(Received for publication, December 31, 1996, and in revised form, January 27, 1997)
From the Dienst Ultrastruktuur, Vlaams Interuniversitair Instituut Biotechnologie, Vrije Universiteit Brussel, Paardenstraat 65, B-1640 Sint-Genesius-Rode, Belgium
It has been established that Tyr-42, Tyr-45, and Glu-46 take part in a structural motif that renders guanine specificity to ribonuclease T1. We report on the impact of Tyr-42, Tyr-45, and Glu-46 substitutions on the guanine specificity of RNase T1. The Y42A and E46A mutations profoundly affect substrate binding. No such effect is observed for Y45A RNase T1. From the kinetics of the Y42A/Y45A and Y42A/E46A double mutants, we conclude that these pairs of residues contribute to guanine specificity in a mutually independent way. From our results, it appears that the energetic contribution of aromatic face-to-face stacking interactions may be significant if polycyclic molecules, such as guanine, are involved.
The energetics of virtually all binding properties in proteins is the culmination of a complex set of intermolecular interactions. The individual contributions and the mutual interdependence of these interactions are currently being probed through protein engineering by many research groups. The most powerful experimental approach is to analyze the effects of single mutations on binding, turnover, or conformational stability and to compare these with the properties of proteins where multiple mutations are combined in one molecule (1-4).
The functional role and cooperative interplay between catalytic residues have been investigated in detail for a number of enzymes including subtilisin (5, 6), staphylococcal nuclease (7, 8), and ribonuclease T1 (9). In each case, the free energy barriers to substrate turnover introduced by mutations of catalytic residues are not additive in the corresponding double (or multiple) mutants. Residues involved in the process of breaking and forming covalent bonds appear to contribute to turnover in a mutually dependent way. Far less information is available on the additivity of molecular interactions involved in substrate binding. In this study, we investigate the mutual dependence of interactions at the guanine-binding site of ribonuclease T1 by protein engineering.
Ribonuclease T1 (RNase T1; EC 3.1.27.3) of the slime mold
Aspergillus oryzae (10) is the best known representative of a large family of homologous microbial ribonucleases with members in
the prokaryotic and eukaryotic worlds (11, 12). RNase T1 has a
pronounced specificity for the base guanine; kinetic studies on the
trans-esterification of dinucleoside phosphates revealed that the specificity constant
(kcat/Km) for
GpN1 substrates is
~106-fold greater than for corresponding ApN substrates
and at least 108-fold greater than for CpN and UpN
substrates (13). The three-dimensional structure of RNase T1 complexed
with the competitive inhibitor 2-GMP (14, 15) provides a structural
basis for understanding the enzyme's specificity. In the complex, the
hydrogen-bonding potential of the guanine base is completely saturated
by complementary donor/acceptor sites on the enzyme involving the
backbone atoms of Asn-43, Asn-44, and Asn-98 and the side chain
carboxyl group of Glu-46 (Fig. 1). The N(1)-H-Glu-46 O-
1,
N(2)-H-Glu-46 O-
2, and the N(2)-H-Asn-98 O hydrogen bonds have
apparent contributions of 2.7, 1.1, and 1.2 kcal/mol to the specificity
of RNase T1 for guanosine, respectively (16). Moreover, the guanine
base is stacked between the phenolic side chains of tyrosines 42 and 45 (14, 15, 17-20).
In this study, Tyr-42, Tyr-45, and Glu-46 (the only residues that interact with the guanine base through their side chains) have been replaced by alanine. The effects of the single mutations are discussed below. To measure the functional cooperativity between these residues, single and multiple mutations have been analyzed by double mutant cycles. The structural implications of the Y42A mutation have been investigated by x-ray diffraction analysis.
Oligonucleotides were bought from Eurogentec. The
3,5
-dinucleoside phosphate substrate GpU was from Sigma. Common
reagents were purchased at the highest purity available.
The overproduction of wild-type RNase T1 and the E46A mutant as secretory proteins in Escherichia coli has previously been described (16, 21). The Y42A and Y45A mutants, the corresponding double mutant, as well as the Y42A/E46A double mutant were constructed via a polymerase chain reaction-based site-directed mutagenesis technique (22). Mutations were identified by DNA sequence determination (23); the entire sequence of each mutant gene was determined to check that no additional unwanted mutations had arisen during the polymerase chain reaction steps. The wild-type and mutant enzymes used in this study are of the isoform containing a lysine at position 25; they were purified to homogeneity as described (24).
Kinetics of Dinucleotide Phosphate trans-EsterificationAll
experiments were performed at 35 °C in 50 mM imidazole,
50 mM NaCl, and 2.5 mM EDTA at pH 6.0 (ionic
strength = 0.1 M). The protein concentrations were
determined spectrophotometrically at 278 nm, where
A0.1% = 1.54 (25). For the mutants lacking one
or two tyrosine side chains, this value was recalculated according to
Pace et al. (26). The obtained A0.1%
values were 1.42 for mutants with one truncated tyrosine and 1.30 for
the mutant with two tyrosine side chains removed. The kinetic
parameters for the trans-esterification of GpU were
determined from initial velocities by measuring the absorbance increase
at 280 nm (27). GpU concentrations varied between 10 µM
and 1 mM. Reactions were started by adding enzyme to final
concentrations ranging from 3 × 1010 to 5 × 10
6 M depending on the enzyme used, except
for the Y42A/E46A mutant, for which a concentration of 397 µM had to be used. For this mutant, the second-order rate
constant (kcat/Km) for the
trans-esterification of GpU was derived from a progress
curve run overnight at a substrate concentration of 408 µM (much lower than the Km). Indeed, the Km for the single E46A mutant exceeds 5 mM (16). In experiments requiring high substrate
concentrations, 0.5-, 0.2-, or 0.1-cm path length cuvettes were used to
diminish the background absorbance. Experimental data were analyzed
with the program GraFit (28).
Single crystals were grown by vapor diffusion in
sitting drops at 20 mg/ml protein in 50 mM sodium acetate
buffer, pH 4.2, containing 0.125% (w/v) 2-GMP and 1.25% (w/v)
CaCl2 using 50% (v/v) 2-methyl-2,4-pentanediol as a
precipitant. X-ray data up to a resolution of 2.3 Å were measured on a
MAR image plate and processed using the CCP4 (Collaborative
Computational Project 4) suite of programs (29). The crystal was found
to belong to space group P212121,
with cell dimensions a = 49.61, b = 48.49, and c = 40.61 Å (see Table I).
|
Because the cell constants of the F100A RNase T1 crystals (30) are very
similar, we used the coordinates of this structure (Protein Data Bank
code pdb1bir.ent) as our initial model for molecular replacement. After
removing all non-protein atoms and the side chain of Tyr-42, rigid body
refinement was carried out using X-PLOR (31), resulting in an
R-factor of 0.316. The structure has been refined to a final
R-factor of 0.216 and an Rfree factor of 0.257 by successive sessions of stereochemically restrained least-square refinements using X-PLOR and manual model revisions using
the program O (32) on a Silicon Graphics Inc. graphics station. The
phenylalanine at position 100 and 2-GMP were fitted manually into the
corresponding difference densities after the first couple of refinement
cycles. 82 water molecules were inserted in spherical difference
densities where suitable hydrogen-bonding partners were available.
The interaction of an individual side
chain with the substrate may be analyzed by comparing the kinetics for
the wild-type enzyme with those for a mutant in which the side chain
has been truncated (2, 3). The obtained apparent binding energy is always a measure of the specificity of binding or catalysis (33). If
the mutation does not induce structural changes and allows the access
of bulk water to the cavity in the enzyme-substrate complex, the
apparent binding energy may be a crude measure of the incremental
binding energy of the interaction, i.e. the net free energy
contribution of the interaction to transfer the substrate from bulk
water to the enzyme complex. Free energy changes calculated from the
specificity constants (G =
RT
ln((kcat/Km)/(kcat/Km)group
0))
equal the apparent contribution to substrate specificity of the group
under investigation. In this study, Tyr-42, Tyr-45, and Glu-46 have
been replaced by alanines. The effects of these mutations on the
steady-state kinetic parameters are discussed below.
This paper also addresses the mutual interplay of amino acid residues
involved in substrate binding. For this purpose, we constructed the
Y42A/Y45A and Y42A/E46A double mutants and measured their kinetics. The
free energy barriers to substrate specificity introduced by the single
and multiple mutations have been analyzed by double mutant cycles (1,
4, 34). The degree to which one mutation affects the contribution of a
second mutation (quantified by the coupling term G)
measures the mutual component of the interaction energy between both
residues and the substrate, provided that none of the mutations gives
rise to a disrupted spatial arrangement of other amino acid residues
(1). Simple additivity (
G = 0) is observed when
the two residues contribute to binding/turnover in a functionally
independent way (4).
To discuss the apparent effects of mutations on the steady-state kinetics in terms of binding energy and specificity (33), it is a requisite to analyze the structural implications of these mutations (see above). In the ideal case, a mutation causes the removal of a simple interaction with no perturbation of the structure. Below, we argue that the mutations we introduced in RNase T1 are non-disruptive deletions (according to the classification of Fersht et al. (35)). Therefore, the apparent energetic contributions calculated from the effects of these mutations are genuine estimates of changes in incremental interaction energies.
The phenolic side chain of Tyr-42 forms the basis of the
guanine-binding site (Fig. 1). The aromatic ring lies on
top of a hydrophobic core involving the side chains of Phe-50, Val-79, Ile-90, and Phe-100. Because of its buried location in the
enzyme-substrate complex and the dramatic effect of the Y42A mutation
on the kinetics (see below), we were concerned that this mutation
induces major changes in the overall structure of the protein. For this
reason, we examined the structural implications of the Y42A mutation by x-ray crystallography. The complex of Y42A RNase T1 with the specific inhibitor 2-GMP has been refined to an R-factor of 0.216 using x-ray diffraction data to 2.3 Å (Table I). The
Y42A mutation does not disrupt the overall structure of the
enzyme-inhibitor complex (overall root mean square deviation = 0.349 Å). The local structural effects caused by the Y42A mutation are
summarized in Fig. 2. The apparent structural
perturbations at lysine 25 and glutamate 28, two solvent-exposed
residues, are due to alternative side chain conformations. Table
II compares the intermolecular hydrogen bonds for
wild-type and Y42A RNase T1 in complex with 2
-GMP. A water molecule
(Wat-134) occupies part of the free space available after removal of
the tyrosine side chain (Fig. 3). The water molecule
takes up the same position as the Tyr-42 O-
does in the wild-type
enzyme and forms a new hydrogen bond of 2.92 Å (Table II) with O-6 of
the base (Fig. 4). The sugar pucker in the Y42A RNase
T1·2
-GMP complex is C-2
endo (Table III),
as in the case of the wild-type complex (15).
|
|
Several crystallographic (15, 18, 19) and NMR (36) studies on RNase T1 indicate that the Tyr-45 side chain is flexible and solvent-exposed. Removal of this side chain is therefore not likely to disturb the enzyme's three-dimensional structure. Glu-46, another solvent-exposed residue in the free enzyme, swings into hydrogen bond the guanine base (see the Introduction) upon substrate binding (19). The invagination resulting from the removal of the Glu-46 side chain is probably sufficiently large to allow access of bulk water (16). All structural data corroborate the view that the Y42A, Y45A, and E46A mutations correspond to non-disruptive deletions.
Tyr-42 and Glu-46 Contribute to Substrate Binding Rather than to TurnoverTable IV lists the steady-state kinetic parameters for the trans-esterification of GpU obtained for the wild-type enzyme and for the Y42A, Y45A, and E46A mutants. The Y45A mutation has minor effects on the steady-state kinetic parameters of the enzyme, indicating that this residue does not contribute significantly to the incremental binding energy of the substrate. Consistent with this notion, the Y45F and Y45W mutations have been found to induce marginal effects on the trans-esterification kinetics of pGpC (37-39).
|
In contrast to the Y45A mutation, the Y42A and E46A deletions were found to affect the steady-state kinetic parameters of the enzyme significantly. The trans-esterification rates of Y42A and E46A are linearly proportional to the substrate concentration over the entire concentration range that is experimentally accessible. From these linear relationships, we calculated the second-order rate constant for the trans-esterification reaction (kcat/Km). We defined a minimum value of Km equal to 1 mM (16). From our data, it appears that substrate binding, rather than turnover, is impaired upon deletion of the Tyr-42 or Glu-46 side chain.
The six-membered aromatic rings of Tyr-42 and the guanine base are
involved in a parallel face-to-face stacking interaction (dihedral
angle = 12°; centroid separation = 5.0 Å). The Y42A mutation has a devastating effect (G =
5.4
kcal/mol) on the enzyme's incremental binding energy toward guanosine.
This effect is much larger than normally observed for intramolecular
aromatic-aromatic interactions between side chains of
phenylalanine, tyrosine, and tryptophan, mostly of the
perpendicular edge-to-face type (40-42). Indeed, simple aromatic rings
prefer to associate via enthalpically favorable edge-to-face
interactions of ~1.5 kcal/mol. This is in contrast to parallel
stacking, which contributes almost zero when the rings are monocyclic.
The large contribution of Tyr-42 to substrate binding suggests that
face-to-face stacking can be significant when polycyclic molecules such
as guanine are involved. The contribution of Glu-46 to substrate
binding and guanine specificity (see the Introduction) has been
discussed in detail previously (16, 43).
Studies on model
compounds indicate that hydrogen bonding and stacking interactions may
mutually reinforce each other (44). The large effects of the E46A and
Y42A mutants could theoretically be explained by such a type of
cooperativity. Hydrogen bonds with the periphery of the base (Glu-46)
might enhance a stacking interaction with this base (Tyr-42) and vice
versa. To investigate this hypothesis, we performed a double mutant
cycle analysis involving Tyr-42 and Glu-46 (Fig. 5). The
effect of the Y42A mutation was measured in the presence and absence of
the Glu-46 side chain. The coupling energy G, measured
by comparing the free energy differences corresponding to parallel
transitions of this cycle, is 0.394 kcal/mol. This is a small value
(<10%) compared with the effects of the individual mutations,
indicating that the contributions of Tyr-42 and Glu-46 to guanine
binding are mutually independent.
Crystallographic data suggest that specific recognition of guanosine by RNase T1 involves sandwich-like parallel stacking of the guanosine base between the phenolic side chains of Tyr-42 and Tyr-45 (14, 15, 17-20). Our data indicate that this view is no longer valid. The Tyr-45 phenolic side chain does not contribute to substrate binding or to substrate turnover. By constructing the Y42A/Y45A double mutant, we also investigated whether the stacking interaction between the guanine base and Tyr-42 depends on the presence of the Tyr-45 phenolic ring. The kcat/Km ratio for the double mutant is very similar to that for Y42A RNase T1 (Table IV), indicating that the strength of the Tyr-42-guanine base interaction does not depend on Tyr-45. Taken together, all data show that the Tyr-45 side chain, often referred to as the lid of the guanine-binding site (15, 38, 39), does not contribute to the incremental binding energy for guanosine.
From our double mutant cycles, it appears that the Tyr-42/Tyr-45 and Tyr-42/Glu-46 pairs contribute to substrate binding in a mutually independent way. This is consistent with a large data base for free energy changes that result when single mutants are combined (4). Indeed, if two residues do not interact with each other by direct or indirect contact, the sum of the free energy changes derived from the single mutations is nearly always equal to the free energy change measured in the multiple mutant. One major exception where such simple additivity does not apply exists for catalytic residues which, instead, act in a highly cooperative way (5, 7-9).
Evolutionary ImplicationsStructural conservation of particular enzyme residues among homologous family members is indicative of their structural or functional importance (45). Among the guanine-specific members of the homologous family of microbial ribonucleases, Tyr-42, Tyr-45, and Glu-46 are conserved (11, 12). Both tyrosines are consistently observed at equivalent positions in all eukaryotic members. In the prokaryotic ribonucleases, a phenylalanine occupies position 42 (RNase T1 numbering), and an arginine is found at position 45. Glu-46 is found throughout the family. The conservation of an aromatic ring (Tyr, Phe) at position 42 (RNase T1 numbering) is compatible with the observation that the guanine base interacts with the phenyl part of the side chain. The tyrosine hydroxyl group makes no direct intermolecular contacts with the guanine base. The strict conservation of Glu-46 is easily explained in terms of two specific hydrogen bonds with the guanidinium part of the guanine base. It has been established that the Glu-46 carboxylate takes part in an invariant structural motif that renders guanine specificity (46).
The fact that Tyr-45 is conserved but does not contribute to substrate binding or to catalysis indicates that it fulfills another function. We can only speculate on its role. The Tyr-45 side chain may be involved in the folding and stability of RNase T1. It may prevent other nucleotides from binding the catalytic site. Comparison of the expression of the authentic RNase T1 gene in Saccharomyces cerevisiae and A. oryzae (47) suggests the existence of an intracellular inhibitor for the enzyme. If A. oryzae has a specific inhibitor for its ribonuclease, Tyr-45 may be essential in RNase inhibitor recognition. Arg-59 (barnase numbering), the equivalent of Tyr-45 in barnase, interacts with five barstar residues and contributes ~5 kcal/mol to the intermolecular interaction energy of the barnase-barstar complex (48, 49).
We thank Remy Loris, Rex Palmer, and Andrew Hemmings for assistance during x-ray data collection.