Aspartic proteinases in Antarctic fish

Aspartic proteinases in Antarctic fish

Marine Genomics 2 (2009) 1–10 Contents lists available at ScienceDirect Marine Genomics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m...

1MB Sizes 9 Downloads 153 Views

Marine Genomics 2 (2009) 1–10

Contents lists available at ScienceDirect

Marine Genomics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m a r g e n

Review

Aspartic proteinases in Antarctic fish Viviana De Luca a, Giovanna Maria a, Gaia De Mauro a, Giuliana Catara a, Vincenzo Carginale a, Giuseppe Ruggiero a, Antonio Capasso a, Elio Parisi a,1, Sebastien Brier b, John R. Engen b, Clemente Capasso a,⁎ a b

CNR, Institute of Protein Biochemistry, Via P. Castellino 111-80131, Naples, Italy Department of Chemistry & Chemical Biology and The Barnett Institute for Chemical & Biological Analysis, Northeastern University, Boston, MA, USA

a r t i c l e

i n f o

Article history: Received 2 October 2008 Received in revised form 9 February 2009 Accepted 2 March 2009 Keywords: Antarctica Aspartic proteinase Cathepsin Pepsin Nothepsin Kinetic analysis

a b s t r a c t The present review surveys several recent studies of the aspartic proteinases from Antarctic Notothenioidei, a dominating fish group that has developed a number of adjustments at the molecular level to maintain metabolic function at low temperatures. Given the unique peculiarities of the Antarctic environment, studying the features of Antarctic aspartic proteinases could provide new insights into the role of these proteins in fish physiology. We describe here: (1) the biochemical properties of a cathepsin D purified from the liver of the hemoglobinless icefish Chionodraco hamatus; (2) the biochemical characterization of Trematomus bernacchii pepsins variants A1 and A2 obtained by heterologous expression in bacteria; and (3) the identification of two closely related, novel aspartic proteinases from the liver of the two Antarctic fish species mentioned above. Overall, the results show that Notothenioidei aspartic proteinases display a number of characteristics that are remarkably different from those of mammalian aspartic proteinases, including high turnover number or high catalytic efficiency. We have named the newly identified aspartic proteinases “Nothepsins” and classified them relative to aspartic proteinases from other species. © 2009 Elsevier B.V. All rights reserved.

Contents 1. 2. 3.

The Antarctic environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aspartic proteinases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cathepsin D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Biochemical characterization of a cathepsin D from the liver of Chionodraco hamatus . . . . . . 3.2. Kinetic analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Findings on the icefish Cathepsin D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Pepsin isoforms in Trematomus bernacchii: production of fish pepsin variants by heterologous expression 4.1. Biochemical properties of recombinant fish pepsins . . . . . . . . . . . . . . . . . . . . . . 4.2. Kinetic of the hydrolysis of Pro-Thr-Glu-Phe-(p-nitro-Phe)-Arg-Leu. . . . . . . . . . . . . . . 4.3. Cleavage preference of fish pepsins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Findings on recombinant pepsin isoenzymes from T. bernacchii. . . . . . . . . . . . . . . . . 5. Nothepsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Findings on the new Antarctic fish proteinase, named Nothepsin . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

⁎ Corresponding author. Tel.: +39 81 6132559; fax: +39 81 6132714. E-mail address: [email protected] (C. Capasso). 1 This review is dedicated to the memory of our wonderful colleague, Dr. Elio Parisi, who recently passed away. 1874-7787/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.margen.2009.03.001

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . in E. coli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

1 2 2 2 3 4 4 5 6 7 7 7 9 9 9 10

2

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

1. The Antarctic environment Antarctica is one of the most isolated regions in the world: originally it was encompassed in the super continent Gondwana that originated from the fragmentation of Pangea and successive continental drift. Following the breakup of Gondwana, Antarctica remained in contact with South America, Australia and New Zealand until the end of the Cretaceous period when it attained its present position at the South Pole (Eastman, 1993). Owning to progressive cooling, the Antarctic continent underwent a drastic climate change; today, the temperature of the seawater is constantly close to −1.9 °C, the equilibrium temperature of seawater with ice. Antarctica is at the moment a region of the globe that has suffered the effects of environmental disturbance caused by anthropic activities. For good reason, it has been considered the largest natural laboratory existing on the planet. Antarctica is gaining increasing attention from the scientific community in relation to the effects of global change and biodiversity preservation. Antarctica became isolated from the other continents about 25 million years (MY) ago, following the opening of the Drake Passage and formation of the Polar Front, an oceanic frontal system running between 50°S and 60°S: this barrier prevents fish migration from and to Antarctica. The geographical isolation of Antarctic fish and the drastic climatic conditions were probably the main reason for the specializations achieved during cold adaptation. Unlike other marine habitats, the Antarctic Ocean is a closed basin, isolated from other areas by the Polar Front, resembling ancient lakes: fish living in such a kind of habitat exhibit monophyly, endemism and speciosity (Johns and Avise, 1998). The continental shelf of Antarctica constitutes a habitat harboring two dominant fish groups, represented by liparids and notothenioids. Notothenioids form a perciform suborder, comprising 120 species mostly endemic to the waters surrounding Antarctica (Eastman, 1993). Phylogenetic relationships among notothenioid taxa have been reconstructed on the basis of morphological (Eastman, 1993) and molecular data (Bargelloni et al., 1994). From all of the above, it is apparent that notothenioids diverged quite recently (10–15 MY ago), likely from a single ancestor, thus providing an ideal system for evolutionary studies at a comparative level. The notothenioid group includes highly cold-adapted and remarkably stenothermal fish: some species, such as the members of hemoglobinless family of Cannichthydae (icefish), can be regarded as the most extreme forms of a series of fish with diminishing hemoglobin content. Antarctic fish have evolved a number of adaptive mechanisms to cope with the hostile conditions of their habitat and these mechanisms are quite interesting from many perspectives. Further, the ecological importance of notothenioids in the Antarctic marine ecosystem and their remarkable adaptation to this extreme environment helps explain why a great amount of effort has been devoted to studying notothenioid physiology, thereby producing a great wealth of data at both functional and biochemical levels.

autocatalytic cleavage leads to the formation of the active enzyme (Kageyama et al., 1989). The catalytic mechanism depends on the presence of two aspartic acids positioned roughly in the center of a deep cleft forming the active site and covered by a hairpin loop (flap) protruding from the N-terminal lobe of the molecule. The active site cleft can accommodate about seven residues of a substrate. These residues are usually designated as P4-P3-P2-P1⁎P1′-P2′-P3′, with the scissile peptide bond between P1 and P1′ indicated by an asterisk and normally flanked by two hydrophobic residues. The corresponding subsites that constitute the topography of the active site cleft in each enzyme are designated accordingly as S4-S3-S2-S1-S1′-S2′-S3′ (Fusek and Vetvicka, 1995). On the basis of their different molecular characteristics and tissue/ cellular localization, aspartic proteinases of vertebrates have been classified as cathepsin D, cathepsin E, gastricsin, pepsin, chymosin and renin. In addition, other types of aspartic proteinases have been isolated from invertebrates (Kay et al., 1996; Francis et al., 1997), plants (Guruprasad et al., 1994; Brodelius et al., 1995), retroviruses (Phylip et al., 1992; von der Helm, 1996) and a number of microbial sources (Toogood et al., 1995; Hill and Phylip, 1997). In our studies on aspartic proteinases in notothenioids, we have focused our attention on cathepsins and pepsins. 3. Cathepsin D Cathepsin D is present in most cells in many species and is probably one of the major factors contributing to lysosomal digestive activity (Metcalf and Fusek, 1993; Tsuji and Akasaki, 1994). Cathepsin D activity was found to be particularly elevated in a wide variety of tissues during remodeling or regression, as well as in apoptotic cells (Yang and Schnellmann, 1996). Although important for the degradation of intracellular proteins, cathepsin D has also been implicated in the processing of antigens and peptide hormones (Diment et al., 1989; van Noort and van der Drift, 1989) and overexpression and secretion of cathepsin D was observed in human breast cancer (Garcia et al., 1996). Cathepsin D has been thoroughly investigated in mammals, but a limited number of studies are available on cathepsin D from other species. In fish, cathepsin D has received some attention because of the potential importance of fish physiology in relation to aquaculture (Gildberg, 1988). Indeed, high cathepsin D activity found in fish tissues, especially during spawning, is considered to be the main factor influencing deterioration of fish products (Gildberg, 1988). However, most of these studies have been aimed at the identification of the enzymatic activity present in various fish tissues, and rarely at elucidating the molecular properties of piscine cathepsin D. Recently, however, the molecular characterization of a cathepsin D from the ovary of rainbow trout has been reported; this enzyme seems to play a role in the processing of yolk proteins in oocytes (Brooks et al., 1997). 3.1. Biochemical characterization of a cathepsin D from the liver of Chionodraco hamatus

2. Aspartic proteinases Aspartic proteinases constitute a widely distributed protein superfamily, whose members accomplish a variety of functions, including protein degradation at low pH (Davies, 1990; Dunn and Corbett, 1992). They share a high degree of sequence similarity and have a characteristic bilobal tertiary structure (Foltmann, 1981; Cooper et al., 1990; Davies, 1990). Individual members of the superfamily differ in the topography of their active sites and in their cellular localization. In turn, these features determine physiological function (Takeda-Ezaki and Yamamoto, 1993; Sato et al., 1995). These enzymes are produced as zymogens in a large variety of organs and tissues. The primary structure of the zymogen includes a signal peptide (or presequence) and the so-called propart, whose

We recently isolated an aspartic proteinase from the liver of C. hamatus, an Antarctic icefish, by anion-exchange chromatography followed by affinity chromatography on concanavalin-A Sepahrose (Capasso et al., 1999). While human cathepsin D, as isolated from tissues such as liver or spleen, is commonly a two-chain enzyme, in contrast, the icefish enzyme had clearly not been subject to proteolytic cleavage since a single chain of approximately 40 kDa was observed by SDS-PAGE under both reducing and non-reducing conditions (Capasso et al., 1999). This molecular weight was in agreement with that obtained by gel permeation analysis of the native enzyme. The amino terminal sequence, determined by automated Edman degradation of the electroblotted enzyme, was found to be APTPETLKNYLDAQYYGEIGLGTPPQPFT. This information enabled a primer to be designed to

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

3

Fig. 1. Alignment of the amino acid sequences of cathepsin D from different sources. Catalytic residues are indicated in boxes. The asterisk (⁎) indicates identity at all aligned positions; the symbol (:) relates to conserved substitutions, while (.) means that semi-conserved substitutions are observed. Multialignment was performed with the program Clustal W, version 1.83. Legend: C. hamatusD, icefish cathepsin D (GenBank accession no. CAA07719); O. mykissD, trout cathepsin D (GenBank accession no. AAC60301); G. gallusD, chicken cathepsin D (GenBank accession no. NP_990508); M. musculusD, mouse cathepsin D (GenBank accession no. NP_034113); R. norvegicusD, Norvay rat cathepsin D (GenBank accession no. NP_59961); H. sapiensD, human cathepsin D (GenBank accession no. CAA28955).

facilitate the preparation of cDNA encoding the C. hamatus aspartic proteinase. The sequence of the 5′-end of the gene was determined by 5′-RACE. The full nucleotide sequence showed an open reading frame encoding a 396 residue polypeptide chain containing all the typical features of an aspartic proteinase. The molecular mass of the mature enzyme from the predicted sequence was 36 kDa. Phylogenetic analysis showed that this icefish proteinase displayed the most similarity with cathepsin D of different species. Alignment of the icefish sequence with those for cathepsin D from several other species (Fig. 1) confirmed the extent of the identity. All the results described above allowed us to identify the purified icefish enzyme as a cathepsin D. 3.2. Kinetic analysis Kinetic parameters were determined for the purified icefish aspartic proteinase using the synthetic chromogenic substrate ProPro-Ile-Phe-(p-nitro-Phe)-Arg-Leu. The specificity constants that were calculated (Table 1) indicate that the optimum of activity was at pH 3.0 as might be expected for an aspartic proteinase. ANOVA tests performed on the results reported in Table 1 showed that the kinetic parameters were dependent from the pH value (DF = 11; F = 274.30; P b 0,0001). Enzyme activity was fully inhibited by pepstatin, but was not affected by a protein inhibitor from Ascaris lumbricoides, a nematode worm. This inhibitor has been documented previously to be a potent inhibitor of human/vertebrate aspartic proteinases, such as cathepsin E and pepsin, but to have no effect on cathepsin D (Samloff et al., 1987).

The temperature dependence was also measured using the same synthetic substrate, and measurements were made in parallel with human cathepsin D (Table 2). ANOVA tests performed on the whole dataset showed that the differences between the calculated kinetic parameters and species are significant (DF = 14; F = 139.86; P b 0.0001). At the lower temperatures evaluated (8, 14, and 27 °C) the values measured for the turnover number (kcat) were similar between the icefish and human enzymes. The Km values were noticeably lower for C. hamatus enzyme. At higher temperatures (37 and 45 °C), however, the values of kcat measured for the icefish enzyme were approximately double those calculated for the human enzyme; since the km values were lower (in all cases). The values for the specificity constant (kcat/Km) were 5–10-fold better for the enzyme from icefish operating at temperatures at and above human physiological values. The C. hamatus cathepsin D started to diminish in catalytic efficiency as the temperature was raised from 45 to 50 °C. This prompted an examination of the stability of the enzymes at the highest temperature (i.e. 50 °C) and at pH 7.5 (Capasso et al., 1999). A Table 1 Effect of pH on the activity of icefish cathepsin D. pH

Km (mM)

Kcat (s− 1)

Kcat/Km (mM− 1s− 1)

2.0 3.0 4.0 5.0

105 25 30 30

48 82 87 33

460 3300 2900 1100

pH dependence of kinetic parameters for the hydrolysis of the synthetic substrate by the icefish cathepsin D. Assay was determined at 37 °C.

4

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

Table 2 Kinetic parameters determined for icefish and human cathepsins D. Icefish cathepsin D

Human cathepsin D

T

Km

Kcat

Kcat/Km

Km

Kcat

°C

(mM)

(s− 1)

(mM− 1s− 1)

(mM)

(s− 1)

Kcat/Km (mM− 1s− 1)

8 14 27 37 45 50

10 17a 12 25 25 85

7 12 30 82 155 190

700 700 2500 3300 6200 230

40 40 55 65 140

6 15 33 41 67

150 375 600 630 480

b

b

b

All measurements were carried out using the synthetic substrate Pro-Pro-Thr-Ile-Phe⁎Nph-Arg-Leu. Values were determined at pH 3.0 at the indicated temperatures. a Inhibition observed at high concentration of substrate. b At 50 °C, the human enzyme inactivated too rapidly for a reliable kinetic measurement to be made.

half-life of 55 min was observed for the icefish enzyme. By comparison, the half-life of human cathepsin D was 3 min. Cathepsin D isolated from the liver of rainbow trout was inactivated completely in less than 2 min at 50 °C (Capasso et al., 1999). A number of recent studies regarding the comparison of Km values for orthologs of enzymes purified from differently thermal adapted species show that for each orthologos an increase in temperature causes an increased km value. However, at the measured temperature the Km value is lowest for the most warm-adapted species and highest for the most cold-adapted species (Hochachka and Somero, 2002). The authors assume that at any single temperature all the possible conformational states that characterize an enzyme will be defined by the inherent structural flexibility of the protein. That means, if states exist that are both permissive and non-permissive of ligand binding, it would be expected, at a common temperature of measurement, an enzyme from a cold-adapted species would exhibit lower affinity for a ligand than an ortholog from a warm-adapted species. The cold-adapted enzymes compensate by evolving relatively higher Km's at any measurement temperature compared to warm-adapted orthologos. In fact, the more flexible structure of the ortholog of the cold-adapted species would allow fluctuation of the protein to occur more readily, and a lower fraction of the total population of enzyme molecules would exist in a binding competent state. Intriguing, our results demonstrate that icefish cathepsin D is more heat-stable and has a lower Km values than the human version. The characteristics of this Antarctic cathepsin D reflect what would normally be considered features of a more warmadapted enzyme. 3.3. Findings on the icefish Cathepsin D Antarctic fish are highly cold adapted and remarkably stenothermal (Cheng and Detrich, 2007), as a consequence of evolution of antifreeze glycoproteins and the higher catalytic efficiency of many enzymes at low temperatures (Mangum and Hochachka, 1998). For this reason, the study of enzymes in these poikilothermic species is of interest, especially in relation to the strategies adopted by these organisms to achieve a normal level of proteolysis at temperatures well below that of homeothermic species. Organisms can follow different strategies to achieve cold adaptation at the metabolic level: a) one way is to increase enzyme production in order to compensate for reduced kinetic efficiency; b) another is via the expression of enzymes with relatively higher substrate turnover and the capacity to maintain ligand-binding properties at low temperature (D'Amico et al., 2002); and c) as described in Section 3.2 adapted enzymes could realize the cold-adaptation by evolving relatively higher Km's at any measurement temperature compared to warm-adapted orthologos (Hochachka and Somero, 2002). Enzymes from cold-adapted species commonly have an optimal activity at relatively modest temperatures; above this value, the loss of

activity is usually attributed to thermal inactivation as a consequence of the increased structural flexibility of the protein. In their recent review on the characteristics of cold-adapted enzymes, Gerday and co-workers (Feller et al., 1996) identified two major criteria that in general, typify such proteins: 1) their catalytic efficiency is higher than that of their mesophilic counterparts over a temperature range from 0 to 30 °C; 2) they have limited stability at moderate temperatures. From the biochemical features of the purified enzyme we found in C. hamatus liver (Section 3) and from the amino acid sequence predicted from that of the cDNA, the C. hamatus liver proteinase appears to be a cathepsin D. This conclusion is supported by the wellestablished presence of cathepsin D as a major lysosomal proteinase in the liver of many species, including fish (Brooks et al., 1997). From our results, it is readily apparent that icefish cathepsin D not only meets the requirements of criterion 1 described by Gerday, but its catalytic efficiency continues to exceed that of human cathepsin D at temperatures as high as 50 °C. Even more clearly, it is evident that the C. hamatus proteinase completely contradicts the requirements of criterion 2 of cold-adapted enzymes, as the half-life of the protein was about 1 h at 50 °C. It has been demonstrated that in psychrophilic organisms, enzyme thermostability is not always inversely related to catalytic efficiency (Narinx et al., 1997). Cold-adapted enzymes are commonly suggested to have a reduction in the number of proline, arginine, and aromatic residues in order to increase flexibility of the protein structure and to enhance interaction with the aqueous surroundings. Comparison of the sequence of the icefish cathepsin D with those of the counterpart enzymes from other mesophilic species indicates that this is not the case for icefish cathepsin D. Furthermore, the icefish enzyme contains only the two predicted glycosylation sites that are conserved in cathepsin D from all other species so far examined. Thus, although it has been proposed (Yang et al., 1997) that the level of glycosylation (a highly flexible carbohydrate chain) may act a heat reservoir to stabilize the conformation of proteins, it would seem unlikely that this explanation can account for the high level of thermostability in the case of fish cathepsin D. It was not possible to determine the nature and extent of glycosylation of cathepsin D since insufficient amounts of the protein could be purified from the liver of the rather scarce C. hamatus caught in Antarctica; to enable such quantitative analysis, exploitation of recombinant DNA technology would allow the production of sufficient amounts of cathepsin D in eukaryotic expression systems for structural studies. Elucidation of a structure for the icefish enzyme, together with site-specific mutagenesis and kinetic analyses, may provide a rational explanation for the remarkable activity and stability showed by cathepsin D from the Antarctic icefish. 4. Pepsin isoforms in Trematomus bernacchii: production of fish pepsin variants by heterologous expression in E. coli Pepsins are a family of aspartic proteinases accomplishing important digestive functions in both invertebrates and vertebrates (Kageyama, 2002). The two major groups of immunogenetically and biochemically distinct pepsins are pepsin A and pepsin C, the latter also being known as gastricsin. These enzymes are active at extremely low pH values (Fruton, 1970; Dunn et al., 1987) in both the stomach and the duodenum, and are quickly denatured when the pH exceeds 5.5 (Foltmann, 1981), thus preventing continuous proteolytic action that may damage these organs. Pepsin C has a slightly higher optimal pH than pepsin A, but substrate specificity is nearly the same for the two enzymes (Fusek and Vetvicka, 1995). Most of our knowledge of pepsin derives from studies on human and mammalian enzymes (Abad-Zapatero et al., 1990; Gildberg et al., 1990; Sanchez et al., 1992; Fujinaga et al., 1995), and much less information is available on enzymes from other vertebrates, including fish (Gildberg, 1988; Tanji et al., 1988; Gildberg et al., 1990).

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

We have previously cloned and sequenced three forms of pepsin A (which were named A1, A2 and A3), and a single form of gastricsin (named pepsin C), from the gastric mucosa of the Antarctic notothenioid Trematomus bernacchii (rock cod) (Carginale et al., 2004). Note that the naming convention used here is tentative, because fish pepsin A is a close relative of mammalian pepsin A, pepsin F and chymosin, whereas fish gastricsin is close to a common ancestor of mammalian gastricsin and pepsin B. Phylogenetic analysis has shown that rock cod pepsin C and pepsin A form two distinct clades, and the three fish pepsin A isotypes result from two rounds of gene duplication leading to the most ancestral pepsin, pepsin A3, and to the most recent forms represented by pepsin A1 and pepsin A2 (Carginale et al., 2004). Molecular modeling studies have unraveled significant structural differences in these enzymes with respect to their mesophilic counterparts. Rock cod pepsin A2 displayed local changes in the substrate-binding cleft, with a significant reduction of the hydropathic character and concomitant increase in the flexibility of this region with respect to the two isoforms A1 and A3, the gastricsin and the other fish pepsins (Carginale et al., 2004). In general, a reduced level of hydropathy and higher molecular flexibility characterize cold-adapted proteins such as the pepsins from rock cod (Marshall, 1997). Moreover, the central role of pepsin A in food digestion makes it an attractive candidate for genetic engineering and identification of the amino acid residues that modulate its activity and stability. To better understand the molecular mechanisms responsible for adaptation of food digestion at temperatures below 0 °C, we produced the two T. bernacchii fish pepsin variants A1 and A2 (hereafter referred to as fish pepsins A1 and A2) by heterologous expression in Escherichia coli. The enzymes were purified, and their biochemical properties were studied in comparison to pepsin A from porcine stomach (Brier et al., 2007). To determine the purity and validate the correct mass of the purified fish pepsinogens, ESI MS was used (Brier et al., 2007). The mass spectra confirmed that the proteins were relatively pure, as the major peaks in the spectra were those of pepsinogen. 4.1. Biochemical properties of recombinant fish pepsins Fish pepsins were prepared by activating purified recombinant pepsinogens by acidification with 2 M glycine-HCl (pH 2.0). The solution

5

was neutralized by dialysis against sodium acetate buffer (pH 5.3). The recovered solution containing the active form of fish pepsin A1 showed a single band on SDS-PAGE, and similar results were observed with activated fish pepsin A2. The conversion of pepsinogen to pepsin is expected to decrease the molecular mass by approximately 4600 Da, owing to the loss of the N-terminal enzyme propart (residues 1-38). The N-terminal sequence determined by Edman degradation of the electroblotted recombinant fish pepsins was YQSGTESMTNDADLSYYGVI for both isoforms, confirming that the mature enzymes were generated by autoactivation of the refolded zymogen. The secondary structure of fish pepsins A1 and A2 was also investigated by far-UV CD spectroscopy. The CD spectra for the two fish pepsin isoforms overlapped, indicating no substantial secondary structure differences. Secondary structure calculations from the far-UV CD spectral data showed that both Antarctic enzymes contained a high proportion of β-sheets (56.1% for A1 and 52.2% for A2), as reported for mammalian pepsins (Cooper et al., 1990). This result indicates that the secondary structure of the recombinant proteins was not altered during protein processing. Sequence alignment of Antarctic fish versus pig pepsins revealed strong sequence identity (52.7%) between the three enzymes (Fig. 2). As compared to pig pepsin, both Antarctic fish enzymes lack the Cterminal sequence GMDVP (shaded box, Fig. 2). Interestingly, some of these residues belong to the subsites that constitute the topography of the active site cleft. Notably, the four amino acids E287, M289, V291 and T293 forming the loop EGMDVPT are residues of the S4-S1 subsites (E287 and M289) and S1′–S3′ subsites (V291 and T293) of pig pepsin that make contact with the residues of an ideal heptapeptide substrate (Sielecki et al., 1990). The enzymatic activity of fish pepsins A1 and A2 was investigated at different pH values and temperatures and compared with that of pig pepsin. ANOVA tests performed on the whole dataset of Fig. 3A and B showed that the calculated kinetic parameters were dependent from pH and species (DF = 14; F = 82.133; P b 0.0001) and from temperature and species (DF = 12; F = 449.04; P b 0.0001), respectively. A broad pH range was tested using hemoglobin as the substrate. The mesophilic porcine enzyme exhibited 100% relative activity at pH 2.0, even though a rapid reduction of activity occurred at pH 2.5. In contrast, both Antarctic isoenzymes were less active at their optimum pH (2.5 for fish pepsin A1 and 2.0 for fish pepsin A2), and showed a slow decline of their relative activity at pH values above their optimum (Fig. 3A). The effect of the temperature on enzyme activity

Fig. 2. Alignment of Antarctic recombinant T. bernacchii isoenzymes with pig pepsin. The two aspartic acid residues present in the catalytic site are indicated in the transparent boxes. The missing residues in fish pepsins are marked in the gray box. Multialignment was performed with the program Clustal W. Legend: T. bernacchiiPEPA1, recombinant pepsinogen A1 (GenBank accession no. AJ550949); T. bernacchiiPEPA, recombinant pepsinogen A2 (GenBank accession no. AJ550950); S. scrofaPEP, swine pepsinogen (GenBank accession no. AAA31095).

6

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

Fig. 3. (A) Effect of pH on the activity of pig pepsin and recombinant Antarctic isoenzymes A1 and A2. (B) Effect of temperature on pig pepsin and recombinant fish enzymes. For studying the influence of pH and temperature on the enzyme activity, the enzymatic activity was determined using denatured hemoglobin as substrate. The relative activity is expressed as percentage of the highest activity over the pH or temperature range examined. ■, activity of commercial pepsin; ▲, activity of fish pepsin A1; ○, activity of fish pepsin A2.

was also examined for fish and pig pepsins. The optimum temperatures (the temperatures at which there was the most activity) for Antarctic fish pepsins A1 and A2 were found to be 50 °C and 37 °C respectively. At lower temperatures (4 °C, 10 °C and 25 °C), pig and fish pepsin A2 showed similar relative activities, whereas fish pepsin A1 was slightly less active (Fig. 3B). Pig pepsin exhibited 100% relative enzyme activity at 60 °C but the activity dropped sharply to 20% at 70 °C. At the temperatures at which they were optimally active, the fish isoforms were less active than the pig enzyme, and they rapidly lost activity above 50 °C. This prompted us to examine the stability of all three enzymes at 50 °C and at pH 5.3: aliquots were removed at appropriate times and the residual activity was measured at pH 2.0 using hemoglobin as the substrate. Under these conditions, a half-life of 310 min was observed for the mammalian enzyme. By comparison, the half-lives of Antarctic fish pepsins A1 and A2 under the same conditions were 270 and 72 min, respectively. As compared to pig pepsin and fish pepsin A1, fish pepsin A2 appeared to be more labile at the temperature at which they are maximally active. The full inactivation observed for pepsin A2 might be a consequence of the increased structural flexibility of the protein. Carginale et al. (Carginale et al., 2004) reported that, apparently, fish pepsin A2 underwent local changes such as reduced hydropathy and increased flexibility at the level of the substrate cleft. Fish and pig pepsin activity was assayed in the presence of increasing concentrations of pepstatin A, using hemoglobin as the substrate. Pepstatin A is a well-known acid protease inhibitor that binds in the active site cleft of the enzyme (Marciniszyn et al., 1976; Fujinaga et al., 1995). Pepsin activity was determined at pH 2.0 and 37 °C for all three enzymes. Fish and pig pepsins were inhibited completely by the equivalent molar amount of pepstatin, but fish

Table 3 Effect of pepstatin A on fish and pig pepsins.

isoforms showed different pepstatin A inhibition behavior. Commercial pig pepsin required an inhibitor:enzyme ratio of 0.3:1 to produce 50% inhibition, whereas fish pepsins A1 and A2 required an inhibitor: enzyme ratio of 0.1:1 to reach about 50% and 60% inhibition, respectively (Table 3). 4.2. Kinetic of the hydrolysis of Pro-Thr-Glu-Phe-(p-nitro-Phe)-Arg-Leu Kinetic parameters were determined for the purified Antarctic pepsins and pig pepsin using the synthetic chromogenic substrate Pro-Thr-Glu-Phe-(p-nitro-Phe)-Arg-Leu (peptide 1) and are shown in Table 4. The Km values were lower for fish enzymes. The measured kcat values for fish pepsins A1 and A2 were 22 and 130 times smaller, respectively, than that of pig pepsin. Although the Km values of fish pepsins were lower, the resultant values for the specificity constant (kcat⁄Km) of fish pepsins A1 and A2 were six and 12 times smaller, respectively, than that of pig pepsin. These data suggest that the differences in the sequence and 3D structure of fish and pig pepsins (see above) may well affect the kinetic behavior of the porcine and the fish enzymes. In porcine pepsin, there is a loop formed by the following residues: EGMDVPT. It contains four amino acids that are residues of the S4-S3′ subsites. This loop is missing in the Antarctic fish enzymes. As a consequence, fish pepsins might accommodate the synthetic substrate in the catalytic pocket better than pig pepsin, due to a reduction in steric hindrance in fish pepsin. This hypothesis is corroborated by the fact that fish pepsins were more sensitive to pepstatin A than was pig pepsin. On the other hand, the specificity constants (kcat ⁄ Km) of fish pepsins were lower than that of the mammalian enzyme. This may be because the differences in subsite residues decrease the rate of hydrolysis. To explore the possibility that differences between fish and pig pepsins result from sequence changes that affect the substrate recognition region of the active site, we analyzed the digestion of two different substrates, as shown below.

Inhibition (%) Pepstatin/Enzyme ratio (mol/mol)

Pig pepsin

Fish isoforms A1

Fish isoforms A2

0.01 0.04 0.1 0.3 1 4

0 5 16 50 99 100

0 18.4 53 75 100 100

3.2 9 63 75 94 100

Enzymes were assayed in the presence of increasing concentration of pepstatin A using 2.5% hemoglobin as substrate. Pepsin activity was determined at pH 2.0 and 37 °C. Pepstatin A is a well-known acid protease inhibitor that binds in the active site cleft of the enzyme.

Table 4 Kinetic constants determined for Antarctic and pig pepsins. Enzyme

Km (mM)

Kcat (s− 1)

Kcat/Km (mM− 1s− 1)

Antarctic isoenzymes A1 Antarctic isoenzymes A2 Pig pepsin

0.074 0.025 0.26

3.32 0.54 72.02

45 22 277

All reactions were performed at 37 °C and pH 2.0, using the chromogenic peptide substrate Pro-Thr-Glu-Phe-(p-nitro-Phe)-Arg-Leu, (indicated as peptide 1 in the text).

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

4.3. Cleavage preference of fish pepsins To determine whether changing the substrate residues in proximity to the scissile peptide bond Phe–Phe would cause a difference in the activity of Antarctic pepsins, we used an HPLC assay involving two substrates: peptide 1 [described above] and peptide 2 [Phe-Gly-His-Phe-(p-nitro-Phe)-Ala-Phe]. These two peptides have different sequences on either side of the Phe–Phe scissile bond residues. Both fish pepsin A1 and fish pepsin A2 were less efficient at cleaving peptide 1 than was pig pepsin (Brier et al., 2007). However, both fish pepsins were unable to cleave peptide 2, and pig pepsin activity towards this substrate was reduced as compared to its digestion of peptide 1. These data suggest that the flanking residues on either side of the P1–P1′ position play a role in the substrate cleavage of the fish enzymes. They are consistent with the hypothesis suggested above, whereby the missing loop in Antarctic pepsins (formed by residues EGMDVPT in pig pepsin) alters the substrate cleavage of the fish pepsins by removing some of the residues near the scissile peptide bond. The specificity of the Antarctic fish pepsins A1 and A2 for residues at the P1 and P1′ positions that flank the scissile peptide bond was further determined by digesting a test protein with both fish pepsins A1 and A2, as well as pig pepsin. The resulting peptic peptides were sequenced by mass spectrometry. The digestion pattern of the test protein, rabbit muscle creatine phosphokinase (CK-MM), was different for digestion with pig versus fish pepsins. The peptides in each part of the HPLC separation were visualized, compared, and then sequenced by tandem MS (MS⁄MS) (Brier et al., 2007). The P1–P1′ amino acid preference for CK-MM was summarized (Fig. 4). In general, the fish pepsins produced much larger pieces than the pig pepsin during the 3.5 min digestion time, again indicating the difference in the overall activity between the mesophilic and the Antarctic enzymes. The results of this analysis indicate that both pig pepsin and fish pepsins A1 and A2 prefer hydrophobic residues in the P1–P1′ positions and that the overall specificity was similar among the three enzymes. However, important differences were apparent. Fish pepsin A1 could tolerate histidine or methionine in the P1 position, whereas fish pepsin A2 tolerated lysine and threonine in the P1 position. Pig pepsin was able to accommodate aspartic and glutamic acids in the P1 position, whereas neither of the fish pepsins produced any peptides that had these charged amino acids in the P1 position. As a result, the overall digestion profiles of the three pepsins on CK-MM were different. Further, pig pepsin produced many fragments in one portion of the substrate protein, whereas the fish pepsins preferred other regions (Brier et al., 2007). The size of the peptide population in this study was too small to allow any definitive conclusions about the

Fig. 4. Specificity of the Antarctic fish and pig pepsins for residues at the P1 and P1′ position. The specificity was determined by digesting the test protein, rabbit muscle creatine phosphokinase (CK-MM). The size of the letters indicates the probability of cleavage.

7

contribution to specificity from amino acids in positions outside P1 and P1′. Taken together, our data indicate that the specificities of cleavage of fish pepsins A1 and A2 are similar but not identical to that of pig pepsin, and we attribute the differences to the key changes in the active site of the fish enzymes. 4.4. Findings on recombinant pepsin isoenzymes from T. bernacchii Multiple forms of type A pepsinogens are known to occur in various animals (Foltmann, 1981), and most of them are known to be products of different genes, as has been shown in humans (Zelle et al., 1988) and cows (Lu et al., 1988). The presence of multiple forms of pepsin A in some species may be correlated with the type of food or to the feeding habitat. The present review describes enzymatic properties of two recombinant pepsinogen isoforms, A1 and A2, from the Antarctic rock cod. The two Antarctic isoenzymes are, to our knowledge, the first Antarctic fish pepsinogens obtained using the strategy of expression in E. coli followed by refolding and purification (Brier et al., 2007). The enzymes used for the analysis have been analyzed by far-UV CD spectroscopy technique. The analysis of the secondary structure tells us that all the enzymes considered are in the active form. Besides, the difference in the relative activity reflects differences in enzyme stability as confirmed by the far-UV CD spectroscopy carried out at the temperature indicated in the Fig. 3B (data not shown). In terms of the cold adaptation of pepsins from Antarctic fish, our study suggests that fish pepsin A2 meets the criterion for a coldadapted enzyme showing limited stability at a moderate temperature (i.e. 50 °C). However, neither isoform meets the criterion that coldadapted enzymes have a specific activity higher than that of their mesophilic counterparts over a temperature range from 0 °C to 30 °C. We now believe that Antarctic rock cod have adopted primarily gene amplification to accomplish efficient food digestion at low temperatures. Gene amplification increases enzyme production to compensate for the reduced kinetic efficiency. Antarctic rock cod have three forms of pepsin A (Carginale et al., 2004) resulting from two gene duplication events: the first event consisted of the duplication of the pepsin A3 ancestor gene; the second duplication led to the appearance of pepsin A1 and A2 genes that contributed to a further increase in enzyme concentration. Through the production of isoenzymes that perform nearly the same type of digestion with similar specificity (Kageyama, 2002), Antarctic fish may accomplish a similar overall rate of digestion as their mesophilic counterparts. 5. Nothepsin We have identified two novel aspartic proteinases from the liver of two Antarctic fish species, the red-blooded Trematomus bernacchii and the haemoglobinless icefish Chionodraco hamatus. At first, it was not immediately clear what these proteins were or if they had any similarity to known aspartic proteinases. The nucleotide sequences of cDNA obtained by RT-PCR of RNA extracted from the two fishes show 90% identity with each other but only 58% identity with aspartic proteinases from other sources (Capasso et al., 1999). The open reading frame of the T. bernacchii gene encodes a 406 amino acid polypeptide chain, while that of C. hamatus encodes a chain of 402 residues (Capasso et al., 1998). The two polypeptide sequences correspond to aspartic proteinases with the two typical aspartyl residues in the active site. The estimated Mr values for both are about 40 kDa. A BLAST analysis, carried out using the sequence of the protein in C. hamatus as the query sequence, showed that the first forty top ranking library sequences are represented by aspartic proteinases. With the exception of the T. bernacchii sequence (optimized score = 4980), all the other sequences have similarity scores falling in the range 1147–775. Cathepsin D (about 60% identity), cathepsin E (57–59% identity), pepsin (58%) and gastricsin (57%) from various

8

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

9

Fig. 6. Bootstrap majority rule consensus tree of the sequences shown in Fig. 5. The tree was constructed using the program PAUP (version 3.1.1). Bootstrap value on 100 replicates are reported at branch points.

species all have comparable identities to the two fish sequences. This does not readily permit assignment of the fish enzymes into one of these sub-families on a statistical basis. However, renin and chymosin have significantly lower scores: still the Antarctic enzymes would not appear to be either renin or chymosin equivalents. On this basis, it was necessary to scrutinize the two fish sequences on a less objective basis. The sequences of the newly identified Antarctic fish aspartic proteinases were aligned with those of other known aspartic proteinases (Fig. 5). The hallmarks characterizing the individual aspartic proteinases are indicated with numbers in the multialignment reported in Fig. 5. In the propart region, the Antarctic sequences show a consensus motif (hallmark 1) typical of cathepsin E. The KY motif (hallmark 2) present in the propart region of pepsin and cathepsin D is missing in the Antarctic sequences. These show an RY motif (hallmark 3) displaced to the right of the KY consensus. Other features common to cathepsin E are: the cysteine residues probably involved in the dimerization of the peptide chain (hallmark 4) and the N-glycosylation site (hallmark 6). Interestingly, the consensus marked 9 present also in cathepsin D. Distinctive features of the Antarctic enzymes were an insertion (hallmark 5) absent in all the other aspartic proteinase for which sequence is available, and the lack of the proline residues in the region known as the polyproline loop in cathepsin D, cathepsin E and renin (Capasso et al., 1998). We constructed a most parsimonious tree (Fig. 6) in order to better investigate the relationships of the Antarctic sequences with different aspartic proteinase species. Following this strategy, the two new aspartic proteinases that were isolated from the Antarctic fish appeared closely associated with each other, but clustered in a branch distinct from that of cathepsin D. It is noteworthy that fish cathepsin D is more closely related to other vertebrate cathepsin D than to those of the Antarctic proteinases. Although closer to the clusters of pepsins and cathepsins E, the Antarctic fish enzymes also appear well separated from these two groups. For this reason, the name “Nothepsin” was chosen for the two

new aspartic proteinases identified for the first time in the Antarctic Notothenioidei. 5.1. Findings on the new Antarctic fish proteinase, named Nothepsin In the present review, we describe the identification of two novel aspartic proteinases from the liver of two Antarctic fish species. Sequence analysis shows features of these Antarctic enzymes that are not present in related enzymes of other organisms. The two aspartic proteinases described are thus difficult to classify. The two Antarctic proteinases have several features typical of mammalian cathepsin E, but in other aspects they do not appear closely related to any other aspartic proteinase known, including fish cathepsin D. Possibly, the new Antarctic proteinases are the piscine equivalent of mammalian cathepsin E, but a direct comparison is not possible at present as no enzyme of this class has yet been described in fish. Alternatively, the proteinase found in the Antarctic fish may be the result of modifications of an ancestral aspartic proteinase gene. The name “Nothepsin” may be appropriate for this enzyme. 6. Conclusions As we have described in this work, aspartic proteinases from the Notothenioidei are in many ways different from the mammalian aspartic proteinases. Life in the harsh conditions of Antarctic waters has resulted in alterations to the enzymatic properties and expression of these proteases. Further work is likely to uncover more differences and lead to a better understanding of cold-adaption in general. Acknowledgments This study was supported by the Italian National Programme for Antarctic Research (P.N.R.A.) and the National Institutes of Health (GM-070590).

Fig. 5. Alignment of the sequences of the aspartic proteinases from different organisms. Numbers on the top of each sequence block indicate enzyme hallmarks (7, 10 and 12 represent consensus motifs common to all aspartic proteinases; 8 is the cathepsin D loop; for other hallmarks see text). Legend: C. hamatus, Antarctic aspartic proteinase (GenBank accession no. AJ223836); T. bernacchii, Antarctic aspartic proteinase (GenBank accession no. Y08567); for M. musculusD, R. norvegicusD, H. sapiensD, G. gallusD, and O. mykissD, see Fig. 1; A. aegypti, mosquito aspartic proteinase (GenBank accession no. Q03168); O. ariesREN, sheep renin (GenBank accession no. AAA69809); M. musculusE, mouse cathepsin E (GenBank accession no. CAA71859); R. rattusE, rat cathepsin E (GenBank accession no. BAA08128); H. sapiensE, human cathepsin E (GenBank accession no. AAH42537); O. cuniculusE, rabbit cathepsin E (GenBank accession no. P43159), C. porcellusE, swine cathepsin E (GenBank accession no. P25796); M. fuscataGAS, monkey gastricsin (GenBank accession no. P03955); H. sapiensGAS, human gastricsin (GenBank accession no. AAB18273); M. mulattaPEP, monkey pepsinogen (GenBank accession no. AAA36902); S. scrofaPEP, (see Fig. 2); O. cuniculusPEP, rabbit pepsinogen (GenBank accession no. AAA85369); O. ariesCHY, sheep chymosin (GenBank accession no. P18276); S. japonicum, blood fluke aspartic proteinase (GenBank accession no. Q7JNB4).

10

V. De Luca et al. / Marine Genomics 2 (2009) 1–10

References Abad-Zapatero, C., Rydel, T.J., Erickson, J., 1990. Revised 2,3, a structure of porcine pepsin: evidence for a flexible subdomain. Proteins 8, 62–81. Bargelloni, L., Ritchie, P.A., Patarnello, T., Battaglia, B., Lambert, D.M., Meyer, A., 1994. Molecular evolution at subzero temperatures: mitochondrial and nuclear phylogenies of fishes from Antarctica (suborder Notothenioidei), and the evolution of antifreeze glycopeptides. Mol. Biol. Evol. 11, 854–863. Brier, S., Maria, G., Carginale, V., Capasso, A., Wu, Y., Taylor, R.M., Borotto, N.B., Capasso, C., Engen, J.R., 2007. Purification and characterization of pepsins A1 and A2 from the Antarctic rock cod Trematomus bernacchii. Febs J. 274, 6152–6166. Brodelius, P.E., Cordeiro, M.C., Pais, M.S., 1995. Aspartic proteinases (cyprosins) from Cynara cardunculus spp. Flavescens cv. cardoon; purification, characterisation, and tissue-specific expression. Adv. Exp. Med. Biol. 362, 255–266. Brooks, S., Tyler, C.R., Carnevali, O., Coward, K., Sumpter, J.P., 1997. Molecular characterisation of ovarian cathepsin D in the rainbow trout, Oncorhynchus mykiss. Gene 201, 45–54. Capasso, C., Lees, W.E., Capasso, A., Scudiero, R., Carginale, V., Kille, P., Kay, J., Parisi, E., 1999. Cathepsin D from the liver of the antarctic icefish Chionodraco hamatus exhibits unusual activity and stability at high temperatures1. Biochim. Biophys. Acta 1431, 64–73. Capasso, C., Riggio, M., Scudiero, R., Carginale, V., di Prisco, G., Kay, J., Kille, P., Parisi, E., 1998. Molecular cloning and sequence determination of a novel aspartic proteinase from Antarctic fish. Biochim. Biophys. Acta 1387, 457–461. Carginale, V., Trinchella, F., Capasso, C., Scudiero, R., Riggio, M., Parisi, E., 2004. Adaptive evolution and functional divergence of pepsin gene family. Gene 333, 81–90. Cheng, C.H., Detrich 3rd, H.W., 2007. Molecular ecophysiology of Antarctic notothenioid fishes. Philos. Trans. R. Soc. Lond. B Biol. Sci. 362, 2215–2232. Cooper, J.B., Khan, G., Taylor, G., Tickle, I.J., Blundell, T.L., 1990. X-ray analyses of aspartic proteinases. II. Three-dimensional structure of the hexagonal crystal form of porcine pepsin at 2.3 A resolution. J. Mol. Biol. 214, 199–222. D Amico, S., Claverie, P., Collins, T., Georlette, D., Gratia, E., Hoyoux, A., Meuwis, M.A., Feller, G., Gerday, C., 2002. Molecular basis of cold adaptation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 357, 917–925. Davies, D.R., 1990. The structure and function of the aspartic proteinases. Annu. Rev. Biophys. Biophys. Chem. 19, 189–215. Diment, S., Martin, K.J., Stahl, P.D., 1989. Cleavage of parathyroid hormone in macrophage endosomes illustrates a novel pathway for intracellular processing of proteins. J. Biol. Chem. 264, 13403–13406. Dunn, R.W., Corbett, R., 1992. Yohimbine-induced seizures involve NMDA and GABAergic transmission. Neuropharmacology 31, 389–395. Dunn, B.M., Valler, M.J., Rolph, C.E., Foundling, S.I., Jimenez, M., Kay, J., 1987. The pH dependence of the hydrolysis of chromogenic substrates of the type, Lys-Pro-XaaYaa-Phe-(NO2)Phe-Arg-Leu, by selected aspartic proteinases: evidence for specific interactions in subsites S3 and S2. Biochim. Biophys. Acta 913, 122–130. Eastman, J., 1993. Antarctic fish biology. Evolution in a unique enviroment. Academic Press, INC., San Diego, California, USA. Feller, G., Narinx, E., Arpigny, J.L., Aiattaleb, M., Daise, E., Genicot, S., Gerday, C., 1996. Enzymes from psycrophilic organisms. FEMS Microbiol. Rev. 18, 189–202. Foltmann, B., 1981. Gastric proteinases: structure, function, evolution and mechanism of action. Essays Biochem. 17, 52–84. Francis, S.E., Sullivan Jr., D.J., Goldberg, D.E., 1997. Hemoglobin metabolism in the malaria parasite Plasmodium falciparum. Annu. Rev. Microbiol. 51, 97–123. Fruton, J.S., 1970. The specificity and mechanism of pepsin action. Advan. Enzymol. Relat. Areas Mol. Biol. 33, 401–443. Fujinaga, M., Chernaia, M.M., Tarasova, N.I., Mosimann, S.C., James, M.N., 1995. Crystal structure of human pepsin and its complex with pepstatin. Protein Sci. 4, 960–972. Fusek, M., Vetvicka, V., 1995. Aspartic proteinases. Physiology and Pathology. CRC press, Boca Raton. Garcia, M., Platet, N., Liaudet, E., Laurent, V., Derocq, D., Brouillet, J.P., Rochefort, H., 1996. Biological and clinical significance of cathepsin D in breast cancer metastasis. Stem Cells 14, 642–650. Gildberg, A., 1988. Aspartic proteinases in fishes and aquatic invertebrates. Comp. Biochem. Physiol. B 91, 425–435. Gildberg, A., Olsen, R.L., Bjarnason, J.B., 1990. Catalytic properties and chemical composition of pepsins from Atlantic cod (Gadus morhua). Comp. Biochem. Physiol. B 96, 323–330.

Guruprasad, K., Tormakangas, K., Kervinen, J., Blundell, T.L., 1994. Comparative modelling of barley-grain aspartic proteinase: a structural rationale for observed hydrolytic specificity. FEBS Lett. 352, 131–136. Hill, J., Phylip, L.H., 1997. Bacterial aspartic proteinases. FEBS Lett. 409, 357–360. Hochachka, P.W., Somero, G.N., 2002. Biochemical Adaptation. Mechanisms and Process in Physiological Evolution. Oxford University press, New York. Johns, G.C., Avise, J.C., 1998. Tests for ancient species flocks based on molecular phylogenetic appraisals of Sebastes rockfishes and other marine fishes. Evolution 52, 1135–1146. Kageyama, T., 2002. Pepsinogens, progastricsins, and prochymosins: structure, function, evolution, and development. Cell. Mol. Life Sci. 59, 288–306. Kageyama, T., Ichinose, M., Miki, K., Athauda, S.B., Tanji, M., Takahashi, K., 1989. Difference of activation processes and structure of activation peptides in human pepsinogens A and progastricsin. J. Biochem. (Tokyo) 105, 15–22. Kay, J., Tyas, L., Humphreys, M.J., Hill, J., Dunn, B.M., Berry, C., 1996. Aspartic proteinases from parasites. Adv. Exp. Med. Biol. 389, 247–250. Lu, Q., Wolfe, K.H., McConnell, D.J., 1988. Molecular cloning of multiple bovine aspartyl protease genes. Gene 71, 135–146. Mangum, C.P., Hochachka, P.W., 1998. New directions in comparative physiology and biochemistry: mechanisms, adaptations, and evolution. Physiol. Biochem. Zool. 71, 471–484. Marciniszyn Jr., J., Hartsuck, J.A., Tang, J., 1976. Mode of inhibition of acid proteases by pepstatin. J. Biol. Chem. 251, 7088–7094. Marshall, C.J., 1997. Cold-adapted enzymes. Trends Biotechnol. 15, 359–364. Metcalf, P., Fusek, M., 1993. Two crystal structures for cathepsin D: the lysosomal targeting signal and active site. Embo. J. 12, 1293–1302. Narinx, E., Baise, E., Gerday, C., 1997. Subtilisin from psychrophilic antarctic bacteria: characterization and site-directed mutagenesis of residues possibly involved in the adaptation to cold. Protein Eng. 10, 1271–1279. Phylip, L.H., Mills, J.S., Parten, B.F., Dunn, B.M., Kay, J., 1992. Intrinsic activity of precursor forms of HIV-1 proteinase. FEBS Lett. 314, 449–454. Samloff, I.M., Taggart, R.T., Shiraishi, T., Branch, T., Reid, W.A., Heath, R., Lewis, R.W., Valler, M.J., Kay, J., 1987. Slow moving proteinase. Isolation, characterization, and immunohistochemical localization in gastric mucosa. Gastroenterology 93, 77–84. Sanchez, L.M., Freije, J.P., Merino, A.M., Vizoso, F., Foltmann, B., Lopez-Otin, C., 1992. Isolation and characterization of a pepsin C zymogen produced by human breast tissues. J. Biol. Chem. 267, 24725–24731. Sato, K., Nagai, Y., Suzuki, M., 1995. Purification and characterization of an acid proteinase from Dirofilaria immitis worms. Adv. Exp. Med. Biol. 362, 299–304. Sielecki, A.R., Fedorov, A.A., Boodhoo, A., Andreeva, N.S., James, M.N., 1990. Molecular and crystal structures of monoclinic porcine pepsin refined at 1.8 A resolution. J. Mol. Biol. 214, 143–170. Takeda-Ezaki, M., Yamamoto, K., 1993. Isolation and biochemical characterization of procathepsin E from human erythrocyte membranes. Arch. Biochem. Biophys. 304, 352–358. Tanji, M., Kageyama, T., Takahashi, K., 1988. Tuna pepsinogens and pepsins. Purification, characterization and amino-terminal sequences. Eur. J. Biochem. 177, 251–259. Toogood, H.S., Prescott, M., Daniel, R.M., 1995. A pepstatin-insensitive aspartic proteinase from a thermophilic Bacillus sp. Biochem. J. 307 (Pt 3), 783–789. Tsuji, H., Akasaki, K., 1994. Identification and characterization of lysosomal enzymes involved in the proteolysis of phenobarbital-inducible cytochrome P450. Biol. Pharm. Bull. 17, 568–571. van Noort, J.M., van der Drift, A.C., 1989. The selectivity of cathepsin D suggests an involvement of the enzyme in the generation of T-cell epitopes. J. Biol. Chem. 264, 14159–14164. von der Helm, K., 1996. Retroviral proteases: structure, function and inhibition from a non-anticipated viral enzyme to the target of a most promising HIV therapy. Biol. Chem. 377, 765–774. Yang, X., Schnellmann, R.G., 1996. Proteinases in renal cell death. J. Toxicol. Environ. Health 48, 319–332. Yang, J., Teplayakov, A., Quail, J.W., 1997. Crystal structure of the aspartic proteinase from Rhizomucor miehei at 2.15 Å resolution. J. Mol. Biol. 268, 449–459. Zelle, B., Evers, M.P., Groot, P.C., Bebelman, J.P., Mager, W.H., Planta, R.J., Pronk, J.C., Meuwissen, S.G., Hofker, M.H., Eriksson, A.W., et al., 1988. Genomic structure and evolution of the human pepsinogen A multigene family. Hum. Genet. 78, 79–82.