Fish Sci (2010) 76:403–410 DOI 10.1007/s12562-010-0228-4 ORIGINAL ARTICLE Fisheries Spatial and temporal variation in the distribution of juvenile southern bluefin tuna Thunnus maccoyii: implication for precise estimation of recruitment abundance indices Ko Fujioka • Ryo Kawabe • Alistair J Hobday • Yoshimi Takao Kazushi Miyashita • Osamu Sakai • Tomoyuki Itoh • Received: November 2009 / Accepted: February 2010 / Published online: 13 March 2010 Ó The Japanese Society of Fisheries Science 2010 Abstract Acoustic tags were used to examine the spatial and temporal distribution of southern bluefin tuna (SBT) in southern Western Australia, which is in a region where fishery-independent acoustic surveys of the recruitment abundance index of SBT have been historically undertaken We investigated patterns of SBT distribution within and inshore of the acoustic survey area during three summer seasons Annual differences in distribution patterns were characterized by two distinctive migration pathways An inshore-migrating pathway was observed in two seasons (2004/2005 and 2006/2007), with a relatively high proportion of tagged SBT (84.5, 65.0%) migrating inshore of the acoustic survey area The other pathway was concentrated along the shelf (2005/2006 season), with an estimated 63.3% of tagged SBT moving within the survey area These variable migration patterns may bias the interannual fluctuations in abundance indices Current survey methods can be modified to include both inshore and continental shelf areas This contribution shows that the accuracy of acoustic surveys can be improved by including ecological patterns Keywords Acoustic telemetry Á Migratory pathway Á Movement patterns Á Recruitment abundance index Á Residence time Introduction K Fujioka (&) Graduate School of Science and Technology, Nagasaki University, Taira-machi, Nagasaki 851-2213, Japan e-mail: d706164e@cc.nagasaki-u.ac.jp R Kawabe Institute for East China Sea Research, Nagasaki University, Taira-machi, Nagasaki 851-2213, Japan A J Hobday CSIRO Wealth from Oceans National Research Flagship and Marine and Atmospheric Research, GPO Box 1538, Hobart Tasmania 7001, Australia Y Takao National Research Institute of Fisheries Engineering, Fisheries Research Agency, Hasaki, Kamisu 314-0408, Japan K Miyashita Field Science Center for Northern Biosphere, Hokkaido University, Minato, Hakodate 041-8611, Japan O Sakai Á T Itoh National Research Institute of Far Seas Fisheries, Fisheries Research Agency, Shimizu, Shizuoka 424-8633, Japan Southern bluefin tuna Thunnus maccoyii (SBT) spawn in the northeast Indian Ocean from August to May [1] Young-of-the-year migrate down the shelf of the western coast of Australia and are found as age-1 SBT in southern Western Australia A fishery-independent survey, which has become a necessity in many regions [2] for assessing the stock abundance of commercially important fish, has been applied to determine the relative abundance of 1-yearold SBT and is referred to as an acoustic recruitmentmonitoring survey (ARS) The survey is a line transect survey using a consistent protocol by means of omniscanning sonar; it has been undertaken since the 1995/1996 season under the Japan–Australia Southern Bluefin Tuna Recruitment Monitoring Program (T Itoh and S Tsuji, unpublished data, 2004) The ARS has been conducted in a specified area between Albany and Esperance, in southern Western Australia, where the width of the continental shelf becomes narrow, because it was assumed that most 1-yearold (and 2-year-old) SBT pass through and along the 123 404 southern western coast of Australia during the summer [3] The observed biomass is then used to determine the recruitment indices of age-1 SBT (T Itoh and S Tsuji, unpublished data, 2002) A common assumption is that the movement pattern of SBT passing through the ARS area does not change over time when the recruitment survey is conducted, so that changes in the recruitment index can be interpreted as a variation in the characteristics of the same fish group, rather than changes brought by movement of different populations through the ARS area Such assumptions may be flawed The reliability of these survey indices may depend on the changes caused by spatial and temporal distribution of SBT A recent study suggests an interannual difference in the spatial and temporal distribution [4] Thus, there is need to examine what causes declines in indices relative to the spatial and temporal distribution (especially the timing of the departure from the ARS area, migratory pathways, and residence times in and out of the ARS area) of summer-resident 1-year SBT in coastal waters of southern Western Australia [3] The study of finer localized migration patterns, such as those in coastal waters (and around small seamounts) or over the continental shelf, requires techniques capable of a much finer (about km) spatial resolution Recent advances in acoustic tagging technology have made available low-cost, submersible receivers that can automatically detect and identify passing fish, such as cod and tuna [4–7] The use of acoustic tracking systems has provided valuable insight into migration pathways [6], residence time on small and large seamounts [8, 9] and habitat usage [10] Some applications using acoustic tracking data have examined migration patterns of individual animals [4, 6, 11] These studies typically look for individuals that pass a specific point or through a line of receivers to define migration metrics These data are then used to calculate the rate of progression through the area and also to examine swimming speed based on movement past known points Similar SBT cross-shelf distributions and habitat utilization in southern Western Australia have been described in previous papers [3, 4] However, these studies did not explicitly address movement patterns and differences in residence time Here, we determined the annual fluctuation of distribution and movement patterns of age-1 SBT out of and within the ARS area using behavioral data obtained during the summer migration of juvenile SBT in 2004/2005, 2005/ 2006, and 2006/2007 seasons The main objective of our study was to identify annual fluctuations in fish distribution and aggregation as a function of temporal variation in a specific habitat and to discuss the design of ARS in relation to the summer migration of juvenile SBT population over the three seasons 123 Fish Sci (2010) 76:403–410 Materials and methods Acoustic receivers Seventy VR2 acoustic receivers (Vemco, Halifax, Canada) were deployed inshore of and within the ARS area in southern Western Australia (Fig 1a, b) Each receiver was fastened to a vertical wire cable on a mooring anchor (125 kg section of railway track) Listening stations with a receiver consisted of time-scheduled electronic releaser, 50 m of release rope in a PVC canister, and four/five floats When deployed, the receivers were set to a depth of 20–25 m, just below the subsurface floats in waters up to 150 m deep [3] Listening stations were deployed in three cross-shelf lines running from the coast to the edge of the continental shelf at each location [Fig 1a; western line from Bald Island (line 1: 20 stations), middle line from Point Henry (line 2: 20 stations), eastern line from West Island (line 3: 21 stations)] In addition, three listening stations were deployed at each of three coastal topographic features (lumps) located between line and line at depths of 40–60 m Some topographic lumps occur in this coastal area and are known to attract pelagic fishes of several species, including juvenile SBT [12] During the summer season in 2004/2005, listening stations were deployed December 3–5, 2004, and retrieved March 15–17 (106 days), 2005 In the 2005/2006 season, listening stations were deployed December 1–3, 2005, and retrieved May 9–11, 2006 (161 days) In the 2006/2007 season, listening stations were deployed December 1–3, 2006, (lines 1–2 and lumps) and January 14 (line 3) and retrieved May 29–31, 2007 (181 days) All receivers thus continuously monitored the passage of any tagged individuals over at least a 3-month period spanning December– March Acoustic transmitters The transmitters used to tag SBT were V8, V9, and V16 coded pingers (Vemco) Each pinger transmits a unique pinging sequence at a frequency of 69 kHz, which is repeated after a random delay of between 20 and 60 s Battery life is rated at 365 (V8/V9) and 700 (V16) days Receivers were separated by approximately 1,500 m This spacing decision was based on a desire to cover the width of the shelf; a tag detection range of up to 450 m (V8) and 800 m (V16) was expected based on detection experiments (Hobday et al unpublished data, 2005) The same protocol used for the capture and selection of SBT for conventional tagging was followed for the acoustic tagging [3, 4] Briefly, fish were caught by polling or trolling at the stern of the vessel (F/V Quadrant) and immediately placed in a tagging cradle Caudal fork length Fish Sci (2010) 76:403–410 Fig Research area in southern Western Australia a Locations of acoustic receivers are represented by white circles (lines 1–3: n = 20–21 receivers per line, lumps: n = receivers on each lump) Receivers were aligned as a ‘‘curtain’’ on the shelf b The area outlined in black shows the acoustic recruitmentmonitoring survey area (ARS area), the recruitment survey for juvenile southern bluefin tuna by omni-scanning sonar carried out since 1995/1996 Study regions within the ARS area are indicated by the dark diagonal area, and study regions out of the ARS area are shown as light diagonal area 405 Esperance (a) 34° 00’ S 41 Line Bremer Bay 61 21 Lumps 68-70 65-67 62-64 Line 40 Albany 35° 00’ S Australia Line 20 25 50km Research area 200 m 1000 m Esperance (b) 34° 00’ S Out of ARS area In ARS area Bremer Bay Acoustic recruitment monitoring survey (ARS) area Albany 35° 00’ S 200 m 1000 m 118° 00’ E (FL) was measured to the nearest centimeter For acoustic transmitters, a 1–1.5 cm horizontal incision was made about 0.5–1 cm off the midline and anterior to the vent by about 2–3 cm The body wall was penetrated until the membranes of the peritoneum were observed The membrane was then torn by a gloved finger, and a space in the visceral cavity (where the transmitter would be inserted) was carefully wedged out to help ensure no damage occurred to internal organs The incisions were closed with one (or two) sutures The entire implantation procedure generally took less than Fish were also tagged with conventional plastic dart tags placed between the pterygiophores adjacent to the insertion of the second dorsal fin All fish were tagged by a single experienced operator For the 2004/2005 season, a total of 79 fish [FL 41– 64 cm (mean ± SD 51.8 ± 6.2)] were tagged in the area between line and line For the 2005/2006 season, a total of 81 fish [FL 43–73 cm (mean ± SD 49.4 ± 6.4)] were also tagged in the area between line and line For the 2006/2007 season, a total of 84 fish [FL 44–93 cm (mean ± SD 57.3 ± 5.8)] were tagged in a wider region to the west of line than in previous years The release locations on the continental shelf (west–east range) each 120° 00’ E 122° 00’ E season were 117.950°E–119.509°E (mean 118.847°E), 117.961°E–119.265°E (mean 118.840°E), and 115.288°E– 118.979°E (mean 118.082°E), respectively Details of the tagging date and the number of tagged fish, FL (cm), and tagging location are given in Table Analyses The site fidelity of tagged fish among seasons was determined by examining the time series of acoustic detections using data from each listening station Two areas are relevant for estimating movement patterns relative to the detection coverage of the acoustic receiver array: (1) an area within the ARS area and (2) an area out of the ARS area (Fig 1b) The area within the ARS area was 4,215.5 km2 (54% of total detection area) and out of the ARS area was 3,613.8 km2 (46%) To compare the spatial usage and migratory direction of tagged SBT between the regions in and out of the ARS area, SBT occupancy in each area was determined using acoustic detection data We calculated the proportion of time spent in and out of the ARS area based on movement time (days) between two receivers that detected the fish In other words, the time 123 406 Fish Sci (2010) 76:403–410 Table Details of tagged southern bluefin tuna Thunnus maccoyii, for the three seasons (2004/2005, 2005/2006, 2006/2007) Season 2004/2005 2005/2006 2006/2007 Tagging date No of tagged fish Mean tagging location Latitude Longitude 7-Dec-2004 22 56.0 ± 6.0 -34.58 118.99 3-Jan-2005 45.0 -35.20 117.95 4-Jan-2005 46.6 ± 5.6 -35.18 118.04 5-Jan-2005 15 53.4 ± 6.6 -34.76 118.70 6-Jan-2005 10 50.0 ± 5.2 -34.66 118.83 7-Jan-2005 54.2 ± 5.4 -34.65 118.82 8-Jan-2005 9-Jan-2005 13 47.8 ± 2.0 48.6 ± 1.3 -34.92 -34.49 119.08 119.42 4-Dec-2005 45.0 -34.54 119.26 5-Dec-2005 45.0 -34.68 118.75 7-Dec-2005 67.0 -34.69 118.76 8-Dec-2005 44.4 ± 1.1 -35.19 118.00 6-Jan-2006 21 46.0 ± 1.0 -34.85 119.04 7-Jan-2006 45.5 ± 0.7 -34.93 118.72 8-Jan-2006 13 46.0 ± 1.2 -34.74 118.83 9-Jan-2006 62.3 ± 10.8 -34.63 118.88 10-Jan-2006 30 51.2 ± 3.7 -34.65 118.84 4-Dec-2006 17 57.6 ± 2.7 -34.57 118.97 5-Dec-2006 15 58.1 ± 4.2 -34.57 118.97 7-Dec-2006 62.0 ± 8.5 -34.66 118.84 8-Dec-2006 49.0 -35.17 117.94 9-Dec-2006 10-Dec-2006 18 53.1 ± 3.4 54.0 -35.11 -34.52 116.60 115.29 8-Jan-2007 57.9 ± 1.8 -34.94 116.08 9-Jan-2007 58.6 ± 15.7 -35.20 117.97 11-Jan-2007 63.0 -34.57 118.97 12-Jan-2007 15 59.9 ± 2.3 -34.57 118.98 ratio indicated the time completing movements in and out of the ARS area Time, distance, and speed (rate of movement) were calculated for each movement recorded between pairs of receivers [6] Time was calculated as the period between the last detection at one receiver and the first detection at the next receiver Distance was measured between the positions of the two relevant receivers Residence times for SBT were based on survival analyses conducted with the Kaplan-Meier method The residence time was defined as the duration in which a tagged fish stayed within detection range of all receivers [13] We used the Breslow-Gehan-Wilcoxon test to determine differences in the data, since the hazard functions of the dataset were not parallel, which is a requirement of the logrank test [14] The statistical analysis was performed using the statistical software package StatView 5.0 (SAS Institute, Cary, NC, USA), and a p value less than 0.05 indicated statistical significance 123 Fork length, cm (mean ± SD) Results Detections and movement A total of 60 (86%), 58 (83%), and 62 (89%) receivers were retrieved each year and provided sufficient spatial coverage to determine SBT movement between receivers The total number of tagged SBT detected at the receivers was 55 (70%) in 2004/2005, 68 (84%) in 2005/2006, and 62 (73%) in 2006/2007 There was no significant difference between the types of tagged and detected fish (U test, p \ 0.05) The total numbers of migrating SBT movements between receivers were 2,744, 416, and 662, respectively The proportion of migration movements that occurred between receivers at lumps was 93% in 2004/2005, 21% in 2005/2006, and 68% in 2006/2007; conversely, 7, 79, and 32% occurred between cross-shelf lines Fish Sci (2010) 76:403–410 407 Spatial movements in ARS area 100 Time spent ratio (%) Tagged SBT in 2004/2005 and 2006/2007 migrated mostly around the lumps (very little shelf migration) while in 2005/2006 a large number of fish movements were not only between lumps and the cross-shelf lines, but also between cross-shelf lines (Table 2) The estimated proportion of time spent out of the ARS area (84.5 and 65.0%) during the 2004/2005 and 2006/2007 seasons was greater than that within the ARS area (15.5 and 35.0%) On the other hand, unlike the 2004/2005 and 2006/2007 seasons, tagged SBT during 2005/2006 spent more time in the ARS area (63.3%) than out of ARS area (36.7%) (Fig 2) 80 60 40 20 In Out In Out In Out 2004/05 2005/06 2006/07 Residence patterns Fig Proportion of time spent in and out of the acoustic recruitment-monitoring survey area (ARS area), calculated as the movement time of tagged fish between pairs of receivers While the tagged SBT were detected at cross-shelf lines and lump receivers in all three research seasons, differences in spatial distribution patterns and residence time for tagged fish were found between years In the 2004/2005 season, tagged SBT were present mostly at the lumps every day during the research period, and 91.3% of all detections (n = 27,855) were recorded by the receivers deployed at the lumps, although a small number were recorded at the cross-shelf lines (Fig 3a) In the 2005/2006 season, tagged SBT were widely distributed across the cross-shelf lines, and there were few fish detected at the lumps The total number of detections was 5,214, of which 93% occurred at cross-shelf lines (Fig 3b) In the 2006/2007 season, 88% of all detections (n = 18,514) occurred at the lumps, and 12% occurred at the cross-shelf lines (Fig 3c) Residence patterns in this region indicate two migration pathways: (1) inshore migration during both the 2004/2005 and 2006/ 2007 seasons, when site fidelity appeared high, and (2) shelf migration during the 2005/2006 season when site fidelity was relatively low The number of tagged fish remaining in the research area declined over time in each season (Fig 4) Some tagged fish were never detected, although half of the detected fish remained in the research area for the next 55.7 days following tagging and then moved out of the area in the 2004/2005 season In contrast, in the 2005/2006 season, half of the detected fish moved out of the area within 14.1 days following tagging; however, 18% of the tagged fish remained for over 100 days after tagging The residence time in the 2004/2005 season was significantly Table The number of movements of tagged juvenile SBT Thunnus maccoyii recorded between acoustic receivers and the time spent in and out of the acoustic recruitment-monitoring survey (ARS) area Migration route 2004/2005 No of migrations Line 1–lumps 40 Line 2–lumps Line 3–lumps 2005/2006 Time (days) In area Out of area No of migrations 2006/2007 Time (days) In area 4.4 210.0 30 23 37.6 126.1 14 49.3 30.6 Line 1–line 11 101.5 42.3 21 140.4 Line 2–line 30.9 38.1 28 529.4 Line 1–line 12.7 4.7 103.7 Between lumps 0.0 Out of area No of migrations Time (days) In area 0.3 Out of area 139.7 28 149.2 25.8 48.7 26 72.9 143.7 103.7 12.4 83.4 111.0 105.4 193.6 77.2 74.7 18 140.9 26.2 12.4 17.7 1.9 88 0.0 287.3 13 0.0 75.1 15 0.0 31.1 104 57.7 189.5 158 262.4 189.3 93 94.9 462.6 Between adjacent receivers along lines 79 1.3 51.6 118 94.6 49.5 88 56.6 43.5 Between adjacent receivers in lumps 2,384 0.0 634.8 30 0.0 22.8 378 0.0 179.2 Total 2,744 295.3 1,615.0 416 1,260.0 730.0 662 660.3 1,225.6 Along lines 123 (a) Fish Sci (2010) 76:403–410 2004/05, n = 27,855 Lumps Line3 70 60 50 40 Line2 30 20 Line1 10 70 60 Line3 50 30 20 Line1 10 (c) 100 2004/05 2005/06 80 2006/07 60 40 20 0 40 2006/07, n = 18,514 Lumps Lnie3 70 60 50 40 Line2 30 20 120 Table Summary of the Kaplan-Meier estimates for survival analysis in three seasons (2004–2007) Season No of SBT Residence time (days) 25% 50% 75% Median value Mean value 2004/2005 55 32.8 55.7 65.5 55.7 50.2 2005/2006 68 5.4 14.1 52.8 14.4 36.3 2006/2007 62 11.3 47.2 81.2 47.3 50.3 12/1 1/1 2/1 3/1 4/1 5/1 6/1 Date Fig Time series of acoustic detections (represented by dots) for all tagged southern bluefin tuna by VR2 receivers a from December 3–5 to March 15–19 in 2004/2005, b from December 1–3 to May 9–11 in 2005/2006, and c from December 1–3 to May 29–31 in 2006/2007 VR2 receiver locations and numbers [no 1–20 (line 1), no 21–40 (line 2), no 41–61 (line 3), and no 62–70 (lumps)] are illustrated in Fig 1a different compared to the 2005/2006 season (BreslowGehan-Wilcoxon test, P \ 0.01) indicating a difference among inshore- and shelf-migration run seasons There were also significant differences between the 2006/2007 season (inshore-migrating) and 2005/2006 season (shelfmigrating) (Breslow-Gehan-Wilcoxon test, P \ 0.05) In 2006/2007 season, half of the detected fish remained in the area for 47.2 days following tagging, which was longer than the shelf-migrating season in 2005/2006 For both inshore-migrating seasons (2004/2005 and 2006/2007), there were no significant differences between the residence time curves (Breslow-Gehan-Wilcoxon test, P [ 0.05) Residence times (days) at 25, 50, and 75% within each season are shown in Table The median values in inshore-migration seasons (55.7 days in 2004/2005, 47.3 days in 2006/2007) were clearly longer than the 123 160 Fig Kaplan-Meier curves for survival analysis of tagged fish in the survey area during three seasons The residence time relationship between the inshore-migrating seasons (open circles 2004/2005, triangles 2006/2007) and the shelf-migrating season (closed circles 2005/2006) was significantly different (Breslow-Gehan-Wilcoxon test, P \ 0.05), however no difference was detected among the two inshore-migrating seasons (Breslow-Gehan-Wilcoxon test, P [ 0.05) Line1 10 80 Days since fish tagged 40 Line2 Receiver location 2005/06, n = 5,214 Lumps (b) Cumulative percentage remaining in area 408 shelf-migration season (14.4 days in 2005/2006) In addition, the release month (December and January) did not influence the residence times of the fish (Breslow-GehanWilcoxon test, P [ 0.05) Discussion The spatial and temporal distribution patterns of juvenile SBT in southern Western Australia could be characterized as two distinctive migration pathways The majority of age1 SBT showed a strong association with coastal topographic features (lumps) in 2004/2005 and 2006/2007 These inshore-migrating SBT mostly moved between the cross-shelf lines and lumps, and among the lumps Conversely, in 2005/2006, shelf migration of SBT occurred at higher rates not only between cross-shelf lines and lumps but also between cross-shelf lines Moreover, residence times in the survey region were characterized by two temporal patterns related to these spatial migration Fish Sci (2010) 76:403–410 pathways The proportion of inshore-migrating SBT that were detected declined progressively with time in the early months of each summer (Dec–Feb); in contrast, shelfmigrating SBT remained on the shelf over a longer period (Dec–Apr) then decreased sharply until about 50% remained in the area Thereafter, SBT left the survey area at a much slower rate Therefore, it is evident that there are major interannual spatial (i.e., migration pathways), and temporal (i.e., residence times) variations in juvenile SBT residence These spatial and temporal differences may contribute to the observed interannual fluctuations in the recruitment abundance indices, given that ARS are conducted in the same survey area every year To monitor the annual recruitment index of age-1 SBT, ARS has been conducted for several summer seasons (1996/1997–2004/2005) For instance, in 2004/2005 season, the sonar transect survey in the ARS area was conducted between January 14 and February 17, 2005 (Itoh et al unpublished data, 2005) When information from the ARS is combined with the spatial and temporal variation observed from acoustic tagging, it is apparent that most of the tagged fish are moving into coastal waters, and hence 84.5% migrated out of the ARS area They would thus be ‘‘hidden’’ from the acoustic survey vessel The recruitment indices calculated by acoustic data of 2004/2005 season were at the lowest level since the 1999/2000 season (T Itoh, unpublished data, 2005) While it should be noted that there are some issues regarding the reliability of the acoustic recruitment indices based on the detection of SBT schools estimated by sonar specialists (T Itoh and S Tsuji, unpublished data, 2004), our results based on tagged fish indicate that the majority of 2004/ 2005 SBT migrated mainly inshore (i.e., out of the ARS area) Therefore, our results suggest that more fish schools would have been detected if the acoustic survey had included more inshore regions In 2006/2007, the majority of the SBT (65.0%) also migrated to inshore areas, which is out of the ARS acoustic survey region, in contrast to 2005/2006, where only a small proportion (36.7%) of SBT migrated out of the survey region Thus, an important factor contributing to variation in the abundance indices could be variation in the proportion of the age-1 SBT population migrating within and out of the ARS area The indices could change on an annual basis as a result of changes in the horizontal distribution of SBT, even with the same total biomass The two different patterns of migration observed in this study could be caused by the water masses of different oceanographic features [4] Fujioka et al [4] showed that the 2004/2005 and 2006/2007 seasons were characterized by a lack of nutrient-rich sub-Antarctic water on the shelf; in contrast, during the 2005/2006 season, the sub-Antarctic water mass intruded from the continental slope into the 409 subsurface layers and had higher surface chlorophyll a concentrations Therefore, during such a season, tagged fish may respond to temporally varying sub-Antarctic water influx, which produces upwelling that in turn would lead to increased abundance of their teleost prey such as sardine and anchovy [15, 16] While large Japanese research vessels (36 m, 315 ton) conducting acoustic surveys have used high performance omni-scan sonar to detect SBT schools in the ARS area, it would be too dangerous for such large vessels to conduct acoustic surveys in the inshore region, as these coastal waters include unsurveyed topographic features An alternative method to estimate SBT recruitment indices, proposed by Itoh et al (unpublished data, 2005), is based on a troll survey This involves detection of SBT via troll-capture along a repeated cross-shelf survey line (‘‘piston line’’) from Bremer Bay (center of the ARS area) to the shelf break, approximately following line of the acoustic receivers (see Fig 1a) These piston line surveys have been conducted since the 2004/2005 season (same effort in all seasons: nine round trips off Bremer Bay during January– February) The results of the three summer seasons (2004/ 2005, 2005/2006, 2006/2007) show an increasing trend in the trolling-based recruitment index [1.20, 2.82, 4.72 (n_school/100 km)] over the three seasons (T Itoh, unpublished data, 2007) However, this trend was different from the tag-based results, which showed the same distribution patterns over the first and third seasons High proportions of migrating SBT around topographic features in 2004/2005 (first) and 2006/2007 (third) seasons may have resulted in lower troll catches and lower indices than during the shelf-migration season in 2005/2006 There was no significant difference in the residence times between the two inshore-migration seasons Therefore, these indices could be interpreted to mean that the abundance of age-1 SBT population in 2006/2007 substantively increased compared with 2004/2005 It is not clear whether the true fish abundance in 2006/2007 was higher than in 2005/2006, as residence times were significantly different during both seasons This hypothesis could be verified by conducting studies over several years regarding fisheries data, such as catch trends by purse seine for SBT of ages 2–4 and by longline for SBT over age This study showed that one cause of interannual fluctuations in abundance estimates could be variation in temporal and spatial distribution patterns It is important to note that the recruitment abundance indices may be underestimated during inshore-migration seasons In order to minimize the interannual bias, we propose that a trolling-based recruitment survey should also be carried out at coastal topographic features and that catch information from both across the shelf and the topographic features be combined into an abundance index 123 410 The abundance estimates may also be affected by the peak time of migration through the survey area In order to correct for variation in the timing of peak population migration across the piston line survey area, the underlying mechanism for the temporal and spatial movement patterns displayed by migratory SBT, perhaps related to oceanographic conditions, must be further investigated Acknowledgments The assistance and cooperation of Geoff Campbell and the crew of F/V Quadrant and St Gerard were greatly appreciated G N Nishihara, Institute for East China Sea Research, Nagasaki University, provided comments that substantially improved the manuscript The research was supported by Fisheries Research Agency, JAMARC, CSIRO Marine and Atmospheric Research and the Australian Fisheries Management Agency as part of the Japan Australia SBT Recruitment Monitoring Program, and by the Japan Society for the Promotion of Science (16255010), and was part of the Japan–Australia Cooperative Research Projects ‘‘Marine Science’’ by the Japan Science and Technology Agency (JST) References Farley JH, Davis TLO (1998) Reproductive dynamics of southern bluefin tuna, Thunnus maccoyii Fish Bull 96:223–236 Greenstreet S, Armstrong E, Mosegaard H, Jensen H, Gibb IM, Fraser H, Scott BE, Holland GJ, Sharples J (2006) Variation in the abundance of sandeels Ammodytes marinus off southeast Scotland: an evaluation of area-closure fisheries management and stock abundance assessment methods ICES J Mar Sci 63:1530– 1550 Hobday AJ, Kawabe R, Takao Y, Miyashita K, Itoh T (2009) Correction factors derived from acoustic tag data for a juvenile southern bluefin tuna abundance index in southern Western Australia In: Nielsen JL, Arrizabalaga H, Fragoso N, Hobday AJ, Lutcavage M, Sibert J (eds) Methods and technology in fish biology and fisheries: tagging and tracking of marine animals with electronic devices Springer, London, pp 405–422 Fujioka K, Hobday AJ, Kawabe R, Miyashita K, Honda K, Itoh T, Takao Y (2010) Interannual variation in summer habitat 123 Fish Sci (2010) 76:403–410 10 11 12 13 14 15 16 utilization by juvenile southern bluefin tuna (Thunnus maccoyii) in southern Western Australia Fish Oceanogr (in press) Klimley AP, Holloway CF (1999) School fidelity and homing synchronicity of yellowfin tuna, Thunnus albacares Mar Biol 133:307–317 Comeau LA, Campana SE, Castonguay M (2002) Automated monitoring of a large-scale cod (Gadus morhua) migration in the open sea Can J Fish Aquat Sci 59:1845–1850 Heupel MR, Semmens JM, Hobday AJ (2006) Automated acoustic tracking of aquatic animals: scales, design and deployment of listening station arrays Mar Freshw Res 57:1–13 Willis J, Hobday AJ (2007) Influence of upwelling on movement of southern bluefin tuna (Thunnus maccoyii) in the Great Australian Bight Mar Freshw Res 58:699–708 Klimley AP, Jorgensen SJ, Muhlia-Melo A, Beavers SC (2003) The occurrence of yellowfin tuna (Thunnus albacares) at Espiritu Santo Seamount in the Gulf of California Fish Bull 101:684–692 Lowe CG, Topping DT, Cartamil DP, Papastamatiou YP (2003) Movement patterns, home range, and habitat utilization of adult kelp bass Paralabrax clathratus in a temperate no-take marine reserve Mar Ecol Prog Ser 256:205–216 Stark KE, Jackson GD, Lyle JM (2005) Tracking arrow squid movements with an automated acoustic telemetry system Mar Ecol Prog Ser 299:167–177 Hobday AJ, Campbell G (2009) Topographic preferences and habitat partitioning by pelagic fishes off southern Western Australia Fish Res 95:332–340 Ohta I, Kakuma S (2005) Periodic behavior and residence time of yellowfin and bigeye tuna associated with fish aggregating devices around Okinawa Islands, as identified with automated receivers Mar Biol 146:581–594 Harrington DP, Fleming TR (1982) A class of rank test procedures for censored survival data Biometrika 69:553–566 Young JW, Bradford RW, Lamb TD, Lyne VD (1996) Biomass of zooplankton and micronekton in the southern bluefin tuna fishing grounds off eastern Tasmania, Australia Mar Ecol Prog Ser 138:1–14 Ward TM, Mcleay LJ, Dimmlich WF, Rogers PJ, McClatchie S, Matthews R, Manpf J, Ruth PDV (2006) Pelagic ecology of a northern boundary current system: effects of upwelling on the production and distribution of sardine (Sardinops sagax), anchovy (Engraulis australis) and southern bluefin tuna (Thunnus maccoyii) in the GAB Fish Oceanogr 15:191–207 Fish Sci (2010) 76:411–434 DOI 10.1007/s12562-009-0201-2 ORIGINAL ARTICLE Biology The role of flatfishes in the organization and structure of the eastern Bering Sea ecosystem Sung Il Lee • Kerim Y Aydin • Paul D Spencer Thomas K Wilderbuer • Chang Ik Zhang • Received: June 2009 / Accepted: 11 November 2009 / Published online: 15 January 2010 Ó The Japanese Society of Fisheries Science 2010 Abstract We evaluated the role of flatfishes in the organization and structure of the eastern Bering Sea ecosystem using the Ecopath/Ecosim approach As basic input data for the Ecopath/Ecosim model, we used estimates of biomass from bottom trawl surveys and age-structured population models, production/biomass (P/B) ratio, consumption/biomass (Q/B) ratio, diet composition (DC), and fisheries harvests for each component of species or species groups We estimated the trophic level of each component, niche overlaps among flatfishes, and the impacts of competition and predation on flatfish species in the eastern Bering Sea ecosystem Based on those estimates, we developed the tropho-dynamic structure of the ecosystem, and the model was used to simulate ecological effects of fishery exploitation patterns No single flatfish species appeared to have a profound and uniquely important role in the organization and structure of the ecosystem Instead, the most important component among the guild of flatfish species appeared to be yellowfin sole Pleuronectes asper, which had greater biomass than other flatfish and a relatively diverse diet among the small flatfish species Pacific halibut Hippoglossus stenolepis, Greenland turbot Reinhardtius S I Lee (&) East Sea Fisheries Research Institute, National Fisheries Research and Development Institute, Gangnung 210-861, Korea e-mail: silee@nfrdi.go.kr K Y Aydin Á P D Spencer Á T K Wilderbuer Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way NE, Bldg 4, Seattle, WA 98115, USA C I Zhang Pukyong National University, Daeyeon 3-dong, Nam-gu, Busan 608-737, Korea hippoglossoides, and arrowtooth flounder Atheresthes stomias were important keystone predators in the eastern Bering Sea ecosystem together with some groups of marine mammals and sea birds Intra flatfish complex cannibalism was not observed, however, substantial diet overlaps were common in the flatfish guild system Keywords Eastern Bering Sea Á Ecopath Á Ecosim Á Flatfishes Á Food web Introduction The eastern Bering Sea has supported considerable numbers of fishes, crustaceans, and marine mammals for commercial harvest over the past century The groundfish complex is the most abundant fisheries resource, totaling more than 21 million t of exploitable biomass and contributing about million t of catch each year [1] However, despite a variety of regulations to protect these biological resources, some species have undergone large population fluctuations, at least partly due to commercial exploitation of mammals, fish, and invertebrates [2] Large-scale international fisheries for groundfish developed rapidly, and since 1960 groundfish resources have been exploited intensively [3] From the mid-1950s to the early 1970s, large populations of both fish and mammals (particularly the large whales) were dramatically reduced On the other hand, walleye pollock Theragra chalcogramma abundance increased dramatically in the Bering Sea during the late 1960s and early 1970s, and again during the early 1980s, along with a number of other groundfish species (e.g., Pacific cod Gadus macrocephalus and flounders) It may have been in some way linked to overexploitation and reduction of these other populations In fact, there is little 123 532 a Standard mixture, 10 µg/ml 0.04 0.02 0.00 Signal intensity (AU) room light The room light was turned off h after the light change, as the room was automatically kept on an 11 h dark:13 h light cycle Matching cages with injected mice were shielded from light throughout the assay The condition of the mice was observed at 3, 6, and 24 h after the injection The second experiment was designed to confirm that pyropheophorbide was required for the onset of symptoms For this, ml of extract from the high-pyropheophorbide homogenate (Oshamanbe, March 2006) or from the lowpyropheophorbide homogenate (Abuta, 26 June 2006) suspended in saline containing 1% Tween 60 was intraperitoneally injected into two mice per cage, and ml of saline containing 1% Tween 60 was also injected into two mice per cage to estimate the effect of the surfactant saline in mice The mice were exposed to strong light as described above The condition of the mice was observed at 3, 6, and 24 h after the injection Fish Sci (2010) 76:529–536 10 20 30 40 20 30 40 b Chlorophilid-a, 0.71 µg/ml 0.01 0.00 10 c Scallops (midgut glands), Oshamanbe, Mar 7, 2006 0.06 0.04 0.02 0.00 10 Results Figure shows the reversed-phase HPLC profiles of six standard pigments and of the extract from the midgut glands of scallops on ODS The standard pigments of chlorophillid-a, pheophorbide-a, pyropheophorbide-a, chlorophyll-a, pheophytin-a, and pyopheophytin-a were eluted in this order from the column with good separation within 30 after sample injection under the conditions employed (Fig 2a, b) The same elution order with a longer elution time on reversed-phase columns has been observed in plant extracts [11] The retention times of five peaks detected in the extract from the midgut glands of scallops (10.11, 12.15, 19.61, 24.39, and 28.87 min, Fig 2c) were in close agreement with those of standard pigments The absorbance spectrum of the predominant component was the same as that of standard pyropheophorbide-a (Fig 3) Pyropheophorbide content in the midgut glands of scallops Figure shows the amount of chlorophyll-a, pyropheophorbide-a, pheophorbide-a, and pheophytin-a in homogenates of the midgut glands of cultured Japanese scallops that were sampled in Funka Bay from 24 January to 26 April 2006 The total content of the four pigments in the Otoshibe samples started to increase at the beginning of February 2006 and reached a peak (637 lg per 1-g homogenate) on 20 February 2006 After the peak, the amounts of pigment began to decrease and by 26 April 123 30 40 Fig HPLC chromatograms of standard pigments (a, b) and of the extract from the midgut glands of scallops (c) recorded at 650-nm wavelength with photodiode-array detector Peak identification: pheophorbide-a, pyropheophorbide-a, chlorophyll-a, pheophytin-a, pyopheophytin-a, chlorophillid-a HPLC conditions as given in text 0.20 (0.4) Absorbance (AU) HPLC separation and absorbance spectra of pigments 20 Time (min) 0.15 (0.3) Scallops (midgut glands), Oshamanbe, Mar 7, 2006 (the peak at 12.15 min, Fig 2c) 0.10 (0.2) Pyropheophorbide, 10 µg/ml (the peak at 12.17 min, Fig 2a) 0.05 (0.1) 0.00 350 400 450 500 550 600 650 700 Wavelength (nm) Fig Comparison with photodiode-array detector between the absorbance spectrum of a standard preparation and that of a sample The spectrum of the peak with the same retention time as pyropheophorbide-a in the 650-nm chromatogram of the extract from scallop midgut glands was similar to that of pyropheophorbide-a The number in parentheses on the vertical axis is the scale of the absorbance for the scallop sample 2006 fell to a level as low as at the start of the sampling Pigment content of the Oshamanbe samples did not peak as clearly as the Otoshibe samples, and the pigment content Contents of pigments (µg/g) Fish Sci (2010) 76:529–536 700 Otoshibe a 600 Phe Chlo pyrPheid Pheid b 500 400 300 510 530 380 200 280 100 280 110 2006 1/24 Contents of pigments (µg/g) 533 700 100 2/6 2/20 3/7 3/27 4/10 330 360 4/26 Oshamanbe 600 500 b 400 a 300 200 100 300 No data 2006 1/24 340 230 95 2/6 2/20 3/7 3/27 4/10 4/25 Sampling date Fig Contents of chlorophyll and three pheopigments in homogenates of midgut glands of scallops sampled for inspections of diarrhetic shellfish poisoning in Funka Bay from 24 January to 26 April 2006 Phe Pheophytin-a, Chlo chlorophyll-a, Pheid pheophorbide-a, pyrPheid pyropheophorbide-a Value in column indicates content (lg) of pyropheophorbide-a per g of the homogenate a Mice intraperitoneally injected with the extract from this sample showed unusual symptoms: piloerection, substantial eye discharge, swelling of the ears and head, and ear scratching, in the DSP inspection b A few mice intraperitoneally injected with the extract from this sample showed only substantial eye discharge in the DSP inspection began to increase approximately weeks later than in the Otoshibe area The sample on March 2006 had the highest total pigment content (399 lg per 1-g homogenate), but the highest value was approximately 40% lower than in the Otoshibe samples Pyropheophorbide-a, which is known to be a photosensitizer, accounted for 68–92% of the total pigment content during the pigment-increasing phase, whereas pheophorbide-a, which is similarly harmful, accounted for only 3–4% of the total pigment content Pheophytin-a and chlorophyll-a, respectively, accounted for 6–21 and 2–12% of the total pigment content in the same period The extracts that induced the apparent symptoms of photosensitive reaction in mice were from scallops collected in both fishing areas on 21 February 2006, i.e., in the high pyropheophorbide period Results of mouse bioassays In the bioassay for the light requirement for the symptoms, mice that were injected with the pyropheophorbide-rich extract (Otoshibe, 510 lg/g of the homogenate) and exposed to light were dead at h after the injection, whereas the mice in all the other groups were alive with no symptoms after 24 h In the bioassay for pyropheophorbide requirement for the symptoms, all of the mice that received extract from the high-pyropheophorbide homogenate (Oshamanbe, 340 lg/g of the homogenate), from the low-pyropheophorbide homogenate (Abuta, 36 lg/g of the homogenate), or saline only (control group) were still alive after h under the light, but the mice in the high-pyropheophorbide group had symptoms of swelling of the ears (Fig 5c) and head (Fig 5d), and ear scratching They were dead at 24 h after the injection, whereas the mice in the other groups were alive with no symptoms after 24 h Pyropheophorbide content in the midgut glands of other edible bivalves The pigment contents of nonscallop bivalves collected in Hokkaido were also analyzed Table shows the amount of each of the pigments in the homogenate of the midgut glands Nonscallop bivalves also contained pyropheophorbide, but, with the exception of two surf clams, Pseudocardium sachalinense, the amounts were less than in the scallops The surf clams sampled in the Otsu area, off the Pacific coast of eastern Hokkaido on February 2007, had approximately 300 lg of total pigment per g of midgut gland homogenate, but the pyropheophorbide content was only 88 lg The surf clams collected in Tomakomai, off the Pacific coast of eastern Hokkaido on 20 February 2007 contained almost the same amount of pyropheophorbide as the Otsu clams Nonscallop bivalves, including the two surf clams, caused no symptoms in mice in the DSP inspections Additionally, pyropheophorbide content of scallops collected in early spring was high, approximately 300–500 lg/g, but the content in summer was 36–100 lg per g of midgut gland homogenate, and the samples induced no symptoms in mice in the DSP inspections Discussion Our study shows that the midgut gland of the scallop Patinopecten yessoensis, which is a suspension feeder, can potentially cause photosensitivity disease in mice This is similar to the response to the midgut glands of abalones or turban shells, which are herbivorous gastropods We 123 534 Fish Sci (2010) 76:529–536 Fig Symptoms in the mice injected with the extract from high-pyropheophorbide homogenate under intense light a Control mouse, b substantial eye discharge, c swelling of the ears, d swelling of the head Table Contents of chlorophyll and pheopigments in midgut glands of edible bivalves Fishing area Fishing date Content (lg/g of homogenate) Total Chlo a Pheid a pyrPheid a Phe a pyrPhe a Sakhalin surf clam Pseudocardium sachalinense Tomakomai 5-Aug-06 – – – – Nemuro 19-Sep-06 12 – \3 12 – – Ootsu 5-Feb-07 294 77 88 93 28 Tomakomai 20-Feb-07 120 78 22 10 38 – – – 18 10 \3 7 – – – – Japanese littleneck Ruditapes philippinarum Notsuke Akkeshi 26-Jun-06 19-Feb-07 Japanese oyster Crassostrea gigas Mori 11-Nov-06 Lake Saroma 9-Jan-07 4 – – – – Lake Saroma 3-Feb-07 \3 \3 – \3 – – 43 Japanese scallop Patinopecten yessoensis Osyamanbe 7-Mar-06 509 \3 17 420 29 Abuta 26-Jun-06 39 \3 \3 36 – Tokoro 26-Jun-06 110 – 100 Abashiri 21-Aug-06 54 – 51 \3 \3 We also analyzed chlorophyllid-a, but it was below the limit of detection in all samples Limits of quantitation were lg/g of homogenate Only the scallops collected in Osyamanbe on March 2006 caused photosensitivity-like symptoms in the inspection for DSP Chlo a Chlorophyll-a, Pheid a pheophorbide-a, pyrPheid a pyropheophorbide-a, Phe a pheophytin-a, pyrPhe a pyropheophytin-a, – below the limit of detection believe that the atypical symptoms originally observed in the DSP tests were of photosensitivity disease caused by pyropheophorbide in midgut glands, but we have no direct 123 evidence of this, as we did not measure the concentration of pyropheophorbide in the blood of the mice However, the HPLC measurements revealed that photosensitivity Fish Sci (2010) 76:529–536 disease symptoms in mice were induced by extracts from scallops collected during the high pyropheophorbide period The pyropheophorbide content was 300–530 lg per g of the homogenate of midgut glands of scallops The weight of the homogenate used for the assay corresponded to 3.28–4.28 g per injection; therefore, the applied dose of pyropheophorbide per mouse with 20 g body weight was 0.99–2.3 mg in the calculation It has been reported that the lethal dose of pyropheophorbide in mice exposed to strong light (10,000 lux) is approximately 3.3–5.0 mg per 100 g body weight by injection into the abdominal cavity [14–17], which equates to 0.66–1.0 mg per 20 g body weight Therefore, the content of pyropheophorbide in our bioassay that we suspect caused the onset of photosensitivity was enough to induce death in a mouse Additionally, from the comparative mouse bioassays, it is clear that a substantial amount of pyropheophorbide and exposure to light are necessary for the onset of the symptoms However, not all mice that were injected with highpyropheophorbide extract, for example, from the Otoshibe or Oshamanbe samples taken March 2006, showed strong photosensitivity in the actual DSP inspections (Fig 4, column indicated with b, only eye discharge) Meanwhile, the content of pyropheophorbide in the Oshamanbe sample from 20 February 2006 was almost equal to that of the above-mentioned Oshamanbe sample from March 2006, but swelling of the ears and head and ear scratching were observed in mice injected with it (Fig 4, column indicated with a) We attribute the failure to cause symptoms to differences in illuminance due to the location of the mouse cage When the illuminance of our cage rack was measured, almost all the cages were indirectly illuminated by approximately 65–400 lux with fluorescent lights on the ceiling, but there were some locations at the end of the rack where part of a cage was exposed to light directly, approximately 600–1,100 lux Our results suggest that pyropheophorbide in the midgut glands of scallops induced photosensitization in mice The total content of four pigments in our samples peaked in late February to early March in Funka Bay The increase reflects a concurrent increase in diatoms in the culture area The cultured scallops in Funka Bay are strongly influenced by the diatom bloom, which occurs near the sea surface, because the scallops are cultured, hanging between 10 and 25 m below the sea surface The diatom bloom in this area commonly occurs between late February and mid-March [18, 19] Baba et al and Kawano et al reported an increased amount of chlorophyll in the seawater at different points in Funka Bay between February and March 2006 [20, 21] The origin of the pyropheophorbide we detected in the midgut glands of scallops might be chlorophyll in the diatoms ingested by scallops, if the chlorophyll is metabolized into mainly 535 pyropheophorbide in the metabolic pathway of scallops Alternatively the pyropheophorbide might be derived mainly from metabolites of phytoplankton feeders converging on the bloom, or from the feces and remains that accumulate there In our study, we not have data about the origin of the pyropheophorbide detected in the midgut glands of scallops To clarify the origin, we need to research metabolic pathways of chlorophyll in scallops and the composition of pyropheophorbide and the related substance in the suspended matter that scallops ingest We also measured pigments in some nonscallop bivalves collected in Hokkaido Pyropheophorbide was detected in almost all of the bivalves measured, but it was present in insufficient quantity to cause photosensitivity disease in mice The sample numbers were insufficient for us to rule out a possible risk of pyropheophorbide in edible bivalves causing photosensitivity to humans; however, we think that eating edible bivalves presents a low risk of photosensitivity disease if they are correctly cooked or eaten in normal quantities The midgut glands of large bivalves such as scallops or surf clams are generally removed before eating If the problematic scallops that we detected are eaten, approximately 270–630 g of the whole body with the midgut glands would need to be consumed to induce photosensitization, since the lowest effective level of pheophorbide and pyropheophorbide is 25 mg per person per day [5] Small bivalves such as littleneck clams or short-necked clams are eaten with the midgut glands, but the risk of photosensitization is also low because the midgut glands are small and they are usually not eaten in large quantities Therefore, we think that a normal intake of edible bivalves would cause little photosensitivity disease in humans However, there is reason to be concerned about photosensitization caused by oysters Oysters are more likely to cause photosensitization because they have comparatively larger midgut glands than littleneck clams or short-necked clams and are often eaten raw with the midgut glands Some people eat a lot of oysters at restaurants that serve oysters in an all-you-can-eat style Moreover, oysters may also be influenced by the diatom bloom strongly because they are cultured using the hanging method similarly to scallops in Japan Our measured values of pigments in Pacific oysters Crassostrea gigas were low, but the samples may have been collected out of the diatom blooming season Therefore, we need more data on the content of pyropheophorbide in oysters It is necessary to attend to the amount of pyropheophorbide in oysters fished during a diatom bloom if oysters tend to accumulate pyropheophoride in the midgut glands in amounts as high as scallops To secure food safety, we need more data on the content of pyropheophorbide in edible bivalves during diatom blooms We have paid little attention to pyropheophorbide 123 536 levels in edible bivalves in food-safety inspections because we have assumed that only herbivorous gastropods such as abalones cause photosensitivity disease However, our study shows that photosensitivity was caused in mice by pyropheophorbide accumulated in the midgut gland of some scallops We think that levels of pyropheophorbide in most bivalves living in their native habitat is generally too low to induce photosensitivity in animals In fact, the amounts of pyropheophorbide in our nonscallop bivalve samples were low and those of scallops were also not always high However, the chemical may be rapidly accumulated in bivalves for some reason Our study suggests that a diatom bloom is one of the factors for the accumulation Artificial changes in the habitat environment of bivalves, such as use of hanging culture, may also accelerate the accumulation To secure food safety, we need to confirm whether pyropheophorbide in edible bivalves increases to harmful levels or not during diatom blooms Acknowledgments We thank the staff members of the Department of Biological Science in Hokkaido Institute of Public Health for their support and advice We also thank the food sanitation division staffs of all health and welfare offices in Hokkaido, and the staffs of Abashiri, Abashiri Tobu, Hiyama, and Iburi Fisheries Technical Guidance Office, for sampling and dispatching samples References Hashimoto Y, Naito K, Tsutsumi J (1960) Photosensitization of animals by the viscera of abalones, Haliotis spp Fish Sci 26:1216–1221 Hashimoto Y, Tsutsumi J (1961) Isolation of a photodynamic agent from the liver of abalone, Haliotis discus hannai Fish Sci 27:859–866 Hashimoto Y, Tsutsumi J (1963) Dietary photosensitization in animals (in Japanese) J Food Hyg Soc Jpn 4:185–191 Tsutsumi J, Hashimoto Y (1964) Isolation of pyropheophorbide a as a photodynamic pigment from the liver of abalone, Haliotis discus hannai Agric Biol Chem 28:467–470 Tamura Y, Maki T, Shimamura Y, Nishigaki S, Naoi Y (1979) Causal substances of photosensitivity dermatitis due to Chlorella ingestion (in Japanese with English abstract) J Food Hyg Soc Japan 20:173–180 123 Fish Sci (2010) 76:529–536 Endo H, Hosoya H, Koyama T, Ichioka M (1982) Isolation of 10-hydroxypheophorbide a as a photosensitizing pigment from alcohol-treated Chlorella cells Agric Biol Chem 46:2183–2193 Pandey RK, Dougherty TJ (2006) Pyropheophorbides and their use in photodynamic therapy U.S Patent US RE39,094 E Henderson BW, Dougherty TJ (1992) How does photodynamic therapy work? Photochem Photobiol 55:145–157 Oleinick NL, Evans HH (1998) The photobiology of photodynamic therapy: cellular targets and mechanisms Radiat Res 55:145–157 10 Ma L, Dolphin D (1999) The metabolites of dietary chlorophylls Phytochemistry 50:195–202 11 Gauthier-Jaques A, Bortlik K, Hau J, Fay LB (2001) Improved method to track chlorophyll degradation J Agric Food Chem 49:1117–1122 12 Ministry of Health and Welfare, Japan (1981) Method of inspecting for diarretic shellfish poison, in notice Kannyu no 37 (in Japanese) Sanit Res 31:565–571 13 Tu AT (1992) Food poisoning (handbook of natural toxins) Marcel Dekker, New York, p 445 14 Yamada K, Nakamura N (1972) Photodynamic agent of pickled green (in Japanese with English abstract) J Jpn Soc Food Nutr 25:466–471 15 Isobe A, Kimura S (1976) Appearance of hypersensitiveness in different species (in Japanese with English abstract) J Jpn Soc Food Nutr 29:221–224 16 Isobe A, Kimura S (1976) Effect of light-intensity and irradiation way on hypersensitiveness by photodynamic agent (in Japanese with English abstract) J Jpn Soc Food Nutr 29:225–227 17 Miki K, Tajima O, Matsuura E, Yamada K, Fukimbara T (1980) Isolation and identification of a photodynamic agent of Chlorella (in Japanese with English abstract) J Agric Chem Soc Japan 54:721–726 18 Shimada H, Nishida Y, Ito Y, Mizushima T (2000) Relationship among growth and survival of cultured scallops (Patinopecten yessoensis Jay), and environmental conditions in the coastal area off Yakumo, Funka Bay, Hokkaido, Japan (in Japanese with English abstract) Sci Rep Hokkaido Fish Exp Stn 58:49–62 19 Baba K, Miyazono A (2006) Heisei 16 nendo jigyou houkokusyo (annual report for 2005) (in Japanese) Hokkaido Hakodate Fisheries Experiment Station, Hakodate, pp 73–75 20 Baba K, Sugawara R (2007) Heisei 17 nendo jigyou houkokusyo (annual report for 2006) (in Japanese) Hokkaido Hakodate Fisheries Experiment Station, Hakodate, pp 74–76 21 Kawano K, Yoshida S (2007) Heisei 17 nendo jigyou houkokusyo (annual report for 2006) (in Japanese) Hokkaido Hakodate Fisheries Experiment Station, Hakodate, pp 134–138 Fish Sci (2010) 76:537–546 DOI 10.1007/s12562-010-0236-4 ORIGINAL ARTICLE Food Science and Technology Effect of extracts from narezushi, a fermented mackerel product, on cholesterol metabolism in Wistar rats Kouji Itou • Yoshiaki Akahane Received: 26 October 2009 / Accepted: 26 February 2010 / Published online: 14 April 2010 Ó The Japanese Society of Fisheries Science 2010 Abstract Narezushi extract was separated into peptide and nonpeptide fractions by ion-exchange column chromatography The narezushi extract and fractions were administered to rats in a diet enriched with lipid and cholesterol for 30 days In the narezushi extract and nonpeptide fraction groups, increases in triglyceride, total cholesterol, and low-density lipoprotein-cholesterol levels in the plasma and accumulation of total lipids and triglyceride in the liver were suppressed, while both lipid and cholesterol fecal excretion were increased In the peptide fraction group, these effects were also observed, except for the suppressing effect on liver lipid accumulation Narezushi extract administration tended to increase fecal bile acids and promoted the activity of cholesterol 7a-hydroxylase, the rate-limiting enzyme in the synthesis of bile acid from cholesterol in the liver However, the activity of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, the rate-limiting enzyme in the cholesterol synthesis system in the liver, decreased due to regulation by the feedback of lipid transportation from diet to the liver These results suggest that both the increase in cholesterol and bile acid fecal excretion and the promotion of cholesterol 7a-hydroxylase activities are related to the hypocholesterolemic effects of narezushi extract Amino acids and organic acids, which are abundantly contained in the nonpeptide fraction, seemed to have more intensive hypocholesterolemic effects than peptides existing in the peptide fraction K Itou (&) Á Y Akahane Department of Marine Bioscience, Faculty of Marine Bioscience, Fukui Prefectural University, Obama, Fukui 917-0003, Japan e-mail: kitou@fpu.ac.jp Keywords Bile acid Á Cholesterol Á Cholesterol 7a-hydroxylase Á Narezushi Á Wistar rat Introduction A hypolipidemic effect of various foods, such as vegetables [1–4], royal jelly [5], and milk products [6], has been reported Hyperlipidemia is one of the risk factors with a close relationship to circulatory organ diseases Recently, the physiological functions of fermented foods have also drawn attention, with extensive research particularly on the hypolipidemic effect of sake cake [7], natto [8], kefir [9], and katsuobushi [10], as have other physiological functions, such as antihypertensive [11–14] and antioxidative effects [8, 15–18], and atherosclerosis-preventing activity [19, 20] We previously reported the antihypertensive [21] and hypolipidemic [22] effects of heshiko and the antihypertensive effect of narezushi [23] Heshiko and narezushi are fermented mackerel products that are abundantly produced in the area around Wakasa Bay in Japan Heshiko is prepared from salted mackerel with a high concentration of NaCl (C20%) through fermentation with rice bran for at least months, while narezushi is prepared from salted mackerel with a relatively low concentration of NaCl (approximately 5%) through fermentation with boiled rice for at least months We reported that a large amount of free amino acids, peptides, and organic acids accumulated in narezushi [24] and heshiko [25] during processing As oral administration of extracts from narezushi and heshiko was previously shown to decrease the systolic blood pressure (SBP) of spontaneously hypertensive rats (SHR) [21, 23], these extracts were separated into two fractions: a peptide-rich fraction (peptide fraction) and a fraction enriched with free amino 123 538 acids and organic acids (nonpeptide fraction) for administration to SHR The peptide fraction with strong angiotensin I-converting enzyme (ACE) inhibitory activity decreased SBP in SHR to the same level as that for the extract, whereas the nonpeptide fraction with little ACE inhibitory activity also decreased SBP in SHR Therefore, we assumed that the nonpeptide fraction was involved in additional blood pressure regulatory systems other than the renin-angiotensin system (which controls the shrinkage of veins), and we subsequently examined the metabolic changes in plasma lipid levels following administration of narezushi extract and its fractions to rats In our previous report [22], we found that the heshiko extract, peptide fraction, and nonpeptide fraction had hypocholesterolemic effects in plasma via the increase both in cholesterol and bile acid contents in feces of Wistar rats The increase in both fecal cholesterol and bile acid levels and the progress of cholesterol 7a-hydroxylase (the rate-limiting liver enzyme in bile acid synthesis from cholesterol) activity were also observed The objective of the present study was to elucidate the effect of narezushi extract on cholesterol metabolism in rats, as narezushi differs from heshiko in both submaterials and processing method Thus, the cholesterol levels in blood, liver, feces, and the enzymatic activity in the liver relating to cholesterol catabolism were examined in Wistar rats fed a diet enriched with fat containing narezushi extract or its fractions Materials and methods Preparation of narezushi extract Extract from narezushi, manufactured in the laboratory, was prepared according to the previously reported method [23, 24] Briefly, 10 g dorsal white muscle of narezushi was homogenized with 100 ml distilled water The homogenate was heated at 100°C for min, cooled to room temperature, and then centrifuged at 10,0009g for 20 at 2°C After separating the first supernatant, 50 ml distilled water was added to the precipitate to obtain a second supernatant in the same manner These two supernatants were mixed and passed through filter paper (no 2; Toyo, Tokyo, Japan), and the filtrate was diluted to 200 ml with distilled water to produce the narezushi extract Fractionation of narezushi extract by ion-exchange chromatography A glass column (diameter, cm; length, 20 cm) packed with a cation-exchange resin (TSK-Gel SP-650C, H? type; Tosoh, Tokyo, Japan) was equilibrated with 5% acetonitrile (ACN) This resin was comprised of a vinyl polymer 123 Fish Sci (2010) 76:537–546 carrier and a sulfopropyl ion-exchange residue A glass column containing the anion-exchange resin (TSK-Gel DEAE-650C, OH- type; Tosoh) was similarly equilibrated This resin comprises a vinyl polymer carrier and a trimethylammonium ion-exchange residue The hot-water extract of narezushi (250 ml; peptide content, mg/ml) was mixed with ACN to a final concentration of 5%, after adjusting the pH to 4.0, and applied to the cation-exchange resin The column was washed with 1.25 l of 5% ACN, and washes were collected as the ACN-1 fraction The sample was then eluted with 1.25 l of M NH4OH containing 5% ACN, and the eluate was collected as the NH4OH fraction The NH4OH fraction was lyophilized, dissolved in 95 ml distilled water, and the pH was adjusted to 9.0 using NaOH ACN was subsequently added to a final concentration of 5%, and the mixture was applied to the anionexchange resin and washed with 1.25 l of 5% ACN Washes were collected as the ACN-2 fraction The sample was eluted with 1.25 l of N HCl containing 5% ACN, and the eluate was collected as the NH4OH–HCl fraction ACN-1 and ACN-2 fractions were combined to yield the ACN fraction Both ACN and NH4OH–HCl fractions were lyophilized and dissolved in distilled water The fractions were adjusted to pH 7.0, diluted to 100 ml with distilled water, and stored at -20°C Measurement of free amino acid and peptide contents Free amino acid content in the narezushi extract and fractions was determined using the Pico-Tag system (600E; Waters, Milford, MA, USA) The peptide content in the narezushi extract and fractions was determined according to the method of Lowry et al [26], using bovine c globulin as the standard Peptide content was also evaluated using the same system after hydrolysis in vacuo with M HCl at 150°C for h Preparation of experimental diet for breeding rats The experimental diet was prepared according to the method of Igarashi et al [4] and its composition is shown in Table The amount of narezushi extract added to the diet was equivalent to 50 mg of peptides/kg body weight/day The quantity of NH4OH–HCl and ACN fractions added to the experimental diet was equivalent to the amount of narezushi extract mixed in the experimental diet The protein and other components in the narezushi extract and fractions were replaced by casein and sucrose, respectively Breeding rats Six-week-old Wistar S/T rats (SLC Japan, Hamamatsu, Japan) were divided into four groups of five individuals Fish Sci (2010) 76:537–546 Table Composition of four kinds of experimental diets mixed with narezushi extract or its fractions for feeding to rats (%) a AIN-93 (Oriental Yeast, Tokyo) was used as the vitamin mixture b AIN-93G (Oriental Yeast, Tokyo) was used as the mineral mixture 539 Control diet Narezushi extract diet NH4OH–HCl fraction diet (peptide fraction) ACN fraction diet (nonpeptide fraction) Narezushi extract – 1.1 – – NH4OH–HCl fraction – – 0.2 – ACN fraction – – – 1.0 Sucrose 21.3 20.7 21.2 20.7 Cornstarch 40.0 40.0 40.0 40.0 Gluten 10.0 10.0 10.0 10.0 Casein 10.0 9.5 9.9 9.6 Corn oil 10.0 10.0 10.0 10.0 Free cholesterol 0.5 0.5 0.5 0.5 Vitamin mixturea 1.0 1.0 1.0 1.0 Mineral mixtureb Sodium chloride 3.3 0.7 3.3 0.7 3.3 0.7 3.3 0.7 Choline chloride 0.2 0.2 0.2 0.2 Water 3.0 3.0 3.0 3.0 each Before breeding with the experimental diet, animals were raised for week with ad libitum access to artificial diet (CE-2; CLEA Japan, Osaka, Japan) and sterilized tap water under the following conditions: temperature, 23 ± 2°C; relative humidity, 55 ± 5%; light/dark cycle, 12:12 h The experimental diet was then provided for 30 days The care and treatment of experimental animals conformed to the guidelines of the Science Council of Japan Preparation of analytical samples for lipid level determination After rats were killed under diethyl ether anesthesia, blood and liver were collected To prevent blood coagulation, 0.4 ml of 3.2% sodium citrate solution was added to 3.6 ml blood and stirred well Plasma was then separated by centrifugation at 7359g for 10 at 5°C Liver weight was measured after washing with cold physiological saline and then cold 1.1% potassium chloride solution to remove any adhered blood Feces, collected for days before the end of breeding and feeding, were lyophilized and crushed to prepare samples for analysis These samples were kept at -80°C until measurement of lipid levels Measurement of lipids in plasma, liver, and feces Levels of total cholesterol, high-density lipoprotein (HDL)cholesterol, and triglyceride in plasma were enzymatically measured using a commercial kit (Wako, Osaka, Japan) The level of low-density lipoprotein (LDL)-cholesterol was calculated by subtracting the HDL-cholesterol level from the total cholesterol level Total lipids in the liver were extracted using the method of Folch et al [27] Approximately g liver was homogenized in a glass homogenizer with 10 ml chloroform– methanol solution (2:1 by volume) Homogenates that had been defatted by chloroform were filtered using filter paper (no 5B; Toyo) The supernatant was diluted to 50 ml with the same chloroform–methanol solution After adding 10 ml of 0.04% CaCl2 to the supernatant and leaving to stand overnight, the upper water layer was removed, and 10 ml chloroform–methanol–water solution (3:48:47 by volume) was added to the lower chloroform layer This procedure was repeated three times to purify the extraction of lipids The lower layer was transferred to another flask, dried to remove the chloroform, and the weight of the residue was measured as total lipid content Extracted lipids were dissolved in isopropyl alcohol containing 10% Triton X-100 and used as the samples for cholesterol and triglyceride measurements Liver cholesterol and triglyceride levels were also measured enzymatically using a commercial kit (Wako, Osaka, Japan) Fecal lipids were also extracted by the method of Folch et al [27] and the weight of the residue dissolved in the chloroform layer was measured as total lipid content after drying Cholesterol and bile acid were extracted according to the method of Sautier et al [28] Next, 10 ml ethanol (heated to 78°C) was added to g feces, and the supernatant was transferred to another beaker After repeating this operation three times, supernatants were mixed, filtered in the same manner as in the extraction of liver lipids, and the extract was diluted to 50 ml with ethanol Total cholesterol, HDL-cholesterol, triglyceride, and bile acid levels were enzymatically measured using a commercial kit (Wako) The LDL-cholesterol level was calculated by subtracting HDL-cholesterol level from total cholesterol level 123 540 Fish Sci (2010) 76:537–546 Measurement of liver enzyme activity The microsome fraction in the liver was prepared according to the method of Yomogida et al [29] Approximately g liver was homogenized with ml of 0.1 M potassium phosphate buffer solution (pH 7.0), and homogenates were centrifuged at 9,0009g for 20 at 4°C The supernatant was recentrifuged at 105,0009g for 60 at 4°C, and the resultant microsome fraction was diluted to 10 ml using the same buffer Cholesterol 7a-hydroxylase activity in the microsome fraction was measured using the method described by Ogishima and Okuda [30] A 100-ll aliquot of microsome fraction was incubated at 37°C for 20 with 250 ll of 0.l M potassium phosphate buffer (pH 7.4) containing 0.1 mM ethylenediaminetetraacetic acid (EDTA), 20 mM cysteamine–HCl, mM MgCl2, mM sodium isocitrate, 0.075 U of isocitrate dehydrogenase, and 0.5 mM nicotinamide adenine dinucleotide phosphate (NADPH) The reaction mixture was incubated again for 10 at 37°C with 30 ll of 5% sodium cholate solution and 20 ll of 2.2 U/ml cholesterol oxidase dissolved in 10 mM potassium phosphate buffer (pH 7.4) containing 20% glycerol and mM dithiothreitol After stopping the reaction by adding 0.3 ml methanol, ml petroleum ether was added to the reaction mixture to extract 7a-hydroxycholesterol A 1.5-ml petroleum layer portion was dried in vacuo The residue was resolved in 250 ll n-hexane–isopropanol mixture (82:18 by volume) and analyzed by high-performance liquid chromatography (HPLC) Analysis was performed using HPLC equipped with a normal-phase column (TSK-Gel Silica-60, 4.6 250 mm; Tosoh) with the thermostat set at 25°C The n-hexane–isopropanol mixture was eluted at a rate of 0.8 ml/min, and absorbance of the effluent at 214 nm was monitored One unit was defined as the amount of cholesterol 7a-hydroxylase required to produce nmol of 7a-hydroxycholesterol per minute 3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase activities were measured according to the method of Sung et al [31] As the blank test, 600 ll sodium phosphate buffer (pH 6.8) containing 400 mM NADPH, 200 mM NaCl, mM EDTA, and 20 mM dithiothreitol was added to 300 ll microsome fraction The change in absorbance of this mixture at 340 nm was monitored for while heating at 37°C As the sample test, 400 ll sodium phosphate buffer (pH 6.8), 200 ll of 400 mM HMG-CoA solution dissolved in the same buffer, and 300 ll microsome fraction were mixed, and changes in absorbance were measured as described above The activity of HMG-CoA reductase was calculated from the difference in the change of absorbance per minute between the sample test and the blank test NADP was used as a standard, and one unit was defined as the amount of enzyme required to oxidize lmol of NADPH for Statistical analysis Experimental data were analyzed statistically by Tukey’s method [32], and significant differences of the means were examined at p \ 0.05 Results Effect of narezushi extract on plasma lipid levels Table shows free amino acid and peptide contents in the narezushi extract and the fractions separated from the extract by ion-exchange column chromatography The NH4OH–HCl and the ACN fractions contained 3.0 and 50.0 mg free amino acid per gram of narezushi muscle, respectively This means that 5.6 and 90.5% of the amino acids were transferred to the NH4OH–HCl fraction and the ACN fraction, respectively Although analytical data are not shown here, about 95% of organic acids such as lactic and acetic acids in the narezushi extract were confirmed to be transferred to the ACN fraction Peptide contents in the NH4OH–HCl and ACN fractions were 23.6 mg/g (83.2%) and 3.5 mg/g (12.4%), respectively, as measured by the Pico-Tag system and 22.4 mg/g (87.7%) and 3.1 mg/g Table Amount and ratio of free amino acid and peptide in two fractions separated from narezushi extract by ion-exchange column chromatography (mg/1 g of narezushi) Narezushi extract (mg/g) a NH4OH–HCl fraction (peptide fraction) ACN fraction (nonpeptide fraction) mg/g mg/g % % 55.2 3.0 5.6 50.0 90.5 a Peptide 28.3 23.6 83.2 3.5 12.4 Peptideb 25.5 22.4 87.7 3.1 12.0 Free amino acid a Measured by Pico-Tag system b Measured by the method of Lowry et al [26] 123 Fish Sci (2010) 76:537–546 541 Table Composition of free amino acids in two fractions separated from narezushi extract by ion-exchange column chromatography (mg/ g of narezushi) Narezushi extract NH4OH–HCl fraction (peptide fraction) ACN fraction (nonpeptide fraction) Asp 3.9 0.2 3.6 Glu 6.1 0.3 5.7 Hyp 0.2 0.0 0.2 Ser 1.9 0.1 1.7 Gly 2.6 0.1 2.5 His 4.1 0.3 3.4 Arg 1.3 0.1 1.1 Thr 2.8 0.1 2.5 Ala 4.6 0.2 4.2 Pro 1.1 0.1 1.0 Tyr 2.1 0.1 1.8 Val 2.5 0.1 2.2 Met 2.1 0.1 1.8 Cys Ile 0.0 2.8 0.0 0.1 0.0 2.6 Leu 6.0 0.5 5.5 Phe 4.1 0.1 3.8 Lys 5.9 0.4 5.4 Tau 1.1 0.1 1.0 Total 55.2 3.0 50.0 (12.0%), respectively, by the method of Lowry et al [26] Therefore, if necessary, the NH4OH–HCl and ACN fractions are described, respectively, as the peptide and nonpeptide fractions Table shows the composition of free amino acids in the narezushi extract and fractions Recovery of each free amino acid from the narezushi extract to the ACN fraction ranged from 82 to 100% The ACN fraction from narezushi extract displayed practically no ACE inhibitory activity (data not shown), similar to that from heshiko extract [22] As shown in Table 4, few differences were seen among the rat groups in terms of food intake, body weight gain, and food efficiency during administration of narezushi extract and its fractions for 30 days Changes in plasma lipid levels in rats after 30 days of administration are shown in Table In the control group, total plasma cholesterol level of rats tended to increase, from 70.8 mg/100 ml at the start (0 day) to 79.2 mg/ 100 ml at day 30, although this increase was not significant The LDL-cholesterol level increased markedly from 33.6 to 52.0 mg/100 ml; the HDL-cholesterol level, however, did not show a significant change during administration Total cholesterol levels at the end of administration in the narezushi extract (58.4 mg/100 ml), NH4OH–HCl fraction (52.3 mg/100 ml), and ACN fraction (59.6 mg/ 100 ml) groups were significantly suppressed compared to that in the control group (79.2 mg/100 ml) The LDLcholesterol levels in these groups were also significantly lower than that in the control group, below 30 mg/100 ml The triglyceride level in the control group at day (29.7 mg/100 ml) increased appreciably to 139.3 mg/ 100 ml at day 30, while those in the narezushi extract (69.8 mg/100 ml), NH4OH–HCl fraction (69.6 mg/ 100 ml), and ACN fraction (55.3 mg/100 ml) groups were significantly suppressed Effects of narezushi extract on liver lipid levels Table shows changes in liver weight and lipid levels in rats between the start (day 0) and end of administration (day 30) Liver weights per body weight ranged from 4.1 to 4.7% among the rat groups during this period and no significant differences were observed among these groups Total liver lipid levels significantly increased from 7.3 to 15.6% in the control group, and to 7.5, 10.0, and 16.4% in the narezushi extract, NH4OH–HCl fraction, and ACN fraction groups, respectively The lipid accumulation in the liver at day 30 was significantly lower in the narezushi extract group and tended to be lower, but not significantly, in the NH4OH–HCl and ACN fraction groups compared to the control group Total liver cholesterol content was as low as 0.7 mg/g at day 0, and it increased to 1.9, 1.3, 1.8, and 1.4 mg/g in the control, narezushi extract, NH4OH– HCl fraction, and ACN fraction groups, respectively, at day 30; the level in the narezushi extract and ACN fraction groups was significantly lower than in the control and NH4OH–HCl fraction groups Triglyceride content in the liver markedly increased from 15.5 mg/g at day to 69.0 mg/g at day 30 in the control group, and to 36.2, 71.0, and 52.7 mg/g in the narezushi extract, NH4OH–HCl fraction, and ACN fraction groups, respectively; the level was significantly lower in the narezushi extract group and tended to be lower in the ACN fraction group than in the control group Effects of narezushi extract on fecal lipid levels Changes in fecal lipid levels in rats after 30 days of administration are shown in Table Total fecal lipid levels in the control group tended to decrease, from 19.3 mg/day at day to 14.6 mg/day at day 30 Conversely, the levels in the narezushi extract and ACN fraction groups significantly increased to 30.6 and 33.4 mg/ day, respectively, compared to the control group Total lipid level in the NH4OH–HCl fraction group increased, but not significantly, to 20.4 mg/day Total fecal cholesterol level in the control group at day 30 was 3.9 mg/day, almost the same as at day (4.6 mg/day) Fecal cholesterol 123 542 Fish Sci (2010) 76:537–546 Table Dietary effect of narezushi extract and its fractions on food intake and growth of rats Control group Food intake (g/day) Narezushi extract group 16.4 ± 0.45 NH4OH–HCl fraction group (peptide fraction) 16.0 ± 0.77 16.1 ± 0.15 ACN fraction group (nonpeptide fraction) 16.4 ± 0.64 Initial body weight (g) 158.2 ± 2.74 158.1 ± 6.92 155.7 ± 1.59 160.0 ± 9.39 Final body weight (g) 359.5 ± 3.64 353.5 ± 23.75 353.1 ± 8.44 357.7 ± 21.97 Body weight gain for 30 days (g) 212.4 ± 35.8 195.4 ± 19.02 197.4 ± 6.85 197.71 ± 13.39 13.9 ± 2.35 13.2 ± 1.28 13.2 ± 0.46 13.0 ± 0.88 Food efficiencya See Table for details of diets fed to rat groups a Body weight gain/food intake Table Dietary effect of narezushi extract and its fractions on plasma lipid levels in rats At day After 30 days of feeding Control group Narezushi extract group NH4OH–HCl fraction group (peptide fraction) ACN fraction group (nonpeptide fraction) Total cholesterol (mg/100 ml) 70.8 ± 6.00ab 79.2 ± 8.14a 58.4 ± 4.62b 52.3 ± 7.06b 59.6 ± 4.47b HDL-cholesterol (mg/100 ml) 37.2 ± 4.78a 27.2 ± 5.65a,xy 34.9 ± 4.50ab,xy 25.5 ± 3.87ab,x 39.8 ± 2.52b,y LDL-cholesterol (mg/100 ml) Triglyceride (mg/100 ml) 33.6 ± 7.45a 29.7 ± 1.86a 52.0 ± 4.63c,x 139.3 ± 51.83b,x 23.1 ± 4.35b,y 69.8 ± 23.02bc,y 26.8 ± 6.31b,y 69.6 ± 14.41c,y 19.8 ± 2.69b,y 55.3 ± 7.25c,y Different letters in each data column indicate significant differences (p \ 0.05) Letters a, b, c indicate a significant difference among all groups containing day Letters x, y indicate significant differences among four groups at day 30 See Table for details of diets fed to rat groups Table Dietary effect of narezushi extract and its fractions on liver lipid levels in rats At day After 30 days of feeding Control group 4.6 ± 0.74 NH4OH–HCl fraction group (peptide fraction) 4.8 ± 0.59 ACN fraction group (nonpeptide fraction) Liver weight (%) 4.1 ± 0.11 Total lipid (%) 7.3 ± 0.32a 15.6 ± 1.15b,x 7.5 ± 0.83a,y 10.0 ± 0.74b,yz 10.8 ± 1.18b,z 0.7 ± 0.12a 15.5 ± 2.63a 1.9 ± 0.09b,x 69.0 ± 16.12b,x 1.3 ± 0.15c,y 36.2 ± 2.14c,y 1.8 ± 0.13b,z 71.0 ± 8.04b,x 1.4 ± 0.22c,yz 52.7 ± 7.19b,xy Total cholesterol (mg/g) Triglyceride (mg/g) 4.7 ± 0.17 Narezushi extract group 4.6 ± 0.33 Different letters in each data column indicate significant differences (p \ 0.05) Letters a, b, c indicate a significant difference among all groups containing day Letters x, y, z indicate the significant differences among four groups at day 30 See Table for details of diets fed to rat groups Table Dietary effect of narezushi extract and its fractions on fecal excretion of lipids of rats Rat groups fed diets in Table At day After 30 days of feeding Control group Total lipid (mg/day) Total cholesterol (mg/day) Triglyceride (mg/day) Free fatty acid (mg/day) Bile acid (mg/day) Narezushi extract group NH4OH–HCl fraction group (peptide fraction) ACN fraction group (nonpeptide fraction) 19.3 ± 0.19a 14.6 ± 0.07a,x 30.6 ± 0.23b,y 20.4 ± 0.16a,z 33.4 ± 0.12b,y 4.6 ± 1.48a 3.9 ± 1.07a,x 16.9 ± 0.68b,y 7.6 ± 2.54a,x 13.4 ± 3.03b,y 4.0 ± 1.04a 11.6 ± 2.06a 0.4 ± 0.16b,x 9.3 ± 3.22a,x 1.8 ± 0.19b,y 21.9 ± 1.86b,y 0.7 ± 0.20b,x 8.3 ± 0.88ab,x 1.0 ± 0.19b,x 17.6 ± 0.80a,y 0.4 ± 0.12a,x 0.9 ± 0.15b,y 0.8 ± 0.24ab,xy 0.4 ± 0.12ab 0.8 ± 0.13ab,xy Different letters in each data column indicate significant differences (p \ 0.05) Letters a, b indicate a significant difference among all groups containing day Letters x, y, z indicate significant differences among four groups at day 30 See Table for details of diets fed to rat groups 123 Fish Sci (2010) 76:537–546 543 level in the narezushi extract and ACN fraction groups increased significantly to 16.9 and 13.4 mg/day, respectively Fecal cholesterol level in the NH4OH–HCl group (7.59 mg/day) also tended to increase, but not significantly, during the 30 days of administration Fecal triglyceride level significantly decreased from 4.0 mg/day at day to 0.4, 1.8, 0.7, and 1.0 mg/day in the control, narezushi extract, NH4OH–HCl fraction, and ACN fraction groups, respectively, at day 30 Free fatty acid level in the feces was 11.6 mg/day at day and 9.3 mg/day at 30 days in the control, and this level increased significantly in the narezushi extract (21.9 mg/ day) and ACN fraction (17.6 mg/day) groups, but was almost the same in the NH4OH–HCl fraction group (8.3 mg/day) The amount of bile acid excreted to the feces at day 30 was 0.4 mg/day in the control group, almost the same level as at day However, fecal bile acid significantly increased to 0.9 mg/day in the narezushi extract group and tended to increase in the NH4OH–HCl fraction (0.8 mg/day) and ACN fraction (0.8 mg/day) groups Effects of narezushi extract on enzyme activity related to cholesterol metabolism Table shows changes in enzyme activity related to liver cholesterol metabolism in rats after 30 days of administration HMG-CoA reductase activity at day (224.3 lmol g-1 min-1) was markedly reduced at day 30 (124.3 lmol g-1 min-1) in the control group Significant reductions were also seen in the narezushi extract and fraction groups, with the activity ranging from 145 to 155 lmol g-1 min-1 Conversely, cholesterol 7a-hydroxylase activity at day (11.0 nmol g-1 min-1) showed little change in the control (10.5 nmol g-1 min-1) and NH4OH–HCl fraction (11.5 nmol g-1 min-1) groups at day 30, but tended to increase in the narezushi extract (14.6 nmol g-1 min-1) and ACN fraction (14.9 nmol g-1 min-1) groups Discussion Although both narezushi and heshiko are fermented mackerel products, the amounts and compositions of extractive components that accumulate in these foods, such as free amino acids, peptides, and organic acids, differ due to differences in their submaterials and processing methods [24, 25] In fact, more components were extracted from narezushi than from heshiko [24, 25] We successfully separated the narezushi extract containing peptides into two equivalent fractions: peptide fraction and nonpeptide fraction by ionexchange column chromatography The peptide fraction and nonpeptide fraction indicate the NH4OH–HCl fraction and ACN fraction, respectively, hereafter, because most of the peptides were extracted into the NH4OH–HCl fraction (peptide fraction) and most of the free amino acids and organic acids into the ACN fraction (nonpeptide fraction) (Tables 2, 3) The narezushi extract and its peptide and nonpeptide fractions were administered to rats at an amount equivalent to 50 mg of peptides/kg body weight/day, and the hypocholesterolemic effect of narezushi was examined using the methods described in a previous paper [22] As cholesterol forms a complex with lipids and proteins and is transported to the whole body as LDL-cholesterol, the LDL-cholesterol level in blood increases corresponding to the increase in cholesterol intake during administration of a diet enriched with fat The control group showed both a decrease in HDL-cholesterol and an increase in LDLcholesterol, resulting in about 20% increase in plasma total cholesterol level compared with that at day (Table 5) On the other hand, the levels of both total and LDL-cholesterol were lower in the narezushi extract and fraction groups at day 30 than at day Since no significant difference in body weight and food intake was observed among the rat groups, the administration of narezushi extract had little influence on rat growth (Table 4) Therefore, the hypocholesterolemic effects observed in the rats were considered to be due to the components contained in the narezushi extract We attempted to compare the results obtained by the administration of narezushi extract with those by heshiko extract, by adjusting the amount of the narezushi extracts and the heshiko extracts administered to rats to an equivalent of 50 mg of peptides/kg body weight/ day After 30 days of administration, the narezushi extract decreased the plasma cholesterol to a lower level than that at day 0, whereas the heshiko extract maintained almost the same plasma cholesterol level [22] These results indicated Table Dietary effect of narezushi extract and its fractions on HMG-CoA reductase and cholesterol 7a-hydroxylase activities in liver of rats At day After 30 days of feeding Control group Narezushi extract group ACN fraction group NH4OH–HCl fraction group (peptide fraction) (nonpeptide fraction) Cholesterol 7a-hydroxylase (nmol g-1 min-1) 11.0 ± 1.19a 10.5 ± 1.67a,x 14.6 ± 1.74a,x 11.5 ± 1.63a,x 14.9 ± 0.57a,x Different letters in each data column indicate significant differences (p \ 0.05) See Table for details of diets fed to rat groups 123 544 that the narezushi extract exhibits a stronger hypocholesterolemic effect than the heshiko extract Narezushi contained a larger amount of extracted components than heshiko [24, 25], resulting in twice the total amounts of free amino acids and organic acids in the narezushi extract administered to the rats as in the heshiko extract In particular, the content of glycine in the diet of narezushi extract group was three times as much as those of the heshiko extract group Differences in the amount of free amino acids and organic acids in rats administered narezushi and heshiko extracts are thought to be related to the stronger hypocholesterolemic effect of narezushi extracts In addition, differences in the composition of free amino acids and organic acids found in narezushi and heshiko [24, 25] might also influence the hypocholesterolemic effect Although the total liver lipid and cholesterol contents were increased in the narezushi extract and the nonpeptide fraction groups after 30 days of administration, the levels were significantly lower than those in the control group (Table 6) As the peptide fraction showed no definite decreasing effect on the liver lipid level, the major components in the narezushi extract and nonpeptide fraction, free amino acids and organic acids, may possess the suppressive effect on liver cholesterol accumulation Moreover, no differences in the liver cholesterol level were observed after 30 days of administration among the heshiko extract, peptide fraction, and nonpeptide fraction groups [22], indicating that differences between narezushi and heshiko in both the amount and composition of free amino acids and organic acids produced during fermentation may also be correlated with the suppression of liver cholesterol level The suppression of cholesterol re-absorption at the intestine or the increase in cholesterol consumption via promotion of catabolism are considered to be related to the decrease in cholesterol levels in the plasma and liver of the narezushi extract group or the nonpeptide fraction groups With regard to the hypocholesterolemic effect of foodstuff, increased fecal cholesterol and bile acid excretion have been also reported [2–4] Increases in total lipid, cholesterol, and bile acid levels in feces during administration were observed in the narezushi extract and nonpeptide fraction groups (Table 7), and similar changes in fecal cholesterol and bile acid occurred with the administration of heshiko extract [22] The result indicates that the extracts from these fermented mackerel products have the effect of increasing fecal cholesterol and bile acid, even if they differ with respect to submaterials and processing method Bile acid, one of the acidic steroids synthesized from cholesterol in the liver and forming a lipid-complex called a micelle, is thought to play an important role in the absorption or emulsification of lipids and cholesterol in the 123 Fish Sci (2010) 76:537–546 intestine The narezushi extract and nonpeptide fraction increased the content of fecal bile acid to a greater extent than the peptide fraction Therefore, both free amino acids and organic acids, as major components in the nonpeptide fraction, were thought to be involved in the decrease in plasma cholesterol via the promotion of bile acid excretion to feces Bile acid, which is circulated between the liver and intestine after its absorption at the intestine, is regulated at a certain level Increases in fecal bile acid thus lead to a shortage in the amount of circulating bile acid and result in the induction of cholesterol 7a-hydroxylase activity, a ratelimiting enzyme in the bile acid synthesis system Increased cholesterol 7a-hydroxylase activity was observed in rats whose plasma cholesterol content was decreased by the administration of heshiko extract [22] and soybean proteins [33] In the narezushi extract and nonpeptide fraction groups observed to have decreased plasma cholesterol levels, cholesterol 7a-hydroxylase activity tended to increase corresponding to the increase in fecal bile acid (Tables 5, 8) Upon administering the narezushi extract and nonpeptide fraction, the excretive promotion of bile acids to feces was suggested to bring about the decrease in plasma cholesterol level via acceleration of bile acid synthesis from cholesterol in the liver Although the increase in cholesterol 7a-hydroxylase activity by administration of narezushi extract was smaller than that by heshiko extract, fecal bile acid levels of both were increased to almost the same level As the cholesterol is used for the synthesis of bile acid, the low cholesterol content in the liver after narezushi extract administration might have suppressed the increase in cholesterol 7a-hydroxylase activity Cholesterol is not only consumed as a substrate for bile acid synthesis but is also utilized as an essential material for constructing cell membranes Therefore, cholesterol is also synthesized in the liver in addition to being available from dietary intake The rate-limiting enzyme in the cholesterol synthesis system is HMG-CoA reductase, and its activity is regulated by a feedback mechanism with respect to the amount of cholesterol transported to the liver from the intestine HMG-CoA reductase activity shows the tendency to decrease in all groups after 30 days of administration, although data were not shown because the other enzyme activity as cholesterol 7a-hydroxylase was not be able to be removed in the measurement of activity by the consumption of NADPH in the assay media From the results of the present study, it was confirmed that narezushi extract had a hypocholesterolemic effect and that the effect was related to both the increase in fecal cholesterol excretion and the promotion of bile acid synthesis via the increase in fecal bile acid The hypocholesterolemic effect of narezushi extract appeared to be stronger than that of heshiko extract, since the narezushi Fish Sci (2010) 76:537–546 extract decreased cholesterol levels in the plasma and the liver more than the heshiko extract Moreover, their nonpeptide fractions were suggested to relate more closely to cholesterol metabolism than the peptide fraction There are several reports on the hypocholesterolemic effect of free amino acids [33–35], such as glycine, taurine, cysteine, and acetic acid [36] As the contents of glycine and acetic acid in the diet of narezushi extract group were times and 2.5 times as much as those of heshiko extract group, respectively, glycine and acetic acid may affect the difference in hypocholesterolemic effect between narezushi and heshiko Although detailed research is now in progress, we are convinced that a large portion of the free amino acids and organic acids contained in the nonpeptide fraction participates in the hypocholesterolemic effects of the narezushi and heshiko extracts References Sugano M, Goto S, Yamada Y, Yoshida K, Hashimoto Y, Matsuo T, Kimoto M (1990) Cholesterol-lowering activity of various undigested fractions of soybean protein in rats J Nutr 120:977–985 Satoh T, Goto M, Igarashi K (1993) Effects of protein isolates from radish and spinach leaves on serum lipids levels in rats J Nutr Sci Vitaminol 39:627–633 Yang L, Kumagai T, Kawamura H, Watanabe T, Kubota M, Fujimura S, Watanabe R, Kadowaki M (2007) Effects of rice proteins from two cultivars, Koshihikari and Shunyo, on cholesterol and triglyceride metabolism in growing and adult rats Biosci Biotechnol Biochem 71:694–703 Igarashi K, Satoh A, Numazawa S, Takahashi E (1997) Effects of cabbage leaf protein concentrate on the serum and liver lipid concentrations in rats J Nutr Sci Vitaminol 43:261–270 Nakasa T, Ueda S, Nakatsuka M, Okinaka O (2003) Effect of protease-treated royal jelly on plasma and liver lipids in rats fed on a high fat plus high cholesterol diet Nippon Shokuhin Kagaku Kougaku Kaishi 50:463–467 Nagaoka S, Kanamaru Y, Kazuya Y (1991) Effects of whey protein and casein on the plasma and liver lipids in rats Agric Biol Chem 55:813–818 Ashida Y, Saito Y, Kawato A, Suginami K, Imayasu S (1997) Effects of dietary sake cake on cholesterol metabolism in rats Nippon Nogeikagaku Kaishi 71:137–143 Iwai K, Nakaya N, Kawasaki Y, Matsue H (2002) Antioxidative functions of natto, a kind of fermented soybeans: effect on LDL oxidation and lipid metabolism in cholesterol-fed rats J Agric Food Chem 50:3597–3601 Choi YM, Bae SH, Kang DH, Suh HJ (2006) Hypolipidemic effect of lactobacillus ferment as a functional food supplement Phytother Res 20:1056–1060 10 Matsumoto J, Enami K, Doi M, Kishida T, Ebihara K (2007) Hypocholesterolemic effect of katsuobushi, smoked-dried bonito, prevents ovarian deficiency-induced hypercholesterolemia J Nutr Sci Vitaminol 53:225–231 11 Wang L, Saito M, Tatsumi E, Li L (2003) Antioxidative and angiotensin I-converting enzyme inhibitory activities of Sufu (fermented tofu) extracts JARQ 37:129–132 12 Je J-Y, Park J-Y, Jung W-K, Park P-J, Kim S-K (2005) Isolation of angiotensin I converting enzyme ACE inhibitor from fermented oyster sauce, Crassostrea gigas Food Chem 90:809–814 545 13 Sun TS, Zhao SP, Wang HK, Cai CK, Chen YF, Zhang HP (2009) ACE-inhibitory activity and gamma-aminobutyric acid content of fermented skim milk by Lactobacillus helveticus isolated from Xinjiang koumiss in China Eur Food Res Technol 228:607–612 14 Nielsen MS, Martinussen T, Flambard B, Sorensen KI, Otte J (2009) Peptide profiles and angiotensin-I-converting enzyme inhibitory activity of fermented milk products: effect of bacterial strain, fermentation pH, and storage time Int Dairy J 19:155–165 15 Yen G-C, Chang Y-C, Su S-W (2003) Antioxidant activity and active compounds of rice koji fermented with Aspergillus candidus Food Chem 83:49–54 16 Jung W-K, Rajapakse N, Kim S-K (2005) Antioxidative activity of a low molecular weight peptide derive from the sauce of fermented blue mussel, Mytilus edulis Eur Food Res Technol 220:535–539 17 Singh B, Chand R (2006) Antioxidative activity of bovine milk fermented with Lactobacillus helveticus Milchwissenschaft 61:63–65 18 Igoshi K, Kondo Y, Kobayashi H, Kabata K, Kawakami H (2008) Antioxidative activity of cheese Milchwissenschaft 63:424–427 19 Omura K, Hitosugi M, Zhu X, Ikeda M, Maeda H, Tokudome S (2005) A newly derived protein from Bacillus subtilis natto with both antithrombotic and fibrinolytic effects J Pharmacol 99:247–251 20 Tsai T-Y, Chu L-H, Lee C-L, Pan T-M (2009) Atherosclerosispreventing activity of lactic acid bacteria-fermented milk-soymilk supplemented with Momordica charantia J Agric Food Chem 57:2065–2071 21 Itou K, Akahane Y (2004) Antihypertensive effect of heshiko, a fermented mackerel product, on spontaneously hypertensive rats Fish Sci 70:1121–1129 22 Itou K, Akahane Y (2009) Effect of extracts from heshiko, a fermented mackerel product, on cholesterol metabolism in Wistar rats Fish Sci 75:241–248 23 Itou K, Akahane Y (2007) Antihypertensive effect of narezushi, a fermented mackerel product, on spontaneously hypertensive rats Fish Sci 73:1344–1352 24 Itou K, Akahane Y (2006) Changes in proximate composition and extractive components in narezushi, a fermented mackerel product, during processing Fish Sci 72:1269–1276 25 Itou K, Akahane Y (2000) Changes in proximate composition and extractive components of rice-bran-fermented mackerel heshiko during processing Nippon Suisan Gakkaishi 66:1051–1058 26 Lowry O, Rosebrough N, Farr A, Randall R (1951) Protein measurement with the folin phenol reagent J Biol Chem 193:265–275 27 Folch J, Lees M, Sloane Stanley GH (1957) A simple method for the isolation and purification of total lipids from animal tissues J Biol Chem 226:497–509 28 Sautier C, Doucet C, Flament C, Lemonnier D (1979) Effects of soybean protein and saponins on serum, tissue and feces steroids in rat Atherosclerosis 34:233–241 29 Yomogida S, Ishikawa M, Sasaki M, Sato S, Sasaki K (2002) Effect of sizofilan, immunomodulator, on hepatic drug metabolizing enzyme activities Bull Kokusai Gakuin Saitama Junior College 23:43–49 30 Ogishima T, Okuda K (1986) An improved method for assay of cholesterol 7a-hydroxylase activity Anal Biochem 158:228–232 31 Sung JH, Choi SJ, Lee SW, Park KH, Moon TW (2004) Isoflavones found in Korean soybean paste as 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors Biosci Biotechnol Biochem 68(5):1051–1058 32 Satou Shin (1985) Scoring method In: Statistical analysis of sensory evaluation JUSE Press, Tokyo, pp 195–198 (in Japanese) 33 Morita T, Oh-hashi A, Takei K, Ikai M, Kasaoka S, Kiriyama S (1997) Cholesterol-lowering effects of soybean, potato and rice 123 546 proteins depend on their low methionine contents in rats fed a cholesterol-free purified diet J Nutr 127:470–477 34 Sugano M (1987) Nutritional studies on the regulation of cholesterol metabolism—the effects of dietary protein J Soc Nutr Sci 40:93–102 35 Sugiyama K, Ohoshi A, Ohnuma Y, Muramatsu K (1989) Comparison between the plasma cholesterol-lowering effects of 123 Fish Sci (2010) 76:537–546 glycine and taurine in rats fed on high cholesterol diets Agric Biol Chem 53:1647–1652 36 Tanizawa H, Sazuka Y, Komatsu A, Takino Y (1983) Acute toxicity of komezu and its effects on lipid metabolism in male mice J Soc Nutr Food Sci 36:283–289 [...]... -1 19 75 8.769 0.322 1.1 05 1976 8.436 0.291 1.013 0 .55 2 1977 7.879 0. 458 0.924 0 .52 4 1978 7.709 0 .55 3 0.914 0 .52 5 1979 9. 755 1.044 0.8 75 1980 17.488 1.982 1981 19. 953 2.638 1982 22. 459 1983 1984 15 -1 16 -1 17 -1 18 -1 19 -1 0. 054 20 -1 25 -1 26 -1 27 -1 1.744 0.371 2.700 2.211 0.387 3.019 0.290 2.699 0.411 3.289 0.370 3.063 0. 459 3 .51 7 0 .54 4 0 .50 7 3.418 0 .51 6 3.6 95 0.842 0 .56 4 0.619 3.911 0 .59 0 3.812... 0 .59 3 0.768 4.211 0.676 3.872 3.271 0.699 0.634 0.890 4 .52 8 0. 752 3.888 0.3 65 0.713 0.320 0.926 0.367 22.1 95 26.768 3 .57 1 3.613 0.617 0 .53 4 0.703 0.776 1.073 1.298 4.667 4 .59 2 0.973 1.083 3.863 3.740 0. 358 0.416 0.6 15 0 .52 9 0.144 0.110 0. 759 0.6 05 0.179 0. 351 19 85 25. 237 3.836 0.496 0.860 1.467 4.470 1.268 3 .56 7 0.332 0.3 95 0.049 0.2 45 0.0 85 1986 26.431 3.720 0.472 0. 956 1.614 4.197 1 .59 4 3.368 0 .55 8... (a) Pacific halibut 19 85 1990 19 95 2000 20 05 (c) Arrowtooth flounder 19 85 1990 19 95 2000 20 05 0.6 0 .5 0.4 0.3 0.2 0.1 0.0 1980 20 05 0.6 0 .5 0.4 0.3 0.2 0.1 0.0 1980 (e) Yellow fin sole 19 85 1990 19 95 2000 0.6 0 .5 0.4 0.3 0.2 0.1 0.0 1980 (b) Greenland turbot 19 85 1990 19 95 2000 20 05 2000 20 05 2000 20 05 (d) Flathead sole 19 85 1990 19 95 (f) Rock sole 19 85 1990 19 95 Year 0.6 0 .5 0.4 0.3 0.2 0.1 0.0 1980... doi:10.1007/s1 256 2-009-0201-2 S I Lee (&) East Sea Fisheries Research Institute, National Fisheries, Research and Development Institute, Gangnung 210-86 1, Korea e-mail: silee@nfrdi.go.kr K Y Aydin Á P D Spencer Á T K Wilderbuer Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way NE, Bldg 4, Seattle, WA 9811 5, USA C I Zhang Pukyong National University, Daeyeon 3-dong, Nam-gu,... Fairbanks, pp 2 25 2 45 Spencer PD, Wilderbuer TK, Zhang CI (2002) A mixed-species yield model for eastern Bering Sea shelf flatfish fisheries Can J Fish Aquat Sci 59 :291–302 Power ME, Tilman D, Estes JA, Menge BA, Bond WJ, Mills LS, Daily G, Castilla JC, Lubchenco J, Paine RT (1996) Challenges in the quest for keystones Bioscience 46:609–620 Pianka ER (1970) On r and K selection Am Nat 104 :59 2 59 7 Fish... 20 05 123 430 Fish Sci (2010) 76:411–434 Table 9 Time series data of the eastern Bering Sea for use in Ecosim, 1926–20 05 Pool codea Type codeb 9 -1 11 -1 13 -1 14 -1 15 -1 16 -1 17 -1 18 -1 19 -1 20 -1 25 -1 26 -1 1926 1927 1928 1929 1930 1931 1932 1933 1934 19 35 1936 1937 1938 1939 1940 1941 1942 1943 1944 19 45 1946 1947 1948 1949 1 950 1 951 1 952 1 953 1 954 2.394 1 955 2.117 1 956 1. 953 1 957 2.001 1 958 ... 2.138 0.201 1.9 15 1.6 15 2.484 3. 255 2.1 45 0.912 0.379 0.083 0.264 0.169 2003 23.373 2.207 0.193 1.9 85 1 .53 9 2.433 3.138 2.149 0. 859 0.433 0.123 0.292 0.2 95 2004 20.171 2.120 0.188 2. 051 1.463 2.403 3.187 2.166 0. 950 0. 458 0.111 0.372 0.287 20 05 19.434 1.969 0.190 2.104 1.4 05 2.349 3.310 2.1 95 1.188 0.483 0.221 0 .58 7 0.247 Pool codea Type codeb 9 -6 11 -6 13 -6 14 -6 15 -6 16 -6 17 -6 18 -6 25 -6 26 -6 27... 1992 23.014 2 .56 2 0.387 1.601 2.266 3.739 3.002 2.642 0.868 0.493 0.317 1.122 0.128 1993 24.2 35 2 .52 7 0.382 1.667 2.278 3 .57 6 3 .53 3 2.603 0.833 0.462 0.182 0.8 45 0.191 1994 28 .58 9 2 .58 0 0.3 65 1.713 2.278 3.498 3.740 2 .55 7 0.921 0.338 0.137 0. 856 0.134 19 95 26. 356 2.680 0.341 1.726 2.237 3.268 4.184 2 .51 6 0.871 0.474 0.112 1.0 95 0.171 1996 22.427 2.484 0.320 1.742 2.180 3.080 4.003 2. 453 0.942 0.420... 1996 2. 651 0 .53 5 0.0 15 0.033 0.039 0.288 0.104 0.036 0.002 0.066 0.013 1997 2.499 0 .57 3 0.016 0.022 0.046 0.403 0. 150 0.047 0.000 0.121 0.014 1998 2.447 0.429 0.020 0.034 0. 055 0.2 25 0.0 75 0.032 0.000 0.246 0.018 1999 2.200 0.387 0.013 0.023 0.041 0. 150 0.090 0.031 0.000 0.186 0.012 2000 2 .51 7 0.4 25 0.0 15 0.029 0.0 45 0.186 0.109 0.032 0.000 0.031 0.008 2001 3.083 0.393 0.012 0.031 0.040 0.141 0.0 65 0.019... 2.180 3.080 4.003 2. 453 0.942 0.420 0.093 1.107 0. 155 1997 21 .50 3 2. 253 0.301 1.741 2.089 3.042 3.776 2.381 0.8 75 0.481 0.0 45 0.9 75 0.211 1998 1999 23.306 22.000 2.013 2.033 0.279 0. 254 1. 757 1.7 75 1.9 85 1.8 75 2.881 2.804 3.773 3.628 2.331 2.263 0.787 0.823 0.427 0.318 0.046 0. 059 0. 752 0.272 0.226 0.186 2000 20.682 2.004 0.236 1.812 1.786 2 .52 3 3 .53 1 2.203 0.723 0.346 0.064 0.460 0.143 2001 20.610