Mechanisms of colour adaptation in the prawn Penaeus monodon

Changes in crustacean colouration can be due to physiological or morphological mechanisms. Physiological mechanisms that influence prawn colour include carotenoid availability in the diet, background substrate colour, photoperiod, light intensity and temperature (Rao, 1985). Such colour changes are often rapid, reversible and rhythmic, and are associated with hormonally controlled expansion and contraction of pigment structures, known as chromatophores, contained within the hypodermal layer, the pigmented layer in between the exoskeleton and abdominal muscle (Kleinholz, 1961; Fingerman, 1966; Rao, 1985; Rao, 2001). Chromatophores strongly influence crustacean colour, particularly in prawns that have thin, opaque shells. Morphological mechanisms that influence prawn colour, in contrast, are considered to involve quantitative modifications of exoskeletal pigment concentration or composition, and hence are thought to be slower and more permanent. These changes have been associated with particular crustacean life history stages (Wade et al., 2008) or interspecific differences in colour and patterning (Wade et al., 2009). The traditional view is that physiological and morphological mechanisms of colour change are independent of each other; however, the relationship between them has not been investigated.

Astaxanthin is the predominant carotenoid in penaeid prawns (Katayama et al., 1971; Katayama et al., 1972) and dietary inclusion levels to produce optimal pigmentation in aquaculture have been determined for many species (Negre-Sadargues et al., 1983; Negre-Sadargues et al., 1993; Yamada et al., 1990; Chien and Jeng, 1992). Colouration is dependent upon the presence of astaxanthin within hypodermal tissues (Menasveta et al., 1993; Boonyaratpalin et al., 2001) and must reach a critical level for optimal colouration. However, total hypodermal carotenoid content does not directly correlate with cooked prawn colour and it is the proportions of free and esterified astaxanthin that change in prawns in response to background substrate colour (Tume et al., 2009).

Within the exoskeleton and hypodermal tissue, astaxanthin is often bound within a multimeric protein complex called crustacyanin (CRCN). This interaction modifies the naturally red carotenoid, producing the diverse array of colours seen in the exoskeleton of crustaceans (reviewed in Chayen et al., 2003). During cooking, this interaction is disrupted, releasing the distinct red colouration of cooked seafood. Two genes have also been identified across a range of crustaceans that encode CRCN-A and CRCN-C (Wade et al., 2009). Expressed predominantly in the outer layer of the hypodermis, the spatial regulation of the CRCN genes is thought to be a major contributor to the colours and patterns that an individual can produce (Wade et al., 2009). Despite the central role that the CRCN protein and gene expression play in crustacean colouration, any potential role of CRCN in adaptive colouration in prawns has not been investigated.

In the present study, we compared the differences in colouration, carotenoid content, CRCN protein abundance and CRCN gene expression between groups of prawns both fed high dietary astaxanthin levels but subject to either long-term black or white background colours. We used this system to investigate the underlying mechanisms of colour production and colour adaptation in crustaceans, and investigate the potential interplay between physiological and morphological mechanisms of colour change.

Live Penaeus monodon (Fabricius 1798) were obtained from commercial farms and maintained at CSIRO Marine and Atmospheric Research (CMAR) laboratories at Cleveland, Australia. For all trials, filtered seawater was heated and then pumped through the tanks at 1.2 l min^–1, maintaining water temperatures at 28°C and salinity at 35 g l^–1. Animals were held indoors in 1.9 m diameter 2000 l tanks covered with twin-walled tinted polycarbonate sheets (Suntuf, Capalaba, QLD, Australia) to reduce light intensity by approximately 80%, and under natural photoperiod, which was approximately 12.5 h:11.5 h light:dark during the course of the experiment.

To induce a strong difference in prawn colour, there were three replicate tanks painted black in colour and three replicate tanks painted white in colour, each containing 20 prawns. At the start of the experiment, animals of 13.9±4.2 g were stocked into experimental tanks and fed commercial prawn pellets (Ridley Grower, Ridley Aquafeed, Narangba, Australia) containing approximately 70 mg kg^–1 Carophyll Pink (DSM Nutritional Products, Basel, Switzerland) to excess twice daily. Animals were maintained in these tanks for 6 weeks, which has been shown to provide optimal colouration in this species at this carotenoid concentration (R. Smullen, Ridley Aquafeed, personal communication). At the end of the trial, animals had grown to 22.8±4.6 g in size with 97% survival. Subsets of these animals were used to assess changes in: (1) colour, (2) hypodermal carotenoids, (3) abundance of hypodermal CRCN protein and (4) gene expression of CRCN in the hypodermis as outlined below.

Penaeus monodon from the same population used to stock the background colouration trial were moult staged according to setal staging and epidermal withdrawal in uropods (Smith and Dall, 1985; Promwikorn et al., 2004). In total, six individuals from each of the seven classifications (A, B, C, D₀, D₁, D₂ and D₃) were sampled. Hypodermal tissue was dissected from each of the animals and used to assess the gene expression of CRCN across the moult cycle as outlined below.

At 6 weeks, five animals from each of the six tanks in the background colouration trial were selected at random for colour assessment using a CR-300 colorimeter (Konica Minolta, Marunouchi, Tokyo, Japan) over a 1 cm diameter circle standardised under D65 illumination that most accurately replicates natural sunlight. The instrument uses the Commission Internationale de l'Eclairage (CIE) ‘Lab’ system of colour notation (CIE, 2004) to measure the absolute colour of a sample on a three-dimensional scale of value, hue and chroma. The value of colour [or lightness (L)] has a scale of 0 (pure black) to 100 (pure white). The hue has two components that distinguish opposing colours. The first is a, which represents the red–green scale, and the other is b, which represents the blue–yellow scale. Chroma (or saturation) indicates the amount of hue: positive a towards red, negative a towards green and positive b towards yellow, negative b towards blue.

Uncooked individuals were measured in triplicate along the lateral and dorsal sides of the first three abdominal segments and these measures were averaged to give one reading for each individual. Individual prawns were boiled for 2 min in water, and then re-measured in the same location to record the corresponding colour change induced by cooking. Subjective grade scoring was made under standardised fluorescent lighting using a Lineal Salmofan (DSM Nutritional Products, Kaiseraugst, Switzerland), which is the accepted international benchmark for prawn colour. Scores recorded for each animal were made in agreement by two researchers.

At 6 weeks, two animals from each of the six tanks in the background colouration trial were selected at random for carotenoid analysis and quantification. After removal of the exoskeleton, hypodermal tissue was dissected from the first three abdominal sections of uncooked prawns, freeze dried using an Alpha 1-2 freeze drier (Martin Christ, Osterode am Harz, Germany) and the dry masses of the tissues were recorded. Carotenoids were extracted from three different components of the target tissue: whole dissected hypodermal tissue, soluble protein extract from dissected hypodermal tissue, and the insoluble pellet from the dissected hypodermal tissue. Carotenoids were triple extracted from freeze-dried tissue by homogenisation, using an Ultraturrax TP 18-10 (IKA, Staufen, Germany), in equal volumes of iso-amyl alcohol and hexane. Pooled extracts were dried under nitrogen and resuspended in 200 μl of hexane, before being separated by HPLC using the methods and equipment of Tume et al. (Tume et al., 2009). Peaks of free astaxanthin, astaxanthin mono-esters or astaxanthin di-esters were quantified using Empower Pro software (Waters Corporation, Milford, MA, USA) with a peak width of 45 s and a threshold of 50–70 absorbance units. Peak areas were normalised against the dry mass of the original tissue extracted.

At 6 weeks, hypodermal tissue was dissected from two individuals selected randomly from each of the six tanks in the background colouration trial. Tissue samples were homogenised into a neutral extraction buffer (150 mmol l^–1 NaCl, 10 mmol l^–1 Tris HCl pH 7.4, 1 mmol l^–1 EDTA, 1 mmol l^–1 EGTA, Complete mini-EDTA free protease inhibitors; Roche, Basel, Switzerland). The soluble fraction was retained and the insoluble material was washed in 5 ml MilliQ water, followed by two further washes in 1 ml MilliQ water. This washed material contained the majority of the blue colouration from the hypodermal tissue, and the proteins in this fraction were precipitated using 50% ammonium sulphate and centrifuged at 20,000 g for 15 min at 4°C. The dark blue pellet (λ_max=634 nm) was exchanged into phosphate-buffered saline using microcon 10,000 MWCO filters (Millipore, Billerica, MA, USA) and added back to the original soluble protein fraction to make the total soluble protein fraction. The insoluble pellet that remained was also used for carotenoid extraction as outlined above, along with the soluble fraction.

A detergent solution (1% Triton X-100, 0.5% IGEPAL CA-630 and 1% sodium dodecyl sulphate) was added to the total soluble protein fraction, which triggered a distinctive red-shift in the colour of the soluble protein fraction (λ_max=484 nm). Protein concentrations of the soluble protein extract were estimated using the bicinchoninic acid assay (Thermo Fisher Scientific, Waltham, MA, USA), and 20 μg of total protein was separated on a precast 4–16% BisTris polyacrylamide gel (Invitrogen, Carlsbad, CA, USA). The Coomassie-stained gel was visualised on a Gel Doc XR gel documentation system (Bio-Rad, Hercules, CA, USA) and bands were quantified using Quantity One analysis software (Bio-Rad). The four bands that had an average intensity of >700 optical density were selected for quantification across samples.

At 6 weeks, RNA was extracted from hypodermal tissue from four individuals from each of the six tanks in the background colouration trial and from six individuals sampled for each of the seven moult stages. Total RNA was extracted using RNeasy RNA Extraction Columns and on-column DNase digestion (Qiagen, Germantown, MD, USA) according to the manufacturer's instructions. Quality and quantity of all RNA samples were assessed by gel electrophoresis and on a Nanodrop spectrophotometer (Thermo Fisher Scientific), and were subsequently diluted to 100 ng μl^–1. An equivalent of 1 μg of total RNA was reverse transcribed following previously published protocols (Callaghan et al., 2010), which incorporated 400 pg of non-endogenous Luciferase RNA (Promega L4561, Venlo, The Netherlands) as an internal control. Expression of pmCRCN-A and pmCRCN-C as well as the control genes glyceraldehyde-3-phosphate dehydrogenase (GAPDH), elongation factor-1 alpha (EF1α) and Luciferase were analysed in all samples by real-time PCR as described below.

Real-time PCR primers specific to pmCRCN-A, pmCRCN-C and control genes (Table 1) were designed with CLC Main Workbench (CLC Bio, Aarhus, Denmark). Real-time PCR reactions were carried out using 1× SYBR Green PCR Master Mix (Life Technologies Corporation, Carlsbad, CA, USA), 0.2 μmol l^–1 of each primer and the equivalent of 7.5 ng of reverse-transcribed RNA. Amplification cycle conditions were 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 40 s at 60°C, and completed by a dissociation melt-curve analysis. Reactions were set up in triplicate using the epMotion 5070 (Eppendorf, Hamburg, Germany) and run on a 7500 Real-time PCR system (Life Technologies Corporation). Removal of gDNA contamination was verified by gene-specific PCR amplification of an equivalent amount of DNase treated RNA sample that had not been reverse transcribed. Optimal template concentration and PCR efficiencies for each primer pair were calculated using a fivefold serial dilution of a mixture of cDNA. Normalisation was performed against the internal Luciferase control gene using the ΔΔC_q method as this method has been shown to be the least biased approach (De Santis et al., 2010; Zhong et al., 2008). For comparison between treatments, expression of targets genes is shown as relative fold change against the Luciferase control.

Exposure to black or white substrates over the experimental period caused a striking difference in colour of uncooked P. monodon (Fig. 1). Using the‘Lab’ colour space, a (positive=red, negative=green) and b (positive=yellow, negative=blue) define the colour, whereas L (lightness) defines the intensity of that colour. For the animals measured in this study, the L value ranged between 30 and 40; the a value was close to zero, indicating very little red or green; and the b value was negative, indicating the presence of blue (Fig. 2A). In general, P. monodon from the dark substrate were bluer and darker than their counterparts held on a white substrate. It was possible to distinguish between treatment groups of P. monodon on the basis of any of the L, a or b values and, when averaged, the measured values were all significantly different (P<0.001; Table 2). Despite this clear difference in average L, a or b values, when individuals were plotted (Fig. 2B–D), only the L value showed a clear delineation between the black and white substrate treatments (Fig. 2C,D). This indicated that the absolute colour of uncooked animals could not accurately differentiate between treatment groups, but the intensity of that colour was consistently different. A linear regression was also used to determine whether the uncooked L, a or b values were able to predict the substrate the animals had been exposed to (P. monodon black versus P. monodon white). We observed a significant regression for the uncooked L value (P<0.001), but this regression was not significant for either a or b.

The visible colour change induced by cooking was large (Fig. 1), with distinct increases in the L (lightness), a (redness) and b (yellowness) values of P. monodon on either the black or white substrate (Fig. 3A, Table 2). The uncooked colour began near zero or low negative values on the a and b scales, but once cooked, these values reached between positive 15 and 30 on both the redness and yellowness scales, reflecting the strong observed colour change. The ΔE was much smaller for P. monodon held on a white background compared with those held on a black background, indicating that these animals did not change colour as much during cooking, which was also reflected in the difference in subjective Salmofan colour chart score (Table 2).

Regardless of whether P. monodon were held on the black or the white background, the direction of this colour change was the same (indicated by the slope of the lines, Fig. 3A), even though the ΔE value was different. This indicates that the colour seen in animals from black and white substrates is produced by the same mechanism, but the intensity of this colour has changed. This is consistent with the L value being the distinguishing component of the colour scale, indicating that the colour itself is very similar but the intensity or lightness of that colour is different.

Carotenoids were extracted from dissected hypodermal tissue and separated by HPLC. Astaxanthin accounted for >90% of the total carotenoids (data not shown), and peaks evaluated were only quantified from the free, mono-ester or di-ester forms of astaxanthin. Results on whole tissue showed that free astaxanthin levels were lower in P. monodon held on a white substrate compared with those on a black substrate (Fig. 4A), although these changes were not significant. There was also a corresponding increase in esterified astaxanthin in prawns exposed to white substrates, with astaxanthin mono-esters being the predominant form in these animals (55%). These trends were similar to those found by previous work in this species (Tume et al., 2009), which showed that exposure to black or white substrates triggered a change in the proportion of free astaxanthin, from 51.9 to 12.8%, with a corresponding increase in mono-esterified astaxanthin, from 20.8 to 66.2%. Discrepancies between these two data sets may be attributable to slight differences in the methodology of carotenoid extraction and analysis, such as not including the exoskeleton with the hypodermal fraction in this study or differences in peak threshold values. The similar trend seen in this study reinforced the conclusions of Tume et al. (Tume et al., 2009), indicating that esterified forms of astaxanthin play an important role in regulating colouration in P. monodon. This response has also been observed in natural colour morphs of the lobster Panulirus cygnus (Wade et al., 2005), suggesting that this mechanism is not restricted to prawns and potentially exists throughout decapod crustaceans.

In order to further investigate the role of free and esterified astaxanthin, the hypodermal tissue was separated into a soluble protein fraction or an insoluble pellet fraction. The pellet fraction showed no difference in quantitative levels of any carotenoid class across the two treatments, and was mostly (>90%) composed of astaxanthin esters (Fig. 4B). In contrast, the free astaxanthin content of the soluble fraction was 3.6-fold lower in prawns exposed to white substrates compared with those exposed to black substrates (P<0.002), and this fraction contained predominantly (80–85%) free astaxanthin (Fig. 4C). These data demonstrate that astaxanthin esters (mono- and di-esters) are contained in an insoluble pellet fraction of the hypodermis, and the majority of the free astaxanthin is found in the soluble protein fraction and shows quantitative differences in response to changes in background substrate colour. This indicates that the soluble protein fraction may be central to the colour changes that are observed in response to background substrate colour in P. monodon.

The soluble protein fraction was isolated from hypodermal tissue from prawns exposed to both black and white substrates. The colour was initially extracted under native conditions and had a peak absorbance in the blue spectrum at 634 nm, which was very similar to the characteristics of the protein previously identified in this species (Nur-E-Borhan et al., 1995) and also to the properties of α-crustacyanin isolated from lobster shells (Wald et al., 1948; Buchwald and Jencks, 1968). This native extract was denatured using detergents, which triggered a distinctive colour shift from 634 to 484 nm, also typical of disrupting the CRCN–carotenoid interaction. Once purified, the soluble protein fraction had a strong difference in the total absorbance at 484 nm (A₄₈₄) mg^–1 protein between the black (0.283±0.035) and the white (0.108±0.010) treatments.

Equal amounts of these protein samples were separated on the basis of size on a polyacrylamide gel, and total proteins were visualised using a total protein stain (Fig. 5). Of the bands quantified, there was a distinct quantitative difference in a 21 kDa protein across prawns exposed to black or white substrates (Fig. 5), with no change in the quantity of bands 65, 43 or 39 kDa in size. This 21 kDa protein was recognised by a specific CRCN antibody (data not shown), and is the same size as CRCN from this species (Nur-E-Borhan et al., 1995) as well as other crustaceans (Zagalsky, 1985). However, it is important to note that the A and C subunits cannot be distinguished, neither on the basis of size using one-dimensional gel electrophoresis, nor by the CRCN antibody, which is likely to detect both subunits with equal affinity. Relative proportions of the subunits within the CRCN complex may be worthy of investigation using more sensitive techniques, such as two-dimensional gel electrophoresis.

These data show a clear difference in the abundance of the colour protein CRCN as a result of exposure to a black or white substrate. Differences in CRCN protein abundance are consistent with the measured difference in free astaxanthin from the soluble protein fraction, indicating that CRCN bound with free astaxanthin is producing a large proportion of the colour in prawn hypodermis, and is central to the regulation of colouration in prawns. At this point it is not possible to say whether the observed difference in CRCN protein abundance is due to degradation or relocation to another tissue. Detection of CRCN protein in other prawn tissues in response to background treatment has not been attempted. No knowledge exists on the rate at which CRCN protein is accumulated or lost during the rapid expansion or contraction of chromatophore pigments.

To investigate the genetic control of colouration in prawns, the expression of the two CRCN genes was measured across the moult cycle and also in response to black or white substrates. Expression of the CRCN genes was stable across the moult cycle in P. monodon (Fig. 6A). Relative fold changes across the moult stages were very small, and both genes were evenly expressed in all stages. The CRCN-A gene showed the greatest variation in expression, with two small peaks of expression at the A and C stages, as well as a short period of lower expression at D₂ stage. Although not as strong, the CRCN-C gene also showed two small peaks in expression at the C stage, and weaker expression at pre-moult stages D₂ and D₃.

Initial evidence using northern blots suggested that there might have been a short post-moult (A stage) activation of CRCN expression in the freshwater prawn Macrobrachium rosenbergii (Wang et al., 2007). Here we demonstrate very little change in expression across the moult cycle or in response to background substrate colour in P. monodon. We conclude that these subtle changes in CRCN gene expression level are unlikely to be attributable to a biological function in prawns. The control gene GAPDH showed a strong spike of expression during pre-moult stage D₁ (Fig. 6A), suggesting that this gene undergoes strong increases in expression at specific times in the moult cycle. This gene is commonly used as a control gene for quantitative PCR, but these data suggest that this gene should not be used for normalisation of gene expression in prawns.

Results in P. monodon adapted to black or white substrates showed that fold-change differences in the expression level of either CRCN-A or CRCN-C were very small, and both genes were evenly expressed (Fig. 6B). There was a slight increase in CRCN-C expression in those animals exposed to the white substrate; however, this was not significant. Despite a marked colour difference and a strong change in CRCN protein abundance, expression of either CRCN gene showed little variation in prawns exposed to either black or white substrates. The ratio of the expression of the two genes also remained constant. The lack of transcriptional regulation of the CRCN gene suggests a post-translational regulation mechanism or translocation of CRCN protein.

The data presented here demonstrate that the basis of the adaptive background colour response, and the regulation of crustacean colouration, goes beyond a simple expansion or contraction of the pigment structures. This study showed that there is clear interplay between physiological and morphological mechanisms of colouration in crustaceans. When adapted to a dark substrate, prawns with dilated chromatophores also accumulate CRCN protein bound with free astaxanthin within hypodermal tissues. Conversely, in white-substrate-adapted animals with constricted chromatophores, the CRCN protein is depleted, but the astaxanthin is retained in the hypodermal tissue in esterified form. This adaptive response to background colour has a profound impact on the intensity of the prawn's external appearance, and this is reflected in the cooked prawn colour and corresponding colour grade score. The abundance of the CRCN protein is central to the adaptive colour response in prawns, yet there is a distinct lack of regulation of this process at the gene expression level.

 
• a

  red–green scale

 
• b

  blue–yellow scale

 
• CIE

  Commission Internationale de l'Eclairage

 
• CMAR

  CSIRO Marine and Atmospheric Research

 
• CRCN

  crustacyanin

 
• CSIRO

  Commonwealth Scientific and Industrial Research Organisation

 
• EF1α

  elongation factor-1 alpha

 
• GAPDH

  glyceraldehyde-3-phosphate dehydrogenase

 
• L

  lightness

 
• ΔE

  distance between two points in the three-dimensional colour space

The authors would like to thank Gold Coast Marine Aquaculture for supply of experimental animals, Nick Polymeris for his assistance with experimental tank setup and animal husbandry, Joanne Mountford for her assistance with carotenoid quantification and Simon Tabrett for his helpful discussion on the regulation of prawn colouration.