Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia under reduced frequency feeding strategies

RESEARCH ARTICLE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia under reduced frequency feeding strategies Scott A. SalgerID1, Jimi Reza2, Courtney A. Deck1, Md. Abdul Wahab2,3, David A. Baltzegar1,4, Alexander T. MurrID1, Russell J. Borski1* 1 Department of Biological Sciences, North Carolina State University, Raleigh, North Carolina, United States of America, 2 Department of Fisheries Management, Bangladesh Agricultural University, Mymensingh, Bangladesh, 3 WorldFish, Dhaka, Bangladesh, 4 Genomic Sciences Laboratory, North Carolina State University, Raleigh, North Carolina, United States of America * russell_borski@ncsu.edu Abstract Feed constitutes 50–70% of total production costs of tilapia, one of the most widely cultured finfishes in the world. We evaluated reduced-feeding strategies for improving production efficiency of Nile tilapia (Oreochromis niloticus). In a 12-week pond trial, fish were fed daily, every other day, every third day, or not at all. Ponds were fertilized to enhance natural foods. In a fifth group fish were fed daily without pond fertilization. Fish fed daily with or without pond fertilization and fish fed every other day had higher specific growth rates, survivability, and net production than the other two treatments. Fish feed efficiency and benefit to cost ratio was highest for treatments fed in a pulsatile manner (i.e. fed every other day or every third day) with fish fed on alternate days providing the best net return among all groups. Fish

fed on alternate days had more moderate gene expression levels of intestinal nutrient trans- porters which may allow for a more balanced and efficient nutrient uptake. Fecal microbe analyses identified 145 families of prokaryotic and 132 genera of eukaryotic organisms in tilapia. The highest diversity of prokaryotes was found in fish fed either every other day or daily in fertilized ponds and the highest diversity of eukaryotes was found in fish fed every other day. These studies indicate feeding Nile tilapia on alternate days along with weekly pond fertilization has no deleterious effects on growth, survivability, or production versus daily feeding regimes, but enhances feed efficiency by 76% and provides the greatest net return on investments. Our studies also suggest for the first time that combining alternate- day feeding with pond fertilization produces the greatest microbial biodiversity in the intes- tine that could contribute to enhanced feed efficiency and overall health of tilapia. Introduction Aquaculture production is one of the fastest expanding agriculture industries in the world with growth rates in excess of 40% [1]. Tilapia and other freshwater species, namely carps, will make up 60% of the fish cultured worldwide by 2025. Global production of farmed Nile tilapia (Oreo- chromis niloticus) has increased exponentially since 1985, with over 3.7 million metric tons PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 1 / 23 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS

Citation: Salger SA, Reza J, Deck CA, Wahab M.A, Baltzegar DA, Murr AT, et al. (2020) Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia under reduced frequency feeding strategies. PLoS ONE 15(7): e0236100. https://doi. org/10.1371/journal.pone.0236100 Editor: Michael H. Kogut, USDA-Agricultural Research Service, UNITED STATES Received: December 30, 2019 Accepted: June 30, 2020 Published: July 23, 2020 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0236100 Copyright: © 2020 Salger et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All Prokaryotic and Eukaryotic microbiome Fastq files can be accessed in the National Center for Biotechnology information (NCBI) under GenBank BioProject

consumed in 2014 [2]. Currently, tilapia and many other fish species are grown predominantly in pond culture under various conditions: intensive culture with high stocking densities, reli- ance on commercial feeds, and often requiring mechanical aeration of the ponds; extensive cul- ture where fish are stocked at low densities and organic and inorganic fertilizers stimulate pond primary food production with little to no additional feed supplementation; and, most fre- quently, semi-intensive culture conditions in which fish are stocked at moderate densities and provided supplemental feeds to complement microorganisms produced naturally with pond fertilization, which can effectively quadruple production over extensive culture systems [3–5]. Feed is the major expense in aquaculture, comprising up to 50–70% of total production costs for tilapia. Methods to limit the amount of feed can help reduce overall costs to producers and has major impacts on small-scale farmers that dominate in under-developed countries. Reductions in feed ration by up to 50% by using alternate-day feeding improves the feed effi- ciency and economic returns of tilapia monocultured at densities of 2–4 fish/m2 with little impact on growth, survival, or production yield [6]. A better understanding of how finfish acquire and utilize nutrient inputs is requisite for future improvements in aquaculture production efficiency. The underlying mechanism explaining how reduced frequency or alternate-day feeding strategies can achieve equivalent

production yields with less feed is poorly understood in finfish. Some evidence suggests that during periods of fasting, nutrient uptake efficiency in the intestine is intrinsically enhanced, leading to a more-efficient uptake of nutrients at the next feeding period [7–10]. Thus, fish being fed a daily regime have lower uptake efficiency and do not receive maximal dietary bene- fit. The increase in nutrient uptake efficiency following periods of decreased nutrient availabil- ity has been postulated, in part, to explain the compensatory growth (CG) response observed in many aquaculture species [7–11]. Additionally, reduced feeding may promote foraging on pri- mary food production within the ponds, leading to a more diverse diet (e.g. algae, insect larvae, plankton), enhancing nutrient recycling within the ponds. This enhanced diversity may directly influence intestinal absorption by promoting colonization and growth of microbes that may aid in nutrient utilization and uptake efficiency that are key to growth and health of the fish [12]. The emerging field of metagenomics has substantial implications for sustainable aquacul- ture, as diet, feeding strategy, and other environmental factors strongly influence the diversity and constitutive abundance of intestinal microbiota in both humans and fish [12–15]. In fin- fish, research has shown that probiotic maintenance of beneficial gut flora can promote growth, greater nutrient availability, and better stock health [16–17]. Early studies in channel

catfish (Ictalurus punctatus) and carp (Cyprinidae) identified several nutrients (e.g. biotin, pantothenic acid, vitamin B12), which are produced by intestinal microbes, but may be limit- ing in lesser-quality feeds [18–19]. Interestingly, proper intestinal flora in tilapia may also posi- tively impact human health as natural flora inoculates could theoretically out-compete non- natural pathogenic microbes. In Nile tilapia cultured in fertilized ponds found to be contami- nated with fecal coliform bacteria in Bangladesh, E. coli comprised up to 10% of fecal coliform bacteria identified by classic plating techniques, which could be passed on to consumers through improper storage and handling practices [20]. Absorption of nutrients from feed is facilitated through membrane bound transporters located within enterocytes of the gastrointestinal lumen [21–22]. Nutrient transporters are sol- ute-linked nutrient carriers with roles in transporting and exchanging nutrients, such as amino acids and sugars, and ions across epithelial cellular membranes [23]. Upregulation of nutrient transporter gene expression has been associated with enhanced utilization efficiencies of nutrients from the gut [24] and could play an important role in increasing uptake in tilapia fed at reduced frequencies [6, 25]. Deficiencies in key nutrients may also lead to upregulation of nutrient transporter gene expression or activity resulting in a greater ability to utilize the PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia

PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 2 / 23 Accession No. PRJNA485292. All other relevant data are within the paper and its Supporting Information files. Funding: This work was supported by the AquaFish Innovation Lab through the United States Agency for International Development/Oregon State University (USAID EPP-A-00-06-00012-00) and by participating institutions, particularly North Carolina State University and Bangladesh Agricultural University and the NSF (IOS-1457040 to RJB). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist.

limited amount of nutrients available [26]. This could play a role in increased feed efficiency in animals initially deprived of specific nutrients or of tilapia fed at reduced frequencies. Here, we evaluated potential underlying mechanisms behind the greater feed efficiency in fish fed on alternate days. It is currently unknown as to how reduced frequency feeding strate- gies may alter the gastrointestinal microbiome and nutrient transporter gene expression of tila- pia resulting in better nutrient utilization and feed efficiencies. Here, we evaluated the practices of alternate-day feeding to both establish whether this practice is effective in other areas of the world and to determine if changes in the tilapia gastrointestinal microbiome and nutrient trans- porter gene expression might account for improved feed efficiency under the management prac- tice. We also sought to assess if reducing the frequency further to every third day might further improve production efficiency of tilapia. We found no difference in growth parameters or sur- vival in tilapia fed on alternate days versus those fed daily. Nile tilapia fed on alternate days had improved feed efficiency and a greater diversity of gut prokaryotic and eukaryotic microorgan- isms, while transcript levels of select nutrient transporters changed little in fish fed on alternate days versus those fed daily, every third day or not at all. Taken together, this is the first study to suggest that enhanced gut microbial diversity may contribute to improved feed utilization under

reduced frequency feeding strategies in the culture of Nile tilapia and perhaps other finfish. Materials and methods All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at North Carolina State University (approval number: 13-139-0). All animals were euthanized using buffered tricaine methanesulfonate (MS-222) followed by cervical transection following standard procedures. Fish and pond sampling The tilapia growth trial was performed at the Fisheries Field Laboratory, Bangladesh Agricul- tural University, Mymensingh, Bangladesh. All-male sex reversed Nile tilapia (~3.5 g) were stocked at 5 fish/m2 in 16 ponds (0.1 ha; 4 replicates per treatment), with weekly pond fertiliza- tion at a rate of 28 kg N and 7 kg P/ha/week for all treatment groups. Twenty ponds were ran- domly divided into one of five treatments (4 ponds per treatment). The treatments were: (T1) commercial diet feeding daily with weekly pond fertilization, (T2) commercial diet feeding every other day with weekly pond fertilization, (T3) commercial diet feeding every third day with weekly pond fertilization, (T4) weekly pond fertilization with no supplemental feeding of commercial diet, and (T5) daily feeding with commercial diet with no pond fertilization. Fish were fed with formulated feed (CP Bangladesh, 30% crude protein) initially at 10% and then down to 3% body weight/day based on a standard tilapia feed schedule. Growth (length and weight) was monitored at 2-week intervals by subsampling of fish over

Scientific, Waltham, MA, USA) at room temperature for shipment to NCSU and subsequently stored at -80˚C until total RNA extraction. All five samples per pond were pooled together prior to total RNA extraction (See Nutrient Transporter Gene Expression, below; n = 4 pooled samples per treatment; N = 20). Tilapia fecal material was collected from the posterior intestine (colon; approximately 5 cm section 1 cm anterior to anus) and rRNA was extracted using a Xpedition Soil/Fecal DNA Miniprep kit (Zymo Research Corp., Irvine, CA, USA) following the manufacturer’s protocol (see Ribosomal RNA extraction, below). Samples from 4 fish per pond were pooled together from all 4 replicate ponds for all treatment groups (n = 4 pooled samples per treatment; N = 20). This pooled sample design was used to offset potential vari- ability of microbiota within individuals, instead focusing on common patterns, which may be more reflective of changes with treatment group among the population (pond) as a whole. Water quality Water temperature (˚C), transparency (cm), pH and dissolved oxygen (mg/L) were measured weekly and total alkalinity (mg/L), ammonia-nitrogen (mg/L), nitrate-nitrogen (mg/L), nitrite-nitrogen (mg/L), phosphate-phosphorus (mg/L) and chlorophyll-a (μg/L) were mea- sured biweekly between 0900–1000 h according to previously established protocols [27]. Collection and quantification of plankton samples Ten liters of water were collected biweekly from five locations in each pond for plankton sam- pling. The collected water was filtered through a 45 μm plankton mesh screen. The filtered

sample was brought to a standard volume of 50 mL using distilled water. The plankton samples were fixed in 5% neutral buffered formalin in 100 mL bottles for subsequent studies. Plankton was quantified using a 1 mL fixed plankton sample in a Sedgewick-Rafter (S-R) cell counting chamber on a Swift M4000 binocular microscope with phase contrast capabilities. Plankton identifications were made at the Family level. Growth and production parameters Growth and production parameters were calculated as below: Weight gainðgÞ ¼ ðMean final weight of fishðgÞÞ ðMean initial weight of fish ðgÞÞ Specific Growth RateðSGR; % body weight=dayÞ ¼ Lnðfinal weightÞ Lnðinitial weightÞ Culture period in days � � � 100 Survival rate % ð Þ ¼ No:of fish harvested No: of fish stocked � � � 100 Feed Conversion RatioðFCRÞ ¼ Total feed given ðgÞ Total weight gainðgÞ Gross production ¼ ðNo:of fish harvestedÞ � ðFinal weight of fish ðgÞÞ Net production ¼ ðNo: of fish harvestedÞ � ðWeight gain of fish ðgÞÞ PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 4 / 23

Benefit to Cost RatioðBCRÞ ¼ Total revenue Total cost Nutrient transporter gene expression From the publicly available tilapia genome assembly [28], we identified 6 candidate transport- ers putatively involved in the digestive transport of amino acids, dietary sugars, and lipids across the intestinal epithelium: (1) facilitated glucose/fructose transporter (slc2a5), (2) facili- tated glucose transporter (slc2a6), (3) long-chain fatty acid transporter (slc27a4), (4) Na+- amino acid transporter 2 (slc38a2), (5) Na+-amino acid transporter 4 (slc38a4), and (6) large neutral amino acid transporter subunit 3 (slc43a1). Tilapia gut transporter mRNA expression was quantified in anterior intestinal tissues (see Fish and Pond Sampling, above, for sampling details) using real-time quantitative PCR (qPCR) performed as previously described with a few modifications [29]. Briefly, total RNA was isolated from the anterior intestine using Tri Reagent (Invitrogen, Carlsbad, CA). One microgram of the total RNA was used to synthesize cDNA using a High Capacity cDNA Syn- thesis kit (Applied Biosystems, Carlsbad, CA) following treatment with Turbo DNA-free (Ambion, Foster City, CA). RNA was quantified and checked for quality at each step using a Nanodrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE) and agarose gel electrophoresis, respectively. Gene expression (mRNA) of the different transporter forms was measured in the tilapia cDNA using SYBR Green chemistry. Transporter gene-specific primers

(S1 Table; Integrated DNA Technologies, Inc., Coralville, IA) were designed using the IDT PrimerQuest Tool (Integrated DNA Technologies, Inc.). Optimization for appropriate anneal- ing temperature, primer concentrations, and cycling parameters was performed using pooled cDNA from the above reverse transcription reactions. One hundred ng of starting total RNA was used for qPCR analysis with Brilliant II QPCR Master Mix (Agilent Technologies, Inc., Clara, CA) containing 1.5 μM gene-specific primers. No template controls and no reverse tran- scription controls were incorporated into the assay. Pooled cDNA was used to produce the cDNA for creating the standard curves. All qPCR assays were run in triplicate wells on a CFX384 Real-Time PCR System (BioRad Laboratories, Inc., Hercules, CA). Cycling conditions were: 1 cycle, 50˚C for 2 min; One cycle, 95˚C for 10 min; 40 cycles, 95˚C for 15 s, 60˚C for 1 min. Melting curve analysis was performed to determine primer specificity. Data was normal- ized to the starting total RNA concentration. Gene copy number was calculated by comparing the mean cycle threshold (Ct) to the serially diluted cDNA standard curve (R2 = 0.98). The PCR efficiencies were between 90% and 110% for all reactions. The gene expression data were then normalized to the expression of 18S ribosomal RNA, whose levels were found to be simi- lar across treatment groups. Ribosomal RNA extraction Extraction of ribosomal RNA from tilapia fecal samples (see Fish and Pond Sampling, above,

for sampling details) was performed using an Xpedition™ Soil/Fecal DNA MiniPrep Kit (Zymo Research, Corp., Irvine, CA) following the included protocol. Up to 0.25 g of fecal sample was placed into a ZR BashingBead Lysis Tube with 750 μL Xpedition Lysis/Stabilization Solution. The tube was secured in an Xpedition Sample Processor and processed for 30 s and stored at room temperature until extraction. The concentration and quality were determined by PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 5 / 23

Nanodrop (Thermo Fisher Scientific, Inc., Waltham, MA). The extracted rRNA was stored at -20˚C for sequencing library preparations. Prokaryotic 16S and eukaryotic 18S rRNA sequencing library preparation Prokaryotic 16S and eukaryotic 18S rRNA gene amplicons were prepared following the 16S Metagenomic Sequencing Library Preparation protocol for the Illumina MiSeq system with some modifications. Primers were designed to amplify the V3 to V5 regions of 16S rRNA [30– 31] and the V9 region of 18S rRNA [32–33] with overhang adapter sequences compatible with the Illumina index and sequencing adapters. This allowed for double indexing to increase the accuracy of the multiplexed reads (S1 Table). Amplicon PCR was used to amplify the region of interest from the gDNA extracted from the tilapia fecal material samples. A Nile tilapia-specific blocking primer (S1 Table) was introduced during the 18S amplification step to reduce the amplification of tilapia 18S rRNA and increase the amplification of lower abundance eukary- otic rRNA present in the samples. The PCR was performed as follows: 1 cycle of 95˚C for 3 min; 25 cycles each of 95˚C for 30 s, 55˚C for 30 s, and 72˚C for 30 s; 1 cycle of 72˚C for 5 min; and hold at 4˚C. Clean-up of the PCR amplicon products to remove free primers and primer dimers was performed using Agencourt AMPure XP beads (Beckman Coulter, Brea, CA). Fresh 80% ethanol was prepared prior to clean-up. Following amplicon clean-up, index PCR was performed to attach indices to the amplicon PCR products. Dual-index primers were

designed so that samples could be multiplexed in one MiSeq lane (S1 Table). Index PCR was performed as follows: 1 cycle of 95˚C for 3 min; 8 cycles each of 95˚C for 30 s, 55˚C for 30 s, and 72˚C for 30 s; 1 cycle of 72˚C for 5 min; and hold at 4˚C. Clean-up of the PCR index prod- ucts was performed as above. All indexed amplicon concentrations were normalized and amplicons pooled into a single tube. The pooled library was verified for quality and quantified using a High Sensitivity DNA chip on the Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA). The library was diluted and combined with a PhiX Control library (v3) (Illumina, San Diego, CA) at 10%. The library was sequenced on an Illumina v3 300PE MiSeq run, using standard sequencing protocols at the Genomic Sciences Laboratory at North Carolina State. Base calls were generated on-instrument during the sequencing run using the MiSeq Real Time Analysis (RTA 1.18.54) software and fastq generation. Demultiplexing, adapter trimming, and quality filtering were performed by the MiSeq Reporter Software (2.4 and 2.5.1). The library was run on two lanes to increase the number of reads for each sample. Sequence and statistical analyses The resulting demultiplexed reads were processed using the QIIME 1.9.1 toolkit [34]. Briefly, the paired end reads were joined together and UCLUST [35] was used to search against Green- genes 13_8 reference database [36] for 16S (prokaryote) analysis and SILVA release 119 refer- ence database [37–39] for 18S (eukaryote) analysis, both filtered at 97% identity. Reads not

matching a reference sequence were removed from analysis. OTUs were assigned based on a database hit of 97% or greater sequence identity and taxonomy was assigned against the appro- priate database. Core diversity analysis was used to perform α-diversity and β-diversity and their rarefaction functions. Weighted and unweighted Unifrac distances [40] were used to compute the between sample diversity which was visualized using principal coordinates analy- sis (PCoA) plots with Emperor [41]. Significant differences between treatment groups where pond was the experimental unit were determined by one-way ANOVA (Analysis of Variance) followed by Tukey’s post-hoc tests or two-way ANOVA analysis followed by pairwise comparisons using t-tests using SPSS (Statistical Package for Social Science) version-16.0, JMP 12 (SAS Institute, Cary, NC), PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 6 / 23

GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA), or R 3.6.3 [42]. Significance was assigned at an α of 0.05. Venn diagrams were produced using InteractiVenn [43]. Results Effect of feeding frequency on tilapia growth performance and production There was no significant difference in weight gained, specific growth rate, survival or yield of Nile tilapia between groups fed daily with (T1) or without (T5) pond fertilization and those fed on alternate days with pond fertilization (T2; Fig 1; Table 1). These treatments had the highest fish growth and net production followed by those fish fed every third day in fertilized ponds (T3; Table 1). There were also no significant differences in length (Fig 1a) or weight (Fig 1b) Fig 1. Mean lengths (a) and weights (b) of Nile tilapia in reduced-feeding regimes throughout the study (error bars have been removed for clarity; see Table 1). � = P < 0.05. �� = P < 0.01. ND = No Difference. https://doi.org/10.1371/journal.pone.0236100.g001 Table 1. Growth and production performances of Nile tilapia (O. niloticus) subjected to reduced-feeding regimens. Fed Daily with Fertilization (T1) Fed Alternate Days with Fertilization (T2) Fed Every Third Day with Fertilization (T3) Fertilization Only (T4) Fed Daily Only (T5) Initial Weight (g) 3.55±0.90 3.55±0.90 3.55±0.90 3.55±0.90 3.55±0.90a Final Weight (g) 127.63±2.75a 120.17±5.44a 85.10±11.13b 43.15±4.28c 129.53±8.59a Weight Gain (g) 124.08±2.75a 116.62±5.44a 81.55±11.13b 39.60±4.28c 125.98±8.59a SGR (%/day) 3.14±0.02a 3.09±0.04a 2.78±0.12b 2.19±0.09c 3.15±0.06a

FCR (Feed Conversion Ratio) 1.64±0.10c 0.93±0.09b 0.68±0.15a - - 1.61±0.10c Survival Rate (%) 93.44±6.26a 91.66±8.00a 90.70±9.74ab 76.79±2.68b 97.71±2.11a Total Yield (kg/ha) 6,282.09±354.52a 5,837.84±527.83a 4,179.08±640.07b 1,950.35±233.89c 6,578.53 ±461.41a Values are mean ± SD. Values with different letters are significantly different (P < 0.05). https://doi.org/10.1371/journal.pone.0236100.t001 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 7 / 23

between groups fed daily with (T1) or without (T5) fertilization throughout the trial (two-way ANOVA; P > 0.05). There was, however, a significant difference between these two treatments (T1 and T5) and the fish fed on alternate days in fertilized ponds (T2) during weeks 7 through 13 (P < 0.05), but no significant difference between these three treatments at weeks 15 and 17 (P > 0.05; Fig 1a). Weight followed similar trends, except that there was also no significant difference between these three treatments (T1, T2, and T5) during weeks 13, 15, and 17 (P > 0.05; Fig 1b). Fish that were grown in fertilized ponds without supplemental feeding (T4) showed the lowest growth among all groups. The fish produced with pond fertilization alone also had the lowest survival relative to the other groups (Table 1), whose survival rates ranged from 92–98%. The estimated feed conversion ratio (FCR) was lowest for fish fed every third day with pond fertilization (T3), followed by fish fed on alternate days with pond fertilization (T2; Table 1). Fish fed daily had the lowest feed efficiency or highest FCRs regardless of whether ponds were fertilized (T1) or not (T5) as these groups had greater feed inputs, but similar growth to those fed on alternate days with pond fertilization (T2). The economic assumption of alternate-day feeding is that it would lead to more cost-effec- tive tilapia aquaculture production, which could enhance income of small-scale farmers and consumption of nutrient rich fish. A marginal cost analysis from this study indicates feeding

on alternate days along with weekly pond fertilization (T2) gives the best return on investment (Table 2) with much of the cost reduction associated with application of half the feed in alter- nate day fed fish. Net return values were also highest for fish fed on alternate days with pond fertilization (T2), followed by those fed every third day with pond fertilization (T3; Table 2). Overall, the group fed on alternate days along with weekly pond fertilization (T2) had the high- est benefit to cost ratio (BCR) but was not significantly different from groups fed every third day or not fed supplemental feeds in fertilized ponds (T3 or T4); all three of these treatments had significantly higher BCRs than fish fed daily regardless of pond fertilization (T1 and T5). Pond water quality and plankton composition All water quality parameters among groups were well within the range for suitable growth of fish in pond culture (S2 Table) [4, 44]. Ammonia was highest in the group fed daily with pond Table 2. Economic analyses of Nile tilapia (O. niloticus) subjected to reduced-feeding regimens. Fed Daily with Fertilization (T1) Fed Alternate Day with Fertilization (T2) Fed Every Third Day with Fertilization (T3) Fertilization Only (T4) Fed Daily Only (T5) Expenditure (USD/pond)�† Fingerlings cost 11.99±0.65 11.99±0.65 11.43±2.30 11.43±2.30 12.64±0.65 Feed cost 113.88±7.20b 59.82±2.16c 29.87±7.64d - - 124.05±8.19a Lime Cost 0.58±0.03 0.58±0.03 0.55±0.11 0.55±0.11 0.61±0.03 Fertilizers cost 4.60±0.25 4.91±0.27 4.67±0.94 4.67±0.94 - -

Total Expenditure 131.05±7.98a 77.29±3.01b 46.52±10.95c 16.65±3.35d 137.30±8.73a Income�† Gross return (USD/ ha) 9,774.19±1,081.58ab 9,082.98±981.20b 5,418.46±659.15c 1,770.14±294.04d 10,235.40 ±1,063.34a Net return (USD/ha) 898.09±563.51c 3,853.63±813.67a 2,250.83±997.06b 610.42±212.27c 1,421.67±612.94c BCR (Benefit Cost Ratio) 1.10±0.06b 1.74±0.16a 1.70±0.35a 1.52±0.18a 1.16±0.07b Values are mean ± SD. Values with different letters are significantly different (P < 0.05). � USD = United States Dollar; BDT = Bangladeshi Taka. Fish were sold at $1.56 USD (120 BDT)/kg [1 BDT = $0.013 USD]. †Leasing and other operational costs such as labor were excluded from these calculations. https://doi.org/10.1371/journal.pone.0236100.t002 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 8 / 23

fertilization (T1) and lowest in the fertilization alone group (T4). Both nitrites and nitrates were higher in fertilized ponds (T1-T4) and lowest in non-fertilized, fed daily ponds (T5). Dis- solved oxygen levels were highest in ponds either fed every 3 days with fertilization or those fertilized only (T3 and T4) and lowest in ponds fed daily independent of whether ponds were fertilized or not (T1 and 5). The pH was highest in the ponds that were fertilized only (T4) and lowest in ponds that were fed a commercial diet (T1-T3, T5). Six Families of eukaryotic phytoplankton and four groups of zooplankton were identified over the course of the pond trials (S3 Table). The eukaryotic phytoplankton Families identified were Bacillariophyceae, Chlorophyceae, Cyanophyceae, Euglenophyceae, Rhodophyceae, and Xanthophyceae and zooplankton were Copepoda, Crustacea, Rotifera, and Cladocera. T5 had significantly higher counts of the phytoplankton Family Euglenophyceae (4035.7 ± 466.8 SEM cells/L) whereas T1 had the lowest counts of this family (1750.0 ± 252.5 SEM cells/L; P = 0.0215). There was no difference in total eukaryotic phytoplankton between treatments (P = 0.0851). For zooplankton, T1 had significantly higher counts of Phylum Rotifera (6482.1 ± 1291.0 SEM cells/L; P < 0.001) and Order Cladocera (1464.3 ± 198.4 SEM cells/L; P = 0.0001) and significantly higher total zooplankton (11571.4 ± 1854.9 SEM cells/L; P = 0.0003) than the other treatments. There was no difference in total plankton between any of the treatments (P = 0.221).

Effect of feeding frequency on intestinal nutrient transporter expression The mRNA levels of six intestinal nutrient transporters were measured via real-time qPCR in the anterior intestine of fish from the growth trial. The nutrient transporters evaluated were: (1) facilitated glucose/fructose transporter (slc2a5), (2) facilitated glucose transporter (slc2a6), (3) long-chain fatty acid transporter (slc27a4), (4) Na+-amino acid transporter 2 (slc38a2), (5) Na+-amino acid transporter 4 (slc38a4), and (6) large neutral amino acid transporter subunit 3 (slc43a1). The efficiency of all real-time qPCR reactions were between 90–110% and the stan- dard curve correlations were greater than 0.97 for all assays performed. Our resulting gene expression values (copy number / ng Total RNA) were normalized to 18S rRNA levels (no sig- nificant difference between treatments was observed in the 18S rRNA gene). There was no significant difference in gene expression of the 6 nutrient transporters ana- lyzed between treatments where supplemental food and pond fertilization was provided (T1—T3) regardless of the frequency of feeding, except that slc38a2 was significantly lower in T2 than T1 and T3 fish (Fig 2). Gene expression for the slc2a5, slc2a6, slc27a4, and slc38a2 transporters was significantly higher in fish from T4 (pond fertilization only) than T5 groups (daily feeding only). There was no difference among treatments in the mRNA abundance of the slc38a4 and slc43a1 transporters (Fig 2e and 2f). Effect of feeding frequency on the gut microbial community—prokaryotic assessment

Following quality filtering, a total of 715,725 paired end sequences (330,883 for 16S; 384,842 paired end sequences for 18S) were obtained from the V3—V5 regions of 16S rRNA (prokary- ote) and the V9 region of 18S rRNA (eukaryote). All resulting Fastq files can be accessed in the National Center for Biotechnology information (NCBI) under GenBank database BioProject Accession No. PRJNA485292. A total of 20 prokaryotic phyla, 43 classes, 92 orders, and 145 families were found in the fecal material of tilapia (100% bacteria, no archaea). The dominant phyla identified from these samples belonged to the Fusobacteria (80.4%), Firmicutes (13.8%), Cyanobacteria (2.4%), Bac- teroidetes (1.3%), and Proteobacteria (1.3%) (Fig 3a, S4 Table). The proportions of the PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 9 / 23

identified microbes varied between treatments with an increase in the proportion of Fusobac- teria (62.6% to 84.6%) and a decline in the proportion of Firmicutes (32.1% to 11.5%) with reductions in feeding frequency from daily to every 3rd day (T1-T3). The fish fed a commer- cial diet only with no fertilization (T5) had the highest proportion of Fusobacteria (90.8%) and lowest proportion of Firmicutes (6.5%) among the treatments. There were 10 unique bacterial OTUs identified in T1, 6 in T2, 1 in T3, 19 in T4, and 4 in T5 (Fig 4, S5 Table). Alpha diversity and rarefaction curves were determined for all treatments evaluated. T1 and T2 had the highest Chao 1 diversity index among the groups (P < 0.05; Chao1 index mean ± SEM; T1, 392.9 ± 108.5 and T2, 371.8 ± 81.0). The bacterial diversity in the fecal mate- rial in tilapia guts declined as the frequency of feeding was reduced to every 3rd day (T3) or where no supplemental feeds were applied (T4) when paired with pond fertilization (Fig 5a, S1a Fig). Although there was no significant difference in species richness due to sample varia- tion within treatments (P = 0.256), treatments with greater access or variety of foodstuff (feed and fertilization) had higher diversity measures. The Chao 1 index was highest for T1 and T2. T3 (187.3 ± 77.6), T4 (244.0 ± 44.2), and T5 (215.5 ± 66.4) had similar diversity indices. Our Fig 2. Gene expression of nutrient transporters in the intestine of Nile tilapia subjected to reduced-feeding

regimes (mean ± SEM). (a) SLC2A5; (b) SLC2A6; (c) SLC27A4; (d) SLC38A2; (e) SLC38A4; (f) SLC43A1. All values reflect copy number / ng Total RNA normalized to 18S rRNA levels. Treatments with different letters are significantly different (P < 0.05). https://doi.org/10.1371/journal.pone.0236100.g002 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 10 / 23

results indicate that there are some community differences between the treatments in our study with PC1 explaining 11.89% of the variation between communities (Fig 6a and 6b). T1 and T5 (both fed daily) had the most similar profiles while T4 (fertilization only) was the most dissimilar. The variation between treatments indicates that there is some distinction between the microbial communities. Fig 3. Relative abundance of prokaryotic (a) and eukaryotic (b) organisms at the level of Phylum in the fecal material of Nile tilapia subjected to reduced-feeding regimes. https://doi.org/10.1371/journal.pone.0236100.g003 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 11 / 23

Effect of feeding frequency on the gut microbial community—eukaryotic assessment All resulting Fastq files can be accessed in the National Center for Biotechnology information (NCBI) under GenBank database BioProject Accession No. PRJNA485292. A total of 8 major eukaryotic groups and 132 genera were associated with the fecal material of tilapia used in this study. T2 (fed alternate days with pond fertilization) had the greatest number of reads for all groups (58,450) followed by T4 (fertilization only; 49,112), T1 (feed every day with fertiliza- tion; 38,378), T5 (feed every day only; 31,758), and T3 (feed every third day with fertilization; 25,752). The dominant groups identified from these samples belonged to the Opisthokonta (Metazoans and Fungi; 41.2%), Archaeplastida (green plants and red algae; 35.4%), and SAR supergroup (Stramenopiles, Alveolates, and Rhizaria, Dinoflagellates, Diatoms, Oomycetes, etc.; 23.3%) (Fig 3b, S4 Table). There were no significant differences in abundance of particular eukaryotic groups among the treatments. Class Chlorophyceae (green algae), Class Mediophy- cae (diatoms), Phylum Rotifera, and group Magnoliophyta (flowering plants) made up the highest proportion of identified eukaryotic organisms (not shown). T1 and T3 had the highest proportions of rotifers (31% and 35%, respectively). The largest proportions of Magnoliophyta were in T2 (19%) and T3 (20%) and Chlorophycae were in T1 (27%) and T4 (24%), while T5 had the highest proportion of Mediophycae (21%) relative to all other treatments. However,

there were no significant differences in abundance of eukaryotes among the treatment groups. Fig 4. Venn diagram showing the number of shared and unique Operational Taxonomic Units (OTUs) between each treatment. Numbers in parenthesis are the total number of OTUs identified in each treatment. https://doi.org/10.1371/journal.pone.0236100.g004 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 12 / 23

The most abundant identified groups overall include the rotifers (50,705 counts), green algae (34,755 counts), and flowering plants (19,951 counts). Alpha and beta diversity measures were also determined for the 18S amplicons. T2 had the highest diversity (Chao1 index, mean ± SEM, 206.9 ± 33.6), followed by T1 (182.0 ± 52.4), T4 (172.0 ± 50.4), and T3 (141.5 ± 59.6) (P < 0.05). T5 had the lowest diversity measure at 123.9 ± 47.1 (Fig 5b, S1b Fig). Again, the results indicate that there are some community differ- ences between the treatments in our study with principal component 1 explaining 22.28% of the variation between communities (Fig 6c and 6d). The groupings of these samples, although with overlap, indicate that there are differences in the microbial communities between treatments. Fig 5. Chao1 diversity boxplots of microbial communities identified from the anterior intestine of Nile tilapia subjected to reduced-feeding regimes (mean ± SEM). (a) 16S rRNA (prokaryotes); (b) 18S rRNA (eukaryotes). Treatments with different letters are significantly different (P < 0.05). https://doi.org/10.1371/journal.pone.0236100.g005 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 13 / 23

Discussion It is estimated that 50–70% of total variable costs for growing tilapia and many other fishes is attributable to formulated feeds. Our study demonstrates that alternate-day feeding along with pond fertilization reduces the costs of feeds for growout of Nile tilapia by 50%, increases feed efficiency by 76%, and has little impact on growth, survival, or yield of fish farmed in ponds. This is noteworthy as both ponds fed every day with (T1) or without (T5) pond fertilization had the highest expenditures in our study while ponds that were fertilized only (T4) with only pond fertilization had the lowest cost input. Tilapia producers can see marked reduction in production costs and the highest benefit: cost ratio (BCR) largely due to the reduced feed costs associated with alternate-day feeding strategies. The overall improved capacity of fish fed on alternate days to utilize nutrients for growth (lower feed conversion ratio, FCR) was accompa- nied by intermediate expression of genes for proximal intestinal solute transporters and ele- vated diversity of gastrointestinal prokaryotes and eukaryotes, suggesting that these factors along with the higher biodiversity of organisms available for consumption (commercial diet and natural pond productivity) may contribute to enhanced feed efficiency while sustaining similar growth and health of tilapia when fed daily. The results suggest that the optimal BCR and net return for tilapia production occurs when fish are fed on alternate days in fertilized ponds. This is contrary to studies performed in chan-

nel catfish. Fingerling growth, weight at harvest, and FCR were all lower in channel catfish fed on alternate days versus those fed every day [45]. Tilapia are omnivorous, and more so than Fig 6. Principal coordinates analysis (PCoA) plots of prokaryotic (a and b) and eukaryotic (c and d) communities found in the fecal material from Nile tilapia subjected to reduced-feeding regimens. Each replicate sample is represented in the plot. https://doi.org/10.1371/journal.pone.0236100.g006 PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 14 / 23

catfish, readily feed on diverse food sources throughout their life, including phytoplankton and zooplankton or other natural pond foods enhanced through fertilization or application of supplemental formulated feeds. Hence, tilapia fed on alternate or even every third day likely forage at a greater rate between meals on pond organisms produced through enhanced nitro- gen and phosphorous inputs from fertilizers or feeds than fish fed daily. Less reliance on sup- plemental feeds will increase overall feed efficiency as demonstrated here with low FCR, albeit at a detriment to growth when animals are fed at a frequency of every three days. Clearly, suffi- cient nutrients, derived in this case from formulated feed, is essential to maximally stimulate growth as we found alternate day fish achieved weights and lengths similar to daily fed fish while those fed every third day did not. The responses to alternate day feeding we observed here with studies in Bangladesh corroborate those previously shown in the Philippines on commercial farms where tilapia were stocked at a 20% lower density and fed a formulated diet composed of high levels of fishmeal and rice bran [6]. Water quality parameters are affected by excess nutrients, but small variations within normal range are unlikely to account for any of the changes in tilapia growth observed in this study. All water quality parameters were well within the suggested range for growing tilapia in ponds (S2 Table) [4, 44]. Collectively, the findings indicate that addition of supplemental feeds can enhance growth of tilapia by 2–4 fold

depending on the frequency of feed application and that alternate day feeding is the most cost- effective feed management method for maximizing growth and yield of tilapia in semi-inten- sive culture, a strategy that is likely to be applicable for pond production of tilapia throughout the world. Reduced feeding strategies can achieve equivalent production yields with less feed; however, the mechanisms are poorly understood [7–11]. Absorption of nutrients from feed is facilitated through membrane bound transporters located within enterocytes of the gastrointestinal lumen [21–22]. We evaluated if changes in expression of key solute-linked nutrient carriers (SLC), whose abundance may impact nutrient absorption efficiency [21] might be associated with the changes in feed efficiency and growth of Nile tilapia shown here under the different feeding protocols. We used the proximal (anterior) portion of the small intestine to measure solute transporter gene expression levels. Tilapia have a long, coiled intestine in order to pro- cess and absorb nutrients from their mainly herbivorous or omnivorous diet. Tilapia lack a pyloric caecum, thus the small intestine is the region of highest nutrient absorption, especially in the proximal area which is the most developed region of the small intestine and exhibits the highest enzymatic activity and absorption of nutrients [46–47]. A number of ATP-binding cas- sette (ABC) efflux transporter genes associated with multi-drug resistance and efflux of xeno-

biotics are also expressed in the proximal small intestine of Nile tilapia [48] and SLCs are expressed in a similar region of the intestine in zebrafish (Danio rerio) [49]. The expression of solute transporters slc2a5, slc2a6, slc27a4, and slc38a2 in the proximal intestine of the Nile tilapia were significantly higher in fish grown in ponds where only fertiliz- ers were applied (T4) than in fish grown in ponds which were fed daily without the addition of fertilizers (T5) (Fig 2). The seemingly low expression of all transporters examined in the intes- tines of fish fed daily without pond fertilization (T5) may in part be the result of receiving simi- lar quantities of feed with the same nutrient profile day in and day out, hence preventing the need for any adaptive response from the intestine [50]. It may also be due to higher rates of intestinal growth in these fish. Increased food consumption can induce lengthening of the intestine and proliferation of the villi to enhance nutrient uptake as well as increase cell turn- over rates. This in turn can increase the proportion of immature, non-transporting cells and cause a non-specific change in overall transporter abundance [50]. Conversely, the relatively high expression of these transporters in fish grown in ponds where only fertilizers were applied (T4) likely indicates that gaining nutrition from natural flora and fauna boosted by fertilization PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 15 / 23

application alone may lack sufficient nutrients necessary to maintain growth and health that would otherwise be available with commercial diets. Thus, gene expression for these transport- ers may be increased due to the lack of nutrients and for preparing the intestinal cells to rapidly take them up once they become available [50]. Interestingly, fish grown in ponds that received feed and fertilization (T1, T2, and T3), particularly those fed every two (T2) or three days (T3), had intermediate mRNA abundances of certain solute transporters (Fig 2). This suggests that they could be more efficient at nutrient uptake and utilization, particularly when commercial feeds are only periodically available, and that they may perhaps have a greater variety of trans- porters expressed at lower quantities due to an increased diversity and abundance of natural food sources in the diet. It is interesting that the fish in T1 (daily feeding with fertilization) had approximately 3-fold higher levels of these transporters than T5 (daily feeding only) as both groups were fed daily, but is likely the result of a more varied diet and greater overall abun- dance of food sources. The solute transporters slc2a5, slc38a2, and slc27a4 exhibited similar expression profiles across the five treatments. When a nutrient yields high metabolizable energy, high concentra- tions will typically stimulate an upregulation of its transporter as is the case for sugars and fatty acids. By contrast, when a nutrient can be toxic at high concentrations, it will normally cause a

reduction in the expression of its transporter [50]. GLUT5, the protein produced from the slc2a5 gene, is a facilitated fructose transporter in vertebrates [23, 51]. Although a member of the facilitated glucose transporter superfamily (SLC2), glucose appears to be a minor compo- nent of its mechanisms. GLUT5 is essential for the transport of fructose across the luminal membrane of the small intestine and transcription of slc2a5 in the intestine is rapidly increased when a high-fructose diet is consumed [52–54]. Thus, the relatively higher levels of this trans- porter in fish from ponds with fertilization only (T4) may indicate a greater abundance of fruc- tose in the natural flora and fauna relative to the commercial diet. Similarly, slc27a4 gene expression was highest in fish grown in fertilized ponds without commercial feed as well as in fish fed at different frequencies along with pond fertilization. These increases may be dictated primarily by the nutrients available in the natural products produced by pond fertilization. Slc27a4 encodes the FATP4 fatty acid transporter. It promotes lipid absorption in the mouse ileum [55] and the gene is located in a growth-related quantitative trait loci (QTL) in the orange-spotted grouper (Epinephelus coioides) [56]. The bacterium Clostridium ramosum increases slc27a4 gene expression with high-fat diets in mice [55]. Although C. ramosum was not directly identified in the guts of the Nile tilapia in this study, several sequences did align to members of Clostridium sp. OTUs.

Contrary to the fructose and fatty acid transporters, gene expression of slc38a2 increases in response to the absence of amino acids. This increase leads to the translation of the gene to make the SNAT2 protein which has been shown to aid the recovery of cell volume following amino acid starvation or hypertonic stress [57–61]. In addition, amino acid transporters may serve as sensors of nutrient supplies to regulate expression of the transporters and thus nutri- ent uptake by the cells [62]. The results show that expression of this gene is upregulated in T4 with only pond fertilization and lower in fed treatments regardless of fertilization. This implies that natural biota increased through fertilization of ponds may lack the full complement of amino acids necessary for the intestine to maintain cellular osmolality. In this case again the slc38a2 would be upregulated to improve the efficiency of absorbing amino acids under limit- ing conditions whereas in fish receiving commercial feeds that typically have adequate protein or amino acids, the transporter is expressed at lower levels to prevent excessive amino acid uptake which can lead to toxicity [50, 63]. This investigation also assessed how the gut microbial flora is altered by feeding/fertilization strategies that could potentially identify microbes beneficial to tilapia growth and health. The PLOS ONE Enhanced biodiversity of gut flora and feed efficiency in pond cultured tilapia PLOS ONE | https://doi.org/10.1371/journal.pone.0236100 July 23, 2020 16 / 23