Bioremediation of organically enriched sediment deposited below fish farms with artificially mass-cultured colonies of a deposit-feeding polychaete Capitella sp. I

Authors


*Tel: 81-96-383-2929.
Fax: 81-96-384-6765. Email: kkyoko@pu-kumamoto.ac.jp

Abstract

ABSTRACT:  For bioremediation of organically enriched sediment deposited below fish farms, the extremely high potential for population growth of a deposit-feeding polychaete, Capitella sp. I, in the organically enriched sediment, and the effect on decomposition of organic matter in the sediment, were examined. A mass-culturing technique was conducted for this species. Bioremediation experiments were conducted on the organically enriched sediment in a fish farm in Kusuura Bay, Japan in 2003–2006. Approximately 1.7 million individuals of the worms were placed on the sediment below one net pen in December 2003, 9.3 million individuals in November 2004, and 2.2 million individuals in November 2005. After the worms were spread on the sediment, they rapidly increased in number and reached the highest densities of approximately 134 000 inds/m2 in February 2004, 527 000 inds/m2 in March 2005 and 103 000 inds/m2 in January 2006. In the process of rapid population growth, the decomposition of the organic matter of the sediment was enhanced markedly. Our results demonstrate that the promotion of population growth by spreading cultured colonies of Capitella can enhance the decomposition rate of organic matter markedly in organically enriched sediment below fish farms. This method is promising for minimization of the negative effects of fish farms.

INTRODUCTION

The development of fish culture has been referred to as ‘the conversion of catching fisheries to rearing fisheries’.1 This new style of coastal fisheries has developed rapidly in various countries throughout the world since the 1980s. Fisheries include salmonid culture in western Europe, Scandinavia and North America and a variety of-non-salmonid species (e.g. grouper, sea bream, sea bass, snapper and yellowtail) in Asian–Pacific waters.2–5 Fish farming using net pens is becoming one of the most important ways to obtain seafood resources from coastal areas.1

Many fish farms, however, have caused organic enrichment of the bottom sediments from the vast discharge of organic matter from the net pens. Even at a conservative estimate, 7–66% of the organic carbon contained in the food for cultured fish reaches the sea floor just below the fish farm.6–11 The organic flux from the net pen to the sea floor ranges between 1.3 and 54.2 g Carbon/m2 per day (g C/m2 per day).12 Although the organic flux changes widely depending on the water depth of the fish farm, it far exceeds that of the natural organic flux, which mainly consists of photosynthetically produced organic matter from phytoplankton and organic debris from terrestrial plants.

Accumulation of organic matter on the sea floor below the net pens has a serious negative effect on the benthic environment and on the fish farms because of the dissolved oxygen deficiency in the bottom water and generation of high levels of hydrogen sulfide in the organically enriched sediment during the warm seasons.7,8,13–16 The development of reducing conditions in the bottom environment causes serious disturbance to benthic communities in the enclosed bays where fish farms reside, and the benthic communities become simpler in species diversity, mainly consisting of several species of small polychaetes, and poorer in abundance and biomass.13,14,17–20 Pearson and Rosenberg21 and Diaz and Rosenberg22 predicted such conditions in their schematic models of benthic faunal changes with gradient of organic enrichment of the sediment. Therefore, we need to find effective measures to prevent further worsening of organic enrichment and to re-establish healthy benthic environments for sustainable management of fish farming in enclosed bays.3,7

In previous studies, several engineering approaches were attempted, including dredging sediment, creation of a sand cover on the sediment, creation of a waterway by digging the sea floor to increase the water exchange and oxygen supply to areas with fish farms, promotion of vertical mixing of the water using a pump23,24 and location of a sludge collection system under the net pens to minimize the organic loading on the sea floor.9 However, these approaches have problems in cost, effect and/or efficiency.

We have attempted to treat the organically enriched sediment by exploiting the biological activities of a deposit-feeding small polychaete Capitella sp. I.25 This species and closely related sibling species are common members of the macrobenthic communities in the extremely organically enriched sediment.21 When the organically enriched sediment is available in laboratory conditions or accessible in the field, Capitella sp. I realize a short life cycle of 4–6 weeks and exhibit very rapid population growth.26–30 Earlier, we showed that the biological activities of the rapidly increasing Capitella population can be applied to bioremediation of organically enriched sediments in experiments in containers,31,32 outdoor pools33 and the innermost areas of an enclosed bay surrounded by heavy industrial areas and a large city.34

We have conducted bioremediation experiments with a mass culture of Capitella sp. I for treatment of organically enriched sediment deposited on the sea floor below a fish farm in an enclosed bay in Amakusa, Kyushu, western Japan, since 2003.35 In this paper we report the results of bioremediation experiments for the organically enriched sediment in the fish farm from 2003 to 2005. We discuss the applicability of this technique using Capitella colonies for bioremediation of organically enriched sediment in fish farms in enclosed coastal seas.

MATERIALS AND METHODS

Study area

Kusuura Bay is an extremely enclosed bay located between Amakusa Kamishima Island and Amakusa Shimoshima Island, Kyushu, western Japan (32°23′N, 130°13′E) (Fig. 1). The water depth at the center of the bay is approximately 16–20 m. In 1973, a fish farm was established in an area of approximately 400 m × 200 m at the center of the bay. Approximately 2000 t of red sea bream Pagrus major and yellowtail Seriola quinqueradiata have been shipped from the fish farms per year. We set three sampling stations (Stns 1–3) for assessment of the benthic conditions on a transect line from the inside (Stn 1) to the outside of the fish farm (Stns 2–3) in Kusuura Bay from May 2003 to October 2006 (Fig. 1). Stn 2 was located beside the edge of the fish farm, and Stn 3 was located approximately 400 m from the edge of the fish farm. From May 2005 to October 2006, we added one sampling station inside the fish farm as a control for Stn 1 (Stn C).

Figure 1.

Study area and sampling site.

Sampling procedure

We conducted quantitative sampling of the macrobenthic animals and assessment of the chemical conditions of the sediment monthly from May 2003 to October 2006. At Stn 1, we collected sediments beside the net pens with an Ekman–Birge grab sampler (20 cm × 20 cm) from a boat, and subsampled ten sediment samples for quantitative sampling of the macrobenthos with a core sampler (5 cm × 5 cm × 5 cm). One sediment sample was collected with an acrylic core sampler (2.9 cm diameter) for chemical analysis of the sediment. From March 2004 to October 2006 at Stns 1 and C, divers collected three sediment samples for quantitative sampling of the macrobenthic animals and one for chemical analysis with an acrylic core sampler (4.9 cm diameter) at five sites just below the four corners plus the center of the net pen. At Stns 2 and 3, we collected sediments with a grab sampler from the boat, and subsampled ten sediment samples for quantitative sampling of the macrobenthos with the core sampler. We also collected one sediment sample for chemical analysis using a gravity corer with 50-cm long, 4.0-cm diameter acrylic pipe from the boat.

Analysis of macrobenthic animals

For quantitative analysis of the macrobenthic animals, we fixed and stained the sediment samples with 10% formalin solution including rose bengal dye. First, we sieved the sediment samples on a 1-mm mesh screen, sorted the red-stained macrobenthic animals from the residues, and identified and counted them. Next, we sieved the remainder through the 1-mm mesh screen again onto a 0.125-mm mesh screen, sorted only Capitella species from the residues under a stereoscopic microscope, and counted the number of individuals to determine the population density. Then, we determined the body size (maximum width of the thoracic segments) of the Capitella with a microscope image analyzing system.35 Finally, we calculated the biomass of the Capitella population as the total wet weight of the population at each sampling station using conversion equations between the body size and dry body weight, and between the dry and wet body weights.35

Bioremediation experiment with Capitella cultured colonies

For the treatment experiment of the organically enriched sediment with mass-cultured colonies of Capitella sp. I, we collected the worms from the sediment just below the fish farm in the study site, and kept the worms in the laboratory as seed colonies for mass culture. We cultured the seed colonies of Capitella in a large tank (10 m2 cross-sectional area) for two months for field experiments. We set an online vertical profiling system of the water beside the net pen at Stn 1, and monitored daily changes in water conditions.36 After we had confirmed the recovery of dissolved oxygen level of the bottom water due to the vertical circulation of the water in autumn, we conducted the bioremediation experiment of the organically enriched sediment in the fish farm. We spread approximately 1 680 000 individuals of cultured worms on the sediment at each of four corners and the center of the area of 12 m × 12 m total area just below the net pen (12 m × 12 m × 8 m) at Stn 1 on 6 December 2003, 9 270 000 individuals on 5 November 2004 and 2 185 000 individuals on 9 November 2005.

Organic matter content of sediment

We sliced three layers of the sediment from the surface of the gravity core samples every 1 cm depth from June 2004 to October 2005; the deepest layer was 2–4 cm in depth. The sediment was freeze-dried once, treated with 2 N HCl to remove inorganic carbonate, and vacuum-dried. We determined total organic carbon (TOC) of the sediment with an elemental analyzer (NA-1500, Fisons, Rodano-Milan, Italy).

Statistical analysis

To evaluate the effect of the spread of the Capitella colonies on the organically enriched sediment, we compared TOC of the sediment at Stn 1 in each year from May until before the spread of the colonies (May–December 2003, May–November 2004 and May–October 2005). We also compared TOC of the sediment at Stn C in each year from May until October as a control of Stn 1. The effects of this treatment were tested by a Tukey–Kramer test (significance level, P < 0.05) using StatView v5 software (SAS Institute, Cary, NC, USA).

RESULTS

Seasonal fluctuations of Capitella population

Seasonal fluctuations of the density and biomass of Capitella between May 2003 and October 2005 are shown in Figure 2. From late spring to summer (May–September), Capitella were at very low densities in all stations (Fig. 2a). At Stn 1 inside the fish farm, the Capitella population started to recover from October. Then, we spread approximately 1 680 000 individuals of the cultured worms on the sediment just below a net pen at Stn 1 in December 2003, 9 270 000 individuals in November 2004 and 2 185 000 individuals in November 2005. Approximately one month later, Capitella increased to 118 640, 128 290 and 232 130 inds/m2 in 2003, 2004 and 2005, respectively. In the winters (december–March) of 2004–2006, Capitella further rapidly increased and reached the highest density of 1 028 000 inds/m2 in January 2006. At Stn C, the Capitella population increased during winter, and reached the highest density of 305 732 inds/m2 in March 2006. At Stn 2 at the edge of the fish farms, the Capitella population also increased in density during the winter; however, population growth was restricted (92 160, 24 480 and 102 160 inds/m2 in February 2004, March 2005 and January 2006, respectively). At Stn 3 outside of the fish farm, the density of Capitella was extremely low throughout the experiments (0–320 inds/m2).

Figure 2.

Seasonal fluctuations in (a) density and (b) biomass of Capitella sp. I at Stn 1 (●), Stn C (○), Stn 2 (◊) and Stn 3 (□).

The biomass of the Capitella population fluctuated seasonally in the same manner as Capitella densities (Fig. 2b). At Stn 1, the biomass of the population just before spreading the cultured colonies of Capitella was 26.0, 45.9 and 0.3 g wet wt/m2 in December 2003, November 2004 and October 2005, respectively. Biomass increased rapidly and reached 120.4, 458.5 and 674.8 g wet wt/m2 in February 2004, March 2005 and January 2006, respectively. At Stn C, the biomass of Capitella population increased in winter, and reached to 200.1 g wet wt/m2 in March 2006. At Stn 2, the biomass of the Capitella population did not increase markedly, and remained less than 62.7, 14.8 and 71.8 g wet wt/m2 in the winters of 2003–2004, 2004–2005 and 2005–2006, respectively. At Stn 3, the biomass of the Capitella population was restricted to less than 0.2 g wet wt/m2 during this study.

Seasonal fluctuations of macrobenthic communities

Seasonal fluctuations in the abundance of the macrobenthic communities are shown in Figure 3. All the density data were obtained by sieving the quantitative sediment samples with a 1-mm mesh screen. At Stn 1, the abundance of the macrobenthic communities repeated cyclic seasonal fluctuations characterized by collapse of the communities from the late spring to early summer and rapid recovery from autumn to winter. The macrobenthic communities of Stn 1 mainly consisted of three species of polychaetes Capitella sp. I (Capitellidae), Neanthes caudata (Nereidae) and Prionospio pulchra (Spionidae), and two species of crustaceans Nebalia japanensis (Neballidae, Leptostraca) and Melita sp. (Melitidae, Amphipoda). These five species occupied 88.1% of the total numbers of the macrobenthic animals. Capitella sp. I reached the highest densities of 40 240, 63 285 and 73 728 inds/m2 in February 2004, March 2005 and January 2006, respectively. Nebalia japanensis increased from April, and reached the highest densities of 35 985, 18 117 and 17 740 inds/m2 in April 2004, May 2005 and June 2006. Melita sp. showed highest densities of 22 144 inds/m2 in May 2004 and 38 059 inds/m2 in June 2006. Macrobenthos at Stn C mainly consisted of three species of polychaetes Capitella sp. I, P. pulchra and Scoletoma longiforlia (Lumbrinereidae) and two species of crustaceans, N. japanensis and Melita sp. These five species occupied 80.0% of the total numbers of the macrobenthic animals. Capitella sp. I increased from autumn to winter, and reached the highest density of 33 404 inds/m2 in January 2006. The dominant species of Capitella, P. pulchra, and S. longiforlia were much less than 6000 inds/m2 each. At Stn 2 there were mainly two species of polychaetes Capitella sp. I, and P. pulchra. These dominant species occupied 77.0% of the total numbers of the macrobenthic animals. The density of Capitella sp. I was usually less than 5000 inds/m2; however, it increased temporarily to greater than 10 000 inds/m2 in February 2004 and January 2006. The abundance of the macrobenthic animals except Capitella sp. I at Stn 2 was much lower than that at Stn 1 (<3000 inds/m2) throughout the year. At Stn 3, the abundance of the macrobenthic communities was also much lower than at Stn 1 (<3000 inds/m2). There were more diverse fauna including a bivalve Theora fragilis, an amphipod Melita sp., and various polychaetes including Chaetozone sp., Nephtys sp. and Magelona sp.

Figure 3.

Seasonal fluctuations in the abundance of the macrobenthic communities at (a) Stn 1, (b) Stn C, (c) Stn 2 and (d) Stn 3. Data are based on sediment samples sieved with a 1-mm mesh screen. P, polychaete, C, crustacean; and B, bivalve.

Seasonal fluctuations of organic matter content of sediment

Seasonal fluctuations in TOC of the three different layers of the sediment at three stations are shown in Figure 4. At Stn 1 where we spread the cultured colonies of Capitella sp. I on the sediment, unique seasonal fluctuation patterns were found in TOC levels of the sediment after the first spread of Capitella in December 2003. Before the spread of the Capitella colonies, the TOC levels of the three layers of the sediment ranged 13.9–24.1 mg/g except in July 2003, and the differences in TOC among these three layers were within 6.6 mg/g. After the first spread of the Capitella colonies, the TOC levels decreased markedly to 2.6–10.0 mg/g in all three layers of the sediment in January 2004, and remained in a low range of 4.3–11.8 mg/g until July 2004. These levels in the sediment were even lower than those in the same layers at Stn 2 at the edge of, and Stn 3 outside, the fish farm. After August 2004, the TOC levels increased markedly in the top layer (0–1 cm in depth) of the sediment at Stn 1 again, and reached 28.5 mg/g in January 2005, while the TOC levels of the deeper layers remained at lower levels. The TOC levels in the second layer (1–2 cm in depth) and the third layer (2–4 cm in depth) of the sediment were 16.8 and 6.7 mg/g, respectively. The difference in TOC between the top and the third layers was 21.8 mg/g. The TOC levels in the top and second layers of the sediment decreased from March 2005 when Capitella increased very rapidly from the second spread in November 2004 and had established dense patches of more than 300 000 inds/m2. The TOC levels at the top and second layers of the sediment decreased to the same levels as the deepest layer of less than 10 mg/g by May 2005. The TOC levels at the top and second layers of the sediment increased from July and August 2005, respectively, again as in the summer of 2004, but that of the third layer remained at 11.3 mg/g. Afterwards, the TOC levels at the top and second layers of the sediment increased in winter and reached 38.5 mg/g in November 2005. However, TOC decreased to 11.6–12.4 mg/g in June. TOC levels in the third layer remained at 7.1–16.2 mg/g. At Stn C, the TOC at the top and second layers of the sediment increased toward winter. TOC levels reached 40.0 mg/g in the top layer in December 2005. After January 2006, the TOC levels in the top and second layers decreased to 18.9–23.3 mg/g in June 2006. The TOC levels at the third layer remained at 13.9–22.0 mg/g. At Stns 2 and 3, the TOC of the sediment in all three layers fluctuated in narrow ranges of 13.9–24.1 and 13.3–18.8 mg/g, respectively, during this study. The difference in TOC among the three layers was only less than 3.0 mg/g at these two stations.

Figure 4.

Seasonal fluctuations in total organic carbon (TOC) of three layers of sediment sectioned from the surface at 0–1 cm (●), 1–2 cm (▵) and > 2 cm (□) at (a) Stn 1, (b) Stn C, (c) Stn 2 and (d) Stn 3. Data are from June 2004 to October 2006. The deepest layer was 2–4 cm depth.

We compared the mean TOC levels of the three layers of the sediment at Stn 1 from May until before the spread of the colonies in 2003–2006 (May–December 2003, May–November 2004, May–October 2005 and June–October 2006) to evaluate the effect of the spread of the Capitella colonies on the organically enriched sediment and their rapid population increase in the sediment during winter (Fig. 5). The mean TOC levels in the top layer of the sediment in 2003–2006 ranged 15.3–20.0 mg/g, and no statistically significant differences were found among them. In the second and third layers, the decrease in the mean TOC levels of the sediment after the first treatment with Capitella colonies in the winter of 2003–2004 was statistically significant (Tukey–Kramer test, P < 0.05). Thus, the effect of the activities of dense patches of Capitella on the sediment appeared in the subsurface layers. Although we did not spread the Capitella colonies at Stn C, Capitella increased from December 2005 to April 2006. Therefore, we compared the mean TOC levels in the three layers of the sediment at Stn C from May to October between 2005 and 2006. The TOC levels in the three layers of the sediment ranged 18.4–26.1 and 15.4–22.9 mg/g in 2005 and 2006, respectively. The TOC level decreased in all layers in 2006, and statistically lower values were found in the third layer (Tukey–Kramer test, P < 0.05).

Figure 5.

Total organic carbon (TOC) of three layers of sediment in (a) each year from May until before the spread of Capitella colonies at Stn 1 and (b) May–October at Stn C. *, significantly different among years 2003–2006 at P < 0.05 (Tukey–Kramer test); error bars, standard deviation.

DISCUSSION

The organic flux from the net pen causes organic enrichment of the sediment just below the net pen.3,6–11 The seasonal changes in the organic flux depend on not only the amount of food spent in the net pen, but also the water column structure. During the warm seasons, the water column tends to be stratified because of low salinity and higher temperature in the surface layer. The organic particles discharged from the net pens disperse widely to outside the fish farm, floating in the surface layer of the low-density regions of the stratified water, while the particles concentrate on the sea floor just below the net pens when the water is vertically well-mixed in autumn and winter.12 The top layer of TOC in the sediment at Stn 1 therefore increased markedly in September in 2004 and 2005 (Fig. 4), just after strong winds and waves caused by typhoons had destroyed the water stratification in this study area.

In this study, we conducted field experiments to treat the organically enriched sediment deposited just below the fish farm by inducing rapid population growth of Capitella with its artificially cultured colonies in the cold seasons. Figure 6 shows the seasonal fluctuations in the organic flux from the net pen to the sea floor12 and the decomposition rate of organic matter in the sediment by the biological activities of Capitella at Stn 1. The decomposition rate was estimated from the biomass data of Capitella in this study and the results of the laboratory experiments on the decomposition potential of organic matter in Capitella culture.31,32 In a laboratory Capitella culture, worms and bacteria decompose the organic matter in the sediment, and the worms stimulate bacterial growth around their burrows.37,38 Therefore, the estimated decomposition rate of the organic matter in the sediment by the biological activities of Capitella are made up of those of Capitella and bacteria.

Figure 6.

Seasonal fluctuations in organic flux from the net pen to the sea floor as total organic carbon (TOC, ○) and decomposition rate of organic matter by biological activities of Capitella (dTOC, ●) at Stn 1. Organic flux data are cited from Tsutsumi et al.12 The decomposition rate of organic matter was estimated from the biomass data (dry weight, DW) of Capitella in this study and the results of laboratory experiments on the decomposition rate of organic matter of the sediment (dTOC) in a Capitella culture from Chareonpanich et al.31,32 dTOC (C/m2 per day) = 0.15 × DW (g) + 0.23, r2 = 0.86; DW (mg) = 0.20 × wet weight (mg) − 1.11, r2 = 0.98.

In the winter of 2003–2004, the decomposition rate of organic matter in the sediment by the Capitella population reached 2.82 mgC/m2 per day in February 2004, and slightly exceeded or almost balanced the organic flux to the sea floor between January and April 2004. In the winter of 2004–2005, the decomposition rate of organic matter in the sediment by the Capitella population reached 10.12 mgC/m2 per day by the establishment of dense patches of more than 500 000 inds/m2 in March 2005, which far exceeded the organic flux to the sea floor in winter. Results of this study demonstrate that the promotion of population growth by spreading the cultured colonies of Capitella can enhance the decomposition rate of organic matter markedly in the organically enriched sediment just below the fish farm. It shows that bioremediation using Capitella is a promising method for minimization of negative effects of fish farms to the benthic environment of the bay by processing the organically enriched precipitated in the fish farm. In this study, we did not spread the Capitella colonies on Stn C; however, Capitella increased at Stn C in winter. These increases were not necessarily reflected in the densities at Stn 1, but the establishment of dense patches may rely on the increase of larval supply from Stn 1.

In the benthic communities inside the fish farm, two small crustaceans Nebalia japanensis and Melita sp., and two polychaetes Neanthes caudata and Prionospio pulchra, were dominant in addition to Capitella (Fig. 3). However, they are not infaunal deposit feeders, and do not have large potentials for population growth. It is difficult to expect these species to show large potential for decomposition of organic matter in the sediment. To understand mechanisms why the establishment of dense patches of Capitella colonies exhibit high decomposition potential of organic matter in the organically enriched sediment, we paid attention not only to the feeding activities of subsurface sediments but also the reworking of the sediment. Tube-dwelling infaunal deposit feeders are often known as a powerful sediment bioturbators. They transport sediment in a vertical direction, bringing sediment from deeper layers to the surface.39Capitella is a subsurface deposit feeder, excreting the sediment as fecal pellets and spouting the subsurface sediment on the sediment surface.40 This reworking potential of the subsurface sediment was estimated from the biomass data of Capitella in the present study and the results of laboratory experiments on burrowing and feeding activities by the worms.40 We estimated that the densest patches of Capitella with 1 028 000 inds/m2 at Stn 1 in January 2006 possessed a potential to rework 827 g dry weight sediment/m2 per day of the subsurface sediment on the sediment surface. It indicates that the Capitella colonies could rework the sediment within less than 1.5 months, assuming that the reworking occurred evenly in the sediment up to 4 cm depth.

In this study, we showed that the TOC levels decreased significantly in the subsurface layers of the organically enriched sediment by the spread of cultured colonies of Capitella and their rapid population growth (Fig. 5). The worms feed the subsurface sediment as a subsurface deposit feeder, and decompose organic matter in the sediment. However, the amount of organic matter that the worms consume as a source for secondary production might be limited.41,42 Madsen et al.43 reported that in laboratory experiments in microcosms without the worms, there was loss of organic sedimentary contaminants which was confined to the oxidized top layer of the sediment, indicating that microbial processes may have been responsible for some of the loss. These results indicate that organic matter in the sediment is finally decomposed through the microbial community that was enhanced by the biological activities of worms. The worm tubes or burrows create an oxidized layer in the sediment, and provide suitable conditions for an increase in aerobic bacteria in the sediment.3,37,44 Kunihiro et al.38 reported a marked increase of aerobic bacteria in organically enriched sediment below a fish farm during winter when the worms established dense patches. Thus, the reworking activities of the sediment by dense patches of Capitella including feeding the subsurface sediment and excreting fecal pellets on the sediment surface, burrowing in the sediment, and spouting the subsurface sediment onto the sediment surface should promote sediment oxidation and provide an oxygen-rich environment suitable for aerobic bacteria in the deeper subsurface sediment. As shown in Figure 5, a large decrease of organic matter content occurred in the subsurface sediment (1–4 cm depth) by the establishment of dense Capitella patches.

We are currently investigating methods for the treatment of organically enriched sediment with cultured colonies of Capitella that lead further rapid population growth. We are also developing bacterial agents that associate with Capitella colonies for decomposition of organic matter to realize higher decomposition rates of organic matter in sediments.

ACKNOWLEDGMENTS

We thank RS Lavin for reviewing the manuscript. This work was supported by a Research and Development Program for New Bio-industry Initiatives of the Bio-oriented Technology Research Advancement Institution (BRAIN) of Japan.

Ancillary