Species specificity of hatching enzyme from medaka and Fundulus
As shown in Figures 1A and 1D, HCE from medaka O. latipes (MHCE) and Fundulus heteroclitus (FHCE1 and FHCE2) swelled the egg envelope of both species. Activities of HCE in the cross-species or xenogenic combinations were somewhat lower (Figure 1B and 1E), but were sufficient to swell the egg envelope. Thus, the action of HCE showed low substrate specificity in both medaka and Fundulus.
In contrast to HCE, LCE could only digest the swollen envelope of the species from which it originated (Figure 1A and 1D). Medaka LCE (MLCE) sharply increased solubilization of the swollen medaka egg envelope with incubation time, whereas MLCE only slightly solubilized that of Fundulus (Figure 1C). Similarly, Fundulus LCE (FLCE) efficiently solubilized the swollen envelope of Fundulus, but not that of medaka (Figure 1 F). These results suggested that LCE activity is substrate-specific in medaka and Fundulus.
Identification of key amino acid residues responsible for the species-specificity of LCE activity
(i) Egg envelope protein
Species-specific egg envelope digestion in Fundulus or medaka would be established by amino acid changes both in egg envelope protein and in hatching enzyme LCE. We first focused on the key LCE cleavage sites in the envelope protein, the mid-ZPd region. We identified six substitutions in 11 amino acid residues of the mid-ZPd regions of both Fundulus and medaka, and synthetic peptides were then designed from the sequences around the mid-ZPd cleavage site (Figure 2C, peptides 1 and 2).
FLCE cleaved the Fundulus mid-ZPd peptide (peptide 1), but not the medaka peptide (peptide 2) (Figure 2A), confirming the species-specificity of egg envelope digestion. MLCE digested the mid-ZPd peptides from both Fundulus and medaka (Figure 2B), suggesting that the species specificity of MLCE does not depend on the peptide sequence around the cleavage site. Because egg envelope has an insoluble macro-molecular structure, the species specificity of MLCE for cleavage of the intact swollen egg envelope may be due to the enzyme-to-substrate interaction in a higher-ordered conformation.
Next, we constructed mutant peptides of the mid-ZPd region whose residue(s) were substituted with either Fundulus- ormedaka-type residues (peptides 3–10). As shown in Figure 2A, FLCE only cleaved peptides 1, 3, and 6, all of which possess the common residue L at the P2 site [22]. L to P substitution at the P2 site of the Fundulus peptide significantly reduced the activity of FLCE. In contrast, P to L substitution at the corresponding site of the medaka peptide enabled FLCE activity. These results suggested that the FLCE-specific digestion could be explained by the presence of L at the P2 site in the FLCE-cleavage site on the substrate.
(ii) Hatching enzyme
It is reasonable to hypothesize that, in addition to the P2 site of the substrate, the species-specific digestion of FLCE depends on the structure of its active site cleft, where FLCE interacts directly with the cleavage site on the egg envelope. The crystal structure of MLCE revealed that 51 residues faced the active site cleft. The residues are shown in Figure 3A. Amongst these residues, 20 substitutions were observed in MLCE or FLCE. To transform the active site cleft from Fundulus-type into medaka-type, we first introduced 20 amino acid mutations into FLCE, and generated recombinant mutant FLCE (rFLCE) named FLCE_mu20 (Figure 3A and 3B). FLCE_mu20 solubilized the swollen medaka egg envelope. Furthermore, the efficiency of FLCE_mu20 towards the medaka mid-ZPd peptide was not significantly different from that of recombinant MLCE (rMLCE) (Figure 4A). Therefore, we concluded that the species specificity of FLCE was caused by some of the 20 amino acid substitutions in the active site cleft.
To identify the most important mutations for species specificity, we reduced the number of mutations introduced into FLCE in a stepwise manner.
-
1)
First, we focused on substitutions between amino acids with distinctly different properties, such as negative, positive, uncharged polar, and/or non-polar amino acids. Among the 20 substitutions, 10 sites showed such differences. These 10 residues in FLCE were substituted with medaka-type residues and named FLCE_mu10 (Figure 3A and C). FLCE_mu10 completely digested the swollen medaka egg envelope (Figure 4A). In addition, FLCE_mu10 had sufficient activity to cleave the medaka mid-ZPd peptide, and its Km value was similar to that of rMLCE (Figure 4E). Therefore, these 10 substitutions were sufficient to cause the FLCE-specific LCE activity.
-
2)
The next strategy was as follows: two of the 10 mutations were returned to Fundulus-types, and named FLCE_mu10. If the returned residue(s) was an important site(s), the peptide cleavage activity of these mutant rFLCEs would decrease, and if not, the activity would not change. We generated five mutant rFLCEs, named FLCE_mu8a to 8e, whose returned sites are shown in Figure 3A. Only FLCE_mu8d did not change the peptide cleavage activity (Figure 4B). These observations suggested that some of the eight mutations (K3M, F74K, R79Y, P91F, N95K, K135H, K138D, and R183T) were important for the species specificity of FLCE.
-
3)
Similar to the above strategy, one of the eight mutations was next returned to the Fundulus-type residue and named FLCE_mu8d. Eight separate mutant rFLCEs were generated, which were named FLCE_mu8d_“returned site”. As shown in Figure 4C, four of these eight mutant rFLCEs had decreased peptide cleavage activity, suggesting that F74K, P91F, K135H, and R183T were important substitutions for FLCE activity and species specificity.
-
4)
Finally, we introduced these four mutations into FLCE and generated recombinant FLCE_mu4 (Figure 3A and B). As shown in Figure 4D, FLCE_mu4 completely solubilized the swollen medaka egg envelope, and its peptide cleavage activity was similar to that of rMLCE (Figure 4D). Therefore, the four residues at positions 74, 91, 135, and 183 were determined to be essential for the species specificity of FLCE.
To determine the contribution of the four FLCE sites to egg envelope digestion, we further individually returned one of the four sites of FLCE_mu4 to the Fundulus type, generating four new mutant rFLCEs, named FLCE_mu4_“returned site”. All the mutant rFLCEs decreased the peptide-cleaving activity (Figure 4D). FLCE_mu4_K74F and FLCE_mu4_T183R lost activity and, in addition, did not solubilize the swollen medaka egg envelope. The Km values of FLCE_mu4_K74F and FLCE_mu4_T183R toward the medaka peptide increased by 1.5- and 3.3-fold compared with that of FLCE_mu4, respectively, suggesting that the affinity of the mutant rFLCEs for the substrate was reduced due to the mutation at either position 74 or 183. Also, in terms of the egg envelope-solubilizing activity, residues at positions 74 and 183 appear to play a critical role in the species specificity of FLCE.
On the other hand, FLCE_mu4_F91P partially digested the swollen medaka egg envelope (Figure 4D and F), and its Km value was similar to that of rMLCE (Figure 4E). Although FLCE_mu4_H135K had the lowest activity toward the medaka mid-ZPd peptide among the four mutants, it completely solubilized the swollen medaka egg envelope. When positions 74 and 183 on rFLCE were substituted to the medaka types, mutant rFLCE (FLCE_F74K&R183T) could not solubilize the swollen medaka egg envelope. This result suggested that positions 91 and 135 are also necessary for the complete egg envelope solubilization. The cleavage of mid-ZPd may be dependent on more complicated enzyme-substrate interaction, in addition to recognition of primary sequence of the cleavage site. We conclude that these four residues are required for FLCE to maintain suitable activity towards the cleavage site peptide and to digest the swollen medaka egg envelope.
Substrate preference of LCEs
We further examined the preference of LCEs towards the P2 site using 13 medaka mid-ZPd peptide mutants. The activity of rFLCE in cleaving the peptide bond between D and S, such as EVQPXD↓SPLSI, was monitored. The minor activities of rFLCE at other cleavage sites, such as EVQPN↓DSPLSI and EVQPD↓DSPLSI, were not further investigated in the present study. rFLCE cleaved five kinds of peptides, whose P2 site residues were L, F, M, Y, or Q (Figure 5A). On the other hand, rMLCE cleaved the majority of the peptides, and its activity was higher than that towards wild-type mid-ZPd peptide with P at the P2 site (Figure 5B). Therefore, we determined that rMLCE had a broader P2 site preference than rFLCE. rFLCE preferentially cleaved the peptides with large hydrophobic residues at the P2 site. Another characteristic difference was in the efficiency of the peptide cleavage. The activity of MLCE was about 10 times greater than that of FLCE. Thus, MLCE and FLCE are different not only in substrate preference, but also in peptide cleavage efficiency.
Next, we evaluated the P2 site preference of LCE by introducing four mutations at positions 74, 91, 135, and 183 that are important for the species specificity of LCE. When medaka-type residues were introduced into Fundulus FLCE (FLCE_mu4), its activity was similar to that of medaka rMLCE (Figure 5D). When Fundulus-type residues were introduced into medaka MLCE (MLCE_mu4), its activity was altered to that of the Fundulus type (Figure 5C). Therefore, we confirmed that the P2 site preference of LCEs was determined by the four amino acid residues in LCE.
A co-evolutionary pathway for hatching enzyme and egg envelope protein
Both genes for hatching enzymes and egg envelope proteins were cloned from various euteleostean species (Additional file 1: Figure S1). We deduced the ancestral sequences using PAML. Figure 6 illustrates the order of changes in amino acids introducing species-specific egg envelope digestion during the evolution of teleostean fishes. The P2 site of the mid-ZPd site in egg envelope protein was predicted to have been occupied by L in the ancestral euteleostean species. These residues were well conserved in most of the euteleostean sequences investigated in this study. The only observed difference occurred during the evolution into Oryziinae, where there was a substitution to V or P.
The ancestral euteleostean LCE was predicted to have had 74 F, 91 F, 135Y, and 183R in its active site cleft. These essential residues were well conserved prior to Atherinomorpha. The ancestor of Atherinomorpha was characterized by a single substitution at position 135. Later, another substitution occurred at position 74 in the ancestor of Beloniformes. Finally, three further substitutions occurred in the ancestor of Oryziinae. These observations suggested that substitutions of the four essential residues in hatching enzyme LCE occurred prior to substitutions in the cleavage site on egg envelope protein.
Activity of ancestral recombinant LCE
Finally, we generated recombinant ancestral LCEs based on the predicted ancestral sequences (Additional file 2: Figure S2) and examined their P2 site preference (Figure 6). The ancestral LCE of Neoteleostei (ancNeotLCE) had strong P2 site preference, similar to that of FLCE. However, the P2 site preference of ancestral LCE of Acanthomorpha (ancAcmoLCE) was less stringent, and the ancestral LCE of Acanthopterygii (ancAcanLCE) had cleavage activity towards various substrates. Again, in the ancestor of Beloniformes (ancBeloLCE), high P2 site preference was revived, but was dramatically altered in the ancestor of Oryziinae, resulting in broader substrate preference. In summary, the MLCE-like activity had been established in the ancestor of Oryziinae, and this ancestral LCE (ancOryzLCE) acquired the ability to cleave the peptide containing P at the P2 site. Therefore, it is reasonable to conclude that the acquisition of this new ability would permit the substitution to P at the P2 site of the ancestor of the medaka egg envelope protein.