Abstract

   The American lobster (Homarus americanus) is an economically important
   species in the western Atlantic and its climate-driven range shift
   northward along the east coast of the United States is well documented.
   The thermal tolerance of lab-reared postlarvae of this species has been
   extensively investigated to better understand settlement and
   recruitment dynamics. However, there have been few studies focused on
   wild-caught postlarvae, and even fewer that have analyzed lab-rearing
   conditions in context of diet. In this study, we investigated gene
   transcriptional changes in postlarvae caught in the wild, as well as
   postlarvae reared in the laboratory on a brine shrimp diet or a
   wild-sourced zooplankton diet. We found between wild-caught and brine
   shrimp-reared larvae 3,682 differentially expressed genes, and between
   wild and zooplankton-reared postlarvae 3,939 differentially expressed
   genes. Between the two lab-reared groups fed different diets 2,603
   genes were differentially expressed. We also exposed individuals in all
   rearing groups to chronic temperature treatments of 8°C and 26°C and
   found that both temperature extremes elicit 68–95% fewer
   transcriptional changes in wild postlarvae compared to either
   lab-reared group. In wild postlarvae, we identified differential
   expression of transcripts within the FoxO signaling pathway, a
   signaling pathway with a central role in cellular physiology, as
   potential molecular markers for cold tolerance in the American lobster.
   These findings contextualize the current literature on lab-reared
   postlarvae as containing conservative estimates for in situ organisms
   and can be used to inform population distribution modeling efforts.
   They also provide evidence for the need to adjust lab-rearing
   techniques or source wild larval crustaceans to augment studies of
   larval biology.

Introduction

   The American lobster (Homarus americanus) is an economically important
   species [[32]1] with a well-documented climate-driven northward range
   shift along the coast of the eastern United States [[33]2, [34]3]. The
   settlement and recruitment success of early life stages subsequently
   dictates population distribution [[35]4–[36]7]. After hatching,
   lobsters go through three planktonic larval stages with varying
   zonation in the water column [[37]8], and a planktonic postlarval
   stage, before settling as stage V juveniles at the benthos. Stage IV
   postlarvae are estimated to spend up to 65% of their time in the upper
   0.5 meters of the water column before settling to molt to stage V
   [[38]9], and their distribution within rapidly warming surface waters
   leaves them more vulnerable to climate change than their
   deeper-dwelling (stages II and III) and benthic (stage V) counterparts
   [[39]10, [40]11].

   In fact, a disconnect between the abundance of egg-bearing females and
   the abundance of juveniles has put pre-settlement life stages, and
   specifically stage IV postlarvae, at the forefront of research on early
   development of this species [[41]9, [42]12–[43]17]. Recent recruitment
   failure has been attributed to sea surface temperature anomalies in the
   Gulf of Maine [[44]18]. Other field-based studies have outlined how
   thermal tolerance, altered molting times, and food
   availability–especially at the extreme thermal ranges of the species’
   distribution–together influence in situ population dynamics [[45]19].
   In particular, the climate-driven decrease in copepod prey species
   Calanus finmarchicus has been investigated as a driving force behind
   recruitment failure [[46]20].

   Lab-based studies have the benefit of isolating individual factors,
   like temperature, to determine their influence on larval fitness and
   success. In such studies, it has been shown that increased seawater
   temperatures in early life are linked to increased mortality and
   decreased development time in crustaceans [[47]21–[48]24]. Further, the
   first signs of sub-lethal stress in lab-reared stage IV postlarvae
   occur from acute exposures to 26–33.8°C [[49]25] and larval performance
   and growth are reduced after chronic exposure to temperatures of 23°C
   and above [[50]21]. More contemporary lab-based studies have
   corroborated these findings, outlining that development time decreases
   with increasing temperature from 10°C to 22°C [[51]26].

   RNA-Seq is a useful and recently popularized tool for probing changes
   in gene expression in larval crustaceans, especially in the context of
   thermal tolerance and climate change [[52]27]. Using this approach in a
   separate study, we have shown that the regulation of transcripts
   related to heat and UV tolerance is altered at different life stages in
   lab-reared American lobster. Stage IV postlarvae exhibit elevated
   expression of DnaJ homologs, which are notable for their involvement in
   the cellular response to heat and UV stress [[53]28]. The ability of
   lab-reared postlarvae to reallocate energy to important physiological
   functions like preventing and repairing cellular damage is compromised
   under projected future climate warming scenarios [[54]15, [55]17].
   Additionally, projected warming scenarios, in conjunction with elevated
   CO[2], alter the regulation of genes related to shell-building
   processes and immune function in lab-reared postlarvae [[56]16].

   Historically, lab-based studies on lobster larvae have utilized larvae
   hatched in the lab and reared on excess, freshly hatched brine shrimp
   (Artemia spp.). These larvae are also typically reared at constant
   temperatures, usually between 15°C and 20°C, with 18°C yielding optimal
   growth rates for American lobster [[57]29]. It is widely accepted that
   using lab-reared organisms reduces variability among individuals.
   Though this practice is commonplace, there is a growing body of
   evidence in crustaceans and other animals, that lab-reared individuals
   of many marine species may not be fully representative of their wild
   counterparts. In lab-reared fish, a diet of Artemia spp. yields
   nutritional deficiencies of essential fatty acids EPA/DHA and vitamin C
   [[58]30]. Lab-reared fish are less efficient at catching prey and
   avoiding predators than their wild counterparts [[59]31, [60]32]. In
   Homarus americanus, similar results have been found, with swimming
   speeds of 18 cm/s observed in wild postlarvae compared to swimming
   speeds of 10 cm/s observed in lab-reared postlarvae [[61]13]. Brine
   shrimp-fed postlarvae are also known to be deficient in polyunsaturated
   fatty acyl chains, and one study suggests that copepods may be worth
   investigating as an alternative feed for lobster hatcheries [[62]14].

   To identify cellular processes driving the observed differences between
   lab-reared and wild-caught Homarus americanus postlarvae, we
   investigated wild-caught postlarvae, lab-reared postlarvae fed the
   traditional diet (freshly hatched brine shrimp), and lab-reared
   postlarvae fed freshly caught zooplankton. We applied transcriptomics
   to identify the reorganization of cellular pathways as a result of
   rearing condition. Additionally, we subjected individuals from each of
   these conditions to a lower (8°C) and upper (26°C) thermal exposure to
   see how each rearing group’s molecular thermal response varied. To the
   best of our knowledge, our study is the first to investigate changes in
   gene expression of H. americanus postlarvae in response to different
   rearing conditions.

Methods

Animal husbandry

   Eight ovigerous female lobsters were sourced from lobster management
   zone E in the Gulf of Maine in the summer of 2022 through the State of
   Maine Department of Marine Resources ventless trap survey. The
   ovigerous lobsters were held under State of Maine Department of Marine
   Resources Special License #2022-19-04 at the Bigelow Laboratory for
   Ocean Sciences in East Boothbay, Maine, in a common flow-through tank
   fed by the Damariscotta River estuary (average tank temperature 14.4°C;
   salinity 30–32 ppt). Hatchlings were fed within 1 to 10 hours of hatch
   and isolated within 24 hours in individual 400 mL glass jars filled
   with 0.45 μm filtered sea water. Jarred individuals were acclimated to
   the documented optimal rearing temperature of 18°C [[63]29], and this
   temperature served as a control in our experiments. Approximately 90%
   of the water in each jar was changed three times per week. Water was
   not aerated, but regular water changes yielded average dissolved oxygen
   levels of 7.2 mg/L.

Rearing condition experiments

   A subset (n = 5) of lab-reared larvae were fed freshly hatched Artemia
   sp. ad libitum, the traditional diet fed to larval lobsters in a
   research setting. Another subset (n = 5) of lab-reared larvae were fed
   a diet of freshly caught zooplankton ad libitum, representative of
   their natural diet. On average, postlarvae fed zooplankton were 13%
   larger and 435% heavier than the postlarvae fed brine shrimp (weight:
   wild-caught 26.3 +/- 9.20 mg, brine shrimp fed 4.9 +/- 0.90 mg,
   zooplankton fed 5.6 +/- 0.40 mg; carapace length wild-caught 6.16 +/-
   0.53 mm, brine shrimp fed 3.94 +/- 0.54 mm, zooplankton fed 3.87 +/-
   0.19 mm). Zooplankton was caught in vertical tows within 48 hours of
   feeding in the same area the mothers were sourced from. These tows, on
   average, were comprised of 62% copepods, including Calanus
   finmarchicus, a potential prey species whose abundance is significantly
   correlated to successful lobster settlement [[64]20]. Additionally,
   fish eggs and other larval crustaceans were captured. Most of the
   contents of our zooplankton tows have been documented as prey items in
   gut content analyses of larval lobsters [[65]12]. Lab-reared lobsters
   from both feeding conditions were sacrificed 48 hours after molt to
   stage IV. Wild stage IV lobsters, identified based on morphology (n =
   5) were also captured in lobster management zone E using a neuston net
   with 1 mm mesh [[66]1] and held in the lab at ambient temperature
   (temperature of the water in the flow-through tank fed by the
   Damariscotta River estuary; average temperature 14.4°C (minimum 11.0°C,
   maximum 19.6°C) for 48 hours before being preserved for analysis. Wild
   postlarvae were, on average, 13% larger than their brine shrimp-reared
   counterparts.

Temperature treatments

   A subset of lobsters from each rearing condition (n = 5 per group) was
   subjected to a low-temperature treatment of 8°C and another subset at a
   high-temperature treatment of 26°C for 48 hours, except for
   zooplankton-fed postlarvae, which were only analyzed for cold
   tolerance. These temperature treatments resemble upper and lower
   extremes that these larvae experience in their habitats. Lab-reared
   lobsters were acclimated to their respective temperature treatments
   within 24 hours after molt to stage IV and maintained at these
   temperatures for 48 hours until sampling. Wild-caught stage IV lobsters
   were immediately acclimated to their respective temperature treatments
   upon being brought back to the lab and maintained at their respective
   temperatures for 48 hours before sampling. Sampling time points were
   kept consistent so lobsters in rearing condition experiments could be
   used as controls.

Sample preservation

   Sampled lobsters (n = 5 per rearing condition, per temperature
   treatment) were rinsed with milli-Q water before preservation in
   DNA-RNA-free microcentrifuge tubes containing 600 μL of RNAlater
   solution (Ambion Inc.). Samples were held at 8°C for 12 hours to allow
   the preservative to permeate the tissue before being transferred to
   -80°C for storage until analysis. Samples were shipped on dry ice to
   Novogene (Sacramento, CA) for extraction, sequencing, and analysis,
   following their internal protocols.

Library preparation

   Library preparation and differential expression analysis were performed
   as described in [[67]28]. In brief: Libraries were prepared using
   poly-A enrichment to select for mRNA following Novogene’s internal
   quality control assays. Sequencing was done using the Illumina NovaSeq
   platform. Samples were sequenced as paired-end reads of 150 base pairs,
   and those reads in the fastq format were processed for downstream
   analyses or “cleaned” by removing reads containing adaptors, reads with
   >10% uncertain nucleotides, and reads with >50% low quality
   nucleotides, defined as Base Quality < 5. Reads were aligned to a
   reference sequence for American lobster [[68]33] using HISAT2 (version
   2.0.5) and assembled into transcripts using StringTie (version 1.3.3b).
   Raw counts of map reads were aggregated using featureCounts (version
   1.5.0-p3). Raw data have been deposited in GenBank Sequence Read
   Archive, SRA, accession number PRJNA1087720.

Differential expression analysis

   Raw read counts were normalized to correct for sequencing depth before
   differential transcript expression was analyzed using DESeq2 software
   [[69]34]. DESeq2 uses the negative binomial as the reference
   distribution and takes a geometric normalization approach. False
   discovery rate (FDR) was controlled by adjusting p-values for multiple
   testing with the Benjamini-Hochberg procedure, and the differential
   transcript screening threshold was |log2(FoldChange)| ≥ 1 and padj ≤
   0.05. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes
   (KEGG) pathway enrichment analysis was performed using clusterProfiler
   (version 3.8.1) [[70]35, [71]36]. The KEGG pathway enrichment analysis
   was restricted solely to Homarus americanus (hame).

Results

Transcriptome quality and mapping

   Samples showed consistent read quality and depth. Across treatments,
   459,499,224 raw reads were generated, of which, an average of 97.4%
   were clean reads and 91.8% were able to be aligned with the reference
   genome ([72]Table 1). Mapping gene IDs to the GO and KEGG databases
   identified 2,921 unique GO terms and 133 unique KEGG pathways
   represented in our transcriptome.

Table 1. RNA-seq read quality parameters of Homarus americanus larvae in
different rearing conditions (n = 5 per treatment group).

   Treatment group Average raw reads Clean reads (%) Reads aligned to
   genome (%)
   Brine shrimp-reared (18°C) 56246590 97.0 92.6
   8°C 67756184 97.4 92.5
   26°C 57717047 97.5 92.5
   Wild caught (ambient) 53874813 97.3 92.4
   8°C 62163368 97.5 92.6
   26°C 55856186 97.6 90.9
   Zooplankton-reared (18°C) 45114401 97.4 91.1
   8°C 60770635 97.4 89.7
   Overall: 459499224 97.4 91.8
   [73]Open in a new tab

Differential transcript expression among treatment groups

   In wild-caught postlarvae compared to brine shrimp-reared postlarvae,
   1,406 transcripts were over-expressed, and 2,276 transcripts were
   under-expressed. In zooplankton-reared postlarvae relative to brine
   shrimp-reared postlarvae, 1,534 transcripts were over-expressed, and
   1,069 transcripts were under-expressed. In zooplankton-reared relative
   to wild-caught, 2,795 transcripts were over-expressed, and 1,144
   transcripts were under-expressed ([74]Fig 1).

Fig 1. Volcano plots show differential gene expression in Homarus americanus
stage IV post larvae after different rearing conditions.

   [75]Fig 1
   [76]Open in a new tab

   A: wild-caught postlarvae vs brine shrimp-reared postlarvae, B:
   zooplankton-reared postlarvae vs brine shrimp-reared postlarvae, C:
   zooplankton-reared postlarvae vs wild-caught postlarvae, all at rearing
   temperatures. The blue dashed line indicates the threshold for
   differential gene screening criteria. Red dots represent up-regulated
   genes and pink dots represent down-regulated genes. Axes are scaled
   differently in each plot for granularity.

   Wild-caught larvae exhibited 95% fewer differentially expressed
   transcripts in response to cold exposure compared to brine
   shrimp-reared postlarvae, and 86% fewer than zooplankton-reared
   postlarvae. Zooplankton-reared postlarvae exhibited 64% fewer
   differentially expressed transcripts in response to cold exposure than
   brine shrimp-reared postlarvae ([77]Fig 2). Across the cold-treated
   rearing groups, 36 transcripts were commonly differentially expressed,
   including transcripts for DNA helicase, methyltransferase, and
   polymerase, among others. The full list of genes, their expression
   values, and results from statistical comparisons among groups can be
   found in [78]S1 File. Variability from sample to sample was relatively
   consistent among groups, and hierarchical clustering of individual
   samples in relation to common differentially expressed genes yielded
   distinct groupings of the known experimental variables. The first tier
   of branching separated samples by temperature treatment. On the second
   tier, cold-treated postlarvae were further separated by rearing
   environment, with one cluster of wild postlarvae and another cluster of
   lab-reared postlarvae. The cluster of lab-reared postlarvae branched
   further into two separate groups defined by their different diets. The
   postlarvae that were not cold-treated were separated by diet first,
   with brine shrimp-reared postlarvae in a separate cluster from
   zooplankton-reared and wild postlarvae. Wild postlarvae all clustered
   together; one zooplankton-reared postlarvae grouped with the brine
   shrimp-reared postlarvae ([79]Fig 3).

Fig 2. Venn diagram of differentially expressed transcripts in Homarus
americanus stage IV postlarvae in response to cold exposure.

   [80]Fig 2
   [81]Open in a new tab

   3,073 transcripts were uniquely differentially expressed in brine
   shrimp-reared postlarvae. 724 transcripts were uniquely differentially
   expressed in zooplankton-reared postlarvae. Only 79 transcripts were
   uniquely differentially expressed in wild-caught postlarvae. 32
   transcripts were commonly expressed.

Fig 3. Clustering heatmap shows the expression of commonly differentially
expressed transcripts in Homarus americanus stage IV postlarvae in response
to cold exposure.

   [82]Fig 3
   [83]Open in a new tab

   At the top of the grid, boxes of varying colors represent samples
   within each experimental group, as follows (noted from left to right):
   cold-treated zooplankton-reared postlarvae in white; cold-treated brine
   shrimp-reared postlarvae in dark blue; cold-treated wild postlarvae in
   yellow; control brine shrimp-reared postlarvae in red; control
   zooplankton-reared postlarvae in green; control wild postlarvae in
   light blue. Mainstream hierarchical clustering was used to cluster the
   fpkm values of genes and homogenize each row (Z-score). The bar at
   right shows that, within the grid, a fourfold increase in expression
   compared to the control is represented by the darkest red, while a
   fourfold decrease in expression compared to the control is represented
   by the darkest blue. All samples within groups cluster together, except
   for one control zooplankton-reared postlarva which clusters with the
   control brine shrimp-reared postlarvae. For individual genes see [84]S1
   File.

   Wild postlarvae exhibited 68% fewer differentially expressed genes in
   response to heat exposure compared to brine shrimp-reared postlarvae
   ([85]Fig 4). In total, 152 transcripts were commonly differentially
   expressed across rearing groups in response to heat exposure. The full
   list of genes, their expression values, and results from statistical
   comparisons among groups can be found in [86]S2 File. Again,
   variability from sample to sample was relatively consistent among
   groups, and hierarchical clustering of individual samples yielded an
   initial branching off of samples within the different temperature
   treatments, and then further branching of samples within different
   rearing groups ([87]Fig 5).

Fig 4. Venn diagram of differentially expressed transcripts in Homarus
americanus stage IV postlarvae in response to heat exposure.

   [88]Fig 4
   [89]Open in a new tab

   495 transcripts were uniquely differentially expressed in wild-caught
   postlarvae. 1,873 transcripts were uniquely differentially expressed in
   brine shrimp-reared postlarvae. 152 transcripts were commonly
   expressed.

Fig 5. Clustering heatmap shows the expression of commonly differentially
expressed transcripts in Homarus americanus stage IV postlarvae in response
to heat exposure.

   [90]Fig 5
   [91]Open in a new tab

   At the top of the grid, boxes of varying colors represent samples
   within each experimental group, as follows (noted from left to right):
   control brine shrimp-reared postlarvae in dark blue; control wild
   postlarvae in white; heat-treated wild postlarvae in light blue;
   heat-treated brine shrimp-reared postlarvae in red. Mainstream
   hierarchical clustering was used to cluster the fpkm values of genes
   and homogenize each row (Z-score). The bar at right shows that, within
   the grid, a fourfold increase in expression compared to the control is
   represented by the darkest red, while a fourfold decrease in expression
   compared to the control is represented by the darkest blue. All samples
   within groups cluster together, except for one control
   zooplankton-reared postlarva which clusters with the control brine
   shrimp-reared postlarvae. For individual genes see [92]S2 File.

GO term and KEGG pathway enrichment analysis

Wild-caught vs. brine shrimp-reared postlarvae

   There were 28 GO terms and 8 KEGG pathways significantly enriched in
   the comparison between wild-caught and brine shrimp-reared postlarvae
   reared at ambient temperature, and these two groups were the least
   similar based on the enrichment analysis. Differentially expressed
   transcripts within the top four significantly enriched GO terms were
   largely under-expressed in wild-caught postlarvae; these terms, in
   order of statistical significance, were (1) structural constituent of
   cuticle, (2) extracellular region, (3) chitin binding, and (4)
   structural molecule activity. Differentially expressed transcripts in
   GO term (5) carbohydrate metabolic process were largely over-expressed
   in wild-caught postlarvae. The top five significantly enriched KEGG
   pathways included (1) mannose type O-glycan biosynthesis, (2) amino
   sugar and nucleotide sugar metabolism, (3) tyrosine metabolism, (4)
   glycosphingolipid biosynthesis—globo and isoglobo series, and (5)
   glycosaminoglycan degradation. Within all of these pathways, the
   majority of differentially expressed transcripts were over-expressed in
   wild-caught postlarvae (Tables [93]2 and [94]3).

Table 2. Summary of GO term enrichment analysis of Homarus americanus
postlarvae in different rearing conditions (n = 5 per treatment group).

   Comparison Total # enriched GO terms Category GO term ID Description
   Gene Ratio p-adjusted # Upregulated Genes # Downregulated Genes
   Rearing condition Zooplankton- vs. brine shrimp-reared 31 MF GO:0008061
   chitin binding 53/826 1.55E-09 24 29
   CC GO:0005576 extracellular region 61/306 9.36E-08 27 34
   BP GO:0005975 carbohydrate metabolic process 50/466 1.49E-07 41 9
   MF GO:0015276 ligand-gated ion channel activity 78/826 2.79E-05 17 61
   MF GO:0022834 ligand-gated channel activity 78/826 2.79E-05 17 61
   Wild-caught vs. brine shrimp-reared 28 MF GO:0042302 structural
   constituent of cuticle 132/1164 1.85E-28 15 117
   CC GO:0005576 extracellular region 102/362 2.76E-28 28 74
   MF GO:0008061 chitin binding 83/1164 4.30E-21 21 62
   MF GO:0005198 structural molecule activity 136/1164 2.99E-16 16 120
   BP GO:0005975 carbohydrate metabolic process 60/613 9.20E-08 45 15
   Zooplankton-reared vs. wild-caught 22 CC GO:0005576 extracellular
   region 90/616 6.78E-05 67 23
   MF GO:0004175 endopeptidase activity 81/1717 1.42E-03 30 51
   BP GO:0007165 signal transduction 146/950 4.37E-03 99 47
   BP GO:0023052 signaling 149/950 4.37E-03 100 49
   BP GO:0007154 cell communication 149/950 4.37E-03 101 48
   Temperature treatment 8°C vs. 18°C (brine shrimp-reared) 6 MF
   GO:0042302 structural constituent of cuticle 132/1256 2.02E-14 114 18
   CC GO:0005576 extracellular region 68/409 1.86E-04 54 14
   MF GO:0008061 chitin binding 56/1256 7.59E-04 46 10
   MF GO:0004930 G-protein coupled receptor activity 63/1256 6.66E-03 56 7
   BP GO:0007186 G-protein coupled receptor signaling pathway 67/674
   8.02E-03 60 7
   26°C vs. 18°C (brine shrimp-reared) 17 MF GO:0042302 structural
   constituent of cuticle 78/638 1.08E-14 2 76
   MF GO:0005198 structural molecule activity 81/638 3.22E-09 2 79
   CC GO:0005576 extracellular region 43/175 8.29E-08 12 31
   MF GO:0008061 chitin binding 34/638 3.54E-04 8 26
   CC GO:0005783 endoplasmic reticulum 11/175 2.70E-03 1 10
   8°C vs. ambient (wild-caught) 1 BP GO:0016051 carbohydrate biosynthetic
   process 3/323 5.30E-03 3 0
   26°C vs. ambient (wild-caught) 0 - - - - - - -
   8°C vs. 18°C (zooplankton-reared) 35 MF GO:0042302 structural
   constituent of cuticle 87/467 1.58E-27 83 4
   MF GO:0005198 structural molecule activity 87/467 4.77E-19 83 4
   CC GO:0005576 extracellular region 49/130 4.56E-17 38 11
   MF GO:0008061 chitin binding 42/467 6.58E-12 37 5
   BP GO:0005975 carbohydrate metabolic process 27/229 5.13E-04 10 17
   [95]Open in a new tab

Table 3. Summary of KEGG pathway enrichment analysis of Homarus americanus
postlarvae in different rearing conditions (n = 5 per treatment group).

   Comparison Total # enriched KEGG pathways KEGG pathway ID Description
   Gene Ratio p-adjusted # Upregulated Genes # Downregulated Genes
   Rearing condition Zooplankton- vs. brine shrimp-reared 6 hame00350
   Tyrosine metabolism 17/306 5.79E-08 16 1
   hame00520 Amino sugar and nucleotide sugar metabolism 22/306 2.45E-05
   17 5
   hame04142 Lysosome 31/306 2.49E-03 26 5
   hame00260 Glycine, serine and threonine metabolism 11/306 1.28E-02 8 3
   hame00040 Pentose and glucuronate interconversions 9/306 3.11E-02 8 1
   Wild-caught vs. brine shrimp-reared 8 hame00515 Mannose type O-glycan
   biosynthesis 20/356 3.55E-07 18 2
   hame00520 Amino sugar and nucleotide sugar metabolism 24/356 2.18E-05
   16 8
   hame00350 Tyrosine metabolism 12/356 3.48E-03 11 1
   hame00603 Glycosphingolipid biosynthesis—globo and isoglobo series
   10/356 1.14E-02 8 2
   hame00531 Glycosaminoglycan degradation 9/356 1.60E-02 7 2
   Zooplankton-reared vs. wild-caught 1 hame04080 Neuroactive
   ligand-receptor interaction 58/643 4.72E-02 35 23
   Temperature treatment 8°C vs. 18°C (brine shrimp-reared) 9 hame03030
   DNA replication 16/459 3.55E-04 0 16
   hame04141 Protein processing in endoplasmic reticulum 37/459 3.55E-04 2
   35
   hame00230 Purine metabolism 28/459 1.78E-03 17 11
   hame00240 Pyrimidine metabolism 15/459 1.84E-02 8 7
   hame01232 Nucleotide metabolism 19/459 2.22E-02 12 7
   26°C vs. 18°C (brine shrimp-reared) 12 hame01200 Carbon metabolism
   23/291 7.30E-04 5 18
   hame00010 Glycolysis / Gluconeogenesis 15/291 7.30E-04 5 10
   hame00531 Glycosaminoglycan degradation 9/291 5.68E-03 5 4
   hame00500 Starch and sucrose metabolism 11/291 8.66E-03 5 6
   hame00520 Amino sugar and nucleotide sugar metabolism 16/291 1.10E-02 5
   11
   8°C vs. ambient (wild-caught) 1 hame04068 FoxO signaling pathway 4/223
   1.16E-02 2 2
   26°C vs. ambient (wild-caught) 0 - - - - - -
   8°C vs. 18°C (zooplankton-reared) 4 hame00010 Glycolysis /
   Gluconeogenesis 10/161 1.07E-02 2 8
   hame03030 DNA replication 8/161 1.09E-02 0 8
   hame00260 Glycine, serine and threonine metabolism 8/161 1.35E-02 3 5
   hame00270 Cysteine and methionine metabolism 7/161 3.83E-02 2 5
   [96]Open in a new tab

Zooplankton-reared vs. brine shrimp-reared postlarvae

   There were 31 GO terms and 6 KEGG pathways significantly enriched in
   the comparison between zooplankton-reared and brine shrimp-reared
   postlarvae reared at ambient temperature. The top five significantly
   enriched GO terms were (1) chitin binding, (2) extracellular region,
   (3) ligand-gated ion channel activity, and (4) ligand-gated channel
   activity, in which the majority of differentially expressed transcripts
   were under-expressed in zooplankton-reared postlarvae, and (5)
   carbohydrate metabolic process, in which the majority of differentially
   expressed transcripts were over-expressed in zooplankton-reared
   postlarvae. The top five most significantly enriched KEGG pathways were
   (1) tyrosine metabolism, (2) amino sugar and nucleotide sugar
   metabolism, (3) lysosome, (4) glycine, serine and threonine metabolism,
   and (5) pentose and glucuronate interconversions, all of which
   contained differentially expressed transcripts which were largely
   over-expressed in zooplankton-reared postlarvae.

Zooplankton-reared vs. wild-caught postlarvae

   There were 22 GO terms and one KEGG pathway significantly enriched in
   the comparison between zooplankton-reared and wild-caught postlarvae
   reared at ambient temperature, and these two groups were the most
   similar based on our enrichment analysis. The top five significantly
   enriched GO terms were (1) extracellular region, (2) endopeptidase
   activity, (3) signal transduction, (4) signaling, and (5) cell
   communication. Of these, endopeptidase activity was the only term
   containing largely under-expressed, rather than over-expressed,
   transcripts in zooplankton-reared postlarvae. The sole significantly
   enriched KEGG pathway, neuroactive ligand-receptor interaction,
   contained differentially expressed transcripts which were largely
   over-expressed in zooplankton-reared postlarvae compared to wild-caught
   postlarvae.

Cold temperature exposure (8°C)

   There were 6 GO terms and 9 KEGG pathways significantly enriched in the
   comparison between brine shrimp-reared postlarvae exposed to the cold,
   8°C temperature treatment and those maintained at their rearing
   temperature (18°C). The top five significantly enriched GO terms were
   (1) structural constituent of cuticle, (2) extracellular region, (3)
   chitin binding, (4) G-protein coupled receptor activity, and (5)
   G-protein coupled receptor signaling pathway. The majority of
   differentially expressed transcripts within these terms were
   under-expressed in cold-treated (8°C) brine shrimp-reared postlarvae.
   The top five significantly enriched KEGG pathways were (1) DNA
   replication and (2) protein processing in endoplasmic reticulum, which
   contained transcripts that were largely under-expressed in response to
   cold exposure. Additionally, pathways for (3) purine metabolism, (4)
   pyrimidine metabolism, and (5) nucleotide metabolism all contained
   transcripts that were largely over-expressed in brine shrimp-reared
   postlarvae in response to cold exposure.

   There were 35 GO terms and 4 KEGG pathways significantly enriched in
   the comparison between zooplankton-reared postlarvae exposed to the
   cold, 8°C temperature treatment and those maintained at their rearing
   temperature (18°C). The top five significantly enriched GO terms were
   (1) structural constituent of cuticle, (2) structural molecule
   activity, (3) extracellular region, and (4) chitin binding. Transcripts
   within these terms were largely over-expressed in the 8°C treatment
   compared to the control. Transcripts within (5) carbohydrate metabolic
   process were largely under-expressed in the 8°C treatment compared with
   the control. Exposure to 8°C in zooplankton-reared postlarvae also
   yielded the overarching under-expression of transcripts within 4 KEGG
   pathways; (1) glycolysis / gluconeogenesis, (2) DNA replication, (3)
   glycine, (4) serine and threonine metabolism, and (5) cysteine and
   methionine metabolism.

   There was only one GO term and one KEGG pathway significantly enriched
   in the comparison between wild-caught postlarvae exposed to the cold,
   8°C temperature treatment and those maintained at their rearing
   temperature (ambient, 11–19°C). The differentially expressed
   transcripts within the GO term, carbohydrate biosynthetic process were
   all over-expressed in response to cold exposure in wild postlarvae. In
   the KEGG pathway for FoxO signaling, two transcripts were
   under-expressed, and two transcripts were over-expressed in response to
   cold exposure in wild postlarvae. Those that were overexpressed were
   further analyzed for potential use as markers for cold tolerance; they
   were both transcripts for phosphoenolpyruvate carboxykinase, an enzyme
   involved in gluconeogenesis (for sequences and fpkm values see [97]S3
   and [98]S4 Files). The differential expression of these two transcripts
   was reduced the further removed rearing conditions were from what is
   experienced in the wild ([99]Fig 6).

Fig 6. Expression of two separate transcripts for a gene in Homarus
americanus coding for phosphoenolpyruvate carboxykinase- an enzyme involved
in gluconeogenesis.

   [100]Fig 6
   [101]Open in a new tab

   121879005 on the left, and 121879006 on the right. Mean normalized read
   counts are on the y-axes, and rearing conditions are on the x-axes;
   control temperature treatments are indicated by black bars, and cold
   temperature treatments are indicated by grey bars. Error bars represent
   standard deviation values. One star indicates a p-value of <0.05. Four
   stars indicate a p-value of <0.0001 (two-way ANOVA multiple
   comparisons, α = 0.05). The over-expression of both transcripts in cold
   treatments was reduced in rearing conditions and further removed from
   natural conditions.

Warm temperature exposure (26°C)

   There were 17 GO terms and 12 KEGG pathways significantly enriched in
   the comparison between brine shrimp-reared postlarvae exposed to the
   warm, 26°C temperature treatment and those maintained at their rearing
   temperature (18°C). The top five significantly enriched GO terms were
   (1) structural constituent of cuticle, (2) structural molecule
   activity, (3) extracellular region, (4) chitin binding, and (5)
   endoplasmic reticulum. Most transcripts within these terms were
   under-expressed in warm-treated (26°C) brine shrimp-reared postlarvae.
   The top five significantly enriched KEGG pathways in this comparison
   were (1) carbon metabolism, (2) glycolysis/gluconeogenesis, (3) starch
   and sucrose metabolism, and (4) amino sugar and nucleotide sugar
   metabolism, all of which contained transcripts which were largely
   under-expressed in response to heat exposure. Genes within the pathway
   (5) glycosaminoglycan degradation were largely over-expressed as a
   result of heat exposure. In stark contrast to the brine shrimp-reared
   postlarvae, exposure of wild-caught postlarvae to 26°C yielded no
   significantly enriched GO terms or KEGG pathways when compared with the
   control (ambient, 11–19°C).

Discussion

   We demonstrated that both diet (traditional lab-rearing diet of brine
   shrimp vs. “natural” diet of zooplankton) and environment (lab setting
   vs. ocean environment) influence the baseline whole-body gene
   expression of Homarus americanus at 18°C in lab-reared postlarvae and
   at ambient in wild-caught, as well as transcriptional responses to
   temperature extremes. One cellular event that may drive differences in
   brine shrimp-reared postlarvae compared to wild-caught postlarvae is
   the under-expression of transcripts related to carbohydrate metabolic
   activity. This pattern likely reflects the brine shrimp-reared
   postlarvae’s reduced energy availability and/or a reduction in
   energetic demands due to decreased predation, room to swim, or need to
   chase after scarce food. Transcripts related to amino sugar and
   nucleotide sugar metabolism, as well as tyrosine metabolism, were also
   under-expressed in brine shrimp-reared postlarvae, again highlighting
   their reduced metabolic demands when compared to wild-caught.
   Therefore, we interpret these findings as lab-reared postlarvae being
   metabolically adapted to the less demanding conditions of their
   upbringing, leading to greater responses to stress, as outlined below.
   Transcripts related to glycosaminoglycan degradation as well as the
   biosynthesis of mannose type O-glycans and glycosphingolipids, notable
   for their importance in cell-to-cell communication [[102]37], were
   largely under-expressed in brine shrimp-reared postlarvae.
   Additionally, there were several transcripts related to shell-building
   processes (i.e., “structural constituents of the cuticle” and
   “chitin-binding”) that were over-expressed in brine shrimp-reared
   postlarvae compared to wild-caught ones. This could be explained by the
   varied time points that wild lobsters were sampled (i.e., wild lobsters
   may have been at a point in their molt cycle in which shell-building
   processes were not upregulated compared with the lab-reared lobsters
   sampled 48 hours post-molt). While we chose to sample lobsters at the
   48-hour time point in the interest of consistency for temperature
   exposure experiments, future studies may benefit from employing one of
   two sampling methods; randomizing the sampling of lab-reared lobsters
   to mimic fishing for wild postlarvae or staging lobsters and sampling
   groups only from the same molt stage, which has been done in other
   crustacean species [[103]38, [104]39]. Additionally, future studies may
   benefit from analyzing smaller, isolated organs (i.e., the heart or
   eyes), as our results are inherently driven by the tissues that make up
   the largest proportion of the postlarvae.

   Similarities in the enrichment analyses of brine shrimp-reared
   postlarvae as compared to both wild-caught postlarvae and
   zooplankton-reared postarvae suggest that diet is one of the mechanisms
   by which brine shrimp-reared postlarvae are less robust than their wild
   counterparts. Transcripts related to carbohydrate metabolic process and
   metabolic pathways for tyrosine, serine and threonine were largely
   under-expressed in brine shrimp-reared postlarvae in comparison to both
   zooplankton-reared and wild-caught postlarvae. We interpret the
   under-expression of these metabolic processes in brine shrimp-reared
   postlarvae to mean that greater energy stores may be derived from a
   natural diet and/or that postlarvae may need to exert more energy to
   capture natural prey items, as copepods and other larval crustaceans
   are more difficult to catch and consume than brine shrimp [[105]40].
   This interpretation complements Castro & Cobb’s 2005 study
   demonstrating disparate swimming speeds in wild and lab-reared
   postlarvae (18 cm/s in wild postlarvae compared to 10 cm/s in
   lab-reared postlarvae).

   Further, tyrosine, serine, and threonine all have important biological
   functions that may be negatively impacted by the traditional lab-reared
   diet of brine shrimp. Tyrosine and serine are precursors for other
   amino acids and are also involved in neurotransmission; specifically,
   tyrosine is a precursor for neurotransmitters L-Dopa, dopamine,
   epinephrine, and norepinephrine. As such, deficiencies in serine and
   tyrosine are linked to impaired function of the nervous system
   [[106]41]. Threonine is involved in energy production, protein
   synthesis, and maintaining healthy gut microbiota [[107]42]. In rats,
   deficiencies in threonine compromise lipid metabolism and force a
   reduction in energy expenditure [[108]43]. Again, this serves to
   further explain reduced energy expenditure in brine-shrimp-reared
   postlarvae.

   Transcripts related to membrane channel activity and chitin binding
   were largely under-expressed in zooplankton-reared postlarvae compared
   to brine shrimp-reared postlarvae. Dampened membrane channel activity
   in zooplankton-fed postlarvae may be the result of a more varied and
   cholesterol-rich diet than that of postlarvae fed freshly hatched
   Artemia spp., as crustaceans are unable to synthesize their own
   cholesterol de novo and thus derive all of their cholesterol from diet
   [[109]44]. A diet of zooplankton has been found to have twice the
   sterol content compared to a diet of Artemia spp. [[110]14]. It has
   been documented that increased membrane cholesterol can suppress
   channel activity, and in particular ligand-gated channels [[111]45],
   which comprise two of the top five significantly enriched GO terms in
   this comparison.

   Transcripts related to endopeptidase activity and neuroactive
   ligand-receptor interaction were over-expressed in zooplankton-fed
   postlarvae relative to wild-caught postlarvae. Transcripts related to
   cell communication and signaling were under-expressed in both
   lab-reared groups compared to the wild-caught group, which suggests
   that reduced cell communication may result from factors inherent to the
   lab outside of diet, such as reduced stimulation or constant
   temperature. In crustaceans, physiological functions like hormone
   secretion are dictated by circadian rhythms [[112]46] which can be
   disrupted by low light levels in the incubator environment.

   Exposure to cold (8°C) yielded very different transcriptional responses
   in postlarvae from each of the rearing conditions. In brine
   shrimp-reared postlarvae, transcripts related to G-protein coupled
   receptor activity, shell-building processes, and DNA replication were
   largely under-expressed after exposure to 8°C, while transcripts
   related to the metabolism of purines, pyrimidines, and general
   nucleotides were largely over-expressed. This pattern of slowed DNA
   replication and increased metabolism of nucleotides has been documented
   in other species exposed to slow growth conditions, such as colder
   temperatures, and may allow for cellular reorganization [[113]47].

   Transcripts related to shell-building processes were largely
   over-expressed in zooplankton-reared postlarvae exposed to 8°C. This
   result, when compared to the trend seen in brine shrimp-reared
   postlarvae, may indicate that a diet of brine shrimp predisposes
   lobster postlarvae to be susceptible to more energetic tradeoffs in a
   cold-stressed state. Under-expression of transcripts related to
   carbohydrate metabolic processes and pathways for metabolism of serine,
   threonine, cysteine, and methionine was also observed in
   zooplankton-reared postlarvae, again illustrating reduced metabolic
   demands at 8°C, in concert with the under-expression of pathways for
   DNA replication and glycolysis / gluconeogenesis.

   Thermally induced transcriptional changes were minimal in wild
   postlarvae compared with either lab-reared group, with only one GO term
   and one KEGG pathway being significantly enriched in response to cold
   exposure. We interpret this decreased cellular response in wild
   postlarvae compared to those reared in a laboratory setting as further
   evidence for decreased fitness of lab-reared organisms, serving as a
   complement to previously described disparities in swimming ability
   [[114]13] and proper nutrition [[115]14]. Transcripts related to
   carbohydrate biosynthetic processes were largely over-expressed in
   response to chronic cold exposure in wild-caught postlarvae. The FoxO
   signaling pathway was the only KEGG pathway that was significantly
   enriched in this comparison (for sequences see [116]S3 File). The FoxO
   signaling pathway has previously been identified in zebrafish as a key
   survival pathway for enhancing cold stress, and thus genes within it as
   molecular markers for thermal tolerance [[117]48]. FoxO has also been
   noted for its importance to growth and stress resistance in arthropods
   (e.g. the brown planthopper Nilaparvata lugens [[118]49]). We found a
   different gene in this pathway, phosphoenolpyruvate carboxykinase, that
   might be relevant as a genetic marker for cold tolerance in lobster.
   While markers for cold tolerance are often targeted in plant species
   for use in selective breeding practices [[119]50, [120]51], they may
   also be valuable tools for studying crustaceans, whether assessing the
   quality of larvae coming from different mothers or the quality of
   larvae coming from a lab. Phosphoenolpyruvate carboxykinase serves in
   the maintenance of glucose homeostasis and converts within the
   gluconeogenesis pathway oxalacetate to phosphoenolpyruvate and carbon
   dioxide [[121]52]. In many vertebrate ectotherms, glucose is a known
   cryoprotectant, preventing the denaturation of important enzymes like
   lactate dehydrogenase [[122]53–[123]55]. In crustaceans, the
   accumulation of other sugars and free amino acids has been linked to
   increased cold tolerance [[124]56]. Thus, enhanced gluconeogenesis
   would increase glucose levels in the tissues and/or hemolymph and aid
   in cold tolerance. This gene’s potential role in the cold tolerance of
   wild-caught lobster postlarvae is therefore fitting, though more
   research is needed to fully understand it.

   Exposure to warm conditions (26°C) yielded very different
   transcriptional responses in the two rearing conditions studied, with
   brine shrimp-reared postlarvae exhibiting signs of cellular stress at
   26°C and wild-caught postlarvae exhibiting no significantly enriched GO
   terms or pathways at all. In brine shrimp-reared postlarvae transcripts
   related to shell-building processes and endoplasmic reticulum were
   largely under-expressed after chronic exposure to 26°C. This is similar
   to their response at 8°C, and it may be that shell-building processes
   are de-prioritized in the face of other energetic demands at
   sub-optimal temperatures. In the wild, reduced shell quality could
   negatively impact predator avoidance as well as settlement, as shell
   quality is necessarily altered in stage IV lobsters for greater
   negative buoyancy and ease of settlement [[125]57, [126]58].
   Additionally, pathways for carbon metabolism,
   glycolysis/gluconeogenesis, starch and sucrose metabolism, and amino
   sugar and nucleotide sugar metabolism were all downregulated in
   response to heat exposure, while the pathway for glycosaminoglycan
   degradation was upregulated. This pattern of gene expression in which
   energy is diverted from non-essential pathways toward life-saving
   processes is contextualized in the framework of the Oxygen- and
   Capacity-Limited Thermal Tolerance hypothesis (OCLTT, [[127]59,
   [128]60]). Within the OCLTT framework critical temperatures (Tc[min]
   and Tc[max]) are defined through the onset of internal anaerobiosis and
   the accumulation of anaerobic endproducts in the presence of sufficient
   oxygen in the environment, due to a failing circulatory or ventilatory
   system. Prior to reaching Tc[min] or Tc[max] a pejus temperature
   threshold (Tp) is characterized by a decline in fitness, growth rate,
   and scope for activity [[129]61]. In the pejus range between Tp and Tc
   the animal’s metabolism accelerates catabolic pathways producing ATP,
   and decelerates anabolic pathways, consuming ATP [[130]62]. Our data of
   downregulating carbon metabolism, glycolysis/gluconeogenesis, starch
   and sucrose metabolism, and amino sugar and nucleotide sugar metabolism
   during heat exposure, while upregulating the pathway for
   glycosaminoglycan degradation indicates that 26°C lies in the pejus or
   sub-optimal temperature range for brine shrimp-reared postlarvae. The
   analysis of wild-caught postlarvae exposed to 26°C, however, did not
   yield any significantly enriched GO terms or KEGG pathways, indicating
   that they are comparatively less stressed at this temperature. Future
   studies should take the discrepancy between lab-reared and wild-caught
   larvae into account when assessing thermal tolerance or impacts of end
   of century temperature projections on marine larvae.

Conclusion

   Through our analysis of gene expression in postlarvae reared in
   different conditions, we characterized the molecular underpinnings of
   observed differences between traditionally lab-reared (fed brine
   shrimp, maintained at 18°C) and wild-caught lobster postlarvae, as well
   as those reared on a natural diet of zooplankton. In addition to
   establishing diet-driven differences in energy and amino acid
   metabolism, we also identified patterns in differential expression that
   seem to be attributable to other characteristics of a lab setting,
   including reduced cell communication. These findings highlight that
   postlarvae reared under traditional laboratory conditions may not
   necessarily be representative of in situ postlarvae; they also suggest
   that changing diet alone is not sufficient to eliminate cellular
   artefacts of lab-rearing. The gene expression profiles of postlarvae
   were decreasingly impacted by both cold (8°C) and warm (26°C)
   temperature challenges the closer their rearing conditions were to what
   is experienced in the wild. This finding contextualizes the results
   from existing research on postlarval lobster thermal tolerance as
   valuable but highly conservative estimations of the thermal tolerance
   of in situ organisms. We identified two gene sequences for
   phosphoenolpyruvate carboxykinase for further investigation as novel
   markers for cold tolerance in American lobster, based on their varied
   upregulation in our decreasingly thermally tolerant rearing groups.
   Further, variability in gene expression was not meaningfully greater in
   wild postlarvae when compared to lab-reared postlarvae, underscoring
   the utility of wild larvae for research purposes. In fact, future
   studies can be strengthened by analyzing wild early-stage organisms to
   better understand thermal thresholds relevant to in situ populations,
   which is of the utmost importance as coastal communities continue to
   gather information to better manage economically important species in
   the face of climate change.

Supporting information

   S1 File. Full list of genes commonly expressed in response to cold
   exposure, their expression values, and results from statistical
   comparisons among groups.

   (XLSX)
   [131]pone.0307169.s001.xlsx^ (45.4KB, xlsx)
   S2 File. Full list of genes commonly expressed in response to heat
   exposure, their expression values, and results from statistical
   comparisons among groups.

   (XLSX)
   [132]pone.0307169.s002.xlsx^ (154.2KB, xlsx)
   S3 File. DNA sequences of two transcripts of PEPCK.

   (DOCX)
   [133]pone.0307169.s003.docx^ (13.7KB, docx)
   S4 File. Phosphoenolpyruvate carboxykinase expression values and
   results from statistical comparisons among treatment groups.

   (XLSX)
   [134]pone.0307169.s004.xlsx^ (9.9KB, xlsx)

Acknowledgments