Dietary protein quality differentially regulates trypsin enzymes at the secretion and transcription level in Panulirus argus by distinct signaling pathways

As in most crustacea, trypsin enzymes play a central role in protein digestion in spiny lobsters (Johnston, 2003; Celis-Gerrero et al., 2004; Perera et al., 2008a). Some information is available on the biochemical characterization of spiny lobster trypsins (Galgani and Nagayama, 1987; Celis-Gerrero et al., 2004; Perera et al., 2008a) and variation in trypsin activity throughout development and molt stages has also been reported (Johnston, 2003; Perera et al., 2008b). However, the effect of diet on these enzymes has been poorly studied and the time course of trypsin activity after ingestion of different diets has only recently been reported for one spiny lobster species (Jasus edwardsii) (Simon, 2009). In addition, virtually nothing is known about the regulatory mechanisms behind protein digestion in spiny lobsters.

In Panulirus argus, trypsin is a polymorphic enzyme (Perera et al., 2008a) and trypsin cDNA showing different expression rates has been cloned (Perera et al., 2010a). Additionally, trypsin isoforms in lobster appear to differ in catalytic properties or specificity as differences in digestion efficiency have been found among three distinct trypsin phenotypes (Perera et al., 2010b). These features of lobster trypsins indicate several points at which regulation of their activity can occur and, thus, makes this species a good model for studying trypsin regulation in crustacea. Despite the many studies on trypsin enzymes in other crustaceans, these regulatory mechanisms are not well known and, in general, limited information is currently available for non-insect invertebrates (Muhlia-Almazán et al., 2008).

There is also great interest in the commercial growout of wild-caught spiny lobsters (Jeffs and Davis, 2003). Research effort over the past decades has considerably increased our knowledge about the nutritional requirements of spiny lobsters (Williams, 2007) but feeding most species with formulated diets on a least-cost basis is still a challenge. Spiny lobster growth performance is relatively poor when they are fed on dry formulated diets unless high levels of krill meal and/or krill hydrolysate are included (Smith et al., 2005; Barclay et al., 2006; Cox and Davis, 2009). It seems that spiny lobsters cannot efficiently use proteins from common aquafeed ingredients (e.g. fish, squid and soybean meals) in dry food. This conflicts with the fact that they are equipped with a wide repertoire of digestive enzymes, especially proteases (Galgani and Nagayama, 1987; Celis-Gerrero et al., 2004; Perera et al., 2008a), and with the results of previous digestibility studies (Ward et al., 2003; Irving and Williams, 2007; Perera et al., 2010b). Poor protein use could be related to some impairment in digestive progression when lobsters are fed on pelleted diets (Simon and Jeffs, 2008; Simon, 2009).

The aim of the present work was to assess the effect of feeding P. argus with different protein sources on trypsin expression and secretion. Our results demonstrate that trypsin enzymes are differentially regulated at the transcription and secretion level by ingested proteins. Additionally, we provide evidence that intact proteins, rather than small peptides or free amino acids, are the major signal eliciting the prandial secretory response in lobster, whereas transcriptional regulation of the major isoform appears to occur by a distinct mechanism.

Spiny lobsters, P. argus (Latreille 1804) (90–150 g), were collected in the Gulf of Batabanó, Cuba, and held under optimal laboratory conditions at the Center for Marine Research of the University of Havana as described before (Perera et al., 2005). Lobsters were fed daily with fresh fish (Opisthonema oglinum) ad libitum and experiments were carried out after animals had been allowed to adapt to captive conditions for 2 weeks and had fed intensively on the food offered. Only intermolt specimens [determined according to Lyle and MacDonald (Lyle and MacDonald, 1983)] were used in experiments because maximum trypsin activity at this stage has previously been reported (Perera et al., 2008b). Soybean, fish and squid meals used as protein sources in diets were prepared as described earlier (Perera et al., 2010b) and experimental formulated diets (Table 1) were prepared by thoroughly mixing the dry ingredients with oil and then adding water until the consistency was that of stiff dough. This was passed through a mincer, dried at 60°C and broken into a convenient pellet size (∼3×10 mm). Diets were stored at –20°C until use.

It has been shown that the time course of digestive enzymes during digestion in the spiny lobster J. edwardsii is better reflected in the gastric fluid, as only secreted enzymes can be measured (Simon, 2009). Thus, we first examined the time course of trypsin secretion in P. argus by sampling the gastric fluid after ingestion. However, this was possible only when fresh fish were used as food, because the ingestion of dry pelleted diets resulted in dough with almost no free gastric fluid. Specimens of P. argus (N=8) were starved for 48 h, fed with fresh fish (O. oglinum) and sampled every hour for 7 h. Gastric fluid samples were obtained through the oral cavity using disposable insulin syringes as described by Vonk (Vonk, 1960) but instead of a fire-polished glass cannula we used a needle with a plastic cannula over the sharp end. Gastric samples were immediately frozen in liquid nitrogen and then stored at –80°C. Lobsters were handled with care and samples were rapidly taken (less than 1 min) to avoid excessive stress.

As the hyperglycemic response to stress is well known in crustaceans (Kleinholz and Little, 1949), in a preliminary trial we tested the influence of this manipulation on glucose levels in P. argus hemolymph (N=5). Hemolymph was taken from the sinus at the base of the third walking leg as described before (Perdomo et al., 2007) and glucose was measured using a clinical RapiGluco-Test (Helfa Diagnostic, Havana, Cuba) according to the manufacturer’s instructions. This preliminary experiment also confirmed that trypsin activity in gastric fluid does not vary as a result of manipulation.

In a second experiment, four groups of five lobsters each were starved for 48 h and fed with three diets named according to the main protein source they contain (soybean, fish and squid). One group was left without food and served as a control. After 4 h of ingestion, lobsters were chilled by immersion in ice-cold water and dissected for digestive gland and gastric fluid extraction. In this study, digestive gland samples were always taken from the medial and superior part of the gland. Both samples for activity assays were immediately frozen in liquid nitrogen and stored at –80°C. Additional digestive gland samples for expression analysis were rapidly transferred to RNAlater® Solution (Ambion, Life Technologies Corporation, Austin, TX, USA), left to stand at 4°C for 24 h, then stored at –20°C.

For assessing whether digestion end-products or intact proteins can stimulate trypsin synthesis and/or secretion, the effects of protein (diet) hydrolysates and bovine serum albumin (BSA) on the digestive gland were studied in vitro. The incubation medium was Panulirus saline (PS) solution (Zhainazarov et al., 1997) composed of (in mmol l^–1): 460 NaCl, 13 KCl, 14 Na₂SO₄, 13 CaCl₂, 10 MgCl₂, 2 glucose and 10 Hepes, but pH was adjusted to 6.6 instead to 7.4, according to the slightly acidic pH in the digestive gland of crustaceans (Bickmeyer et al., 2008). The solution was filtered using Whatman cellulose nitrate membrane filters (0.45 μm pore size) and aerated at room temperature (25°C) during the 30 min before the experiment.

The hydrolysates were obtained as follows. The formulated diets (soybean, fish and squid) were shaken in PS (0.5 g ml^–1) for 15 min and then incubated with lobster gastric fluid (3:1) for 4 h at room temperature. The mixtures were then centrifuged at 1000 g for 5 min to eliminate insoluble materials. The supernatants were boiled for 5 min, in order to abolish enzyme activity and precipitate proteins, and centrifuged again for 30 min under similar conditions. Although this treatment should leave only free amino acids and peptides in the supernatants, we filtered the supernatants through Amicon filters with a cut-off of 10,000 Da. The resultant solutions were stored at –20°C and are referred to as diet hydrolysates. The total amino acid composition of diets and free amino acids in hydrolysates was determined (Instituto de la Grasa analytical service, Sevilla, Spain). Peptide composition of filtered hydrolysates was assessed by Tricine-PAGE.

Digestive gland samples (30 mg) were taken from 48 h starved lobsters (N=5) as described above and washed in 1 ml of PS to remove hemolymph and digestive enzymes from the surface of the tissue, and then transferred to 1.5 ml Eppendorf tubes containing 200 μl of filtered and aerated PS. Then, 100 μl of the following solutions were added to five samples (tubes) from each individual: soybean hydrolysate, fish hydrolysate, squid hydrolysate, BSA and PS. After 20 min of incubation, digestive glands were transferred to RNAlater® solution, left to stand at 4°C for 24 h and finally conserved at –20°C. The incubating medium was frozen in liquid nitrogen and stored at –80°C for determination of the trypsin activity released.

Digestive glands were homogenized in reaction buffer (200 mmol l^–1 Tris HCl pH 7.5) and centrifuged at 1000 g at 4°C for 30 min. Crude extracts of digestive glands or gastric fluid were diluted with reaction buffer to measure initial rates of enzyme activity. Trypsin activity was measured using 1.25 mmol l^–1N-benzoyl-dl-arginine p-nitroanilide (BApNA) in 200 mmol l^–1 Tris HCl pH 7.5 as described before (Perera et al., 2008a). One unit of trypsin activity was defined as the amount of enzyme that catalyzed the release 1 μmol of pNA per minute, using the appropriate molar extinction coefficient. Under our assay conditions, the molar extinction coefficient of pNA was calculated to be 2.563 l mmol cm^–1. Trypsin activity was expressed per ml or per g of tissue as appropriate.

Proteins were measured according to Bradford (Bradford, 1976) using BSA as standard.

All kits in this study were used following the manufacturer’s instructions. Total RNA was purified using the NucleoSpin® RNA II kit (Macherey-Nagel, Düren, Germany) including RNAse-free DNAse treatment. RNA was quantified by absorbance at 260 nm (A₂₆₀) and its quality was assessed by its A_260/280 value and using the Agilent RNA 6000 Nano Assay Kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). cDNAs were synthesized using the qScript™ cDNA Synthesis Kit (Quanta BioSciences, Gaithersburg, MD, USA) and then used as templates for RT-qPCR on a Mastercycler ep realplex (Eppendorf, Hamburg, Germany) using PerfeCTa™ SYBR® Green Fast-Mix™ (Quanta BioSciences) in white wells twin.tec real-time PCR plates 96 (Eppendorf). Four previously isolated lobster trypsin cDNAs (Perera et al., 2010a) were studied: PaTry1a (GenBank accession no. GU338026), PaTry2 (GU338028), PaTry3 (GU338029) and PaTry4 (GU338030). Specific primers (Biomers.net, Ulm, Germany) used for lobster trypsins were successfully employed before (Perera et al., 2010a) and are shown in Table 2. The efficiency of target amplification for each primer set was optimized in the previous study by using 4 pmol μl^–1 of primers and 60°C of annealing/extension (Perera et al., 2010a); thus, these conditions were employed here. By means of calibration curves (10-fold dilutions, corresponding to cDNA in reactions from 20 ng to 0.2 pg) all primer pairs were checked to produce similar efficiencies, except PaTry4. The reaction volume was 10 μl, containing 4 μl cDNA (4 ng total cDNA), 5 μl of PerfeCTa SYBR Green Fast Mix (2×) and 0.5 μl of each primer. Cycling conditions were as follows: initial denaturation/activation step for 5 min at 95°C, followed by 40 cycles of denaturation (95°C for 30 s) and annealing/extension (60°C for 45 s). Control reactions with DEPC water and RNA instead of cDNA were included to ensure the absence of contamination or sample genomic DNA. Specificity was checked by melting curve analysis. Additionally, PCR products were verified by nucleotide sequencing. All samples were run in duplicate. Replicate PCR reactions generated highly reproducible results. The intra-assay coefficient of variation (CV) for Ct values varied from 0 to 0.004 and the inter-assay CV from 0.011 to 0.027.

Four full-length lobster trypsin cDNAs were amplified as detailed before (Perera et al., 2010a) and cloned into plasmids using the TOPO TA Cloning® Kit (Invitrogen Ltd, Paisley, UK). Plasmids were extracted from Transformed One Shot® TOP10 competent Escherichia coli cells using the GenElute™ Five-Minute Plasmid Miniprep Kit (Sigma-Aldrich, St Louis, MO, USA). Clones containing inserts of the expected size were identified by PCR analysis (T3 and T7 primers of TOPO TA Cloning® Kit) followed by agarose gel electrophoresis, and sequenced from both directions to ensure their correspondence with previously reported lobster trypsins (PaTry1 to PaTry4). Plasmids containing the trypsin variants of interest were quantified using a spectrophotometer. Absolute calibration curves were constructed using circular plasmid vectors (pCR®4-TOPO®) containing each of the four full-length lobster trypsin cDNAs, serially diluted with DEPC water from 3×10⁷ to 30 copies. Targets and calibrators showed similar amplification efficiencies near 100%, in all cases but PaTry4. The number of transcript molecules was calculated from the linear regression of the standard curve and analyses were done with the Mastercycler ep realplex software. Results are expressed as means + s.e.m., after being normalized to ng of total RNA.

The software MEGA5 was used for all tests performed (Tamura et al., 2007). Pairwise comparisons between PaTry1 to PaTry3 nucleotide sequences were carried out to calculate the ratio of non-synonymous to synonymous substitution (K_a/K_s) as an indicator of the pattern of evolution (Hurst, 2002) of these enzymes. The best model of nucleotide substitution was obtained in MEGA5 and resembled that of Tamura (Tamura, 1992) because of the transition–transversion and G+C content biases. Then, taking into account the calculated transition–transversion bias (R=1.21) and applying the Jukes–Cantor correction for multiple substitutions at the same site, we computed the K_a/K_s ratio for each mature trypsin pair using the modified Nei and Gojobori method with 1000 bootstraps for variance estimation. Statistical significance of selection was tested by the Z-test and Fisher’s exact test. A codon-based selection test using the likelihood method was used for detecting positive selection sites.

All data were checked for normality and homogeneity of variance using Kolmogorov–Smirnov and Levene’s tests, respectively, with P≤0.05. Data from the time course of trypsin activity after ingestion and data from protein and activity comparison among diets, hydrolysates and BSA were subjected to one-way ANOVA (P≤0.05). Expression data were analyzed by two-way ANOVA (P≤0.05), with diet and trypsin being the two sources of variation. In all cases, N refers to the number of lobsters used and duplicate measurements were performed. The Tukey test (P≤0.05) was used to determine differences among means. The software package Statistica 7.0 (StatSoft Inc., Tulsa, OK, USA) was used for all tests performed and figures were generated by OriginPro 8 (OriginLab Corporation, Northampton, MA, USA).

A preliminary trial testing the effects of gastric fluid sampling procedure on stress level showed that mean glucose levels in hemolymph did not vary as a result of manipulation (F=0.32, P>0.05, 0.2±0.027 mmol l^–1), and no variation in trypsin activity occurred in the gastric fluid within the 7 h of sampling in non-fed lobsters (F=1.18, P>0.05, 0.7±0.04 U ml^–1).

Trypsin activity in the gastric fluid of fasting lobsters was highly dependent on the protein concentration in the fluid and this relationship was better described by a straight line [trypsin activity=0.1525(protein concentration)+0.044] with a high determination coefficient (R²=0.94). This indicates that 94% of trypsin activity in the gastric fluid can be explained by proteins from digestive gland secretion. Dietary proteins remaining in the gastric chamber after 48 h of fasting and non-trypsin enzymes in the gastric fluid (Perera et al., 2008a) account for the variation observed. Thus, the time course of trypsin activity in gastric fluid after ingestion was followed as being indicative of trypsin secretion by the digestive gland. Trypsin activity in the gastric fluid varied significantly (F=4.16, P≤0.01) after ingestion. Following feeding, trypsin activity progressively decreased in gastric fluid, reaching significantly lower levels after 3 h (Fig. 1). However, between the 3rd and the 4th hour after ingestion a significant enhancement of trypsin activity was observed (Fig. 1), returning rapidly to prefeeding levels. Afterwards, trypsin activity remained stable until 7 h post-feeding.

As trypsin secretion peaked 4 h after ingestion, trypsin activity of the gastric fluid and digestive gland at this time was compared in P. argus feeding on diets of different protein composition. No differences were observed in the trypsin activity of digestive glands between lobsters fed with the different diets and fasting animals (Fig. 2A). However a significant reduction (F=3.96, P≤0.05) in protein content of the digestive gland was observed in individuals ingesting fish and squid diets (Fig. 2B). This reduction in soluble protein in the digestive gland corresponds with the significantly higher (F=8.59, P≤0.05) levels of trypsin activity in the gastric fluid of these lobsters (Fig. 2C). The secretion elicited by fish and squid diets is comparable to that found after the ingestion of fresh fish (Fig. 1), whereas lobsters ingesting the soybean-based diet showed no change in trypsin activity in gastric fluid (Fig. 2C).

Expression was studied by means of RT-qPCR for four trypsin variants (PaTry1 to PaTry4) previously reported in P. argus (Perera et al., 2010a). For absolute quantification, standard curves were generated with plasmids containing the different trypsin cDNAs (Fig. 3) and checked to produce similar amplification efficiencies (E) to their targets PaTry1 (E=0.99, R²=1.0, slope=–3.357), PaTry2 (E=0.97, R²=0.999, slope=–3.401) and PaTry3 (E=1.02, R²=0.999, slope=–3.284). PaTry4 was not included in the statistical analysis because there were so few individuals expressing this transcript in most treatments and because of the small difference in amplification efficiency (E=0.92, R²=1.0, slope=–3.542) obtained. Significant differences in trypsin expression were found among diets (F=3.10, P≤0.05) and among trypsin variants (F=19.35, P≤0.001). PaTry3 was always found to be the most abundant transcript (Fig. 4). No variation was found in the expression of PaTry1 and PaTry2 among feeding treatments, whereas the most abundant variant (PaTry3) significantly increased its expression in lobsters ingesting fish and squid diets (Fig. 4).

Under our assay conditions, incubation of lobster digestive glands with diet hydrolysates (<10 kDa) did not produce significant changes in trypsin secretion into the incubating media (Fig. 5A). Digestive glands incubated with squid diet hydrolysate seemed to have slightly raised trypsin secretion, but a significant increase above controls could only be detected for glands incubated with BSA (Fig. 5A).

PaTry3 was the only trypsin variant induced by dietary stimulation, resembling the results obtained in vivo. Only squid diet hydrolysate increased the copy number for PaTry3 transcripts (Fig. 5B). Total amino acid composition of the test diets was similar except for some differences in Val and Lys content (Table 3). Free amino acid composition differed among diet hydrolysates, reflecting differences in the digestion of animal vs vegetal protein sources. The Val and Glu+Gln content of soybean hydrolysate was lower and the content of Arg and Lys was higher than in the other digests (Table 3). Electrophoresis analysis of hydrolysates showed strong differences in peptide composition between the soybean and the fish/squid digests (Fig. 6).

The K_a/K_s ratio for PaTry1/PaTry2 and PaTry1/PaTry3 pairs was 0.70 and 0.66, respectively, while K_a/K_s for the PaTry2/PaTry3 pair was 1.03. Both Z-test and Fisher’s exact test of selection gave no statistical indication of positive selection. However, the maximum likelihood approach revealed 25 codons evolving under positive selection (dN–dS>0). From these sites, 72% contain unique substitutions in PaTry3. Interestingly, 55% of these PaTry3 substitutions occur in important regions for enzyme function (Table 4).

Trypsin activity has been positively correlated with digestion efficiency and growth in fishes (Rungruangsak-Torrissen et al., 2006; Savoie et al., 2011) but conflicting results have been obtained in crustacea (Le Vay et al., 1993). The effect of dietary protein on trypsin activity has been reported in shrimp larvae (Le Moullac et al., 1994), juveniles (Muhlia-Almazán et al., 2003) and adults (Le Moullac et al., 1996) but in general, the effect of proteins on trypsin enzymes has been poorly studied in crustacea in comparison to vertebrates and other groups of invertebrates.

Trypsin secretion has been reported in fasting mammals (Konturek et al., 2003), shrimp (Lehnert and Johnson, 2002) and insects (Moffatt et al., 1995) [except in batch digesters such as mosquitoes (Lehane et al., 1995)]. In accordance, a significant amount of trypsin activity was observed in fasting lobsters. After ingestion of fish, trypsin activity in the gastric fluid of P. argus dropped gradually, as reported before for the lobster J. edwardsii (Simon, 2009) and assumed to be due to the drinking of water (Simon and Jeffs, 2008). In the case of dry diets, food absorbed most of the gastric fluid. This highlights a great problem with the use of dry diets in P. argus as proteins must be dissolved before being attacked by proteolytic enzymes. Recent studies have pointed out that the solubility of dry diets would determine protein digestion efficiency in lobsters (Simon, 2009; Perera et al., 2010b).

Previous studies have found different secretory patterns in the digestive gland of crustaceans: (i) three phases (at 0–15 min, 1–2 h and 3.5–5 h after a meal) in the lobster Homarus gammarus (Barker and Gibson, 1977), (ii) two peaks within 6 h of feeding in the crayfish Astacus leptodactylus (Hirsch and Jacobs, 1928) and the shrimp Litopenaeus vannamei (Muhlia-Almazán and García-Carreño, 2002), and (iii) only one phase of secretion 1–4 h after feeding in the prawn P. semisulcatus (Al-Mohanna et al., 1985). In spiny lobsters, a significant amount of trypsin enzymes from the digestive gland have attained the gastric chamber after 4 h of ingestion [J. edwardsii (Simon, 2009); P. argus, this work].

In our study no statistical differences in trypsin activity of the digestive gland were found in P. argus 4 h after ingestion. Studies on the effects of diet on digestive enzyme activity in crustacea have yielded contradictory results and this has mainly been attributed to the use of the digestive gland as the examined tissue. After collection, glands are usually disrupted to obtain homogenates, in which stored and secreted enzymes are mixed. In spite of this limitation, it is interesting to note that the non-significant trend observed in the digestive gland for trypsin activity indicates more enzyme secretion in fish- and squid-fed lobsters than in fasting or soybean-fed specimens. The high amount of trypsin enzymes remaining in the gland after 4 h of digestion could explain the lack of significance in our results. In contrast, soluble protein content of the gland significantly decreased after feeding lobsters with fish and squid diets (but not with soybean diet), perhaps as a result of secretion of other enzymes in addition to trypsin. Consequently, by examining the gastric fluid, we found that soybean-based diets lack stimulatory capacity in the lobster digestive gland (or some components of the soybean meal block signal transduction into secreting cells), while fish and squid diets elicited a secretory response similar to that observed with fresh fish. These results can be taken as definite evidence that the nature of ingested protein affects the released of trypsin enzymes from the digestive gland of P. argus and that the use of fish and squid proteins in dry feeds is not limited by the digestive response of lobsters, but probably by diet solubility hampering digestibility as suggested before (Simon and Jeffs, 2008; Simon, 2009; Perera et al., 2010b). Conversely, in adult shrimp, Le Moullac and colleagues found that only casein increased trypsin activity, while squid meal and fish soluble concentrate had no effect (Le Moullac et al., 1996).

Our results indicate that trypsin enzymes in P. argus are also regulated transcriptionally by dietary proteins, as reported before for shrimp (Le Moullac et al., 1996; Muhlia-Almazán et al., 2003), mosquito (Noriega et al., 1994) and rats (Lhoste et al., 1994; Hara et al., 2000). Interestingly, the most abundant transcript in P. argus (PaTry3) was the only one for which dietary up-regulation could be demonstrated while all other trypsins appear to be expressed in a constitutive fashion. To our knowledge, this is the first time that different trypsin isoforms have been shown to differ in their responsiveness to dietary proteins for a crustacean species. Differences in trypsin isoform expression have been found in Daphnia magna but in response to protease inhibitors in the diet (Schwarzenberger et al., 2010). It is noteworthy that in the present work, RNA extraction was carried out from tissue biopsies with no regard for the different cell types contained in the digestive gland. Trypsin synthesis in crustacean digestive gland has been demonstrated only in the F cells (Lehnert and Johnson, 2002), while expression values are given herein on a total RNA basis. Therefore, our results probably underestimate actual trypsin expression in F cells of P. argus.

Early studies (Hirsch and Jacobs, 1928) revealed that digestive enzymes in crustaceans are secreted upon feeding but the mechanism was unknown. In dogs (Meyer and Kelly, 1976) and humans (Thimister et al., 1996) amino acids and peptides are more effective than intact proteins in stimulating pancreatic exocrine secretion. However, trypsin secretion is more stimulated by intact proteins than by hydrolyzed proteins or amino acids in rats (Green and Miyasaka, 1983), fishes (Cahu et al., 2004) and insects (Blakemore et al., 1995; Lehane et al., 1995). In order to study the possible signals for trypsin induction/secretion in P. argus, we performed in vitro assays in which digestive glands were exposed to an intact protein (BSA) and diet hydrolysates (<10 kDa). Our results show that trypsin secretion did not change upon incubation with hydrolysates, while 0.01% BSA significantly stimulated trypsin secretion. Trypsin secretion in insect opaque zone preparations occurs at concentrations of BSA as low as 0.0001% (Blakemore et al., 1995). It seems that in P. argus the signals for trypsin secretion are similar to those in rats, fishes and insects (intact proteins) and this information has been not available for crustaceans until now.

Postprandial enzyme secretion in mammals is mostly (70%) stimulated by cholecystokinin (CCK) through neural pathways (Konturek et al., 2003). Although the mechanism of action is not as well understood as in mammals, CCK is also recognized as a major regulator of trypsin secretion in fishes (Buddington and Krogdahl, 2004). There is strong evidence for the presence of CCK-like peptides in crustaceans (van Wormhoudt et al., 1989; Favrel et al., 1991; Resch-Sedlmeier and Sedlmeier, 1999) although their role in digestive enzyme secretion has not been clarified. However, in our in vitro assays BSA was able to elicit enzyme secretion by acting directly over the digestive gland (prandial mechanism), suggesting that, as in insects (Lehane et al., 1995; Noriega et al., 1996; Noriega and Wells, 1999; Lu et al., 2006; Brandon et al., 2008; Graf et al., 1998), peripheral signals (neural and endocrine) are not obligatory though may have modulator roles.

We found that squid diet hydrolysate was able to induce transcription of the major trypsin (PaTry3) after feeding while intact proteins like BSA had no effect. This result indicates that induction signals are different to those for secretion and are probably mediated by the appearance of free amino acids in the gland. [Note the absence of a significant amount of small peptides in squid and fish hydrolysates (Fig. 6).] This result suggests a stepwise regulation of trypsin enzymes during digestion.

Lys, Arg and Met are considered to be the most limiting amino acids for growth of crustaceans (Akiyama et al., 1992). Digestibility of these amino acids (and His and Tyr) appears to be high in the three tested diets (but the soybean diet was supplemented with free Met; Table 1). It is known that these amino acids largely appear as free amino acids in casein (Savoie, 1994) and soybean (Henn and Netto, 1998) hydrolysates. No major disparity was found in the amino acid composition of the squid and fish hydrolysates. Thus, the difference in their capacity for stimulating trypsin expression could result from differences in amino acid concentration. The release of free amino acids from squid meal far exceeds that from fish meal during early digestion (Perera et al., 2010b). Previous studies have also suggested a specific effect of squid proteins on digestive enzymes (Le Moullac et al., 1996; Perera et al., 2005).

As the current results suggest that free amino acid can increase trypsin expression in the digestive gland of P. argus (and potentially protein digestion efficiency), further studies are needed on the selection of rapidly digestible protein sources. However, we argue for the inclusion of protein hydrolysates to counteract the poor digestibility of dietary proteins. The effect of protein hydrolysates other than in terms of the attractiveness of diets is not yet well understood.

The very marked differences in expression among trypsins in P. argus encouraged us to perform a preliminary analysis on the pattern of evolution for three P. argus trypsins (PaTry1 to PaTry3). Analysis of K_a/K_s for PaTry1/PaTry2 and PaTry1/PaTry3 pairs suggests purified selection (K_a/K_s<1) although values are unusually high, whereas neutral selection is suggested by K_a/K_s for PaTry2/PaTry3 (K_a/K_s≈1). However, because averaging K_a/K_s for all sites ignores the type of selective pressure that applies to individual amino acids, we used the maximum likelihood approach for detecting positively evolving sites in a background of purifying/neutral selection.

Our results indicate that most sites of P. argus trypsins are under purifying selection and thus that these proteins are subject to functional constraints. In addition, most non-synonymous changes resulted in the substitution of similar amino acids, with a slight change in charge or hydrophobicity while conserving the volume of residues. However, it is noteworthy that several positively evolving sites are within (or in close vicinity to) functionally relevant motifs (Table 4). Loop 37 in crayfish trypsin is known to be important in hydrophobic interactions with extended subsites of inhibitors (Fodor et al., 2005) and probably proteinaceous substrates. In P. argus trypsins, Trp and Ser residues of PaTry1 and PaTry2 loop 37 are both replaced by the more hydrophobic Phe in PaTry3. In the calcium binding site, Asn residues in PaTry1 and PaTry2 are replaced by the more charged Asp in PaTry3. Two, one and two positive selected sites occur within loops 1, 2 and 3, respectively. These three loops determine trypsin specificity (Hedstrom, 1996). It is not possible to anticipate the impact of these substitutions on enzyme activity but one can hypothesize that highly expressed PaTry3 has distinct functional roles. Although high levels of expression have been correlated with lower rates of protein evolution (Drummond et al., 2005), other studies have provided some evidence for recent adaptive evolution of protein-coding regions in highly expressed genes (Holloway et al., 2007). Studies on the gene sequence of lobster trypsins would provide information on regulatory elements (e.g. cis-acting elements in 5′ regions) to support our findings.

Trypsin isoform sequences from other spiny lobster species are also required to truly determine whether some trypsins are under selective pressure at certain sites to undergo adaptive evolution (or are evolving independently) while other trypsins are evolving in a concerted fashion, as occurs in Drosophila (Wang et al., 1999).

The authors express their gratitude to the crew of the research vessel ‘Felipe Poey’ for their assistance during animal collection. We thank Y. S. Wunderink for useful comments on sequence evolution mechanisms and E. García for comments on an earlier version of the manuscript. Comments from reviewers significantly improved this work. We specially thank the Editor for his support.