To test, at the level of individual amino acids, the conformation of an exchangeable apolipoprotein in aqueous solution and in the presence of an osmolyte trimethylamine-N-oxide (TMAO), six synthetic peptide analogues of human apolipoprotein C-1 (apoC-1, 57 residues) containing point mutations in the predicted α-helical regions were analyzed by circular dichroism (CD). The CD spectra and the melting curves of the monomeric wild-type and plasma apoC-1 in neutral low-salt solutions superimpose, indicating 31 ± 4% α-helical structure at 22 °C that melts reversibly with Tm,WT = 50 ± 2 °C and van't Hoff enthalpy ΔHv,WT(Tm) = 18 ± 2 kcal/mol. G15A substitution leads to an increased α-helical content of 42 ± 4% and an increased Tm,G15A = 57 ± 2 °C, which corresponds to stabilization by δΔGapp = +0.4 ± 1.5 kcal/mol. G15P mutant has ∼20% α-helical content at 22 °C and unfolds with low cooperativity upon heating to 90 °C. R23P and T45P mutants are fully unfolded at 0–90 °C. In contrast, Q31P mutation leads to no destabilization or unfolding. Consequently, the R23 and T45 locations are essential for the stability of the cooperative α-helical unit in apoC-1 monomer, G15 is peripheral to it, and Q31 is located in a nonhelical linker region. Our results suggest that Pro mutagenesis coupled with CD provides a tool for assigning the secondary structure to protein groups, which should be useful for other self-associating proteins that are not amenable to NMR structural analysis in aqueous solution. TMAO induces a reversible cooperative coil-to-helix transition in apoC-1, with the maximal α-helical content reaching 74%. Comparison with the maximal α-helical content of 73% observed in lipid-bound apoC-1 suggests that the TMAO-stabilized secondary structure resembles the functional lipid-bound apolipoprotein conformation.

The mechanism by which the type Iα regulatory subunit (RIα) of cAMP-dependent protein kinase is localized to cell membranes is unknown. To determine if structural modification of RIα is important for membrane association, both beef skeletal muscle cytosolic RI and beef heart membrane-associated RI were characterized by electrospray ionization mass spectrometry. Total sequence coverage was 98% for both the membrane-associated and cytosolic forms of RI after digestion with AspN protease or trypsin. Sequence data indicated that membrane-associated and cytosolic forms of RI were the same RIα gene product. A single RIα phosphorylation site was identified at Ser81 located near the autoinhibitory domain of both membrane-associated and cytosolic RIα. Because both R subunit preparations were 30–40% phosphorylated, this post-translational modification could not be responsible for the membrane compartmentation of the majority of RIα. Mass spectrometry also indicated that membrane-associated RIα had a higher extent of disulfide bond formation in the amino-terminal dimerization domain. No other structural differences between cytosolic and membrane-associated RIα were detected. Consistent with these data, masses of the intact proteins were identical by LCQ mass spectrometry. Lack of detectable structural differences between membrane-associated and cytosolic RIα strongly suggests an interaction between RIα and anchoring proteins or membrane lipids as more likely mechanisms for explaining RIα membrane association in the heart.

One hundred-forty-five full-length aldehyde dehydrogenase-related sequences were aligned to determine relationships within the aldehyde dehydrogenase (ALDH) extended family. The alignment reveals only four invariant residues: two glycines, a phenylalanine involved in NAD binding, and a glutamic acid that coordinates the nicotinamide ribose in certain E-NAD binary complex crystal structures, but which may also serve as a general base for the catalytic reaction. The cysteine that provides the catalytic thiol and its closest neighbor in space, an asparagine residue, are conserved in all ALDHs with demonstrated dehydrogenase activity. Sixteen residues are conserved in at least 95% of the sequences; 12 of these cluster into seven sequence motifs conserved in almost all ALDHs. These motifs cluster around the active site of the enzyme. Phylogenetic analysis of these ALDHs indicates at least 13 ALDH families, most of which have previously been identified but not grouped separately by alignment. ALDHs cluster into two main trunks of the phylogenetic tree. The largest, the “Class 3” trunk, contains mostly substrate-specific ALDH families, as well as the class 3 ALDH family itself. The other trunk, the “Class 1/2” trunk, contains mostly variable substrate ALDH families, including the class 1 and 2 ALDH families. Divergence of the substrate-specific ALDHs occurred earlier than the division between ALDHs with broad substrate specificities. A site on the World Wide Web has also been devoted to this alignment project.

We have identified regions in poly(A) polymerases that interact with ATP. Conditions were established for efficient cross-linking of recombinant bovine and yeast poly(A) polymerases to 8-azido-ATP. Mn2+ strongly stimulated this reaction due to a 50-fold lower Ki for 8-azido-ATP in the presence of Mn2+. Mutations of the highly conserved Asp residues 113, 115, and 167, critical for metal binding in the catalytic domain of bovine poly(A) polymerase, led to a strong reduction of cross-linking efficiency, and Mn2+ no longer stimulated the reaction. Sites of 8-azido-ATP cross-linking were mapped in different poly(A) polymerases by CNBr-cleavage and analysis of tryptic peptides by mass spectroscopy. The main cross-link in Schizosaccharomyces pombe poly(A) polymerase could be assigned to the peptide DLELSDNNLLK (amino acids 167–177). Database searches with sequences surrounding the cross-link site detected significant homologies to other nucleotidyltransferase families, suggesting a conservation of the nucleotide-binding fold among these families of enzymes. Mutations in the region of the “helical turn motif” (a domain binding the triphosphate moiety of the nucleotide) and in the suspected nucleotide-binding helix of bovine poly(A) polymerase impaired ATP binding and catalysis. The results indicate that ATP is bound in part by the helical turn motif and in part by a region that may be a structural analog to the fingers domain found in many polymerases.

Here we describe the NMR conformational study of a 20-residue linear peptide designed to fold into a monomeric three-stranded antiparallel β-sheet in aqueous solution. Experimental and statistical data on amino acid β-turn and β-sheet propensities, cross-strand side-chain interactions, solubility criteria, and our previous experience with β-hairpins were considered for a rational selection of the peptide sequence. Sedimentation equilibrium measurements and NMR dilution experiments provide evidence that the peptide is monomeric. Analysis of 1H and 13C-NMR parameters of the peptide, in particular NOEs and chemical shifts, and comparison with data obtained for two 12-residue peptides encompassing the N- and C-segments of the designed sequence indicates that the 20-residue peptide folds into the expected conformation. Assuming a two-state model, the exchange kinetics between the β-sheet and the unfolded peptide molecules is in a suitable range to estimate the folding rate on the basis of the NMR linewidths of several resonances. The time constant for the coil-β-sheet transition is of the order of several microseconds in the designed peptide. Future designs based on this peptide system are expected to contribute greatly to our knowledge of the many factors involved in β-sheet formation and stability.

Direct thermodynamic and kinetic investigations of the binding of nucleotides to the nucleoside monophosphate (NMP) site of NMP kinases have not been possible so far because a spectroscopic probe was not available. By coupling a fluorescent N-methylanthraniloyl- (mant) group to the β-phosphate of CDP via a butyl linker, a CDP analogue [(Pβ)MABA-CDP] was obtained that still binds specifically to the NMP site of UmpKdicty, because the base and the ribose moieties, which are involved in specific interactions, are not modified. This allows the direct determination of binding constants for its substrates in competition experiments.

The X-ray structure of the Escherichia coli aspartate transcarbamoylase with the bisubstrate analog phosphonacetyl-l-aspartate (PALA) bound shows that PALA interacts with Lys84 from an adjacent catalytic chain. To probe the function of Lys84, site-specific mutagenesis was used to convert Lys84 to alanine, threonine, and asparagine. The K84N and K84T enzymes exhibited 0.08 and 0.29% of the activity of the wild-type enzyme, respectively. However, the K84A enzyme retained 12% of the activity of the wild-type enzyme. For each of these enzymes, the affinity for aspartate was reduced 5- to 10-fold, and the affinity for carbamoyl phosphate was reduced 10- to 30-fold. The enzymes K84N and K84T exhibited no appreciable cooperativity, whereas the K84A enzyme exhibited a Hill coefficient of 1.8. The residual cooperativity and enhanced activity of the K84A enzyme suggest that in this enzyme another mechanism functions to restore catalytic activity. Modeling studies as well as molecular dynamics simulations suggest that in the case of only the K84A enzyme, the lysine residue at position 83 can reorient into the active site and complement for the loss of Lys84. This hypothesis was tested by the creation and analysis of the K83A enzyme and a double mutant enzyme (DM) that has both Lys83 and Lys84 replaced by alanine. The DM enzyme has no cooperativity and exhibited 0.18% of wild-type activity, while the K83A enzyme exhibited 61% of wild-type activity. These data suggest that Lys84 is not only catalytically important, but is also essential for binding both substrates and creation of the high-activity, high-affinity active site. Since low-angle X-ray scattering demonstrated that the mutant enzymes can be converted to the R-structural state, the loss of cooperativity must be related to the inability of these mutant enzymes to form the high-activity, high-affinity active site characteristic of the R-functional state of the enzyme.

The extracellular portions of the chains that comprise the human type I interferon receptor, IFNAR1 and IFNAR2, have been expressed and purified as recombinant soluble His-tagged proteins, and their interactions with each other and with human interferon-β-1a (IFN-β-1a) were studied by gel filtration and by cross-linking. By gel filtration, no stable binary complexes between IFN-β-1a and IFNAR1, or between IFNAR1 and IFNAR2 were detected. However, a stable binary complex formed between IFN-β-1a and IFNAR2. Analysis of binary complex formation using various molar excesses of IFN-β-1a and IFNAR2 indicated that the complex had a 1:1 stoichiometry, and reducing SDS-PAGE of the binary complex treated with the cross-linking reagent dissucinimidyl glutarate (DSG) indicated that the major cross-linked species had an apparent M>r consistent with the sum of its two individual components. Gel filtration of a mixture of IFNAR1 and the IFN-β-1a/IFNAR2 complex indicated that the three proteins formed a stable ternary complex. Analysis of ternary complex formation using various molar excesses of IFNAR1 and the IFN-β-1a/IFNAR2 complex indicated that the ternary complex had a 1:1:1 stoichiometry, and reducing SDS-PAGE of the ternary complex treated with DSG indicated that the major cross-linked species had an apparent Mr consistent with the sum of its three individual components. We conclude that the ternary complex forms by the sequential association of IFN-β-1a with IFNAR2, followed by the association of IFNAR1 with the preformed binary complex. The ability to produce the IFN-β-1a/IFNAR2 and IFN-β-1a/IFNAR1/IFNAR2 complexes make them attractive candidates for X-ray crystallography studies aimed at determining the molecular interactions between IFN-β-1a and its receptor.

Fold assignments for proteins from the Escherichia coli genome are carried out using BASIC, a profile–profile alignment algorithm, recently tested on fold recognition benchmarks and on the Mycoplasma genitalium genome and PSI BLAST, the newest generation of the de facto standard in homology search algorithms. The fold assignments are followed by automated modeling and the resulting three-dimensional models are analyzed for possible function prediction.

Close to 30% of the proteins encoded in the E. coli genome can be recognized as homologous to a protein family with known structure. Most of these homologies (23% of the entire genome) can be recognized both by PSI BLAST and BASIC algorithms, but the latter recognizes an additional 260 homologies. Previous estimates suggested that only 10–15% of E. coli proteins can be characterized this way. This dramatic increase in the number of recognized homologies between E. coli proteins and structurally characterized protein families is partly due to the rapid increase of the database of known protein structures, but mostly it is due to the significant improvement in prediction algorithms.

Knowing protein structure adds a new dimension to our understanding of its function and the predictions presented here can be used to predict function for uncharacterized proteins. Several examples, analyzed in more detail in this paper, include the DPS protein protecting DNA from oxidative damage (predicted to be homologous to ferritin with iron ion acting as a reducing agent) and the ahpC/tsa family of proteins, which provides resistance to various oxidating agents (predicted to be homologous to glutathione peroxidase).

Recent advances in attempts to unravel the protein folding mechanism have indicated the need to identify the folding intermediates. Despite their transient nature, in a number of cases it has been possible to detect and characterize some of the equilibrium intermediates, for example, the molten globule (MG) state. The key features of the MG state are retention of substantial secondary structure of the native state, considerable loss of tertiary structure leading to increased hydrophobic exposure, and a compact structure. NMR, circular dichroism, and fluorescence spectroscopies have been most useful in characterizing such intermediates. We report here a new method for structural characterization of the MG state that involves probing the exposed hydrophobic sites with a hydrophobic photoactivable reagent—2[3H]diazofluorene. This carbene-based reagent binds to hydrophobic sites, and on photolysis covalently attaches itself to the neighboring amino acid side chains. The reagent photolabels α-lactalbumin as a function of pH (3–7.4), the labeling at neutral pH being negligible and maximal at pH 3. Chemical and proteolytic fragmentation of the photolabeled protein followed by peptide sequencing permitted identification of the labeled residues. The results obtained indicate that the sequence corresponding to B (23–34) and C (86–98) helix of the native structure are extensively labeled. The small β-domain (40–50) is poorly labeled, Val42 being the only residue that is significantly labeled. Our data, like NMR data, indicate that in the MG state of α-lactalbumin, the α-domain has a greater degree of persistent structure than the β-domain. However, unlike the NMR method, the photolabeling method is not limited by the size of the protein and can provide information on several new residues, for example, Leu115. The current method using DAF thus allows identification of stable and hydrophobic exposed regions in folding intermediates as the reagent binds and on photolysis covalently links to these regions.

Ni2+ affinity columns are widely used for protein purification, but they carry the risk that Ni2+ ions may bind to the protein, either adventitiously or at a physiologically important site. Dialysis against ethylenediaminetetraacetic acid (EDTA) is normally used to remove metal ions bound adventitiously to proteins; however, this approach does not always work. Here we report that a bacterial endonuclease, the DNase domain of colicin E9, binds Ni2+ acquired from Ni2+ affinity columns, and appears to bind [Ni(EDTA)(H2O)n]2− at low ionic strength. NMR was used to detect the presence of both Ni2+ coordinated to amino acid side chains and [Ni(EDTA)(H2O)n]2−. Dialysis against ≥0.2 M NaCl was required to remove the [Ni(EDTA)(H2O)n]2−. The NMR procedure we have used to characterize the presence of Ni2+ and [Ni(EDTA)(H2O)n]2− should be applicable to other proteins where there is the possibility of binding paramagnetic metal ions that are present to expedite protein purification. In the present case, the binding of Ni2+ seems likely to be physiologically relevant, and the NMR data complement recent X-ray crystallographic evidence concerning the number of histidine ligands to bound Ni2+.

The serine protease subtilisin BPN′ is a useful catalyst for peptide synthesis when dissolved in high concentrations of a water-miscible organic co-solvent such as N,N-dimethylformamide (DMF). However, in 50% DMF, the kcat for amide hydrolysis is two orders of magnitude lower than in aqueous solution. Surprisingly, the kcat for ester hydrolysis is unchanged in 50% DMF. To explain this alteration in activity, the structure of subtilisin 8397+1 was determined in 20, 35, and 50% (v/v) DMF to 1.8 Å resolution. In 50% DMF, the imidazole ring of His64, the central residue of the catalytic triad, has rotated approximately 180° around the Cβ-Cγ bond. Two new water molecules in the active site stabilize the rotated conformation. This rotation places His64 in an unfavorable geometry to interact with the other members of the catalytic triad, Ser221 and Asp32. NMR experiments confirm that the characteristic resonance due to the low barrier hydrogen bond between the His64 and Asp32 is absent in 50% DMF. These experiments provide a clear structural basis for the change in activity of serine proteases in organic co-solvents.

NAD+ facilitates high-yield reactivation of clostridial glutamate dehydrogenase (GDH) after unfolding in urea. The specificity of this effect has been explored by using analogues and fragments of NAD+. The adenine portion, unlike the nicotinamide portion, is important for reactivation. Alteration in the nicotinamide portion, in acetylpyridine adenine dinucleotide, has little effect, whereas loss of the 6-NH2 substitution on the adenine ring, in 6-deamino NAD, diminishes the effectiveness of the nucleotide in promoting refolding. Also ADP-ribose, lacking nicotinamide, promotes reactivation whereas NMN-phosphoribose, lacking the adenine, does not.

Of the smaller fragments, those containing an adenosine moiety, and especially those with one or more phosphate groups, impede the refolding ability of NAD+, and are able to bind to the folding intermediate though unable to facilitate refolding.

These results are interpreted in terms of the known 3D structure for clostridial glutamate dehydrogenase. It is assumed that the refolding intermediate has a more or less fully formed NAD+-binding domain but a partially disordered substrate-binding domain and linking region. Binding of NAD+ or ADP-ribose appears to impose new structural constraints that result in completion of the correct folding of the second domain, allowing association of enzyme molecules to form the native hexamer.

The structural organization of the B1 domain of streptococcal protein G (PGA) has been probed using molecular dynamics simulations, with a particular emphasis on the role of the solvent exposed Ile6 residue. In addition to the native protein (WT-PGA), three single-mutants (I6G-PGA, I6F-PGA, and I6T-PGA), one double-mutant (I6T,T53G-PGA), and three isolated peptide fragments (corresponding to the helix and the two β-hairpins) were studied in the presence of explicit water molecules. Comparative analysis of the various systems showed that the level of perturbation was directly related to the hydrophobicity and the size of the side chain of residue 6, the internal rigidity of the proteins decreasing in the order I6T-PGA > I6G-PGA > WT-PGA > I6F-PGA. The results emphasized the importance of residue 6 in controlling both the integrity of the sheet's surface and the orientation of the helix in relation to the sheet by modulation of surface/core interactions. The effects of mutations were delocalized across the structure, and glycine residues, in particular, absorbed most of the introduced strain. A qualitative structural decomposition of the native fold into elementary building-blocks was achieved using principal component analysis and mechanical response matrices. Within this framework, internal motions of the protein were described as coordinated articulations of these structural units, mutations affecting mostly the amplitude of the motions rather than the structure/location of the building-blocks. Analysis of the isolated peptidic fragments suggested that packing did not play a determinant role in defining the elementary building-blocks, but that chain topology was mostly responsible.

Cellulases are increasingly being used for industrial purposes, particularly in washing powders, yet little is known of the factors governing the stability of proteins in detergent solutions. We present a comparative analysis of the behavior of the cellulase Cel45 from Humicola insolens in the presence of the denaturant guanidinium chloride and the anionic detergent C12-LAS. Although Cel45 unfolds in GdmCl according to a simple two-state model under equilibrium conditions, it accumulates a transient intermediate during refolding. The four disulfide bonds do not contribute detectably to the stability of the native state. Cel45 is unfolded by very low concentrations of C12-LAS (1–4 mM). An analysis of 16 mutants of Cel45 shows a very weak correlation between unfolding rates in denaturant and detergent; mutants that have the same unfolding rate in GdmCl (within a factor of 1.5) vary 1,000-fold in their unfolding rates in C12-LAS. The data support a simple model for unfolding by detergent, in which the introduction of positive charges or removal of negative charges greatly increases detergent sensitivity, while interactions with the hydrophobic detergent tail contribute to a smaller extent. This implies that different detergent-mediated unfolding pathways exist, whose accessibilities depend on individual residues. Double-mutant cycles reveal that mutations in two proximal residues lead to repulsion and a destabilization greater than the sum of the individual mutations as measured by GdmCl denaturation, but they also reduce the affinity for LAS and therefore actually stabilize the protein relative to wild-type. Ligands that interact strongly with the denatured state may therefore alter the unfolding process.

To understand general aspects of stability and folding of c-type cytochromes, we have studied the folding characteristics of cytochrome c553 from Desulfovibrio vulgaris (Hildenborough). This cytochrome is structurally similar but lacks sequence homology to other heme proteins; moreover, it has an abnormally low reduction potential. Unfolding of oxidized and reduced cytochrome c553 by guanidine hydrochloride (GuHCl) was monitored by circular dichroism (CD) and Soret absorption; the same unfolding curves were obtained with both methods supporting that cytochrome c553 unfolds by an apparent two-state process. Reduced cytochrome c553 is 7(3) kJ/mol more stable than the oxidized form; accordingly, the reduction potential of unfolded cytochrome c553 is 100(20) mV more negative than that of the folded protein. In contrast to many other unfolded cytochrome c proteins, upon unfolding at pH 7.0 both oxidized and reduced heme in cytochrome c553 become high-spin. The lack of heme misligation in unfolded cytochrome c553 implies that its unfolded structure is less constrained than those of cytochromes c with low-spin, misligated hemes.

Urea-induced denaturations of RNase T1 and reduced and carboxyamidated RNase T1 (RTCAM) as a function of temperature were analyzed using the linear extrapolation method, and denaturation m values, ΔCp, ΔH, ΔS, and ΔG quantities were determined. Because both ΔCp and m values are believed to reflect the protein surface area newly exposed on denaturation, the prediction is that the ratio of m values for RNase T1 and RTCAM should equal the ΔCp ratio for the two proteins. This is not the case, for it is found that the m value of RTCAM is 1.5 times that of RNase T1, while the denaturation ΔCp's for the two proteins are identical. The paradox of why the two parameters, m and ΔCp, are not equivalent in their behavior is of importance in the interpretations of their respective molecular-level meanings. It is found that the measured denaturation ΔCp's are consistent with ΔCp's calculated on the basis of empirical relationships between the change in surface area on denaturation (ΔASA), and that the measured m value of RNase T1 agrees with m calculated from empirical data relating m to ΔASA. However, the measured m of RTCAM is so much out of line with its calculated m as to call into question the validity of always equating m with surface area newly exposed on denaturation.

The trifluoroethanol (TFE)-induced structural changes of two proteins widely used in folding experiments, bovine α-lactalbumin, and bovine pancreatic ribonuclease A, have been investigated. The experiments were performed using circular dichroism spectroscopy in the far- and near-UV region to monitor changes in the secondary and tertiary structures, respectively, and dynamic light scattering to measure the hydrodynamic dimensions and the intermolecular interactions of the proteins in different conformational states. Both proteins behave rather differently under the influence of TFE: α-lactalbumin exhibits a molten globule state at low TFE concentrations before it reaches the so-called TFE state, whereas ribonuclease A is directly transformed into the TFE state at TFE concentrations above 40% (v/v). The properties of the TFE-induced states are compared with those of equilibrium and kinetic intermediate states known from previous work to rationalize the use of TFE in yielding information about the folding of proteins. Additionally, we report on the properties of TFE/water and TFE/buffer mixtures derived from dynamic light scattering investigations under conditions used in our experiments.

A database of functional sites for proteins with known structures, SITE, is constructed and used in conjunction with a simple pattern matching program SiteMatch to evaluate possible function conservation in a recently constructed database of fold predictions for Escherichia coli proteins (Rychlewski L et al., 1999, Protein Sci 8:614–624). In this and other prediction databases, fold predictions are based on algorithms that can recognize weak sequence similarities and putatively assign new proteins into already characterized protein families. It is not clear whether such sequence similarities arise from distant homologies or general similarity of physicochemical features along the sequence. Leaving aside the important question of nature of relations within fold superfamilies, it is possible to assess possible function conservation by looking at the pattern of conservation of crucial functional residues. SITE consists of a multilevel function description based on structure annotations and structure analyses. In particular, active site residues, ligand binding residues, and patterns of hydrophobic residues on the protein surface are used to describe different functional features. SiteMatch, a simple pattern matching program, is designed to check the conservation of residues involved in protein activity in alignments generated by any alignment method. Here, this procedure is used to study conservation of functional features in alignments between protein sequences from the E. coli genome and their optimal structural templates. The optimal templates were identified and alignments taken from the database of genomic structural predictions was described in a previous publication (Rychlewski L et al., 1999, Protein Sci 8:614–624). An automated assessment of function conservation is used to analyze the relation between fold and function similarity for a large number of fold predictions. For instance, it is shown that identifying low significance predictions with a high level of functional residue conservations can be used to extend the prediction sensitivity for fold prediction methods. Over 100 new fold/function predictions in this class were obtained in the E. coli genome. At the same time, about 30% of our previous fold predictions are not confirmed as function predictions, further highlighting the problem of function divergence in fold superfamilies.

Computer analysis using profiles generated by the PSI-BLAST program identified a superfamily of proteins homologous to eukaryotic transglutaminases. The members of the new protein superfamily are found in all archaea, show a sporadic distribution among bacteria, and were detected also in eukaryotes, such as two yeast species and the nematode Caenorhabditis elegans. Sequence conservation in this superfamily primarily involves three motifs that center around conserved cysteine, histidine, and aspartate residues that form the catalytic triad in the structurally characterized transglutaminase, the human blood clotting factor XIIIa′. On the basis of the experimentally demonstrated activity of the Methanobacterium phage pseudomurein endoisopeptidase, it is proposed that many, if not all, microbial homologs of the transglutaminases are proteases and that the eukaryotic transglutaminases have evolved from an ancestral protease.