S. iniae (Bernard and Ryniewicz of the Caribbean isolates were recovered from wild reef fish, 1984; Facklam et al. 2005). suggesting less selective pressure for the emergence of antimi- Antimicrobial susceptibility profiles of S. iniae are important crobial resistance. This finding further stresses the importance considering that beta-lactam antibiotics such as penicillin are of- of responsible antimicrobial use in aquaculture to avoid inciden- ten used to treat bacterial infections in humans (Lau et al. 2003) tal selection for resistance, especially in bacteria with zoonotic as well as commonly being used in aquaculture in some regions potential. of the world (Park et al. 2012). Susceptibility testing indicated no Although useful in identifying phenotypically similar strains significant differences in MIC between isolates recovered in the of bacteria, commercial biochemical test kits and antimicro- western hemisphere for 14 of the compounds tested (Table 2). bial susceptibility profiles lack the resolution of molecular tech- However, significantly higher MICs for tigecycline, tetracycline, niques in differentiating between largely similar, phenotypically oxytetracycline, florfenicol, and chloramphenicol were found ambiguous bacterial strains. To date, several methodologies for the North and Central American isolates (P < 0.05); two of have been employed to genetically differentiate between S. iniae these antibiotics (florfenicol and oxytetracycline) are currently isolates recovered from different hosts in different geographic approved by the U.S. Food and Drug Administration for use locations, with varying degrees of success. Eldar et al. (1997) HETEROGENEITY AMONG STREPTOCOCCUS INIAE ISOLATES 269 Downloaded by [Department Of Fisheries] at 00:20 17 November 2014

FIGURE 2. UPGMA dendrograms based on Dice coefﬁcient matrices generated from PCR ampliﬁcation of S. iniae DNA using GTG5 primers. The S. iniae samples formed three distinct clusters, referred to in this study as GTG5 genogroups A, B, and C.

compared S. iniae isolates recovered from the USA and Israel (ATCC 29178T) isolated from a freshwater dolphin. Alterna- by restriction fragment length polymorphism (RFLP) based on tively, Fuller et al. (2001) used pulsed field gel electrophore- PCR amplified 16S rDNA and ribotyping with Hind III and sis to genotype S. iniae isolates recovered from fish and hu- EcoR I. The RFLP was not useful for intraspecies comparisons; mans, indicating genetic and virulence diversity associated with however, Hind III and EcoR I ribotypes of S. iniae differen- the host of origin when utilizing a mouse model (Fuller et al. tiated between the western hemisphere and Israeli strains as 2001). Dodson et al. (1999) used rapidly amplified polymorphic well as differentiating fish isolates from the S. iniae type strain DNA (RAPD) analysis and repetitive-element-palindromic PCR 270 CHOU ET AL.

(rep-PCR; BOX) to characterize S. iniae from fish and mammals the Americas clustered together in groups A and B, indicating in North America, finding that neither RAPD nor rep-PCR had a shared origin. The patterns generated by the BOX primers the discriminatory power to distinguish between human and fish consisted of a small number of analogous bands, demonstrating isolates or between isolates recovered from diseased and asymp- a lack of discriminatory resolution as the resulting dendrogram tomatic fish. Kvitt and Colorni (2004) used a similar approach (data not shown) clustered isolates with no correlation to phe- when comparing S. iniae isolates from Israel with isolates recov- notype or geographic location. ered from North America, Australia, Barbados, and Japan. The The results of this study suggest that the same strain of S. iniae authors reported that amplified fragment length polymorphism is responsible for at least two large disease outbreaks among analyses provided better discriminative power than RAPD anal- reef fish in the Caribbean (Ferguson et al. 2000; Keirstead et al. yses and provided evidence of genetic variability among S. iniae 2014). Comparably, there appears to be more than one genotyp- isolates recovered from different geographic locations. Nawawi ically distinct strain associated with disease outbreaks among et al. (2008) utilized pulsed-field gel electrophoresis (PFGE) cultured fish in the USA. By understanding the fundamental dif- separation of Sma I digests of genomic DNA to compare S. iniae ferences between isolates associated with disease outbreaks in isolates recovered from barramundi (also known as barramundi wild fish in the Caribbean and cultured fish in Central and North perch) Lates calcarifer from different locations in Australia America, we can develop a knowledge base on the epizootiol- with isolates recovered from humans in North America. They ogy of S. iniae in both wild and cultured fish. Future work in this reported clear similarities between the strains isolated from fish area should include challenge studies to assess the pathogenic- and humans regardless of geographic origin, though there were ity of the various strains in susceptible species, along with the also marked differences that did not correlate with geographic development of biological models of antimicrobial treatments. origin or host species. More recently, Erfanmanesh et al. (2012) reported high genetic similarity of S. iniae isolates recovered from diseased farmed Rainbow Trout Oncorhynchus mykiss in ACKNOWLEDGMENTS Iran as determined by RAPD analysis. This work was supported by the Ross University School of Previous work has demonstrated that rep-PCR profiles are Veterinary Medicine, Center for Conservation Medicine and generally concordant with PFGE profiles in identifying clonal Ecosystem, and the Mississippi State University College of Vet- relationships from bacterial isolates (Malathum et al. 1998; Ross erinary Medicine and Mississippi Agricultural and Forestry Ex- et al. 2005; Phasipol et al. 2013). Given that rep-PCR is less time- periment Station. We thank Andrew Goodwin and Emily Mare- consuming and less costly than PFGE and other more compli- caux from the University of Arkansas–Pine Bluff for their kind cated techniques, rep-PCR is more suitable for routine use and contribution of isolates and their critical review of this paper. is an increasingly useful technique for discriminating between phenotypically ambiguous strains of bacteria. This current work REFERENCES supports the findings of Dodson et al. (1999), demonstrating the Agnew, W., and A. C. Barnes. 2007. Streptococcus iniae: an aquatic pathogen limited utility of using the BOX primer to discriminate between of global veterinary significance and a challenging candidate for reliable S. iniae isolates recovered from different hosts in different geo- vaccination. Veterinary Microbiology 122:1–15. graphic locations. However, the GTG5 primer offered adequate Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. Journal of Molecular Biology 215:403–410. resolution. Based on GTG5 genogroups, the Caribbean isolates Aviles, F., M. M. Zhang, J. Chan, J. Delamare-Deboutteville, T. J. Green, C. were largely clonal, which is consistent with phenotypic obser- Dang, and A. C. Barnes. 2013. The conserved surface M-protein SiMA of vations and suggests that this group is largely homologous. Streptococcus iniae is not effective as a cross-protective vaccine against differ- Figures 1 and 2 demonstrate the utility of the GTG5 primer ing capsular serotypes in farmed fish. Veterinary Microbiology 162:151–159.

Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 in determining the regional origins of isolates, suggesting com- Baiano, J. C. F., and A. Barnes. 2009. Towards control of Streptococcus iniae. mon ancestry due to similar geographic origins. Several isolates Emerging Infectious Diseases 15:1891–1896. Bernard, P., and H. Ryniewicz. 1984. Evaluation of the API 20 Strep system for from cultured tilapia (MSU 107-7; MSU 109-3; MSU 110-4) species identification of Streptococci isolated from bovine mastitis. Journal showed genetic similarities to the Caribbean isolates; however, of Clinical Microbiology 19:213–214. these isolates had different API 20 codes. Although there are CLSI (Clinical and Laboratory Standards Institute). 2008. Performance stan- some correlations between genetic group and phenotype they dards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animal: approved standards, 3rd edition. CLSI, Wayne, Pennsyl- are not absolute, suggesting that regions amplified by the GTG5 vania. primer are not associated with biochemical pathways. However, Dodson, S. V., J. J. Maurer, and E. B. Shotts. 1999. Biochemical and molecular all of the Caribbean isolates were collected either from the north- typing of Streptococcus iniae isolated from fish and human cases. Journal of eastern or southeastern Caribbean basin in 1999 or 2008 (see Fish Diseases 22:331–336. Table 1). Although collected during two different disease out- Dorsch, M., and E. Stackebrandt. 1992. Some modifications in the procedure of direct sequencing of PCR-amplified 16S rDNA. Journal of Microbiological breaks, based on the GTG5 genogroups these isolates formed a Methods 16:271–279. distinct clade separate from the majority of the North and Cen- Eldar, A., S. Lawhon, P. F. Frelier, L. Assenta, B. R. Simpson, P. W. Varner, tral American isolates, suggesting that the Caribbean isolates and H. Bercovier. 1997. Restriction fragment length polymorphisms of have a common ancestry. Similarly, most of the isolates from 16S rDNA and of whole rRNA genes (ribotyping) of Streptococcus iniae HETEROGENEITY AMONG STREPTOCOCCUS INIAE ISOLATES 271

strains from the United States and Israel. FEMS Microbiology Letters 151: Lau, S. K. P., P. C. Y. Woo, H. Tse, K. Leung, S. S. Y. Wong, and K. Yuen. 155–162. 2003. Invasive Streptococcus iniae infections outside North America. Journal Erfanmanesh, A., M. Soltani, E. Pirali, S. Mohammadian, and A. Taherim- of Clinical Microbiology 41:1004–1009. irghaed. 2012. Genetic characterization of Streptococcus iniae in diseased Malathum, K., K. Singh, G. Weinstock, and B. E. Murray. 1998. Repetitive farmed Rainbow Trout (Oncorhynchus mykiss) in Iran. Scientific World Jour- sequence-based PCR versus pulsed-field gel electrophoresis for typing of nal [online serial] 2012:article 594073. Enterococcus faecalis at the subspecies level. Journal of Clinical Microbiol- Facklam, R., J. Elliott, L. Shewmaker, and A. Reingold. 2005. Identification ogy 36:211–215. and characterization of sporadic isolates of Streptococcus iniae isolated from Miller, J. D., and M. N. Neely. 2005. Large-scale screen highlights the impor- humans. Journal of Clinical Microbiology 43:933–937. tance of capsule for virulence in the zoonotic pathogen Streptococcus iniae. Ferguson, H. W., V. S. St. John, C. J. Roach, S. Willoughby, C. Parker, and Infection and Immunity 73:921–934. R. Ryan. 2000. Caribbean reef fish mortality associated with Streptococcus Nawawi, R. A., J. Baiano, and A. C. Barnes. 2008. Genetic variability amongst iniae. Veterinary Record 147:662–664. Streptococcus iniae isolates from Australia. Journal of Fish Diseases 31: Fulde, M., and P. Valentin-Weigand. 2013. Epidemiology and pathogenicity 305–309. of zoonotic Streptococci. Current Topics in Microbiology and Immunology Park, Y. H., S. Y. Hwang, M. K. Hong, and K. H. Kwon. 2012. Use of antimicro- 368:49–81. bials in aquaculture in Antimicrobial resistance in animal and public health. Fuller, J. D., D. J. Bast, V. Nizet, D. E. Low, and J. C. de Azavedo. 2001. Scientific and Technical Review 31:189–197. Streptococcus iniae virulence is associated with a distinct genetic profile. Phasipol, P., N. Chokesajjawatee, and T. Taradon. 2013. Comparison of repet- Infection and Immunity 69:1994–2000. itive sequence-based polymerase chain reaction (rep-PCR) and pulsed-field Griffin, M. J., M. J. Mauel, T. E. Greenway, L. H. Khoo, and D. J. Wise. 2011. A gel electrophoresis (PFGE) for genetic characterization of Arcobacter spp. real-time polymerase chain reaction assay for quantification of Edwardsiella Thai Journal of Veterinary Medicine 43:307–311. ictaluri in catfish pond water and genetic homogeneity of diagnostic case Prescott, J. F. 2004. Antimicrobial chemotherapy. Pages 26–46 in D. C. Hirsh, isolates from Mississippi. Journal of Aquatic Animal Health 23:178–188. N. J. MacLachlan, and R. L. Walker, editors. Veterinary microbiology, 2nd Griffin, M. J., S. M. Quiniou, T. Cody, M. Tabuchi, C. Ware, R. C. Cipriano, M. J. edition. Blackwell Scientific Publications, Ames, Iowa. Mauel, and E. Soto. 2013. Comparative analysis of Edwardsiella isolates from Ross, T. L., W. G. Merz, M. Farkosh, and K. C. Carroll. 2005. Comparison fish in the eastern United States identifies two distinct genetic taxa amongst of an automated repetitive sequence-based PCR microbial typing system organisms phenotypically classified as E. tarda. Veterinary Microbiology to pulsed-field gel electrophoresis for analysis of outbreaks of methicillin- 165:358–372. resistant Staphylococcus aureus. Journal of Clinical Microbiology 43:5642– Keirstead, N. D., J. W. Brake, M. J. Griffin, I. Halliday-Simmonds, M. A. 5647. Thrall, and E. Soto. 2014. Fatal septicemia caused by the zoonotic bacterium Stevens, D. L. 1999. The flesh-eating bacterium: what’s next? Journal of Infec- Streptococcus iniae during an outbreak in Caribbean reef fish. Veterinary tious Diseases 179:S366–74. Pathology 51:1035–1041. Versalovic, J., T. Koeuth, and J. R. Lupski. 1991. Distribution of repetitive Koh, T. H., L. H. Sng, S. M. Yuen, C. K. Thomas, P. L. Tan, S. H. Tan, and DNA sequences in eubacteria and application to fingerprinting of bacterial N. S. Wong. 2009. Streptococcal cellulitis following preparation of fresh raw genomes. Nucleic Acids Research 19:6823–6831. seafood. Zoonoses and Public Health 56:206–208. Versalovic, J., M. Schneider, F. J. de Bruijn, and J. R. Lupski. 1994. Genomic Kvitt, H., and A. Colorni. 2004. Strain variation and geographic endemism in fingerprinting of bacteria using repetitive sequence-based polymerase chain Streptococcus iniae. Diseases of Aquatic Organisms 61:67–73. reaction. Methods in Molecular and Cell Biology 5:25–40. Downloaded by [Department Of Fisheries] at 00:20 17 November 2014 This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:21 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Evaluation on Activity of Cytochrome P450 Enzymes in Turbot Via a Probe Drug Cocktail Zhi-Qiang Changa, Jian Lia & Qian-Qian Zhaia a Key Laboratory of Sustainable Development of Marine Fisheries, Ministry of Agriculture, Yellow Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, 106 Nanjing Road, Qingdao 266071, China Published online: 04 Nov 2014.

To cite this article: Zhi-Qiang Chang, Jian Li & Qian-Qian Zhai (2014) Evaluation on Activity of Cytochrome P450 Enzymes in Turbot Via a Probe Drug Cocktail, Journal of Aquatic Animal Health, 26:4, 272-277, DOI: 10.1080/08997659.2014.938868 To link to this article: http://dx.doi.org/10.1080/08997659.2014.938868

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:272–277, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.938868

ARTICLE

Evaluation on Activity of Cytochrome P450 Enzymes in Turbot Via a Probe Drug Cocktail

Zhi-Qiang Chang, Jian Li,* and Qian-Qian Zhai Key Laboratory of Sustainable Development of Marine Fisheries, Ministry of Agriculture, Yellow Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, 106 Nanjing Road, Qingdao 266071, China

Abstract Cytochrome P450s (CYPs) are the main catalytic enzymes for metabolism by a variety of endogenous and exogenous substrates in mammals, fish, insects, etc. We evaluated the application of a multidrug cocktail on changes in CYP1, CYP2, and CYP3 activity in Turbot Scophthalmus maximus. The probe drugs were a combination of caffeine (5 mg/kg body weight), dapsone (5 mg/kg), and chlorzoxazone (10 mg/kg). After a single intraperitoneal injection of the cocktail, the concentration of all three probe drugs in the plasma increased quickly to a peak and then decreased gradually over 24 h. Pharmacokinetic profiles of the three probe drugs were determined using a noncompartmental analysis, and the typical parameters were calculated. In the assay for CYP induction, pretreatment with rifampicin significantly reduced the typical pharmacokinetic metrics for caffeine and chlorzoxazone, but not dapsone, indicating that the activity of CYP1 and CYP2 in turbot were induced by rifampicin.

The Turbot Scophthalmus maximus is a demersal flatfish and pheromones (Werck-Reichhart and Feyereisen 2000). native to marine or brackish waters of the North Atlantic, Baltic Because CYPs are relatively conserved across vertebrate lin- Sea and the Mediterranean Sea. It is a favorite seafood in China eages, fish share all CYP families with mammals, except due to its good taste and therefore is an important economic CYP39. In various CYPs, CYP2s are the least conserved family marine fish species and has been successfully cultivated in China with extensive subfamily diversification (Nelson 2003). Up to since 1992. According to data of China Modern Agriculture now, CYPs in mammals, especially human beings, have been Research System Flatfish Center, the total Turbot production deeply investigated, not only for their protein function but also of China has increased remarkably in recent years, reaching for their genome information. Human CYP3A4, 2D6, 2C9, and

Downloaded by [Department Of Fisheries] at 00:21 17 November 2014 57,000 metric tons in 2009, which accounts for 80% of the 1A2 are the main CYPs, associated with 50, 25, 15 and 5%, re- world production (CARSFC 2010). spectively, of drug metabolism (Hemeryck and Belpaire 2002). The cytochrome P450 (CYP) enzymes are particularly However, the full catalytic function of many fish CYPs is still known for the oxidative metabolism of a variety of endogenous unknown. and exogenous substrates in mammals, fish, insects, etc. (Nelson Several in vitro and in vivo techniques are available to mea- et al. 1996).They are able to increase the polarity of xenobiotics, sure the activity of these enzymes, but only the phenotyping such as drugs and environmental toxicants, producing metabo- based on probe drug pharmacokinetics permits measurement of lites that are more amenable to excretion (Curcio et al. 2010). real-time enzyme activity that can detect the combined influ- Among the biotransformation of marketed drugs in human be- ence of genetic, disease, and environmental factors on enzyme ings, over 75% are related to CYPs (Guengerich 2008). Mean- activity (Streetman et al. 2000). Therefore, a new probe drug while, many CYPs also participate in the metabolism of endoge- cocktail containing caffeine, dapsone, and chlorzoxazone was nous signal molecules, such as steroid hormones, eicosanoids, developed in our laboratory to provide a feasible method for

*Corresponding author: [email protected] Received November 25, 2013; accepted May 29, 2014

272 CYTOCHROME P450 ENZYMES IN TURBOT 273

evaluating CYP1, CYP2, and CYP3 activities simultaneously drug injection; otherwise the procedure followed was identical in flatfish, although it probably applies to other fish as well. to that of the controlled experiment. Determination of drug concentration.—The caffeine, dapsone, and chlorzoxazone concentrations in plasma was simul- METHODS taneously determined by HPLC, using an Agilent 1200 HPLC Chemicals and reagents.—Caffeine and rifampicin were pur- system comprising a quaternary pump, a 5-µm Hypersil ODS chased from the Aladdin Industrial Inc. (Shanghai, China). Dap- column (250 × 4.6 mm), an automatic injector, and an Agilent sone and chlorzoxazone were purchased from Sigma-aldrich G1314A UV detector. The mobile phase consists of acetonitrile Co., (St. Louis, Missouri). Methanol and acetonitrile were (ACN) and 20-mM phosphate buffer (pH 6.0), and a multistep HPLC grade and purchased from Merck KGaA, (Darmstadt, gradient procedure was used as follows: 15% ACN for 0–4 min, Germany). Other chemicals used in the study were analytical 15– 40% ACN for 4–8 min, 40% ACN for 8–13 min, 40–15% grade unless specifically described. To make a working solu- ACN for 13–15 min, and 15% ACN for 15–16 min. The flow rate tion of the cocktail, caffeine, dapsone and chlorzoxazone were was set at 1 mL/min. The UV absorbance was set at 275 nm for initially mixed with Tween80 in a sterilized mortar, and then 0–8 min, 295 nm for 8–12 min, and 280 nm for 12–16 min. As diluted with sterilized phosphate buffered saline. to sample pretreatment, 200 µL of plasma was thawed at room Turbot processing.—Healthy Turbot weighing 100 g (SD, 8) temperature and then mixed with 1 mL dichloromethane. The were obtained from the Haiyang Yellow Sea Fisheries Co., Ltd. sample mixture was vortex mixed at high speed for 5 min, and and transferred into 1-m3 fiberglass reinforced plastic tanks with then centrifuged at 5,000 revolutions/min for 10 min. The lower aerated seawater at 10 fish/tank. The seawater temperature was part was transferred into a clean tube and dried under nitrogen 20◦C (SD, 2), salinity was 31.7‰, and pH was 8.1. The subjects gas. Residue was redissolved in 200 µL mobile phase (15% ACN were acclimatized for 1 week prior to the experiment and fed in phosphate buffer), and 20 µL sample was injected into HPLC with commercial feed pellet two times per day. system. The limit of detection of this analytical method was Clinical study design and sampling.—In order to evaluate the calculated to be 0.03 µg/mL for caffeine, 0.02 µg/mL for dap- apparent CYP enzyme activity of Turbot, a controlled experi- sone, and 0.05 µg/mL for chlorzoxazone. ment and an induction experiment were performed; 50 Turbot Pharmacokinetics analysis.—Pharmacokinetic measures of were used for each. In the controlled experiment, each fish was caffeine, dapsone, and chlorzoxazone were determined via non- weighed and then the probe drug cocktail solution—caffeine compartmental analysis of plasma concentrations and real-time 5 mg/kg of body weight, dapsone at 5 mg/kg, and chlorzoxa- values using the DAS software, (ver. 2.0, Drug Clinical Re- zone at 10 mg/kg—was injected into abdominal cavity using a search Center of Shanghai University of T.C.M. and Shanghai 1-mL syringe. Then, five fish were sampled at 5, 10, 15, 30, and BioGuider Medicinal Technology Co., Ltd., China). The peak 45 min and at 1, 2, 6, 8, 12, and 24 h after drug administra- plasma concentration (Cmax), the time to reach Cmax (Tmax), tion, when 500 µL of blood was collected from the caudal vein. and time corresponding to the last concentration above the limit Plasma was obtained by centrifugation at 3,000 revolutions/min of quantification (tlast) were obtained directly from experimen- for 10 min and stored at −20◦C until HPLC analysis. For the tal observations. The area under the plasma concentration–time induction experiment, rifampicin was injected into abdominal curve (AUC) extrapolated to infinity was calculated according cavity at 50 mg/kg of body weight at 30 min before the probe to the following equation: AUC = AUClast + (Clast/λz), where Downloaded by [Department Of Fisheries] at 00:21 17 November 2014

FIGURE 1. HPLC chromatogram of blank Turbot plasma sample (lower blue signal) and the sample spiked simultaneously with caffeine (peak 1), dapsone (peak 2) and chlorzoxazone (peak 3) at intervals of 10µg/mL (upper red signal). [Color ﬁgure available online.] 274 CHANG ET AL.

of the plasma concentration–time curve on a semilogarithmic scale.

RESULTS The analytical conditions were optimized by adjusting the pH of buffer, programming the mobile phase composition, and changing the UV absorbance, in order to effectively separate the peak of the three probe drugs and increase their response in HPLC chromatogram. The probe drugs—caffeine, dapsone, and chlorzoxazone—in the plasma samples were detected simultaneously (Figure 1). All their retention times were longer than 5 min, which is necessary for avoiding the interference of plasma’s basic components. In the base experiment, the concentration of three probe drugs in plasma increased quickly to peak concentration within 1 h after single intraperitoneal injection of the probe cocktail alone, and then decreased gradually over 24 h. The mean Cmax of caffeine was 9.6 µg/mL at Tmax = 0.2 h, dapsone was 1.3 µg/mL at 0.9 h, and chlorzoxazone was 5.0 µg/mL at 0.25 h. Chlorzox- azone had not been found in plasma at 24 h, which might have been due to the relative higher detection limit for chlorzoxazone (Figure 2). To show the inductive effect of rifampicin on CYP enzymes, the probe drug concentration were examined over time for the Turbot pretreated with rifampicin for 30 min (Figure 2). The change on the concentration–time data also showed similar trend, a sharp increasing and a long decreasing phase. The mean Cmax of caffeine was 4.6 µg/mL at Tmax = 0.5 h, dapsone was 1.7 µg/mL at 0.5 h, and chlorzoxazone was 4.6 µg/mL at 0.1 h. The pharmacokinetic proﬁles of the three probe drugs were also estimated using a noncompartmental analysis, and metrics were also compared between the controlled and the induction experiments. This indicated that pretreatment with rifampicin could signiﬁcantly reduce the Cmax,AUClast, AUC of caffeine, the Tmax of dapsone; it also reduced the Tmax,AUClast,AUC, and MRT of chlorzoxazone (Table 1).

DISCUSSIONS

Downloaded by [Department Of Fisheries] at 00:21 17 November 2014 The assessment of CYP activity has been and remains an important topic in physiological, toxicological, and pharmaceutical studies. As a catalytic enzyme for endogenous and exogenous substrates, the activity of CYPs can also be affected by various factors, including genetic background, sex, and diet (Glue and Clement 1999; Pelkonen et al. 2008). Except for di- FIGURE 2. Plasma concentration versus time profile of caffeine (upper panel), rect measurement with specific antibodies, purified enzymes, dapsone (middle panel), and chlorzoxazone (below panel) in Turbot intraperitoneally injected with the probe drug cocktail alone or together with inducer or cDNA-expressed enzymes, the phenotyping of CYP activity rifampicin. with various probe drugs has also been used commonly, and a variety of catalytic substrates of CYPs have been selected to assess the activity of specific CYP isoforms (Yuan et al. 2002). AUClast was calculated using the trapezoidal method from time Furthermore, various probe drug cocktail approaches have also 0 to the real time tlast. The terminal half-life (t1/2z) was deter- been developed to measure the activity of several CYP en- mined according to the following equation: t1/2z = 0.693/λz, zymes simultaneously in order to reduce both time and cost where λz is the slope of the regression line of the terminal phase of analysis. They were found very useful for investigating the CYTOCHROME P450 ENZYMES IN TURBOT 275

TABLE 1. Summary of pharmacokinetic (PK) parameters for probe drug cocktail administered to Turbot alone or together with the inducer rifampicin. Probe + rifampicin means within rows marked with asterisks differed from the probe-alone cocktail.

Mean ± SD Probe drug PK parametersa Unit Probe alone Probe + rifampicin

Caffeine Cmax µg/mL 9.6 ± 2.3 4.6 ± 1.2* Tmax h0.2± 0.1 0.5 ± 0.5 −1 AUClast µg/mL·h 70.6 ± 12.6 37.4 ± 4.3* AUC µg/mL·h−1 92.0 ± 17.9 48.9 ± 6.5* t1/2z h12.0± 1.2 11.7 ± 1.0 MRT h 16.3 ± 1.9 16.0 ± 1.0 Dapsone Cmax µg/mL 1.3 ± 0.3 1.7 ± 0.3 Tmax h0.9± 0.1 0.5 ± 0.3* −1 AUClast µg/mL·h 6.4 ± 1.9 5.5 ± 1.2 AUC µg/mL·h−1 6.5 ± 1.8 5.7 ± 1.3 t1/2z h3.7± 1.9 5.4 ± 4.8 MRT h 5.6 ± 1.4 5.4 ± 2.2 Chlorzoxazone Cmax µg/mL 5.0 ± 0.6 4.6 ± 1.2 Tmax h0.25± 0.0 0.10 ± 0.04* −1 AUClast µg/mL·h 15.2 ± 6.0 7.6 ± 2.1* AUC µg/mL·h−1 15.5 ± 6.0 7.7 ± 2.1* t1/2z h2.3± 0.1 2.1 ± 0.4 MRT h 3.0 ± 0.2 2.5 ± 0.2*

a Abbreviations: AUC is plasma concentration–time curve and AUClast was calculated using the trapezoidal method from time 0 to the real time tlast, t1/2z is terminal half-life, T is time to reach the maximum concentration, Cmax is maximum concentration, and MRT is mean residence time.

metabolic pathways of various substrates and for predicting the ultraviolet detector, which is available in most analytical labo- drug–drug interaction as well. The “Pittsburgh cocktail,” which ratory. contains caffeine, mephenytoin, debrisoquine, chlorzoxazone, The use of caffeine, chlorzoxazone, and dapsone in cocktail and dapsone, has been validated and used to simultaneously for probing CYP1A2, CYP2E1, and CYP3A simultaneously in access activities of CYP1A2, 2C19, 2D6, 2E1, and 3A4 (Frye humans has already been widely accepted because of the nearly et al.1997; Adedoyin et al. 1998). The “Cooperstown cocktail,” ubiquitous consumption, relative safety, and favorable pharma- containing caffeine, dextromethorphan, omeprazole, and mida- cokinetic characteristics, such as complete absorption, relatively zolam has also been developed to evaluate the phenotyping of high clearance, and short half-life (Miners and Birkett 1996; CYP1A2, CYP2C19, CYP2D6, and CYP3A initially (Street- Sharma et al. 2004). The presence of CYP2E-like enzymes was man et al. 2000). Later, another typical probe drug warfarin was also proposed in Winter Flounder Pseudopleuronectes ameri- added into this cocktail to assess the activity of CYP2C9 at the canus (Wall and Crivello 1998) and in viviparous fish (Kaplan same time (Chainuvati et al. 2003), and it was successfully used et al. 2001) by measuring the hepatic microsomal metabolism to investigate a common oral contraceptive preparation on the of chlorzoxazone, but a piscine CYP2E gene ortholog has so Downloaded by [Department Of Fisheries] at 00:21 17 November 2014 activity of drug-metabolizing enzymes in humans (Shelepova far not been reported. The results of our study have shown that et al. 2005) and the drug interaction potential of aplaviroc, a these three probe drugs have good pharmacokinetic characteris- novel human immunodeficiency virus entry inhibitor (Johnson tics in Turbot as well, and are suitable for evaluating the CYPs et al. 2006). However, some of the previously reported cocktails activities in flatfish. that contain four or more probe drugs are time and labor inten- We found the plasma concentrations of caffeine and chlor- sive because the probe drugs need to be detected by different zoxazone in the elimination phase were significantly decreased methods. by the pretreatment with rifampicin, and the typical pharmacoki- Considering the lack of molecular detection method for flat- netic AUC was significantly smaller for both drugs than that of fish CYPs and the knowledge that fish share many CYP fam- cocktail-alone treatment. However, the AUC of dapsone was ilies with mammals, we recently selected caffeine, chlorzox- not significantly changed by the pretreatment with rifampicin, azone, and dapsone as probe drug and developed a cocktail although the few plasma concentration data points obtained be- to simultaneously assess CYP1, CYP2, and CYP3 in flatfish. yond 6 h decreased slightly. More importantly, all three probe drugs in the plasma sample Pretreatment of Turbot with rifampicin could induce both can be detected at the same time by a simple HPLC method with CYP1A and CYP2E or their homologous enzymes (but not 276 CHANG ET AL.

CYP3A), which in turn accelerate the elimination of caffeine CARSFC (China Modern Agriculture Research System Flatfish Center). 2010. and chlorzoxazone by increasing their metabolism. Rifampicin Annual report 2009 of national technology system for flatfish culture industry. is an antibiotic typically used to treat Mycobacterium infec- China Ocean University Press, Qingdao, China. (In Chinese). Chainuvati, S., A. N. Nafziger, J. S. Leeder, A. Gaedigk, G. L. Kearns, E. Sellers, tions such as tuberculosis. It has also been known as an inducer Y. Zhang, A. D. Kashuba, E. Rowland, and J. S. Bertino Jr. 2003. Combined for many CYP family members, including CYP2B6, CYP2C8, phenotypic assessment of cytochrome p450 1A2, 2C9, 2C19, 2D6, and 3A, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and even CYP3A4/5 N-acetyltransferase-2, and xanthine oxidase activities with the “Cooperstown in humans (Pelkonen et al. 2008). In our study, however, ri- 5 + 1 cocktail”. Clinical Pharmacology and Therapeutics 74:437–447. fampicin administration has shown remarkable inductive effect Curcio, R., R. Nicoli, S. Rudaz, and J. L. Veuthey. 2010. Evaluation of an in-capillary approach for performing quantitative cytochrome P450 activity on CYP1A-like activity in Turbot, but not CYP3A-like activ- studies. Analytical and Bioanalytical Chemistry 398:2163–2171. ity. The intrinsic differences between fish and mammals might Ekins, S., E. J. Reschly, L. R. Hagey, and M. D. Krasowski. 2008. Evolution account for this scenario. Previous studies have shown differ- of pharmacologic specificity in the pregnane X receptor. BMC Evolutionary ences even among various mammal species with regards to iso- Biology [online serial] 8:103. form composition, expression, and catalytic activities of CYP Frye, R. F., G. R. Matzke, A. Adedoyin, J. A. Porter, and R. A. Branch. 1997. Validation of the five-drug “Pittsburgh cocktail” approach for assessment of enzymes, in which CYP2E1 shows no large differences. How- selective regulation of drug-metabolizing enzymes. Clinical Pharmacology ever, the species-specific isoforms of CYP1A, −2C, −2D and and Therapeutics 62:365–376. −3A, show appreciable interspecies differences (Martignoni Glue, P., and R. P. Clement. 1999. Cytochrome P450 enzymes and drug et al. 2006). Also, Miranda et al. (1998) have evaluated the metabolism basic concepts and methods of assessment. Cellular and Molec- relative capacity of several inhibitors of mammalian CYP to ular Neurobiology 19:309–323. Guengerich, F. P. 2008. Cytochrome p450 and chemical toxicology. Chemical inhibit CYP-mediated monooxygenase activities of liver micro- Research in Toxicology 21:70–83. somes in vitro in Rainbow Trout Oncorhynchus mykiss.They Hemeryck, A., and F. M. Belpaire. 2002. Selective serotonin reuptake inhibitors found some of inhibitors such as ketoconazole, miconazole, and and cytochrome P-450 mediated drug-drug interactions: an update. Current clotrimazole were still potent inhibitor of trout CYPs, but oth- Drug Metabolism 3:13–37. ers were inactive. Pharmaceuticals induce the transcription of Johnson, B. M., I. H. Song, K. K. Adkison, J. Borland, L. Fang, Y. Lou, M. M. Berrey, A. N. Nafziger, S. C. Piscitelli, J. S. Bertino Jr. 2006. Evaluation of specific CYP genes in mammals mainly via a suite of nuclear the drug interaction potential of aplaviroc, a novel human immunodeficiency receptors, in which pregnane X receptor (PXR) is responsible virus entry inhibitor, using a modified cooperstown 5 + 1 cocktail. Journal of for CYP3 induction. All these receptors, with the exception of Clinical Pharmacology 46:577–587. constitutive and rostane receptor (CAR), have been identified Kaplan, L. A., E. Fielding, and J. F. Crivello. 2001. The genetic regulation in fish and are assumed to regulate the piscine CYPs in similar of liver microsomal CYP2E1 activity among strains of the viviparous fish Poeciliopsis. Comparative Biochemistry and Physiology Part C Toxicology pathways as well (Smith and Wilson, 2010). However, study and Pharmacology 128:143–152. on Zebrafish Danio rerio PXR has revealed that it has a smaller Martignoni, M., G. M. Groothuis, and R. de Kanter. 2006. Species differ- ligand-binding pocket than other mammalian PXRs (Ekins et al. ences between mouse, rat, dog, monkey and human CYP-mediated drug 2008) and cannot be activated by rifampicin. Together with the metabolism, inhibition and induction. Expert Opinion on Drug Metabolism results from our study, it is worth noting that rifampicin does and Toxicology 2:875–894. Miners, J. O., and D. J. Birkett, 1996. The use of caffeine as a metabolic probe not appear to be a ligand for the fish PXR. for human drug metabolizing enzymes. General Pharmacology 27:245–249. Taken together, it is feasible to simultaneously evaluate the Miranda, C. L., M. C. Henderson, and D. R. Buhler. 1998. Evaluation of chem- catalytic activity of CYP in Turbot using the probe drug cock- icals as inhibitors of trout cytochrome P450s. Toxicology and Applied Phar- tail. However, further studies on identification of specific CYP macology 148:237–244. isoforms in flatfish, which are responsible for the probe drug Nelson, D. R., 2003. Comparison of P450s from human and fugu: 420 million years of vertebrate P450 evolution. Archives of Biochemistry and Biophysics metabolism are still needed. 409:18–24. Downloaded by [Department Of Fisheries] at 00:21 17 November 2014 Nelson, D. R., L. Koymans, T. Kamataki, J. J. Stegeman, R. Feyereisen, D. J. Waxman, M. R. Waterman, O. Gotoh, M. J. Coon, F. P. Guengerich, R. W. ACKNOWLEDGMENTS Estabrook, I. C. Gunsalus, F. J. Gonzales, and D. W. Nebert. 1996. P450 The work was financially supported by the National Natural superfamily: update on new sequences, gene mapping, accession numbers, Science Fund (31101298), the Special Scientific Research Funds and nomenclature. Pharmacogenetics 6:1–42. for Central Non-Profit Institutes, Chinese Academy of Fishery Pelkonen, O., M. Turpeinen, J. Hakkola, P. Honkakoski, J. Hukkanen, and H. Raunio. 2008. Inhibition and induction of human cytochrome P450 enzymes: Sciences (2014A09XK02), and the Special Scientific Research current status. Archives of Toxicology 82:667–715. Funds for Central Non-Profit Institutes, Yellow Sea Fisheries Sharma, A., S. Pilote, P. M. Belanger,´ M. Arsenault, and B. A. Hamelin. 2004. Research Institute (20603022011016). A convenient five-drug cocktail for the assessment of major drug metabo- lizingenzymes: a pilot study. British Journal of Clinical Pharmacology 58: 288–297. REFERENCES Shelepova, T., A. N. Nafziger, J. Victory, A. D. Kashuba, E. Rowland, Y.Zhang, Adedoyin, A., R. F. Frye, K. Mauro, and R. A. Branch. 1998. Chloroquine E. Sellers, G. Kearns, J. S. Leeder, A. Gaedigk, and J. S. Bertino Jr. 2005. modulation of specific metabolizing enzymes activities: investigation with Effect of a triphasic oral contraceptive on drug-metabolizing enzyme activity selective five drug cocktail. British Journal of Clinical Pharmacology 46: as measured by the validated Cooperstown 5 + 1 cocktail. Journal of Clinical 215–219. Pharmacology 45:1413–1421. CYTOCHROME P450 ENZYMES IN TURBOT 277

Smith, E. M., and J. Y. Wilson. 2010. Assessment of cytochrome P450 Wall, K. L., and J. Crivello. 1998. Chlorzoxazone metabolism by Winter Floun- fluorometric substrates with Rainbow Trout and killifish exposed to der liver microsomes: evidence for existence of a CYP2E1-like isoform in dexamethasone, pregnenolone-16alpha-carbonitrile, rifampicin, and beta- teleosts. Toxicology and Applied Pharmacology 151:98–104. naphthoflavone. Aquatic Toxicology 97:324–333. Werck-Reichhart, D., and R. Feyereisen. 2000. Cytochromes P450: a success Streetman, D. S., J. F. Bleakley, J. S. Kim, A. N. Nafziger, J. S. Leeder, A. story. Genome Biology 1:3003. 1–3003.9. Gaedigk, R. Gotschall, G. L. Kearns, and J. S. Bertino Jr. 2000. Com- Yuan, R., S. Madani, X. X. Wei, K. Reynolds, and S. M. Huang. 2002. Evaluation bined phenotypic assessment of CYP1A2, CYP2C19, CYP2D6, CYP3A, of cytochrome P450 probe substrates commonly used by the pharmaceutical N-acetyltransferase-2, and xanthine oxidase with the “Cooperstown cock- industry to study in vitro drug interactions. Drug Metabolism and Disposition tail”. Clinical Pharmacology and Therapeutics 68:375–383. 30:1311–1319. Downloaded by [Department Of Fisheries] at 00:21 17 November 2014 This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:22 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Bioaccumulation of Heavy Metals in Blood and Tissue of Striped Mullet in Two Italian Lakes Francesco Fazioa, Giuseppe Piccionea, Katia Tribulatob, Vincenzo Ferrantellic, Giuseppe Giangrossoa, Francesca Arfusoa & Caterina Faggiod a Department of Veterinary Science, Polo Universitario dell'Annunziata, University of Messina, 98168 Messina, Italy b Regional Agency for the Protection of the Environment–Sicily, Department of Messina, Italy c Experimental Zooprophylaxy Institute of Sicily, 90100 Palermo, Italy d Department of Biological and Environmental Science, University of Messina, Viale Ferdinando Stagno d’Alcontres 31, 98166 Messina, Italy Published online: 04 Nov 2014.

To cite this article: Francesco Fazio, Giuseppe Piccione, Katia Tribulato, Vincenzo Ferrantelli, Giuseppe Giangrosso, Francesca Arfuso & Caterina Faggio (2014) Bioaccumulation of Heavy Metals in Blood and Tissue of Striped Mullet in Two Italian Lakes, Journal of Aquatic Animal Health, 26:4, 278-284, DOI: 10.1080/08997659.2014.938872 To link to this article: http://dx.doi.org/10.1080/08997659.2014.938872

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:278–284, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.938872

ARTICLE

Bioaccumulation of Heavy Metals in Blood and Tissue of Striped Mullet in Two Italian Lakes

Francesco Fazio and Giuseppe Piccione* Department of Veterinary Science, Polo Universitario dell’Annunziata, University of Messina, 98168 Messina, Italy Katia Tribulato Regional Agency for the Protection of the Environment–Sicily, Department of Messina, Italy Vincenzo Ferrantelli Experimental Zooprophylaxy Institute of Sicily, 90100 Palermo, Italy Giuseppe Giangrosso and Francesca Arfuso Department of Veterinary Science, Polo Universitario dell’Annunziata, University of Messina, 98168 Messina, Italy Caterina Faggio Department of Biological and Environmental Science, University of Messina, Viale Ferdinando Stagno d’Alcontres 31, 98166 Messina, Italy

Abstract Water and sediment from Faro and Ganzirri lakes were tested for cadmium, mercury, and lead. These heavy metals have been assessed in serum and muscles of Striped Mullet Mugil cephalus. A haemogram was performed to find effects of heavy metals on haematological variables. Student’s t-test showed higher Pb levels in sediments of Ganzirri Lake than Faro Lake. Two-way analysis of variance showed higher Cd and Pb concentration in muscles samples of Striped Mullet from Ganzirri Lake than from Faro Lake. Significant differences were found among haematological parameters as well. Mullet from Ganzirri Lake had lower white and red blood cell and trombocyte counts and higher mean corpuscular haemoglobin and mean corpuscular haemoglobin concentration than those from Faro Lake. Our results indicate a significant presence of lead in the environmental produces an accumulation of this metal in fish muscles. Downloaded by [Department Of Fisheries] at 00:22 17 November 2014

Faro and Ganzirri lakes, situated on the north and northeast- icity and bioaccumulation in the food chain (Gochfeld 2003). ern of Messina, Italy, are particularly vulnerable to the accumu- Natural and anthropogenic sources including industrial and do- lation of toxic substances dangerous to aquatic organisms and mestic sewage, harbor activities, and atmospheric deposits, rep- their consumers (Licata et al. 2003, 2004). This is due to the resent the sources of heavy metals in aquatic ecosystem (Nair input of pollutants from urban and agricultural sources as well et al. 2006). The concentrations of heavy metals accumulated in as low water exchange rates. living organisms reﬂects the degree of environmental pollution During the last years pollution by heavy metals in aquatic (Canbek et al. 2007). Among aquatic species, ﬁsh are a suit- environment gained great importance because of the risk of tox- able biomarker of environmental pollution (Padmini and Usha

*Corresponding author: [email protected] Received April 2, 2014; accepted June 16, 2014

278 BIOACCUMULATION OF HEAVY METALS IN MULLET 279

Rani 2008) because of their position in the aquatic food chain caught with bottom set nets from three stations of Faro Lake (Mansour and Sidky 2002). The digestive tract and the gill sur- (F1, F2, and F3) and Ganzirri Lake (G1, G2 and G3) selected face represent the main two input of metal in fish (Ptashynski randomly (Figure 1). Distances between stations were 200 m in et al. 2002). Among metals, cadmium, mercury, and lead are Faro Lake and 500 m in Ganzirri Lake. After capture, all Striped extremely toxic (Ayed et al. 2011; Squadrone et al. 2013). Mullet considered as healthy (i.e., no external signs of abnor- Because haematological variables are closely dependent on malities or infestation) were immediately transferred to a tank environmental changes, they are considered valuable indicators and anaesthetized (MS-222 at 0.3 g/L of water) prior to blood of metal pollution, particularly, in aquatic environment. These sampling blood collected by venipuncture of the caudal vein variables are valuable tools for monitoring stress response and via a sterile plastic syringe (2.5 mL) and transferred into two predicting systemic relationships and physiological adaptations different tubes. One tube (Miniplast 0.6 mL; LP Italiana Spa, of animals (Danabas et al. 2010). The Striped Mullet Mugil Milano) contained 1.26 mg of ethylenediamine tetraacetic acid cephalus is a sentinel species that feeds on benthic invertebrates (EDTA) as an anticoagulant agent; that sample was used for the and decaying material. Their long gut length possibly increases determination of haematological profile. The other tube (MM absorption efficiency of contaminants ingested (Odum 1970). Conical Test Tubes, 4.5 mL), had no EDTA; it was centrifuged Striped Mullet significantly accumulate heavy metals and other for 10 min at 3000 revolutions/min, and the serum obtained was pollutants and are therefore a suitable indicator of its environ- stored at −80◦C until analyzed for Cd, Pb, and Hg. After blood ment (Ferreira et al. 2005). Previous studies carried out in Faro was sampled all fish were brought to the laboratory on the same and Ganzirri lakes showed heavy metals in water and sediment day, where they were weighed and measured (FL), and condi- samples and in aquatic organisms (Scoglio et al. 2000; Licata tion factors were calculated: (weight × 100)/FL3. On the basis et al. 2003; Licata et al. 2004).The objectives of this study were of their weight and length, all fish, of both groups, were consid- to evaluate the presence of heavy metals (Cd, Hg and Pb) in ered sexually mature and ages 2–4 (McDonough et al. 2005). water and sediment from Faro and Ganzirri lakes, their accumu- At the end of measurements, the fish were deeply anaesthetized lation in muscles and serum of Striped Mullet, and the influence (MS-222 at 0.7 g/L of water) and sacrificed by a sharp blow to of these heavy metals on their haematological variables. the head. All fish were stored at −80◦C. Haematological analysis.—All haematological samples were analyzed in triplicate by the same technician, and no METHODS significant difference among the three measures was found. Fish sampling and analysis.—We examined 20 Striped The blood samples collected in EDTA tubes were used for Mullet from Faro and Ganzirri lakes in December 2012. They the determination of haematological profile using an automated were divided into two equal groups: Group A (10 from Faro haematology analyzer (HeCo Vet C, SEAC, Florence, Italy) Lake) and Group B (10 from Ganzirri Lake). All fish were with special lysing reagent for fish (SEAC, Code 71010460), Downloaded by [Department Of Fisheries] at 00:22 17 November 2014

FIGURE 1. Location of Faro and Ganzirri lakes in Sicily and the three sampling stations in each lake. 280 FAZIO ET AL.

previously used to investigate haematological profile in Striped rial, analytes were first solubilized by gentle refluxing with nitric Mullet (Fazio et al. 2013) and other fish species (Fazio et al. and hydrochloric acids. After cooling, the sample was made up 2012a, 2012b). Evaluation of the haemogram involved the de- to volume and mixed and allowed to settle overnight prior to termination of the white blood cell (WBC) and red blood cell analysis. According to the method describes the multi-element (RBC) counts, haematocrit (Hct), hemoglobin concentration determination of trace elements by ICP-MS (EPA 200.8.1994), (Hgb), mean corpuscular volume (MCV), mean corpuscular the sample is made ready for analysis by the appropriate addi- haemoglobin (MCH), mean corpuscular haemoglobin concen- tion of nitric acid, and then diluted to a predetermined volume tration (MCHC), and thrombocyte count (TC). To evaluate the and mixed before analysis. Sample material in solution is in- reliability of the automatic haematological analysis, the evalu- troduced by pneumatic nebulization into radiofrequency plasma ation of WBC, RBC, and TC were carried out with both man- where energy transfer processes cause desolvation, atomization, ual and automated methods, and no statistical differences were and ionization. The ions are extracted from the plasma through evident. The reliability of automatic haematological analysis a differentially pumped vacuum interface and separated on the has been demonstrated as for Gilthead Seabream Sparus aurata basis of their mass : charge ratio by a quadrupole mass spec- (also known as Gilthead Bream) (Fazio et al. 2012b) and Striped trometer having a minimum resolution capability of 1 atomic Mullet (Faggio et al. 2013). mass unit (amu) peak width at 5% peak height. An electron mul- Serum and muscle heavy metals analysis.—The determina- tiplier or Faraday detector and the ion information processed by tion of heavy metals in fish sample (serum and muscle) was a data handling system detect the ions. Interferences relating performed by inductively coupled plasma - mass spectrometry to the technique must be recognized and corrected for. Such (ICP-MS). For this analysis 2 mL of serum and a representa- corrections must include compensation for isobaric elemental tive a 0.5-g portion of muscle from each fish was collected, as interferences and interferences from polyatomic ions derived weighed by analytical balance (precision of 0.1 mg or 0.1 mL). from the plasma gas, reagents or sample matrix. Instrumental To each sample, 5 mL of nitric acid at 65% (Suprapur) and 1 mL drift as well as suppressions or enhancements of instrument re- of hydrochloric acid at 37% (Suprapur) were added and then sponse caused by the sample matrix must be corrected for by mineralization in the Multiwave 3000 (Anton Paar) microwave the use of internal standards. oven. The obtained solution was filled up to the volume of 50 mL Statistical analysis.—Prior to statistical analysis, all data with Milli-Q water and processed using ICP-MS 7700X (Agi- were tested for normality of distribution using the Kolmogorov– lent). For the quantitative determination two calibration curves Smirnov test at α = 0.05. All data were normally distributed and with 5 points for each element was built (r2 > 0.9800). The con- statistical analysis was performed on mean values. centration of the analyte (Cc; i.e., Pb, Ca, or Hg in mg/kg) was Unpaired Student’s t-test was used to determine significant calculated: (α = 0.05) differences in biometric and haematological variables of groups A and B and to determine significant differ- C = (C × V/1000 × p × R¯ )D; c s ences in chemico-physical water variables and heavy metals content in water and sediment of Faro and Ganzirri lakes, and C = concentration of the analyte in the solution (mg/L) s 95% confidence intervals were calculated. A two-way ANOVA calculated in the linear fit, was used to determine significant differences of Cd, Hg, and V = volume (mL) of the sample solution, Pb concentrations in muscle and serum of group A versus B. P = sample weight (g), Bonferroni’s multiple comparisons test was used for post hoc D = dilution factor, and comparisons. All statistical analyses were performed using the R¯ = average recovery, statistical software program Prism v. 5.00 (Graphpad Software

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 To check the instrumental performance, a 200-µg/L Erbium Ldt., California). solution was used as internal standard, in a confidence range between 70% and 120% for the recovery. RESULTS Water analysis.—For both sites of collection we measured Unpaired t-tests showed no statistical significant difference chemical, physical variables and water and sediments content in mean weight, FL, and condition factors for groups A versus of Cd, Hg, and Pb. Water sampling at each lake and station was B (Table 1). Unpaired t-tests showed no significant statistical carried out on the same day as fish sampling at three randomly differences in temperature, salinity, pH, and dissolved oxygen selected sampling stations at each location. Distances between stations were about 3 m. Niskin bottle (General Oceanics, Inc., ± Miami, Florida) were used for sampling, and a multiparamet- TABLE 1. Mean SD biometric data for Striped Mullet from Faro Lake (Group A) and from Ganzirri Lake (Group B). ric probe YSI 85 system was used for temperature, salinity, and pH. Water and sediment samples were screened for evaluation of Biometric variables Group A (n = 10) Group B (n = 10) heavy metals by means of ICP-MS. An aliquot of a well-mixed ± ± and homogeneous aqueous sample and solid was accurately FL (cm) 29.45 4.20 31.32 5.93 ± ± measured for sample processing. For total recoverable analysis Weight (g) 194.91 7.28 198.11 10.75 ± ± of an aqueous and solid sample containing undissolved mate- Condition factor 0.88 0.56 0.76 0.36 BIOACCUMULATION OF HEAVY METALS IN MULLET 281

TABLE 2. Chemo-physical water quality metrics for the two site of Striped Mullet collection sites (Faro and Ganzirri lakes).

Faro Lake Ganzirri Lake

25th–75th 25th–75th Water Mean Mean percentile Mean Mean percentile variables ± SD Median ± 95% CI range ± SD Median ± 95% CI range Temperature 12.44 ± 0.17 12.40 12.34 ± 12.55 12.38 ± 12.43 12.23 ± 0.44 12.16 11.96 ± 12.50 12.10 ±−12.41 (◦C) Salinity (‰) 34.21 ± 2.53 35.76 32.68 ± 35.73 31.41 ± 35.77 32.79 ± 1.95 32.02 31.61 ± 33.97 32.02 ± 32.53 pH 6.95 ± 0.01 6.94 6.94 ± 6.95 6.94 ± 6.95 7.15 ± 0.55 7.31 6.82 ± 7.49 7.30 ± 7.31 Dissolved 10.81 ± 1.17 10.37 10.10 ± 11.51 10.36 ± 10.38 11.50 ± 0.64 11.71 11.12 ± 11.89 11.63 ± 11.73 oxygen (mg/L)

between Faro Lake and Ganzirri Lake (Table 2) and no sig- found in sediment and in fish from Ganzirri Lake. In particu- nificant differences in sediment Pb levels between Faro and lar, Cd and Pb muscle content were higher in group B than A, Ganzirri lakes (Table 3). Two-way ANOVA showed statistical whereas no statistical differences were found in serum heavy significant differences in Cd and Pb in muscle samples form the metals. Moreover, no significant difference between serum and two groups, Group B bring higher than Group A. No significant muscle heavy metal contents in both groups was found; how- difference between serum and muscle heavy metals content was ever, serum Cd levels were higher than muscle Cd levels in both found in Group B (Figure 2). Unpaired t-test showed significant groups, and serum Pb levels were higher than muscle Pb levels statistical differences in some haematological variables between in group A. Higher levels of these metals in the serum sam- groups A and B. In particular, statistical analysis showed lower ples than in muscle samples might indicate that fish have been WBC, RBC, and TC values and higher MCH and MCHC values recently exposed to these metals and that an acute response oc- in group B than in group A (Table 4). curred. Effectively, after the absorption, metals in fish are then transported through bloodstream to the organs and tissues where they are accumulated (Adeyemo et al. 2010). However, the re- DISCUSSION lationship between exposure duration and biological effect can We found no significant difference in temperature, salin- be complicated (Gordon et al. 2012). Effectively, metals accu- ity, pH and dissolved oxygen between Faro and Ganzirri lakes. mulation is influenced by both physiological (such as ingestion Heavy metals concentration in water from Faro and Ganzirri rate and gut passage time) and biochemical (metal partitioning lakes were relatively low compared with water quality stan- and subcellular metal distribution) factors. Biokinetic studies dards designed by decree (Italian Ministry of the Environment on heavy metals assimilation in teleosts provided that trophic 2010) and to previous findings for these lakes (Scoglio et al. transfer was more important than water exposure in contributing 2000; Ferrando et al. 2005). Heavy metals measured in the to the overall metals accumulation. In addition, the significance sediment sampled from both sites were higher than sediment of water chemistry effects, such as dissolved organic carbon quality standards designed by decree (Italian Ministry of the and pH, on heavy metals uptake in fish should not be ignored. Environment 2010). This may be due to the significant urban- These factors may explain the great variation in metals accu-

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 ization around Faro and Ganzirri lakes since it is known that mulation by ﬁsh in various natural systems (Wang et al. 2010). these heavy metals are directly connected to human impact (Li- Biometric indices represent another important factor affecting cata et al. 2003). The highest heavy metal concentrations were heavy metal absorption and accumulation. While it is known

TABLE 3. Mean heavy metal metrics measured in water and sediment of Faro and Ganzirri lakes. Water and sediment standards are by decree (Italian Ministry of the Environment 2010).

Water Sediment Faro Lake Ganzirri Lake Faro Lake Ganzirri Lake Heavy Metals (mean) (mean) Decree standard (mean ± SD) (mean ± SD) Decree standard Cadmium (µg/L) <0.10 <1 5 0.87 ± 0.01 0.88 ± 0.01 0.3 Mercury (µg/L) <0.10 <0.1 1 0.56 ± 0.01 0.71 ± 0.12 0.3 Lead (µg/L) <0.10 <1 10 161.3 ± 1.07a 218.1 ± 1.95a 30

aFaro and Ganzirri lakes differed signiﬁcantly in bottom sediment lead (P < 0.001). 282 FAZIO ET AL.

FIGURE 2. Mean + SD heavy metal (Cd, Hg and Pb) concentrations recorded in muscle and serum samples of Striped Mullet from Faro and Ganzirri lakes. Signiﬁcant differences between the two lakes for a given metal are noted with an asterisk.

TABLE 4. Mean haematological variables studied in Striped Mullet from Faro (Group A) and Ganzirri (Group B) lakes. Where values between the two groups differed signiﬁcantly the P-value is shown; NS = not signiﬁcant. Variables include white blood cell (WBC) and red blood cell (RBC) counts, haematocrit

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 (Hct), hemoglobin concentration (Hgb), mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC), and thrombocyte count (TC).

Mean ± SD (n = 10) Haematological variable Group A Group B Signiﬁcant difference WBC ( × 103/µL) 24.24 ± 9.85 12.62 ± 4.01 P < 0.001 RBC ( × 106/µL) 3.65 ± 0.77 2.95 ± 0.54 P < 0.05 Hct (%) 44.39 ± 9.24 41.83 ± 4.42 NS Hgb (g/dL) 11.58 ± 1.98 10.28 ± 1.74 NS MCV (fL) 121.8 ± 8.40 146.6 ± 43.01 NS MCH (pg) 32.03 ± 2.05 35.93 ± 5.31 P < 0.05 MCHC (g/dL) 320.3 ± 20.51 359.3 ± 53.11 P < 0.05 TC ( × 103/µL) 38.78 ± 11.83 28.38 ± 5.67 P < 0.01 BIOACCUMULATION OF HEAVY METALS IN MULLET 283

that fish’s length and weight could influence the amount of The mechanism of Pb toxicity occurs by ion-regulatory dis- heavy metals accumulated in tissues (Zyadah 1999; Blahova´ ruptions with little changes in the haematological parameters. et al. 2010), the fish we used were homogeneus for both weight Shah and Altindag (2005) reported significant increase in im- and length. Thus, a linkage between heavy metals concentration munological metrics following Pb exposure, which suggests that in the sediment and heavy metals muscle content in group B Pb may weaken the immune system, resulting in increased sus- may be due to metals intake through trophic chain. The sed- ceptibility to infections. However, although the stimulation of iment constitutes the most important reservoir of metals and the immune system causes an increase in lymphocytes by an in- other toxicants in aquatic environment. Fish feeding on benthic jury or tissue damage, a prolonged or continuous stimulus may organism, as do Striped Mullet, are directly exposed to con- cause the exhaustion of this capacity, resulting in a decrease in taminated sediments (Zrnciˇ c´ et al. 2013). Heavy metals content lymphocytic and subsequently in total WBC count (Adeyemo in Striped Mullet muscle differed from the findings on Golden et al. 2010). Moreover, the decrease in WBC levels may be the Gray Mullet Liza aurata sampled from Ganzirri Lake (Licata result of bioaccumulation of the tested heavy metals in kidney et al. 2003) and from the findings on Striped Mullet from Lig- and liver (Kori-Siakpere et al. 2006). This together with the urian Sea, Italy (Squadrone et al. 2013). This difference might higher concentration of heavy metals found in muscle samples be due to the different period of investigation and to the different than in serum samples could be indicative of a chronic metal habitat, respectively. Heavy metals measured in serum and mus- accumulation in group B. cles samples of groups A and B were within the limits set by the European Commission (2001, 2006), except for Pb, which ex- CONCLUSION ceeded the maximum limit. Muscle is the main tissue consumed Our results showed that Striped Mullet are directly exposed by people; thus, it arouses great interest in routine monitoring of to trace metal concentrations as results of feeding and the in- heavy metal contamination (Sthanadar et al. 2013; Tanee et al. gestion of contaminated sediment and detritus. The changes in 2013). However, this tissue usually has a lower accumulation the studied hematological variables indicated that they could be potential of heavy metals than tissues such as liver and kidney used as indicators of metal pollution conditions in fish. Further (Licata et al. 2003). Therefore, further research on heavy metals investigations are necessary to evaluate heavy metals in other evaluation in parenchymatous organs (liver and kidney) would tissues, such as parenchymatous organs (liver and kidney), and be useful. Fish exposed to environmental pollutants exhibit a va- in hematotological variables to obtain a useful complete pro- riety of physiological responses, among which haematological file and facilitate use this fish species as possible biomarker of profile variation is the most timely and representative (Danabas pollution. et al. 2010). Change in haematological variables are frequently evaluated in clinical diagnoses of fish physiology to determine the effects of external stressors. Fish live in closely contact with ACKNOWLEDGMENTS their environment, and the strong association between a fish’s The authors disclaim any financial support or relationships circulatory system and the external environment makes haema- that may pose conflicts of interest. The authors would like to tological variables useful and reliable indicators of stress and thank the Provincia of Messina for allowing our sampling. Pro- toxic substance presence, including metals (Kori-Siakpere et al. tocols of fish and experimentation were reviewed and approved 2006). It is well known that fish haematological variation in in accordance with the standards recommended by the Guide for response to metals exposure occurs following osmotic changes the Care and Use of Laboratory Animals and Directive 63/2010 that cause hemodilution or hemoconcentrations (Tort and Tor- CEE. res 1988). We found statistical significant differences in some

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 haematological metrics between groups A and B. In particular RBC, WBC, and TC values were statistically lower in group B REFERENCES Adeyemo, O. K., O. B. Adedeji, and C. C. Offor. 2010. Blood lead level as than A, and MCH and MCHC values were statistically higher biomarker of environmental lead pollution in feral and cultured African catﬁsh in group B. Lower RBC, WBC, and TC values in group B may (Clarias gariepinus). Nigerian Veterinary Journal 31:139–147. be due to the toxic effect of heavy metals and, in particular, of Ayed, N., E. Faure, J. Quignard, and M. Trabelsio.` 2011. Determination of P, Cd and Pb, which in muscle samples was statistically higher in Ca, Zn, Cd and Pb concentrations in muscle, gills, liver, gonads and skeletons group B. The RBC in teleosts are produced from hematopoietic of two natural populations of Atherina Lagunae in north Tunis Lake, Tunisia. Journal of Water Resource Protection 3:421–428. tissues of kidney and spleen; decreased values of these cells Blahova,´ J., M. Havelkova,´ K. Kruz´ıkova,´ K. Hilscherova,´ R. Halouzka, H. could result from internal bleeding by damage to the kidney Modra,´ R. Grabic, J. Hal´ırova,´ J. Jurc´ıkova,´ T. Ocelka, D. Harustiakova,´ (Heath 1995; Kori-Siakpere et al. 2006). In addition, it is known and Z. Svobodova.´ 2010. Assessment of contamination of the Svitava and that damaged gills and impaired osmoregulation occur during Svratka rivers in the Czech Republic using selected biochemical markers. Cd exposure and cause haemodilution leading to a decrease in Environmental Toxicology and Chemistry 29:541–549. Canbek, M., A. D. Temir, U. Mustafa, B. Gokhan, E. Ozgur, and A. Naime. RBC (Kori-Siakpere et al. 2006). Though histopathology of the 2007. Preliminary assessment of heavy metals in water and some cyprinidae gills was not investigated in this study, Cd has been implicated species from the Porsuk River. Turkey Journal of Applied Biological Sciences as responsible for gill lesions (Gardner and Yevich 1970). 1:91–95. 284 FAZIO ET AL.

Danabas, D., N. C. Yildirim, A. K. Gulec, N. Yildirim, and O. Kaplan. 2010. Mytilus galloprovincialis collected from Lake Faro (Sicily, Italy). Environ- An investigation on some haematological and biochemical parameters in mental International 30:805–810. Capoeta trutta (Heckel 1843) from Munzur River (Tunceli, Turkey). Journal Mansour, S. A., and M. M. Sidky. 2002. Ecotoxicological studies. 3. Heavy of Animal Veterinary Advances 9:2578–2582. metals contaminating water and fish from Fayoum Governorate, Egypt. Food Italian Ministry of the Environment. 2010. Ministerial Decree 260 (2010): min- Chemistry 78:15–22. istry of the environment and territorial protection and sea. Italian Ministry of McDonough, C. J., W. A. Roumillat, and C. A. Wenner. 2005. Sexual differen- the Environment, Rome. tiation and gonad development in Striped Mullet (Mugil cephalus L.) from European Commission. 2001. Regulation 420/2001: setting maximum levels for South Carolina estuaries. U.S. National Marine Fisheries Service Fishery certain contaminants in foodstuffs. European Commission, Brussels. Bulletin 103:601–619. European Commission. 2006. Regulation 1881/2006: setting maximum levels Nair, M., K. V. Jayalakshmy, K. K. Balachandran, and T. Joseph. 2006. Bioac- for certain contaminants in foodstuffs. European Commission, Brussels. cumulation of toxic metals by fish in a semi-enclosed tropical ecosystem. Faggio, C., S. Casella, F. Arfuso, S. Marafioti, G. Piccione, and F. Fazio. Environmental Forensics 7:197–206. 2013. Effect of storage time on haematological parameters in mullet, Mugil Odum, W. E. 1970. Utilization of the direct grazing and plant detritus cephalus. Cell Biochemistry and Function 31:412–416. food chains by the Striped Mullet Mugil cephalus. Page 222–240 in Fazio, F., C. Faggio, S. Marafioti, A. Torre, M. Sanfilippo, and G. Piccione. A. Berkeley and J. J. Steele, editors. Marine food chains. University of 2012a. Comparative study of haematological profile on Gobius niger in two California. different habitat sites, Faro Lake and Tyrrhenian Sea. Cahiers de Biologie Padmini, E., and M. Usha Rani. 2008. Impact of seasonal variation on HSP70 ex- Marine 53:213–219. pression quantitated in stressed fish hepatocytes. Comparative Biochemistry Fazio, F., F. Filiciotto, S. Marafioti, V. Di Stefano, A. Assenza, F. Placenti, G. Part B 151:278–285. Buscaino, G. Picione, and S. Mazzola. 2012b. Automatic analysis to assess Ptashynski,M.D.,R.M.Pedlar,R.E.Evans,C.L.Baron,andJ.F.Klaverkamp. haematological parameters in farmed Gilthead Sea Bream (Sparus aurata 2002. Toxicology of dietary nickel in Lake Whitefish (Coregonus clu- Linneaus,1758). Marine and Freshwater Behaviour and Physiology 45:63– peaformis). Aquatic Toxicology 58:229–247. 73. Scoglio, M. E., F. Munao,` A. Di Pietro, and I. Picerno. 2000. Elementi in traccia Fazio, F., S. Marafioti, A. Torre, M. Sanfilippo, M. Panzera, and C. Faggio. nel complesso lagunare Ganzirri-Faro. Rischio sanitario. [Trace elements in 2013. Haematological and serum protein profiles of Mugil cephalus, effect Ganzirri and Faro lakes.] Igiene Moderna 114:243–260. of two different habitat. Ichthyological Research 60:36–42. Shah, S. L., and A. Altindag. 2005. Alterations in the immunological parameters Ferrando, S., T. Ferrando, L. Girosi, A. Mauceri, S. Fasulo, and G. Tagliafierro, of Tench (Tinca tinca L. 1758) after acute and chronic exposure to lethal and 2005. Apoptosis, cell proliferation and serotonin immunoreactivity in gut sublethal treatments with Mercury, Cadmium and Lead. Turkish Journal of of Liza aurata from natural heavy metal polluted environments, preliminary Veterinary Animal Science 29:1163–1168. observations. European Journal of Histochemistry 49:331–340. Squadrone, S., M. Prearo, S. Gavinelli, M. Pellegrino, R. Tarasco, A. Benedetto, Ferreira, M., P. Moradas-Ferreira, and M. A. Reis-Henriques. 2005. Oxidative and M. C. Abete. 2013. Heavy metals in Mugil cephalus (Mugilidae) from stress biomarkers in two resident species, mullet (Mugil cephalus) and floun- the Ligurian Sea (north-west Mediterranean, Italy). Food Additives and Con- der (Platichthys flesus), from a polluted site in River Douro Estuary, Portugal. taminants Part B 6:134–138. Aquatic Toxicology 71:39–48. Sthanadar, I. A., A. A. Sthanadar, M. Shah, P. Asmat, M. Yousaf, A. Muham- Gardner, G. R., and P.P.Yevich. 1970. Histological and haematological response mad, and Z. Muhammad. 2013. White muscle as a bio-indicator of cadmium of an estuarine fish to cadmium. Journal of the Fisheries Research Board of (Cd) pollution across Kalpani River Mardan, Khyber Pakhtunkhwa Pakistan. Canada 27:2185–2196. International Journal of Biosciences 3:105–116. Gochfeld, M. 2003. Cases of mercury exposure, bioavailability, and adsorption. Tanee, T., A. Chaveerach, C. Narong, M. Pimjai, P. Punsombut, and R. Sud- Ecotoxicology and Environmental Safety 56:174–179. moon. 2013. Bioaccumulation of heavy metals in fish from Chi River, Gordon, A. K., S. K. Mantel, and N. W. J. Muller. 2012. Review of toxicological Maha Sarakham Province, Thailand. International Journal of Biosciences 3: effects caused by episodic stressors exposure. Environmental Toxicology and 159–167. Chemistry 31:1169–1174. Tort, T., and P. Torres. 1988. The effects of sublethel concentrations of cadmium Heath, A. G. 1995. Water pollution and fish physiology, 2nd edition. Lewis on hematological parameters in the dogfish, Scyliorhinus canicula. Journal Publishers, CRC Press, Boca Raton, Florida. of Fish Biology 32:277–282. Kori-Siakpere, O., J. E. G. Ake, and U. M. Avworo. 2006. Sublethal effects Wang, R., M. Wong, and W. Wang. 2010. Mercury exposure in the of cadmium on some selected haematological parameters of heteroclarias (a freshwater tilapia Oreochromis niloticus. Environmental Pollution 158: hybrid of Heterobranchus bidorsalis and Clarias gariepinus). International 2694–2701. Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 Journal of Zoological Research 2:77–83. Zrnciˇ c,´ S., D. Oraic,´ M. Caleta,´ Z. MIhaljevic,´ D. Zanella, and N. Bilandziˇ c.´ Licata, P., G. Di Bella, G. Dugo, and F. Naccari. 2003. Organochlorine pesticides, 2013. Biomonitoring of heavy metals in fish from the Danube River. Envi- PCBs and heavy metals in tissues of the mullet Liza aurata in Lake Ganzirri ronmental Monitoring and Assessment 185:1189–1198. and Straits of Messina (Sicily, Italy). Chemosphere 52:231–238. Zyadah, M. A. 1999. Accumulation of some heavy metals in Tilapia zil- Licata, P., D. Trombetta, M. Cristani, D. Martino, and F. Naccari. 2004. lii organs from Lake Manzalah, Egypt. Turkish Journal of Zoology 23: Organochlorine compaunds and heavy metals in the soft tissue of the mussel 365–372. This article was downloaded by: [Department Of Fisheries] On: 17 November 2014, At: 00:22 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Aquatic Animal Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uahh20 Dose-Dependent Effects of the Clinical Anesthetic Isoflurane on Octopus vulgaris: A Contribution to Cephalopod Welfare Gianluca Polesea, William Winlowb & Anna Di Cosmoa a Department of Biology, University of Napoli Federico II, Via Cinthia, Building 7, Campus MSA, 80126 Napoli, Italy b University of Liverpool, Liverpool, Merseyside L69 3BX, UK; and University of Central Lancashire, Fylde Road, Preston, Lancashire PR1 2HE, UK Published online: 04 Nov 2014.

To cite this article: Gianluca Polese, William Winlow & Anna Di Cosmo (2014) Dose-Dependent Effects of the Clinical Anesthetic Isoflurane on Octopus vulgaris: A Contribution to Cephalopod Welfare, Journal of Aquatic Animal Health, 26:4, 285-294, DOI: 10.1080/08997659.2014.945047 To link to this article: http://dx.doi.org/10.1080/08997659.2014.945047

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Journal of Aquatic Animal Health 26:285–294, 2014 C American Fisheries Society 2014 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2014.945047

ARTICLE

Dose-Dependent Effects of the Clinical Anesthetic Isoﬂurane on Octopus vulgaris: A Contribution to Cephalopod Welfare

Gianluca Polese Department of Biology, University of Napoli Federico II, Via Cinthia, Building 7, Campus MSA, 80126 Napoli, Italy William Winlow University of Liverpool, Liverpool, Merseyside L69 3BX, UK; and University of Central Lancashire, Fylde Road, Preston, Lancashire PR1 2HE, UK Anna Di Cosmo* Department of Biology, University of Napoli Federico II, Via Cinthia, Building 7, Campus MSA, 80126 Napoli, Italy

Abstract Recent progress in animal welfare legislation relating to invertebrates has provoked interest in methods for the anesthesia of cephalopods, for which different approaches to anesthesia have been tried but in most cases without truly anesthetizing the animals. For example, several workers have used muscle relaxants or hypothermia as forms of “anesthesia.” Several inhalational anesthetics are known to act in a dose-dependent manner on the great pond snail Lymnaea stagnalis, a pulmonate mollusk. Here we report, for the first time, on the effects of clinical doses of the well-known inhalational clinical anesthetic isoflurane on the behavioral responses of the common octopus Octopus vulgaris. In each experiment, isoflurane was equilibrated into a well-aerated seawater bath containing a single adult O. vulgaris. Using a web camera, we recorded each animal’s response to touch stimuli eliciting withdrawal of the arms and siphon and observed changes in the respiratory rate and the chromatophore pattern over time (before, during, and after application of the anesthetic). We found that different animals of the same size responded with similar behavioral changes as the isoflurane concentration was gradually increased. After gradual application of 2% isoflurane for a maximum of 5 min (at which time all the responses indicated deep anesthesia), the animals recovered within 45–60 min in fresh aerated seawater. Based on previous findings in gastropods, we believe that the process of anesthesia induced by isoflurane is similar to that previously observed in Lymnaea. In this study we showed that Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 isoflurane is a good, reversible anesthetic for O. vulgaris, and we developed a method for its use.

Anesthetics have long been used to diminish pain and suffer- dling of an animal or to allow surgical manipulations. However, ing in human beings, and the assumption is made that they have pain in invertebrates is as yet a largely unexplored ﬁeld to which similar actions on other vertebrate animals. Although these ani- experimental biologists have paid little attention during exper- mals are not able to communicate directly with the investigators, imental procedures. Therefore, the procedures for invertebrate assumptions about their level of anesthesia are based on their anesthesia, analgesia, and euthanasia that are available at the behavioral responses. In vertebrates, anesthetics have been used moment are more focused on the scientists’ experimental needs to abolish the perception of pain, either to allow humane han- than on the effective reduction of pain and suffering and show

*Corresponding author: [email protected] Received February 14, 2014; accepted July 7, 2014 285 286 POLESE ET AL.

little regard for the animal’s well-being (Cooper 2011; Mather 2011). However, the recent growth of interest in invertebrate welfare (Moltschaniwskyj et al. 2007; Elwood 2011; De Lisa et al. 2012; Andrews et al. 2013; Horvath et al. 2013; Magee and Elwood 2013) necessitates the use of agents such as isoflurane with known anesthetic properties (Eger 1981) during invasive procedures, along with the establishment protocols for their use that are suitable for individual species. Among marine invertebrates the most developed and intel- ligent animals are probably cephalopod mollusks and decapod crustaceans (Advocates for Animals 2005). Cephalopods dis- play advanced cephalization of the ganglionic masses, which FIGURE 1. Anesthetic delivery system consisting of an electric pump (A), a are associated with hierarchical organization of function and flowmeter (B), and a specific (isoflurane) vaporizer (C) terminating in an air the development of advanced cognitive capabilities. Based on stone (D) in a tank (E) filled with 1.6 L of seawater. The ambient temperature these capabilities and structures as well as their complexity, of the water was 18◦C. A recording camera (F) was placed just over the tank to individuality (Mather and Anderson 1993), adaptability, and record the movements and color changes of the animal. The whole system was ability to learn, it should be assumed that cephalopods feel pain. placed in an enclosure (H) and air was expelled to the outside using an extractor fan (G). Although they respond to noxious stimuli (Crook and Walters 2011), their central neural circuits for pain perception are unknown at present apart from Gray’s suggestion (Gray 1970) that their pain center is located in the subvertical lobe. ment. Other muscle relaxants used with cephalopods include Recently introduced legislation stresses the necessity of treat- gallamine (Flaxedil) and tricaine methanesulfonate (or mesy- ing cephalopods humanely. This has resulted from changes to late [MS-222]) (Mooney et al. 2010). It should be noted that laws for the protection of animals in the United Kingdom in although MS-222 is used as an anesthetic and sedative for fish 1993 (when the Animals [Scientific Procedures] Act of 1986 (Palmer and Mensinger 2004), it has only been shown to act was amended to include Octopus vulgaris; UK Statutory Instru- like a local anesthetic (Frazier and Narahashi 1975) or a mus- ments 1993), and in Australia (Sharman 2004) and the European cle relaxant in cephalopods. While muscle relaxants may have Union (European Parliament 2010) to cover all live cephalopod a role in the induction of anesthesia, none of these substances mollusks. However, there is no protection of cephalopods under have anesthetic actions in themselves. The application of gal- U.S. law (Latham 2012). lamine to the longfin inshore squid Doryteuthis peleii resulted Previous attempts to anesthetize cephalopods have used a in the death of the animals, as did that of the local anesthetic variety of substances, including muscle relaxants such as mag- benzocaine and the analgesic clove oil (Mooney et al. 2010). nesium chloride (Messenger et al. 1985; Mooney et al. 2010; Other substances have also proved problematical; solutions of Gonçalves et al. 2012) and low temperatures (Andrews and the intravenous anesthetic urethane (ethyl carbamate) were used Tansey 1981), resulting in paralysis of the animals rather than to anesthetize cephalopods in the past (Messenger 1968; Young anesthesia. A recent study by Gonçalves et al. (2012) compares 1971). However, urethane fell into disfavor when it was found several different agents commonly used to immobilize common to be carcinogenic and endangered laboratory personnel, though cuttlefish Sepia officinalis for short-term handling and indicates it could still be used with appropriate precautions as an anes- that hypothermia causes severe stress reactions during the re- thetic agent for laboratory animals when procedures that did not

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 covery phase; the authors recommend MgCl2 as an appropriate entail recovery were to be carried out (Field and Lang 1988). “anesthetic” agent because it causes little stress to the animal Ethanol has also been tried as a cephalopod anesthetic, but it under these conditions. However, MgCl2 is not an anesthetic tends to cause the escape behaviors of jetting and inking on agent, and neither of these treatments is acceptable as a form initial immersion of the animal (Andrews and Tansey 1981)—a of anesthesia because the loss of consciousness cannot be as- phenomenon that is not always reported (Froesch and Marthy sumed in the absence of movement without detailed analysis 1975; Harms et al. 2006). Mooney et al. (2010) described of the behavioral signs of anesthesia. Magnesium salts such as similar phenomena when D. peleii were placed in ethanol, MgSO4 may be used as adjuncts to anesthetics and have been along with dramatic changes in color not normally seen in the found to decrease the time to onset of neuromuscular blockage animal. induced by clinical blocking agents such as rocuronium (Ro- Ideally, an anesthetic would be a good muscle relaxant, have tava et al. 2013). Additionally, cephalopods are relatively large both analgesic and anesthetic properties, induce unconscious- animals and it is unlikely that immersion in magnesium chlo- ness, and have no side effects. This ideal is difﬁcult to achieve, ride would have rapid effects on the central nervous system as which is why muscle relaxants are used in the induction phase asserted by Messenger et al. (1985). Furthermore, the octopi of clinical or veterinary anesthesia to minimize the voluntary subjected to MgCl2 treatment only survived for 2 d posttreat- and involuntary excitation and struggling that would otherwise EFFECTS OF ISOFLURANE ON OCTOPUS VULGARIS 287

be observed. These behaviors result in an increase in both res- ether (Guedel 1937) but not normally observable under cur- piratory rate and heart rate during induction. rent practices due to the use of muscle relaxants. We chose the Inhalational anesthetics are widely used clinically, along with inhalational anesthetic isoflurane to develop an appropriate pro- a wide variety of adjunctive agents, because their physiological tocol for anesthetization of O. vulgaris because it is very stable effects are well known and relatively easy to monitor and reverse (Ebert 2006), reliable, easily available, relatively inexpensive, (Ebert 2006). Inhalational anesthetics are delivered in a carrier and “can produce adequate muscle relaxation for any surgical stream of air or oxygen using an anesthetic vaporizer to ensure procedure” (Eger 1981). Here we describe for the first time a that the subject remains adequately oxygenated throughout the method for the successful use of isoflurane as an anesthetic for procedure. Among mollusks, the gastropod Lymnaea stagnalis cephalopods. has proved to be a useful model for the study of anesthetics (Girdlestone et al. 1989; Winlow et al. 1995, 1998). The clini- METHODS cal inhalational anesthetics halothane, enflurane, and isoflurane Animals.—Sixteen specimens of Octopus vulgaris (body have all been used to induce anesthesia in Lymnaea in a dose- weight, 0.38–0.43 kg), of which eight were male and eight dependent manner, reducing L-type calcium (Yar and Winlow female, were captured in the Bay of Naples and bought from 1993) currents and potassium currents (Moghadam and Winlow the local fish market, transported to our animal facility, and kept 1995). There is evidence from cell culture that such anesthet- in captivity following the protocol of our ethical committee ics block excitatory chemical synapses more effectively than (2012/0124283-08/11/2012). The test specimens were fed fresh inhibitory synapses (Spencer et al. 1996). In addition, during crabs (Carcinus aestuarii) for 1 week before the experiments. the anesthetization of Lymnaea an early excitatory phase is fol- Octopi that fully recovered from the experimental procedures lowed by suppression (McCrohan et al. 1987), similar to the were kept in the Department of Biology’s animal facility and phases of anesthesia observed in humans under anesthesia by later employed in other experiments. Downloaded by [Department Of Fisheries] at 00:22 17 November 2014

FIGURE 2. Normalized respiratory rate (%) as determined by the number of mantle contractions per minute in 10 animals. Readings were taken at 5-min intervals. After 50 min the seawater in the tank was refreshed and the isoﬂurane ﬂow was switched off, allowing the animal to be normally aerated. Error bars = SDs. 288 POLESE ET AL.

FIGURE 3. Relative strength of the withdrawal of octopus arms in response to a touch stimulus (6 = strong, 4 = medium, 2 = low, and 0 = none) in 10 animals as the isoﬂurane concentration was progressively increased and later reduced to zero.

Anesthesia setup.—In our study an inhalational anesthesia Criteria for anesthesia.—To determine the concentrations of delivery system commonly used for mammals was adapted for isoflurane that can be equated with anesthesia in octopi we used aquatic animals. The seawater surrounding the animal was aer- two criteria, physiological (the respiratory rate as judged by the ated by a power pump and regulated by a flowmeter (1.8 L/min) frequency of respiratory pumping and behavioral (changes of the in series with an isoflurane vaporizer and delivered to the bath via chromatophore pattern and withdrawal of the arms and siphon). an air stone (Figure 1). Except for the delivery of the anesthetic Protocol for anesthesia.—The concentration of anesthetic in into the aquatic medium, this is a standard clinical technique that the airflow started at 0.5% (volume basis) and was increased

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 allows the isoflurane to be delivered to the animal with air as by 0.5% increments at 10-min intervals to a maximum of 2% the carrier gas, thus ensuring that the animal remains fully oxy- in 10 animals. Some animals were less sensitive to isoflurane genated throughout the procedure. All inhalational anesthetics than others, and in four animals we increased the isoflurane are delivered into the animal’s system via an air–water interface, concentration gradually to 2.5%. In the early stages of our work the respiratory epithelium in both vertebrates and octopi, but in we applied 2.5% isoflurane immediately to two animals, but this case it was via the aquatic medium. The bath was kept in an this proved to be lethal and was not repeated. To determine the enclosure and vented externally. It contained 1,600 mL of sea- depth of anesthesia we applied the criteria above and recorded water and was kept at 18◦C throughout the experiments. Each the responses with a web camera (Apple iSight, 1/4-in; 640 × animal was left to acclimatize to the bath for at least 10 min be- 480 resolution, internal auto-focus lens system, shutter speed fore we started to deliver the air and isoflurane gas mixture via up to 30 frame/s): touch stimuli induced withdrawal of the arms the air stone. Previous experiments using gas–liquid chromatog- and siphon, and we monitored the changes in the respiratory rate raphy have demonstrated equilibration times of about 10 min and in the chromatophore pattern throughout the experiment. for inhalation anesthetics delivered in this manner (Girdlestone The withdrawal responses were elicited by gentle application of et al. 1989). a blunt Plexiglas probe to the arms and siphon. For analytical EFFECTS OF ISOFLURANE ON OCTOPUS VULGARIS 289

FIGURE 4. Relative strength of the withdrawal of octopus siphons in response to a touch stimulus (6 = strong, 4 = medium, 2 = low, and 0 = none) in 10 animals as the isoﬂurane concentration was progressively increased and later reduced to zero.

purposes we categorized the responses as strong, medium, weak, signiﬁcance of differences between the data at time 0 and those or abolished and these were given numerical values of 6, 4, 2, at other times. and 0, respectively. The respiratory rate was taken every 5 min, just before the RESULTS touch test was performed, to avoid any change due to stimula- Although the octopi studied behaved as individuals, ex- tion. Animals had widely varying precontrol respiratory rates hibiting somewhat different responses, in categorizing those (24–45 cycles per minute) as judged by the pumping move-

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 responses we found that after 10 min in the well-aerated exper- ments of the mantle, so for analytical purposes we normalized imental bath respiratory rates were the same as before removal the respiratory rate, with 100% being the observed precontrol from the aquaria, the overall chromatophore pattern was again value. dark, and both the arm and siphon withdrawal responses were Using imageJ software (Collins 2007) we measured a spot 10 rapid. In the 10 final study animals the responses to isoflurane pixels in size from the interbrachial membrane on frames taken detailed below were observed during anesthesia and recovery. every 5 min throughout the experiments. The color intensity of each spot was measured based on the additive red, green, blue 0.5% Isoflurane (RGB) color model whereby a zero intensity (value, 0) for each Few responses were observed in the first 5 min after initiation component gives the darkest color (no light, considered to be of the flow of isoflurane, but after 10 min (20 min into the black) and full intensity for each component gives white (value, experiment overall) the mean respiratory rate fell to 30% (SD, 255). 18%; df = 2, F = 151.23; P < 0.0001) of the initial value Statistical analyses were performed using the JMP software (Figure 2), while the arm and the siphon withdrawal responses package (version 10.0; SAS Institute, Cary, North Carolina). remained normal (Figures 3, 4). There was no observable change One-way analysis of variance (ANOVA) was used to assess the in color pattern. 290 POLESE ET AL.

Figures 5, 6) and the ﬂashing behavior of the chromatophores had increased.

1.5% Isoflurane Five minutes after administration of 1.5% isoflurane the respiratory rate rose to 88% (SD, 16%; df = 5, F = 12.7, P < 0.0001) of the normalized control value (Figure 2), and arm and siphon withdrawal responses were weaker (arms: df = 5, F = 98.5, P < 0.0001; siphon: df = 5, F = 16.38, P < 0.0001; Figures 3, 4) and the chromatophores were flashing in a completely uncontrolled manner. After 10 min, the respiratory rates were at 95% (SD, 10%) of and not significantly different from the normalized control values (Figure 2) and the withdrawal responses had weakened further (arms: df = 6, F = 108.8, P < 0.0001; siphon: df = 6, F = 19.86, P < 0.0001; Figures 3, 4). The patterning of the chromatophores left the animals looking almost white (df = 6, F = 52.99, P < 0.0001; Figures 5, 6).

2.0% Isoflurane During the first 5 min after increasing the isoflurane concentration to 2% the animals became hyperexcitable, characterized by a sequence of uncontrolled movements. Immediately after the animals lost their reactivity to touch stimuli, with minimal or nonexistent withdrawal responses (arms: df = 7, F = 62.22, P < 0.0001; siphon: df = 7, F = 26.55, P < 0.0001; Figures 3, 4). The respiratory rates were still normal (Figure 2), but the overall color pattern showed no further change although ran- dom chromatophore flashing continued at a lower frequency than previously. Thus, after the administration of 2% isoflurane for 10 min the animals were clearly relaxed, unresponsive, and anesthetized. At this point the water in the bath was replaced with fresh, continuously aerated seawater at 18◦C and the animals were allowed to recover.

Recovery Period No significant behavioral changes were detected during the first 15 min after the flow of anesthetic was stopped. After FIGURE 5. Example of the gross color pattern changes in an octopus being subjected to and then recovering from isoflurane anesthesia. The light blue 20 min, most of the animals started to show small to medium triangles indicate exposure to plain seawater; the pink triangles indicate exposure withdrawal responses of both the siphon and the arms (Figures 3,

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 to isoflurane at various percentages. 4), respiration was maintained normally (Figure 2), and the normal chromatic repertoire started to return (Figures 5, 6). 1.0% Isoflurane After 5 min in 1% isoflurane, the normalized respiratory 2.5% Isoflurane Overdose rates recovered to 68% (SD, 38%; df = 3, F = 21.06, P < In four animals anesthetization was incomplete at 2.0% 0.0001) of control values; arm and siphon withdrawal reflexes isoflurane (Figures 7, 8A, 8B) and the concentration was raised were still normal. Color was slightly paler (df = 3, F = 8.14, to 2.5% for 5 min, after which arm and siphon withdrawal re- P < 0.0001; Figures 5, 6), and the chromatophores had started sponses ceased although the animals were moved immediately to flash randomly. to fresh aerated seawater. Further, respiratory movements first After 10 min the respiratory rate was within the control range declined and then ceased 15 min after the anesthetic was flushed (df = 4, F = 18.9, P < 0.0001) (Figure 2), but the withdrawal out. The animals were then artificially respired for 25 min by responses were weaker (arms: df = 4, F = 40, P < 0.0001; massaging their mantles until respiration restarted; respiration siphon: df = 4, F = 10, P < 0.0001; Figures 3, 4). The color gradually returned to the control frequency within 20 min, dur- pattern of the animals changed markedly; the interbrachial area ing which period the withdrawal responses were also gradually was paler than previously (df = 4, F = 94.622, P < 0.0001; restored (Figures 7, 8A, 8B). EFFECTS OF ISOFLURANE ON OCTOPUS VULGARIS 291

FIGURE 6. Color intensity of the interbrachial membrane as the concentration of isoﬂurane was progressively increased and then decreased. Color intensity was measured using the RGB model whereby a zero intensity (value, 0) for each component gives the darkest color (no light, considered to be black) and full intensity for each component gives white (value, 255). Thus, the higher the recorded value the paler the interbrachial membrane and vice versa.

Sudden Onset of 2.5% Isoflurane depth of anesthesia in humans and other animals and are dimin- Early in our studies two animals were subjected to a sudden ished in a dose-dependent manner by inhalational anesthetics application of 2.5% isoflurane. This caused immediate hyper- in the pulmonate mollusk Lymnaea stagnalis (Girdlestone et al. excitability in both animals, whose respiration ceased and could 1989). The second behavioral criterion used was color change, not be reactivated. In the first case isoflurane was applied for which is common in cephalopods and is known to be under 10 min with little immediate effect on withdrawal responses. We central motor control. Using these criteria, we concluded that then returned the animal to aerated seawater, but its respiration in most of our specimen octopi (10 of 14) it was best to slowly ceased and the animal died. In the second case we reduced the increase the concentration of the anesthetic from 0.5% to 2.0% time of 2.5% isoflurane application to 5 min after 5 min with over a period of about 40 min. The animals usually started to recover within 20 min of flushing out and fully recovered in

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 1% isoflurane. In this case we tried to artificially respire the animal in aerated seawater, but to no avail. Thus, rapid exposure less than an hour. However, octopi are highly individualistic to a high dose of isoflurane was lethal to the animal, causing (Mather and Anderson 1993) and some animals (4 of 14) were irreversible respiratory failure. not quite anesthetized by 2% isoflurane under the conditions described here. From our observations, it seems that a further DISCUSSION 5 min during which the isoflurane concentration was increased To develop a method for inducing anesthesia in Octopus to 2.5% would prove fatal to the animals unless they are arti- vulgaris with isoflurane, it was first necessary to determine ap- ficially respired for 25 min due to respiratory failure. Shorter propriate criteria for the changes in the animal’s physiology and periods of exposure to the higher isoflurane concentration might behavior. The most obvious physiological criterion was the res- solve this problem. However, all four of the animals studied here piratory rate as judged by the frequency of respiratory pumping, survived the isoflurane “overdose” without apparent harm. Fur- particularly as most clinical anesthetics depress minute ventila- ther experiments will be needed to determine an appropriate tion in mammals (Berge and Warner 2000). We also applied two maintenance concentration of isoflurane. behavioral tests. One was the withdrawal response to stimula- While it is difficult to directly compare the octopus’s respi- tion of the arms. Withdrawal responses are used to test for the ratory system with that of vertebrates, the former has a clear 292 POLESE ET AL.

FIGURE 7. Normalized respiratory rate (%) as determined by the number of mantle contractions per minute in 4 animals for which the metered isoflurane concentration was raised to 2.5% (volume basis). Readings were taken at 5-min intervals. After 55 min the seawater in the tank was refreshed and the isoflurane flow was switched off, allowing the animals to be normally aerated. However, in these animals respiratory pumping ceased altogether and aeration of the gills could only be maintained by artificial respiration using gentle massage of the mantle to resuscitate the animals.

respiratory rhythm from which it may be inferred that there is tory frequency at low isoflurane concentrations is that while some form of central respiratory pattern generator within its it is “conscious” the animal senses the compound as possibly system that is presumably acted upon by a number of external harmful or unpleasant and reduces its respiratory frequency to factors (such as external oxygen concentration) and that inter- compensate, while at higher concentrations the animal has lost acts with the cardiovascular system. In mammals, inhaled anes- consciousness and the isoflurane has an uncompensated effect thetics depress the volume of respiration but not necessarily its similar to that in humans (Eger 1981). This interpretation is frequency, although irreversible anesthesia results in respiratory supported by the fact that the removal of isoflurane from the ex- Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 arrest and circulatory failure in a matter of minutes. At safe con- ternal medium results in a fall in respiratory frequency, but this centrations isoflurane tends to increase respiratory frequency in fall could also be due to the fact that the isoflurane is still in the humans but produces profound respiratory depression by pro- process of penetrating the octopus’s respiratory centers (which gressively decreasing the tidal volume with increasing depth may be located in the basal lobe system, particularly the median of anesthesia (Fourcade et al. 1971; Pauca and Dripps 1973). basal lobes of the brain; Nixon and Young 2003). This possi- Furthermore, the excitatory phase of anesthesia in mammals bility is supported by the observation that in octopi that were (Guedel 1937) can result in struggling behavior and a rise in “overdosed” with isoflurane respiration ceased entirely and had heart rate, which may explain the rise in respiratory frequency to be maintained artificially (a result that may have been due observed in our experiment with 1–2% isoflurane (Figure 2). to global suppression of the octopus’s respiratory center and In our study octopi, we were only able to monitor the changes associated respiratory motor neurons, as has been observed in in respiratory frequency as anesthesia proceeded. There was a mammals; Berge and Warner 2000). Care must be taken not fall in frequency at low concentrations of isoflurane, but fre- to cause complete respiratory suppression when anesthetizing quency returned to the normal range as the concentration in- octopi. It should be noted that the hyperexcitability observed at creased (Figure 2). One interpretation of the reduced respira- exposure to 2% isoflurane (which occurs just before the loss of EFFECTS OF ISOFLURANE ON OCTOPUS VULGARIS 293

wood 2011). Thus, to determine the actions of anesthetics on octopi will depend on a more detailed understanding of the brain centers involved in consciousness. Withdrawal responses are protective reflexes and take prece- dence over other behaviors (Winlow and Polese 2014); the lack of responsiveness to touch stimuli was one indicator of anesthesia in humans before muscle relaxants (neuromuscular blocking agents) came into common use 70 years ago (Griffith and Johnson 1942). In mollusks, well-studied examples are the gill withdrawal reflex in sea slugs Aplysia spp. (Kandel 1976) or whole-animal withdrawal in Lymnaea (Winlow et al. 1992). The Lymnaea whole-animal withdrawal response has been used as an indicator of anesthesia following application of the volatile anesthetics halothane, enflurane, and isoflurane (Girdlestone et al. 1989). In the current study, siphon and arm withdrawal declined in a dose-dependent manner, indicating that both are good indicators of anesthetic state. This is supported by the fact that movements of the siphon and arms are hierarchically controlled (Boycott and Young 1950). In cephalopods color change and patterning are known to be under direct motor control from the brain (Hanlon and Mes- senger 1996) and respond to visual inputs to the anterior and posterior chromatophore lobes via the optic lobes and lateral basal lobes. Stimulation of the anterior chromatophore lobe results in unilateral darkening of the skin of the arms and the head, while stimulation of the posterior lobe causes darkening of the mantle skin via motor neurons to the radial muscles of the chromatophores (Hanlon and Messenger 1996). The gradual dose-dependent paling of octopi in the presence of isoflurane suggests that normal motor coordination of color change is lost as anesthesia proceeds and thus is a further indicator of the progression of anesthesia. FIGURE 8. Normalized withdrawal responses (6 = strong, 4 = medium, 2 = low, and 0 = none) of (A) siphons and (B) arms in response to a touch stimulus We believe that the process of anesthesia induced by isoflu- during a progressive increase in isoflurane concentration up to 2.5% followed rane is similar to that previously observed in Lymnaea and that by flushing with clean oxygenated seawater. it is dose dependent. We have shown that isoflurane appears to be a good reversible anesthetic for Octopus vulgaris, and all reactivity to touch stimuli) signals that the animal is about to we have developed a method for its use. Further experiments succumb to the anesthetic at a depth sufficient to allow surgical are required to determine the appropriate maintenance dose for manipulation. surgical procedures.

Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 It is difficult to obtain a median effective dose (ED50) for isoflurane with the small number of animals used in this study. ACKNOWLEDGMENTS However, it is clear from our data that the level of anesthetization is dose dependent and that 70% (10 of 14) animals were com- We thank F. P. Ulloa Severino for his assistance in managing pletely anesthetized in 2% isoflurane, suggesting that the ED50 and recording the video during the experiments. This work was is rather less than 2%. In mollusks such as Lymnaea stagnalis supported by a Single Center Research Grant in Neuroscience the ED50 was found to be 1.09%, while in humans it was found from Compagnia di San Paolo (Protocol 29-11). to be 1.15% and in cats Felis cattus it was 1.63% (Girdlestone et al. 1989), all of which indicates a degree of species specificity. REFERENCES At this stage it is difficult to know exactly what concentration Advocates for Animals. 2005. Cephalopods and decapod crustaceans: their of isoflurane reaches the central nervous system of octopi under capacity to experience pain and suffering. Advocates for Animals, Edinburgh, the conditions described here. However, Wells (1978) indicates Scotland. Andrews, P. L. R., A. S. Darmaillacq, N. Dennison, I. G. Gleadall, P. Hawkins, J. that specific octopus brain areas are specialized for specific B. Messenger, D. Osorio, V.J. Smith, and J. A. Smith. 2013. The identification functions such as memory, learning, sensory integration, and and management of pain, suffering, and distress in cephalopods, including decision making (akin to the human cerebral cortex) and exhibit anaesthesia, analgesia, and humane killing. Journal of Experimental Marine a level of learning ability that “might signal consciousness” (El- Biology and Ecology 447:46–64. 294 POLESE ET AL.

Andrews, P. L. R., and E. M. Tansey. 1981. The effects of some anaesthetic Mather, J. A. 2011. Philosophical background of attitudes toward and treatment agents in Octopus vulgaris. Comparative Biochemistry and Physiology C of invertebrates. ILAR (Institute for Laboratory Animal Research) Journal 70:241–247. 52:205–212. Berge, K. H., and D. O. Warner. 2000. Drugs that affect the respiratory system. Mather, J. A., and R. C. Anderson. 1993. Personalities of octopuses Pages 491–498 in H. Hemmings and P. Hopkins, editors. Foundations of (Octopus rubescens). Journal of Comparative Psychology 107:336– anesthesia. Basic and clinical sciences. Mosby, London. 340. Boycott, B. B., and J. Z. Young. 1950. The comparative study of learning. McCrohan, C. R., D. Girdlestone, and W. Winlow. 1987. Effects of halothane Symposia of the Society for Experimental Biology 4:432–453. on feeding motor-activity in the snail Lymnaea stagnalis. Comparative Bio- Collins, T. J. 2007. ImageJ for microscopy. Biotechniques 43:25–30. chemistry and Physiology C 86:55–62. Cooper, J. E. 2011. Anesthesia, analgesia, and euthanasia of invertebrates. ILAR Messenger, J. B. 1968. The visual attack of the cuttlefish, Sepia officinalis. (Institute for Laboratory Animal Research) Journal 52:196–204. Animal Behaviour 16:342–357. Crook, R. J., and E. T. Walters. 2011. Nociceptive behavior and physiology Messenger, J. B., M. Nixon, and K. P. Ryan. 1985. Magnesium chloride as of molluscs: animal welfare implications. ILAR (Institute for Laboratory an anesthetic for cephalopods. Comparative Biochemistry and Physiology C Animal Research) Journal 52:185–195. 82:203–205. De Lisa, E., A. De Maio, L. L. Moroz, F. Moccia, M. R. F. Mennella, and A. Moghadam, H. F., and W. Winlow. 1995. Pentobarbital inhibits K + current in Di Cosmo. 2012. Characterization of novel cytoplasmic PARP in the brain of cultured neurons. Journal of Physiology (London) 483:191. Octopus vulgaris. Biological Bulletin 222:176–181. Moltschaniwskyj, N. A., K. Hall, M. R. Lipinski, J. E. A. R. Marian, M. Ebert, T. J. 2006. Inhalation anesthesia. Pages 384–420 in G.Barash,B.F. Nishiguchi, M. Sakai, D. J. Shulman, B. Sinclair, D. L. Sinn, M. Staudinger, Cullen, and R. K. Stoelting, editors. Clinical anesthesia. Lippincott Williams R. van Gelderen, R. Villanueva, and K. Warnke. 2007. Ethical and welfare and Wilkins, Philadelphia. considerations when using cephalopods as experimental animals. Reviews in Eger, E. I. 1981. Isoflurane: a review. Anesthesiology 55:559–576. Fish Biology and Fisheries 17:455–476. Elwood, R. W. 2011. Pain and suffering in invertebrates? ILAR (Institute for Mooney, T. A., W. J. Lee, and R. T. Hanlon. 2010. Long-duration anestheti- Laboratory Animal Research) Journal 52:175–184. zation of squid (Doryteuthis pealeii). Marine and Freshwater Behaviour and European Parliament. 2010. Council on the protection of animals used for scien- Physiology 43:297–303. tific purposes. Directive 2010/63/EU L276. Official Journal of the European Nixon, M. J., and J. Z. Young.2003. The brains and lives of cephalopods. Oxford Union 53:33–79. University Press, New York. Field, K. J., and C. M. Lang. 1988. Hazards of urethane (ethyl carbamate): a Palmer, L. M., and A. F. Mensinger. 2004. Effect of the anaesthetic tricaine review of the literature. Laboratory Animals 22:255–262. (MS-222) on nerve activity in the anterior lateral line of the oyster toadfish, Fourcade, H. E., W. C. Stevens, C. P. Larson Jr., T. H. Cromwell, S. H. Bahlman, Opsanus tau. Journal of Neurophysiology 92:1034–1041. R. F. Hickey, M. J. Halsey, and E. Eger Jr. 1971. Ventilatory effects of Forane, Pauca, A. L., and R. D. Dripps. 1973. Clinical experience with isoflu- new inhaled anesthetic. Anesthesiology 35:26–31. rane (Forane): preliminary communication. British Journal of Anaesthesia Frazier, D. T., and T. Narahashi. 1975. Tricaine (MS-222): effects on ionic 45:697–703. conductances of squid axon membranes. European Journal of Pharmacology Rotava, P., I. L. Cavalcanti, L. Barrucand, L. A. Vane, and N. Vercosa. 2013. 33:313–317. Effects of magnesium sulphate on the pharmacodynamics of rocuronium in Froesch, D., and H.-J. Marthy. 1975. The structure and function of the ovidu- patients aged 60 years and older, a randomised trial. European Journal of cal gland in octopods (Cephalopoda). Proceedings of the Royal Society of Anaesthesiology 30:599–604. London B 188:95–101. Sharman, K. 2004. Animal law in Australia. Available: http://animallaw.info/ Girdlestone, D., S. G. H. Cruickshank, and W. Winlow. 1989. The actions article/animal-law-australia. (October 2014). of 3 volatile general-anesthetics on withdrawal responses of the pond snail Spencer, G. E., N. I. Syed, K. Lukowiak, and W. Winlow. 1996. Halothane Lymnaea stagnalis (L). Comparative Biochemistry and Physiology C 92:39– affects both inhibitory and excitatory synaptic transmission at a single 43. identified molluscan synapse, in vivo and in vitro. Brain Research 714: Gonçalves, R. A., C. Aragao,˜ P. A. Frias, and A. V. Sykes. 2012. The use of 38–48. different anaesthetics as welfare promoters during short-term human ma- UK Statutory Instruments. 1993. The Animals (Scientific Procedures) Act nipulation of European cuttlefish (Sepia officinalis) juveniles. Aquaculture (Amendment) Order 1993 Number 2103. UK Government, London. 370-371:130–135. Wells, M. J. 1978. Octopus: physiology and behaviour of an advanced inverte- Gray, E. G. 1970. The fine structure of the vertical lobe of octopus brain. brate. Chapman and Hall, London. Philosophical Transactions of the Royal Society of London B 258:379–394. Winlow, W., P. M. Hopkins, H. F. Moghadam, I. A. Ahmed, and T. Yar. 1995. Downloaded by [Department Of Fisheries] at 00:22 17 November 2014 Griffith, H. R., and G. E. Johnson. 1942. The use of curare in general anesthesia. Multiple cellular and subcellular actions of general anaesthetics on cultured Anesthesiology 3:418–420. molluscan neurones. Acta Biologica Hungarica 46:381–393. Guedel, A. E. 1937. Inhalational anesthesia; a fundamental guide. Macmillan, Winlow, W., L. L. Moroz, and N. I. Syed. 1992. Mechanisms of behavioural New York. selection in Lymnaea stagnalis. Pages 52–72 in J. Kien, C. R. McCrohan, and Hanlon, R. T., and J. B. Messenger. 1996. Cephalopod behaviour. Cambridge W. Winlow, editors. Neurobiology of motor programme selection. Pergamon, University Press, Cambridge, UK. Oxford, UK. Harms, C. A., G. A. Lewbart, R. McAlarney, L. S. Christian, and C. Lemons. Winlow, W., and G. Polese. 2014. A neuroplastic network underlying behaviour 2006. Surgical excision of mycotic (Cladosporium sp.) granulomas from the and seasonal change in Lymnaea stagnalis: a neuroecological standpoint. mantle of a cuttlefish (Sepia officinalis). Journal of Zoo and Wildlife Medicine Pages 145–176 in A. Di Cosmo and W. Winlow, editors. Neuroecology and 37:524–530. neuroethology in molluscs: the interface between behaviour and environment. Horvath, K., D. Angeletti, G. Nascetti, and C. Carere. 2013. Invertebrate welfare: Nova Science Publishers, New York. an overlooked issue. Annali dell’Istituto Superiore di Sanita` 49:9–17. Winlow, W., G. E. Spencer, N. I. Syed, and M. M. Qazzaz. 1998. Modula- Kandel, E. R. 1976. Cellular basis of behavior: an introduction to behavioral tion of reconstructed peptidergic synapses and electrical synapses by general neurobiology. Freeman, San Francisco. anaesthetics. Toxicology Letters 100/101:77–84. Latham, S. R. 2012. U.S. law and animal experimentation: a critical primer. Yar, T., and W. Winlow. 1993. Halothane depressed a high-voltage-activated cal- Hastings Center Report 42:S35–S39. cium current in identified cultured neurons from Lymnaea stagnalis. Journal Magee, B., and R. W. Elwood. 2013. Shock avoidance by discrimination learning of Physiology (London) 459:12. in the shore crab (Carcinus maenas) is consistent with a key criterion for pain. Young, J. Z. 1971. The anatomy of the nervous system of Octopus vulgaris. Journal of Experimental Biology 216:353–358. Clarendon Press, Oxford, UK.