J. Mol. Biol. (1990) 214, 327-335
Structural Consequences of the Replacement of Glu239 by Gln in the Catalytic Chain of Escherichia coli Aspartate Transcarbamylase Patrick Taut’,
‘Institut
Patrice Vachette’, Steven A. Middleton and Evan R. Kantrowitz3
d’Enzy mologie,
‘Laboratoire Bdtiment
C.N.R.S.,
pour I’Utilisation 209d, Universite’
3Department
(Received
91190, Gij Sur Yvette, France
du Rayonnement Electromagne’tique Paris-Sud, 91405 Orsay, France
of Chemistry, Boston College, Chestnut Hill MA 02167. U.S.A. 9 May
1989; accepted 5 March
1990)
Low-angle X-ray scattering in solution has been used to probe the quaternary structure of a mutant version of Escherichia coli aspartate transcarbamylase in which Glu239 of the by site-directed mutagenesis. X-ray catalytic chain was replaced by glutamine crystallographic studies of the wild-type enzyme have shown that one set of intersubunit interactions involving Glu239 are lost, and are replaced by another set of intrachain interactions when the enzyme undergoes the allosteric transition from the T to the R state. Functional analysis of the mutant enzyme with glutamine in place of Glu239 indicates that homotropic co-operativity is lost without altering the maximal specific activity. The radius of gyration of the unligated mutant enzyme is larger than the unligated wild-type, indicating an alteration in quaternary structure of the mutant. However, the radius of gyration of the mutant enzyme in the presence of N-(phosphonoacetyl)-L-aspartate (PALA) is identical with the value for the wild-type enzyme in the presence of PALA. X-ray scattering at larger angles indicates that the mutant enzyme is in a new structural state different from the wild-type T and R structures. The scattering pattern in the presence of saturating concentrations of PALA is identical with that of the wild-type R structure. Saturating concentrations of carbamyl phosphate alone are sufficient to convert most of the mutant enzyme to the R structure, in the absence of aspartate. CTP shifts the scattering pattern of the mutant enzyme in the presence of saturating carbamyl phosphate towards the scattering curve of the unligated enzyme, but CTP has no effect on the scattering curve in the absence of carbamyl phosphate or in the presence of subsaturating PALA. However, in the presence of subsaturating PALA, ATP causes a strong shift towards the R structure. Neither ATP nor CTP has any effect on the activity of the mutant enzyme. These data suggest that the replacement of Glu239 by glutamine results in a new quaternary structure. These data also explain, on a structural basis, why co-operativity is lost in this mutant enzyme.
trimers of 34,000 M, catalytic chains carrying the active sites, and three dimers of 17,000 MT regulaAspartate transcarbamylase (EC 2.1.3.2) from tory chains carrying the effector binding sites Escheriehia coli catalyzes the first step of pyrimidine (for reviews, see Gerhart, 1970; Jacobson & Stark, biosynthesis, the condensation of carbamyl phos1973; Schachman, 1974; Kantrowitz et al., 1980a,b; phate and L-aspartate to form N-carbamyl-LKantrowitz & Lipscomb, 1988). aspart’ate. The E. coli enzyme shows homotropic coTwo functional and structural states of the operativity for aspartate, is feed-back inhibited by enzyme have been well characterized by a variety of CTP and activated by ATP. This extensively techniques; a less active T state, stabilized by CTP studied allosteric enzyme is composed of two and a more active R state stabilized by the 327
1. Introduction
0022-2836/90/130327-09
$03.04/O
0 1990 Academic
Press Limited
328
P. Taut et al.
Figure 1. Stereo view of the Cl-C4 interface of asp&ate transcarbamylase in the T (Kim et al., 1987) and R states (Krause et al., 1987). The Cl and C4 catalytic chains are in different catalytic subunits of the holoenzyme. In the T-state (top), the 240s loops of Cl and C4 are side by side, allowing the intersubunit interaction between Glu239 with Tyr165. In the R-state (bottom), the 240s loop of Cl and C4 are on top of one another. The intersubunit link between Glu239 and Tyr165 is lost, and is replacedby an intrechain link betweenGlu239and both Lys164 and Tyr165 that helpsto stabilize the position of the 240s loop in this position. Residue numbers followed by A correspond to the Cl catalytic chain and the residue numbers followed by B correspond to the C4 catalytic chain.
substrates or their analogs, for instance N-(phosphonoacetyl)-L-aspartate (PALAT). The crystal structure of the enzyme in both forms has been determined by Lipscomb & co-workers (Kim et al., 1987; Krause et al., 1987). In the transition from T to R, the enzyme expands by 12 -J! (1 A =@I nm) along the 3-fold axis while subunits and domains undergo rotational movements. This quaternary rearrangement is accompanied by tertiary changes in both the catalytic and regulatory chains. In particular, the loop at residues 230 to 245 of the catalytic chain moves significantly in the process, establishing two different sets of intra- and intermolecular bonds that contribute to the stabilization of each structure (seeFig. 1). This suggeststhat the 230-245 loop plays an important role in the regulatabbreviations used: PALA, N-(phosphonoacetyl)CP, carbamyl phosphate; Glu239-+Gln, the mutant enzyme with Gln in place of Glu239 in the catalytic chain of aspartate transcarbamylase; [S],.,, the aspartate concentration at half the maximal observed specific activity.
tory properties of the enzyme. To test this hypothesis, site-directed mutagenesis was used to modify residues within the loop. Specifically, Glu239 of the catalytic chain was selected for replacement, since it is involved in different sets of interactions in the two structures: in the T state, Glu239 of Cl$ interacts with both Lys164 and Tyr165 of C4 (belonging to the other catalytic trimer: Kim et al., 1987); it is also involved in an intrachain hydrogen bond with Asp236, which could contribute to the stabilization of the short stretch of helix H9 (E. Gouaux, personal communication). In the R state, the two intersubunit interactions are disrupted and replaced by intrachain bonds with Lys164 and Tyr165 of Cl; the intrachain link to Aap236 is broken. Glutamine was substituted for glutamic acid at position 239 and the functional consequences of the mutation were determined. The mutant enzyme shows a
L-aapartate;
$R and C, followed by a number, e.g. Rl, C4, refers to a particular polypeptide chain in aspartate transcarbamylase as specified in Fig. 6 of Honzatko et al. (1982).
Structural
Consequences of a Mutation
complete loss of co-operativity and a decreased [S],., for aspartate along with full retention of catalytic activity (Ladjimi 6 Kantrowitz, 1988). This suggested that either the T state of the mutant enzyme or the T=R equilibrium has been altered by the amino acid substitution. An alteration in the structure of the Glu239+Gln enzyme is suggested by preliminary crystallographic data on the unligated enzyme. These data itidicate an elongation of the unit, cell along the axis that corresponds to the molecular S-fold axis (Kantrowitz & Lipscomb, 1988). The solution X-ray scattering curve of unligated wild-type enzyme (T structure) shows secondary minima and maxima characteristic of its quaternary structure. Upon addition of substrates, it undergoes major changes reflecting the transition from T to R (Moody et al., 1979) and constitutes a sensitive and specific probe of the quaternary structure of the enzyme as was demonstrated by HervB et al. (1985). We present here a study of the structural consequences of the replacement of Glu239 by glutamine as determined by solution X-ray scattering.
2. Materials
and Methods
(a) Materials ATP, CTP, carbamyl phosphate, N-cerbamyl-L-aspartate, L-aspartate and Tris were purchased from Sigma Chemical Co. The carbamyl phosphate was purified by precipitation from 50% (v/v) ethanol, and stored desiccated at -20°C (Gerhart & Pardee, 1962). The replacement of Glu by Gln at position 239 in the catalytic chain of aspartate transcarbamylase was performed by site-specific mutagenesis as described (Ladjimi & Kantrowitz, 1988). Wild-type and Glu239+Gln aspartate transcarbamylases were isolated as described (Nowlan & Kantrowitz, 1985), from E. coli strain EK1104 [F- ara, thi, Apro-kc, ApyrB, pyrF*, rpsL], containing the plasmid pEK2 (Smith et al., 1986) or pEK56, respectively. (b) Methods The transcarbamylaae activity was measured at 25”~: by either a calorimetric (Pastra-Landis et al., 1981) or pHstat method (Wu & Hammes, 1973). pH-stat assays were carried out with a Radiometer TTTSO titrator and an ABUBO autoburette. All calorimetric assays were performed in duplicate and the data points shown in the Figures are the average. Enzyme samples for the X-ray scattering experiments were prepared from a stock solution in 50 mM-Tris-borate buffer (pH %3), @l mM-EDTA, @l mM-dithioerythritol as described (HervB et al., 1985). X-ray scattering curves were recorded on the small-angle scattering instrument D24 using synchrotron radiation at LURE-DCI, Orsay. The instrument has been described (Depautex et al., 1987), as well as the data acquisition system (Bordas et al., 1980). The scattering curves of dilute enzyme solutions (4 to 8 mg/ml) were recorded in the 8 range @0012 to O-012 A-‘. The scattering parameter s is defined, s= 2 sin Ojn, where 20 is the angle through which the X-rays are scattered, and 1 is the wavelength (1.608 A, K-edge of Co). Data were recorded in the angular range @Ol to
in Aspartate
329
Transcarbamylase
I 0.2
31
0
I 0.4
I 0.6
( 8
lo4xs2&*)
Figure 2. Guinier plots. The curves have been arbitrarily shifted along the vertical axis for the sake of clarity. From top to bottom: (0) unligated wild-type enzyme; (0) unligated Glu239+Gln enzyme; (0) wildtype enzyme in the presence of 2 molar excess PALA; ( q ) Glu239-+Gln enzyme in the presence of 2 molar excess PALA.
@05 8-l using 100 mg/ml by Hervb et al. (1985).
enzyme solutions
as reported
3. Results (a) Radius of gyration The small-angle pattern of dilute solutions of the wild-type and of the Glu239-*Gln enzyme were recorded in the absence of substrates and in the presence of a twofold molar excess of PALA (two PALA molecules/active site). Due to the high aflinity of PALA for the active site, this concentration is enough for all active sites to be occupied. The value of the radius of gyration was derived from a Guinier analysis (Fig. 2 and Table 1: and seeGuinier & Fournet, 1955). The radius of gyration of the unligated Glu239+Gln enzyme is only slightly smaller than its value in the presence of PALA,
Table Radii of gyration
Wild-type Glu239+Gln
for
1
the wild-type and Glu239+Gln enzymes
No substrates (4
Saturating PALA (4
46.6 +@3 485*@3
493kO.2 491 kO.2
330
P. Taut et al. 4.3
r
4.12
3.94 z In G B 1 3.76
3.58
3.4
I 0.01
I 0.018
I 0.026
I 0,034
s- 2 sin B/X
I 0.042
3.4 0.65
(L-9
Figure 3. Solution X-ray scattering spectra of the wildtype and Glu239-*Gln enzyme. (0) Unligated wild-type enzyme; (0) unligated Glu239-+Gln enzyme; (0) wildtype enzyme in the presence of 2 molar excess PALA; ( q ) Glu239-tGln enzyme in the presence of 2 molar excess PALA.
which is the same for both the mutant and wildtype enzymes within error bars of 1 standard deviation. (b) X-ray scattering spectra of unligated and
PALA
saturated enzymes
Figure 3 shows the X-ray scattering curve of the Glu239-tGln and wild-type enzymes in the absence and presence of saturating concentrations of PALA. The curves for the wild-type and the Glu239+Gln enzymes in the presence of PALA are indistinguishable. On the contrary, the curve for the unligated Glu239+Gln enzyme is significantly different from the unligated wild-type curve. Indeed, the intensities in the region of the first subsidiary minimum and maximum are intermediate between the intensities of the two curves of the wild-type T curve and R curve, while the position of the first maximum is much closer to that of the R curve than to that of the T curve. Could the proposed destabilization of the T state caused by the mutation lead to a shift in an equilibrium between the two extreme quaternary structures T and R, so that a mixture of the two structures is observed, even in the absence of substrates? In such a case, the scattering curve would be a linear combination of the wild-type T and R curves. Attempts at fitting the experimental
o-01
I 0.018
I 0.026
I 0.034
I 0.042
0.05
S=2 sin B/X (&-I)
Figure 4. Solution X-ray scattering spectrum of’ the Glu239-+Gln enzyme and linear combinations of wildenzyme extreme patterns. (0) IJnligated type Glu239dGln enzyme; (0) 55% R, 45% T: (0) 3096 R. 70% T.
data (0 in Fig. 3) are shown in Figure 4. The hypothetical curve corresponding to 55% R and 45% T fits well around the first maximum but not near the first minimum nor around the second minimulh and maximum. This last part of the curve is well approximated by the hypothetical curve composed of 30%R and 70%T, but this linear combination does not fit the region near the first minimum and maximum. The inability to fit the scattering curve of the Glu239+Gln enzyme in the absence of PALA to a linear combination of the wild-type T and R state scattering curves indicates conclusively that, the solution of the unligated mutant, is not a mixture of T and R structures. Therefore, the unligated Glu239-+Gln enzyme is in a new quaternary structure, different from both the wild-type T and R structures.
(c) Effect of carbamyl phosphate The reaction catalyzed by aspartate transcarbamylase involves an ordered mechanism in which carbamyl phosphate binds first and induces a conformational change that promotes the binding of aspartate or its analogs (Porter et al., 1969; Wedler & Gasser, 1974; Hsuanyu & Wedler, 1987). The scattering pattern of the Glu239-+Gln enzyme was recorded in the presence of a saturating concentration of carbamyl phosphate (25 mM). As seen in Figure 5, the scattering pattern of the mutant
Structural
3.41 0.01
I 0~018
I 0.026 s=2sin
Consequences of a Mutation
I 0.034 8/X
I 0.042
0.05
(ii-‘)
Figure 5. Solution X-ray scattering spectra of the enzyme. (0) Unligated Glu239+Cln enzyme; (a) Glu239-*Gln enzyme in the presenceof 2 molar excessPALA; (0) Glu239-tGln enzyme in the presenceof 25 mM-carbamyl phosphate;(a) Glu239-+Gln enzyme in the presence of 25 mM-carbamyl phosphate$10 mM-CTP.
Glu239-*Gln
enzyme is very similar to the R curve obtained in the presence of saturating PALA, showing that most of the enzyme is in the R structure. The experimental curve can best be fit with a 10~o-90~/o linear combination of the unligated and PALA saturated Glu239-rGln curves, respectively. This is very different from the case of the wild-type enzyme, where the addition of saturating amounts of carbamyl phosphate has almost no effect on the scattering pattern (P.T. and P.V., unpublished results). Subsequent addition of succinate, which shifts the scattering pattern of the wild-type enzyme from T to R (HervB et al., 1985), has no detectable effect with the Glu239-+Gln enzyme (data not shown). This is to be expected, since carbamyl phosphate alone can shift the curve all the way to the R state. (d) EJect of ATP and CTP (i) In the absenceof substrates The Glu239+Gln enzyme is unable to reach the T structure in the absence of substrates. Could the addition of CTP, which is known to stabilize the T structure of the wild-type enzyme, shift the curve of the mutant enzyme further toward the T structure? The scattering curve of the Glu239-+Gln enzyme, recorded in the absence and presence of 10 mi%-CTP
in Aspartate
0 .oi
331
Transcarbamylase
0~018
O-026
0.034
0.042
0~05
s=2 sin B/X (3)
Figure 6. Solution X-ray scattering spectra of the Glu239-tGln enzyme. (0) Unligated Glu239-rGln enzyme; (0) Glu239+Gln enzyme in the presenceof 10mM-CTP.
is shown in Figure 6. The addition of CTP causesno significant change in the scattering curve, and certainly not the decrease in intensity around the first maximum associated with a shift towards the T structure. (ii) In the presenceof saturating carbamyl phosphate For the Glu239+Gln enzyme, carbamyl phosphate alone is able to shift the quaternary structure almost all the way to the R structure. Could the structural alteration induced by the binding of carbamyl phosphate be reversed by addition of CTP? Figure 5 shows the scattering curve obtained in the presence of 25 m&I-carbamyl phosphate and 10 mM-CTP. This scattering curve is intermediate between the curve obtained in the presence of saturating carbamyl phosphate and the curve obtained in the absence of CTP. So, CTP can, at least partially, reverse the effect of carbamyl phosphate and shift the scattering curve towards the unligated state of the Glu239dGln enzyme. (iii) In the presenceof subsaturating PALA The addition of subsaturating amounts of PALA shifts the scattering curve towards the R state, as it does with the wild-type enzyme. In the latter case, it has been shown that further addition of the activator ATP has no detectable effect on the scattering pattern, while the inhibitor CTP causes a slight but significant shift towards the unligated pattern (HervB et al., 1985). What about the Glu239+Gln enzyme? As seen in Figure 7, CTP
332
P. Taut et al. 4.3
-2 u) z
L”
3.94
I ,5 x C ., B I.0 a, ? “0 0, K 0.5
3.76
0
FT -I
0.2
0.4
0.6
0.8
I.0
[NTP](mM)
3.58
3.4
L 0.01
I
I
I
I
0.018
0.026
0.034
0,042
s=2 sin 8/X
(
Figure 8. Influence of the effecters ATP and CTY on the act,ivity of the wild-type and the Glu239+Gln enzymes. The reactions were carried out at 25°C in 50 mM-Tris-acetate buffer (pH 8.3) in the presence of saturating levels of carbamyl phosphate (4.8 mM). For each enzyme? the aspartate concentration was at 0.5 x [S],., for aspartate. ATP effect on the (0) wild-t,ype and (0) Glu239-+Gln enzymes. CTP effect on the (m) wild-type and (0) Gln239+Gln enzymes.
(ii-l)
Figure 7. Solution X-ray scattering spectra of the Glu239+Gln enzyme. (0) In the presenceof 018 molecule of PALA/active site; (0) in the presenceof @18 moleculeof PALA/active site+ 10mM-CTP; (6) in the presence of @18 molecule of PALA/active site + 10 mM-ATP.
shows no detectable effect, while ATP causes a strong shift towards the R state. More precisely, the distribution shifts from about 30% R structure to about 90% R structure. The effects of ATP and CTP on the Glu239+Gln enzyme are in complete contrast to the scattering curves obtained for the wild-type enzyme. (e) Functional behavior of ATPICTP Glu239+Gln enzyme
with the
In order to correlate the structural changes induced by the allosteric effecters with the function of the Glu239+Gln enzyme, ATP and CTP saturation curves were determined. As seen in Figure 8, neither ATP nor CTP alter the activity of the Glu239-tGln enzyme, even at high concentrations. This behavior is in contrast to the activation of the wild-type enzyme observed with ATP and the inhibition of the wild-type enzyme observed with CTP. 4. Discussion The functional analysis of the Glu239-+Gln enzyme showed that it is devoid of co-operativity with respect to aspartate. Furthermore, its activity was inhibited by low concentrations of PALA at low and saturating concentrations of aspartate carbamyl phosphate (Ladjimi & Kantrowitz, 1988),
suggesting that the active sites of the Glu239+Gln enzyme are in the active R form, even a,t low concentrations of aspartate (with saturating carbamyl phosphate). Therefore, kinetically, the Glu239+Gln enzyme appears to be in an R-like state. The determination of the radius of gyration of the Glu239-rGln enzyme provides direct structural evidence to support this conclusion. The value of the radius of gyration of unligated Glu239--+Gln enzyme is close to the value obtained for the Glu239+Gln enzyme saturated with PALA. For the mutant enzyme, the binding of PALA seems t’o cause almost no structural transition. The scattering curves at larger angles show that the situation described above is incomplete. The Glu239-+Gln enzyme, in the absence of all ligands, is in neither the wild-type T nor R structure, since the scattering curve of the mutant is different from the two curves of the wild-type enzyme (seeFig. 3). How can this be reconciled with the radius of gyration data indicating that the mutant enzyme is almost entirely in the R structure? The contradiction is only apparent. Indeed, the position of the first subsidiary maximum is correlated with the distance between the two catalytic trimers along the 3-fold axis, on each side of the central cavity. This position on the curve of t’he unligated mutant is close to the position on the R curve, showing that the two catalytic trimers are almost as far apart as in the wild-type R structure. Since the catalytic subunits account for two-thirds of the molecular mass of the enzyme, their distance determines the value of the radius of gyration. Thus, the R, values are consistent with the scattering curves at larger angles. Changes in the scattering pattern of aspartat,e transcarbamylase at s-values lower than 0.05 A ’
Structural
Consequences of a Mutation
have been shown to be specifically sensitive to quaternary structure changes (Altman et al., 1982; Herve et al., 1985). The scattering curve of the unligated Glu239+Gln enzyme differs from the scattering curves of both the unligated (T) and PALA-saturated (R) structures of the wild-type enzyme. This suggests that either the solution of unligated Glu239+Gln enzyme contains a mixture of molecules in T and R structures or that the mutant enzyme exists in an entirely new structural state. The fact that this curve can not be fit to any linear combination of the wild-type T and R scattering curves strongly supports the second hypothesis. Furthermore, the radius of gyration of the mutant; enzyme along with the position of the first subsidiary maximum indicates that the structure of the unligated Glu239+Gln enzyme is more similar to the R structure than to the T structure of the wild-type enzyme with an elongation along the 3-fold axis. This structure will be hereinafter designated as the R’ structure. The fact that, in the absence of ligands, the Glu239+Gln enzyme does not exist in a wild-type like T structure, suggests that the Glu239 interactions are essential for the stabilit#y of the T structure. Other mutations have been made in the Cl-C4 interface, Tyrl65+Ser (Robey & Schachman, 1984) and Tyrl65--+Phe (Wales et al., 1988), which do not produce enzymes with R-like kinetic properties. The catalytic subunit of the Tyrl65-Ser enzyme has reduced activity and affinity, while the TyrlSEi-tPhe enzyme has a much lower affinity for aspartate. A plausible interpretation is that the Glu239 to Lysl64 intersubunit link is essential for the stability of the T state, together with the (~1~239 to Asp236 intrachain bond, while the Glu239 to Tyrl65 intrasubunit bond contributes to the stability of the R-state. The Glu239-+Gln enzyme is not the first case to be reported of a mutant version of aspartate transcarbamylase adopting a new quaternary structure in the absence of ligands. The pAR5 enzyme is a mutated form of aspartate transcarbamylase in which residues 145 to 153 at the C terminus of the R chain are replaced by a new sequence of six residues (Ladjimi et al., 1985) that contribute to the Rl-C4 interfa,ce. The quaternary structure of the unligated pAR5 enzyme has been shown to be different from both the T and R structures of the wild-type enzyme (Cherfils et al., 1987). It is also different from the structure of the unligated Glu239+Gln enzyme; more precisely, the scattering pattern of the unligated pAR5 enzyme is much closer to the wild-type T curve than is the curve of the Glu239+Gln enzyme. In the presence of saturating amounts of PALA, the scattering curves of the wild-type and of the Glu239+Gln enzyme are indistinguishable. So, upon substrate binding, the mutant enzyme undergoes the R’-+R structural transition. Are there other, intermediary quaternary structures accessible to the Glu239+Gln enzyme? All the scattering curves of the Glu239+Gln enzyme presented here
in Aspartate
Transcarbamylase
333
can be satisfactorily fit by a linear combination of the two extreme scattering patterns of R and R’. So, the structural behavior of the mutant enzyme under the variety of conditions investigated can be accounted for with only two quaternary structures, R and R’. Similar observations were made for two other mutant enzymes, Tyr240-+Phe enzyme (Cherfils et aE., 1989) and the pAR5 enzyme. The addition of saturating carbamyl phosphate does not cause a significant change in the scattering pattern of the wild-type enzyme, while it causes considerable modifications in the case of the Glu239+Gln enzyme. In fact, carbamyl phosphate alone can convert most of the Glu239-+Gln enzyme to the R structure. The binding of carbamyl phosphate is known to occur first and to cause tertiary conformational changes in the carbamyl phosphate domain (Collins & Stark, 1969; Griffin et al., 1972) that help to create the aspartate-binding site. For the wild-type enzyme, the intersubunit interactions involving Glu239 keep the enzyme in the T structure even when carbamyl phosphate binds. It is binding of aspartate after the binding of carbamyl phosphate that triggers the allosteric quaternary conformational change (Kantrowitz & Lipscomb, 1988). In the case of the Glu239+Gln enzyme, the (presumably) same conformational change in the carbamyl phosphate domain results in the conversion of most (approx. 90%) of the enzyme to the R structure. In other words, the value of the allosteric constant L, as defined in the Monod, Wyman & Changeux (Monod et aZ., 1965), of the enzyme-CP complex is about 61; this very low value explains the absence of co-operativity observed with aspartate (Ladjimi & Kantrowitz, 1988). A similar, though weaker, effect of CP was observed with the pAR5 enzyme. Neither ATP nor CTP alter the activity of the Glu239+Gln enzyme (see Fig. 8), although both are able to alter the structure of the mutant enzyme. In the absence of carbamyl phosphate and aspartate, CTP is unable to shift the structure towards the wild-type T state. However, in the presence of carbamyl phosphate, which itself has converted most of the mutant enzyme into the R structure, CTP is able to promote a return towards the R structure, bringing the curve from 907; R to about 45% R. This first proves that CTP does bind to the Glu239-+Gln enzyme; the absence of effect of CTP on the unligated Glu239+Gln enzyme can not thus be explained by the absence of binding. These data suggest also that carbamyl phosphate is not very effective in stabilizing the R structure, since CTP can, at least partially, reverse its effect. However, they can not be correlated directly with the activity measurements, since the latter were performed in the presence of a significant amount of aspartate, which could have stabilized the R structure. To test this possibility, experiments were performed at subsaturating concentrations of CP; they show neither inhibition by CTP nor activation by ATP (data not shown). This suggests that the structural conversion of the enzyme population from mostly R
P. Taut et al.
334
to half R half R’ by CTP is insufficient to inhibit the activity of the enzyme. A possible interpretation could be that the mutant enzyme in the R’ structure has the same affinity for aspartate as in the R structure. In fact, we have no direct indication of the affinity of the unligated Glu239+Gln enzyme for aspartate, since the only available data pertain to the ATCase-CP complex, which is mostly in the R structure. l?or the wild-type enzyme, subsaturating PALA has been used to shift the TtsR equilibrium towards the R structure. Under these conditions, ATP has no detectable influence on the scattering curve of the wild-type enzyme, while CTP has only a small effect (HervB et al., 1985). At subsaturating concentrations of PALA, CTP does not shift the scattering curve of the Glu239--+Gln mutant toward the R’ structure. This result suggests that PALA binding to the mutant enzyme stabilizes the R structure sufficiently to prevent the mutant enzyme from reverting back t)o the R’ structure upon the addition of CTP. On the other hand, ATP is able to shift the scattering curve, bringing a population about equally distributed between R’ structure and R structure mostly into the R structure. Together with the absence of effect observed with CTP under the same conditions, this supports the view that ATP and CTP operate in different ways. The role of Glu239 for co-operativity in aspartate transcarbamylase and in maintaining the enzyme in the T structure has been suggested: Cl-C4 intersubunit interactions with Lysl64 and Tyr165 (Kantrowitz & Lipscomb, 1988) and an intrachain hydrogen bond with Asp236 (E. Gouaux, personal communication). The scattering data reported here substantiate this interpretation. The mutant enzyme is unable to adopt the T structure. Furthermore, the Cl-C4 interface in the R’ structure that it adopts instead is less stable than in the T structure of the wild-type enzyme. In order to establish the importance of the R’ structural state observed in the Glu239-+Gln enzyme for co-operativity of aspartate transcarbamylase, additional functional and structural studies are in progress. After this paper was submitted, the crystallographic enzyme
structure
of the
unligated
Glu239-+Gln
was reported by Gouaux et al. (1989). The quaternary structure of the enzyme appears to be different from both the T and R structures. However, this structure seemsto be much too close to the T structure to be compatible with our results in solution. This discrepancy, which could be due either to the constraints from the crystal lattice or to the low pH (pH 58, close to t,he p1 of the enzyme) at which crystals are grown, will be further investigated by calculating scattering curves from crystallographic co-ordinates.
It is a pleasure to acknowledge the help from the technical staff from LURE-DC1 and from the computing center of LURE. We thank E. Gouaux for attracting our attention towards the role of Asp236, Wei Xu for help with some of the activity measurements, W. N. Lipscomb
for providing X-ray co-ordinates of the enzyme and Jo#l Janin for critical reading of’ the manuscript. E.R.K. acknowledges support from USPHS research grants GM-26237 and DK- 1429.
References Altman,
R. B., Ladner,
J. E. & Lipscomb, W. N. (1982). c’ommun. 108, 592-595. Bordas, cJ., Koch, M. H. J., Clout, P. N., Dorrington. K:.. Boulin. C. & Gabriel, A. (1980). J. Whys., E: 8ci. In&rum. 13. 938-944. Cherfils. J.. Vachette, P., Taut, P. dt Janin. J. (1987). Biochem.
EMBO
Biophys.
Res.
J. 6, 2843-2841.
Cherfils, J.. Sweet, R. M.. Middleton, S. A., Kantrowitz. E. R., Taut, P. & Vachette, P. (1989). PEBS Letters, 247,
361-366.
Collins, K. D. & Stark. U. R. (1969). J. Hiol. Chrvn. 244. 1869~-1877. Depautex, C., Desvignes, C.. Leboucher. P.. Lemonnier, 81.. Dagneaux, D.. Benoit, ,J. P. & Vachette. P. (1987). LURE Annual Report. Laboratoire pour I’Utilisation du RayonnementElectromagnBtique,Paris. Gerhart. ,J. C. (1970). CWT. Top. Cell. Reg. 2, 275~3%. Gerhart. ,J. C. & Pardee. A. H. (1962). J. Biol. C’hevn. 237, 891.-896. Gouaux, J. E., Stevens, R. C.. Ke, H. M. & Lipscomb. W. N. (1989). Proc. ,Vat. Acad. Ski., fT.8.A. 86. 8212-8216. Griffin, J. H., Rosenbusch. J. P.. Weber. K. K. b Blout. E. R. (1972). J. Biol. Chem. 247, 6482%6490. Guinier, -4. & Fournet, G. (1955). Srn.all-Angle Scattering of X-rays, Wiley. New York. HervB, U., Moody. M. F.. Taut, P., Vachette. 1’. Cyr. Jones. P. T. (1985). J. iuol. Riol. 185, 189-199. Honzatko. R. B., Crawford, J. L., Monaco, H. L., Ladner. J. E.. Edwards, B. F. P.. Evans. D. It.. Warren. 8. 0.. Wiley, D. (‘., Ladner, R. C. & Lipscomb. W. N. (1982). J. Mol. Biol. 160, 219-263. Hsuanyu, Y. & Wedler. F. (‘. (1987). Arch. Nioch.rm Biophys
259,
,Jacobson,
316--330.
G. R. & St,ark.
G. R. (1973).
Cnzyvnes
9.
225308.
Kantrowitz, E. R. bz Lipscomb, W. N. (1988). Science. 241, 669-674. Kantrowitz, E. R., Pastra-Landis. S. (‘. & Lipscomb. W. N. (1980a). Trend4 Biochem. Sci. 5. 124-128. Kantrowitz. E. R.; Pastra-Landis, S. C. & Lipscomb. W. N. (198Ob). 7’rendB B&hem. Sci. 5, 150--163. Kim. K. H.. Pan. Z., Honzatko, R. B.. Ke. H. M. & Lipscomb, W. N. (1987). ,/. Mol. Biol. 196, 853-875. Krause, K. I,.. Volz, K. W. 6t Lipscomb, W. N. (1987). J. Mol. Biol. 193, 527ki53. Ladjimi, M. & Kantrowitz. E:. R. (1988). Biochemistry, 27, 276-283.
Ladjimi, M.. Ghklis, (1.. Feller, A.. Cunin, R.. Ulansdorff, S.. PiBrard, A. & HervB. G. (1985). J. Mol. Riol. 186. 715. 724. Monod, ,J.. Wyman. ,J. & (Ihangeux,
Structural Consequences of the Replacement of Glu239 by Gln in the Catalytic Chain of Escherichia coli Aspartate Transcarbamylase Patrick Taut’,
‘Institut
Patrice Vachette’, Steven A. Middleton and Evan R. Kantrowitz3
d’Enzy mologie,
‘Laboratoire Bdtiment
C.N.R.S.,
pour I’Utilisation 209d, Universite’
3Department
(Received
91190, Gij Sur Yvette, France
du Rayonnement Electromagne’tique Paris-Sud, 91405 Orsay, France
of Chemistry, Boston College, Chestnut Hill MA 02167. U.S.A. 9 May
1989; accepted 5 March
1990)
Low-angle X-ray scattering in solution has been used to probe the quaternary structure of a mutant version of Escherichia coli aspartate transcarbamylase in which Glu239 of the by site-directed mutagenesis. X-ray catalytic chain was replaced by glutamine crystallographic studies of the wild-type enzyme have shown that one set of intersubunit interactions involving Glu239 are lost, and are replaced by another set of intrachain interactions when the enzyme undergoes the allosteric transition from the T to the R state. Functional analysis of the mutant enzyme with glutamine in place of Glu239 indicates that homotropic co-operativity is lost without altering the maximal specific activity. The radius of gyration of the unligated mutant enzyme is larger than the unligated wild-type, indicating an alteration in quaternary structure of the mutant. However, the radius of gyration of the mutant enzyme in the presence of N-(phosphonoacetyl)-L-aspartate (PALA) is identical with the value for the wild-type enzyme in the presence of PALA. X-ray scattering at larger angles indicates that the mutant enzyme is in a new structural state different from the wild-type T and R structures. The scattering pattern in the presence of saturating concentrations of PALA is identical with that of the wild-type R structure. Saturating concentrations of carbamyl phosphate alone are sufficient to convert most of the mutant enzyme to the R structure, in the absence of aspartate. CTP shifts the scattering pattern of the mutant enzyme in the presence of saturating carbamyl phosphate towards the scattering curve of the unligated enzyme, but CTP has no effect on the scattering curve in the absence of carbamyl phosphate or in the presence of subsaturating PALA. However, in the presence of subsaturating PALA, ATP causes a strong shift towards the R structure. Neither ATP nor CTP has any effect on the activity of the mutant enzyme. These data suggest that the replacement of Glu239 by glutamine results in a new quaternary structure. These data also explain, on a structural basis, why co-operativity is lost in this mutant enzyme.
trimers of 34,000 M, catalytic chains carrying the active sites, and three dimers of 17,000 MT regulaAspartate transcarbamylase (EC 2.1.3.2) from tory chains carrying the effector binding sites Escheriehia coli catalyzes the first step of pyrimidine (for reviews, see Gerhart, 1970; Jacobson & Stark, biosynthesis, the condensation of carbamyl phos1973; Schachman, 1974; Kantrowitz et al., 1980a,b; phate and L-aspartate to form N-carbamyl-LKantrowitz & Lipscomb, 1988). aspart’ate. The E. coli enzyme shows homotropic coTwo functional and structural states of the operativity for aspartate, is feed-back inhibited by enzyme have been well characterized by a variety of CTP and activated by ATP. This extensively techniques; a less active T state, stabilized by CTP studied allosteric enzyme is composed of two and a more active R state stabilized by the 327
1. Introduction
0022-2836/90/130327-09
$03.04/O
0 1990 Academic
Press Limited
328
P. Taut et al.
Figure 1. Stereo view of the Cl-C4 interface of asp&ate transcarbamylase in the T (Kim et al., 1987) and R states (Krause et al., 1987). The Cl and C4 catalytic chains are in different catalytic subunits of the holoenzyme. In the T-state (top), the 240s loops of Cl and C4 are side by side, allowing the intersubunit interaction between Glu239 with Tyr165. In the R-state (bottom), the 240s loop of Cl and C4 are on top of one another. The intersubunit link between Glu239 and Tyr165 is lost, and is replacedby an intrechain link betweenGlu239and both Lys164 and Tyr165 that helpsto stabilize the position of the 240s loop in this position. Residue numbers followed by A correspond to the Cl catalytic chain and the residue numbers followed by B correspond to the C4 catalytic chain.
substrates or their analogs, for instance N-(phosphonoacetyl)-L-aspartate (PALAT). The crystal structure of the enzyme in both forms has been determined by Lipscomb & co-workers (Kim et al., 1987; Krause et al., 1987). In the transition from T to R, the enzyme expands by 12 -J! (1 A =@I nm) along the 3-fold axis while subunits and domains undergo rotational movements. This quaternary rearrangement is accompanied by tertiary changes in both the catalytic and regulatory chains. In particular, the loop at residues 230 to 245 of the catalytic chain moves significantly in the process, establishing two different sets of intra- and intermolecular bonds that contribute to the stabilization of each structure (seeFig. 1). This suggeststhat the 230-245 loop plays an important role in the regulatabbreviations used: PALA, N-(phosphonoacetyl)CP, carbamyl phosphate; Glu239-+Gln, the mutant enzyme with Gln in place of Glu239 in the catalytic chain of aspartate transcarbamylase; [S],.,, the aspartate concentration at half the maximal observed specific activity.
tory properties of the enzyme. To test this hypothesis, site-directed mutagenesis was used to modify residues within the loop. Specifically, Glu239 of the catalytic chain was selected for replacement, since it is involved in different sets of interactions in the two structures: in the T state, Glu239 of Cl$ interacts with both Lys164 and Tyr165 of C4 (belonging to the other catalytic trimer: Kim et al., 1987); it is also involved in an intrachain hydrogen bond with Asp236, which could contribute to the stabilization of the short stretch of helix H9 (E. Gouaux, personal communication). In the R state, the two intersubunit interactions are disrupted and replaced by intrachain bonds with Lys164 and Tyr165 of Cl; the intrachain link to Aap236 is broken. Glutamine was substituted for glutamic acid at position 239 and the functional consequences of the mutation were determined. The mutant enzyme shows a
L-aapartate;
$R and C, followed by a number, e.g. Rl, C4, refers to a particular polypeptide chain in aspartate transcarbamylase as specified in Fig. 6 of Honzatko et al. (1982).
Structural
Consequences of a Mutation
complete loss of co-operativity and a decreased [S],., for aspartate along with full retention of catalytic activity (Ladjimi 6 Kantrowitz, 1988). This suggested that either the T state of the mutant enzyme or the T=R equilibrium has been altered by the amino acid substitution. An alteration in the structure of the Glu239+Gln enzyme is suggested by preliminary crystallographic data on the unligated enzyme. These data itidicate an elongation of the unit, cell along the axis that corresponds to the molecular S-fold axis (Kantrowitz & Lipscomb, 1988). The solution X-ray scattering curve of unligated wild-type enzyme (T structure) shows secondary minima and maxima characteristic of its quaternary structure. Upon addition of substrates, it undergoes major changes reflecting the transition from T to R (Moody et al., 1979) and constitutes a sensitive and specific probe of the quaternary structure of the enzyme as was demonstrated by HervB et al. (1985). We present here a study of the structural consequences of the replacement of Glu239 by glutamine as determined by solution X-ray scattering.
2. Materials
and Methods
(a) Materials ATP, CTP, carbamyl phosphate, N-cerbamyl-L-aspartate, L-aspartate and Tris were purchased from Sigma Chemical Co. The carbamyl phosphate was purified by precipitation from 50% (v/v) ethanol, and stored desiccated at -20°C (Gerhart & Pardee, 1962). The replacement of Glu by Gln at position 239 in the catalytic chain of aspartate transcarbamylase was performed by site-specific mutagenesis as described (Ladjimi & Kantrowitz, 1988). Wild-type and Glu239+Gln aspartate transcarbamylases were isolated as described (Nowlan & Kantrowitz, 1985), from E. coli strain EK1104 [F- ara, thi, Apro-kc, ApyrB, pyrF*, rpsL], containing the plasmid pEK2 (Smith et al., 1986) or pEK56, respectively. (b) Methods The transcarbamylaae activity was measured at 25”~: by either a calorimetric (Pastra-Landis et al., 1981) or pHstat method (Wu & Hammes, 1973). pH-stat assays were carried out with a Radiometer TTTSO titrator and an ABUBO autoburette. All calorimetric assays were performed in duplicate and the data points shown in the Figures are the average. Enzyme samples for the X-ray scattering experiments were prepared from a stock solution in 50 mM-Tris-borate buffer (pH %3), @l mM-EDTA, @l mM-dithioerythritol as described (HervB et al., 1985). X-ray scattering curves were recorded on the small-angle scattering instrument D24 using synchrotron radiation at LURE-DCI, Orsay. The instrument has been described (Depautex et al., 1987), as well as the data acquisition system (Bordas et al., 1980). The scattering curves of dilute enzyme solutions (4 to 8 mg/ml) were recorded in the 8 range @0012 to O-012 A-‘. The scattering parameter s is defined, s= 2 sin Ojn, where 20 is the angle through which the X-rays are scattered, and 1 is the wavelength (1.608 A, K-edge of Co). Data were recorded in the angular range @Ol to
in Aspartate
329
Transcarbamylase
I 0.2
31
0
I 0.4
I 0.6
( 8
lo4xs2&*)
Figure 2. Guinier plots. The curves have been arbitrarily shifted along the vertical axis for the sake of clarity. From top to bottom: (0) unligated wild-type enzyme; (0) unligated Glu239+Gln enzyme; (0) wildtype enzyme in the presence of 2 molar excess PALA; ( q ) Glu239-+Gln enzyme in the presence of 2 molar excess PALA.
@05 8-l using 100 mg/ml by Hervb et al. (1985).
enzyme solutions
as reported
3. Results (a) Radius of gyration The small-angle pattern of dilute solutions of the wild-type and of the Glu239-*Gln enzyme were recorded in the absence of substrates and in the presence of a twofold molar excess of PALA (two PALA molecules/active site). Due to the high aflinity of PALA for the active site, this concentration is enough for all active sites to be occupied. The value of the radius of gyration was derived from a Guinier analysis (Fig. 2 and Table 1: and seeGuinier & Fournet, 1955). The radius of gyration of the unligated Glu239+Gln enzyme is only slightly smaller than its value in the presence of PALA,
Table Radii of gyration
Wild-type Glu239+Gln
for
1
the wild-type and Glu239+Gln enzymes
No substrates (4
Saturating PALA (4
46.6 +@3 485*@3
493kO.2 491 kO.2
330
P. Taut et al. 4.3
r
4.12
3.94 z In G B 1 3.76
3.58
3.4
I 0.01
I 0.018
I 0.026
I 0,034
s- 2 sin B/X
I 0.042
3.4 0.65
(L-9
Figure 3. Solution X-ray scattering spectra of the wildtype and Glu239-*Gln enzyme. (0) Unligated wild-type enzyme; (0) unligated Glu239-+Gln enzyme; (0) wildtype enzyme in the presence of 2 molar excess PALA; ( q ) Glu239-tGln enzyme in the presence of 2 molar excess PALA.
which is the same for both the mutant and wildtype enzymes within error bars of 1 standard deviation. (b) X-ray scattering spectra of unligated and
PALA
saturated enzymes
Figure 3 shows the X-ray scattering curve of the Glu239-tGln and wild-type enzymes in the absence and presence of saturating concentrations of PALA. The curves for the wild-type and the Glu239+Gln enzymes in the presence of PALA are indistinguishable. On the contrary, the curve for the unligated Glu239+Gln enzyme is significantly different from the unligated wild-type curve. Indeed, the intensities in the region of the first subsidiary minimum and maximum are intermediate between the intensities of the two curves of the wild-type T curve and R curve, while the position of the first maximum is much closer to that of the R curve than to that of the T curve. Could the proposed destabilization of the T state caused by the mutation lead to a shift in an equilibrium between the two extreme quaternary structures T and R, so that a mixture of the two structures is observed, even in the absence of substrates? In such a case, the scattering curve would be a linear combination of the wild-type T and R curves. Attempts at fitting the experimental
o-01
I 0.018
I 0.026
I 0.034
I 0.042
0.05
S=2 sin B/X (&-I)
Figure 4. Solution X-ray scattering spectrum of’ the Glu239-+Gln enzyme and linear combinations of wildenzyme extreme patterns. (0) IJnligated type Glu239dGln enzyme; (0) 55% R, 45% T: (0) 3096 R. 70% T.
data (0 in Fig. 3) are shown in Figure 4. The hypothetical curve corresponding to 55% R and 45% T fits well around the first maximum but not near the first minimum nor around the second minimulh and maximum. This last part of the curve is well approximated by the hypothetical curve composed of 30%R and 70%T, but this linear combination does not fit the region near the first minimum and maximum. The inability to fit the scattering curve of the Glu239+Gln enzyme in the absence of PALA to a linear combination of the wild-type T and R state scattering curves indicates conclusively that, the solution of the unligated mutant, is not a mixture of T and R structures. Therefore, the unligated Glu239-+Gln enzyme is in a new quaternary structure, different from both the wild-type T and R structures.
(c) Effect of carbamyl phosphate The reaction catalyzed by aspartate transcarbamylase involves an ordered mechanism in which carbamyl phosphate binds first and induces a conformational change that promotes the binding of aspartate or its analogs (Porter et al., 1969; Wedler & Gasser, 1974; Hsuanyu & Wedler, 1987). The scattering pattern of the Glu239-+Gln enzyme was recorded in the presence of a saturating concentration of carbamyl phosphate (25 mM). As seen in Figure 5, the scattering pattern of the mutant
Structural
3.41 0.01
I 0~018
I 0.026 s=2sin
Consequences of a Mutation
I 0.034 8/X
I 0.042
0.05
(ii-‘)
Figure 5. Solution X-ray scattering spectra of the enzyme. (0) Unligated Glu239+Cln enzyme; (a) Glu239-*Gln enzyme in the presenceof 2 molar excessPALA; (0) Glu239-tGln enzyme in the presenceof 25 mM-carbamyl phosphate;(a) Glu239-+Gln enzyme in the presence of 25 mM-carbamyl phosphate$10 mM-CTP.
Glu239-*Gln
enzyme is very similar to the R curve obtained in the presence of saturating PALA, showing that most of the enzyme is in the R structure. The experimental curve can best be fit with a 10~o-90~/o linear combination of the unligated and PALA saturated Glu239-rGln curves, respectively. This is very different from the case of the wild-type enzyme, where the addition of saturating amounts of carbamyl phosphate has almost no effect on the scattering pattern (P.T. and P.V., unpublished results). Subsequent addition of succinate, which shifts the scattering pattern of the wild-type enzyme from T to R (HervB et al., 1985), has no detectable effect with the Glu239-+Gln enzyme (data not shown). This is to be expected, since carbamyl phosphate alone can shift the curve all the way to the R state. (d) EJect of ATP and CTP (i) In the absenceof substrates The Glu239+Gln enzyme is unable to reach the T structure in the absence of substrates. Could the addition of CTP, which is known to stabilize the T structure of the wild-type enzyme, shift the curve of the mutant enzyme further toward the T structure? The scattering curve of the Glu239-+Gln enzyme, recorded in the absence and presence of 10 mi%-CTP
in Aspartate
0 .oi
331
Transcarbamylase
0~018
O-026
0.034
0.042
0~05
s=2 sin B/X (3)
Figure 6. Solution X-ray scattering spectra of the Glu239-tGln enzyme. (0) Unligated Glu239-rGln enzyme; (0) Glu239+Gln enzyme in the presenceof 10mM-CTP.
is shown in Figure 6. The addition of CTP causesno significant change in the scattering curve, and certainly not the decrease in intensity around the first maximum associated with a shift towards the T structure. (ii) In the presenceof saturating carbamyl phosphate For the Glu239+Gln enzyme, carbamyl phosphate alone is able to shift the quaternary structure almost all the way to the R structure. Could the structural alteration induced by the binding of carbamyl phosphate be reversed by addition of CTP? Figure 5 shows the scattering curve obtained in the presence of 25 m&I-carbamyl phosphate and 10 mM-CTP. This scattering curve is intermediate between the curve obtained in the presence of saturating carbamyl phosphate and the curve obtained in the absence of CTP. So, CTP can, at least partially, reverse the effect of carbamyl phosphate and shift the scattering curve towards the unligated state of the Glu239dGln enzyme. (iii) In the presenceof subsaturating PALA The addition of subsaturating amounts of PALA shifts the scattering curve towards the R state, as it does with the wild-type enzyme. In the latter case, it has been shown that further addition of the activator ATP has no detectable effect on the scattering pattern, while the inhibitor CTP causes a slight but significant shift towards the unligated pattern (HervB et al., 1985). What about the Glu239+Gln enzyme? As seen in Figure 7, CTP
332
P. Taut et al. 4.3
-2 u) z
L”
3.94
I ,5 x C ., B I.0 a, ? “0 0, K 0.5
3.76
0
FT -I
0.2
0.4
0.6
0.8
I.0
[NTP](mM)
3.58
3.4
L 0.01
I
I
I
I
0.018
0.026
0.034
0,042
s=2 sin 8/X
(
Figure 8. Influence of the effecters ATP and CTY on the act,ivity of the wild-type and the Glu239+Gln enzymes. The reactions were carried out at 25°C in 50 mM-Tris-acetate buffer (pH 8.3) in the presence of saturating levels of carbamyl phosphate (4.8 mM). For each enzyme? the aspartate concentration was at 0.5 x [S],., for aspartate. ATP effect on the (0) wild-t,ype and (0) Glu239-+Gln enzymes. CTP effect on the (m) wild-type and (0) Gln239+Gln enzymes.
(ii-l)
Figure 7. Solution X-ray scattering spectra of the Glu239+Gln enzyme. (0) In the presenceof 018 molecule of PALA/active site; (0) in the presenceof @18 moleculeof PALA/active site+ 10mM-CTP; (6) in the presence of @18 molecule of PALA/active site + 10 mM-ATP.
shows no detectable effect, while ATP causes a strong shift towards the R state. More precisely, the distribution shifts from about 30% R structure to about 90% R structure. The effects of ATP and CTP on the Glu239+Gln enzyme are in complete contrast to the scattering curves obtained for the wild-type enzyme. (e) Functional behavior of ATPICTP Glu239+Gln enzyme
with the
In order to correlate the structural changes induced by the allosteric effecters with the function of the Glu239+Gln enzyme, ATP and CTP saturation curves were determined. As seen in Figure 8, neither ATP nor CTP alter the activity of the Glu239-tGln enzyme, even at high concentrations. This behavior is in contrast to the activation of the wild-type enzyme observed with ATP and the inhibition of the wild-type enzyme observed with CTP. 4. Discussion The functional analysis of the Glu239-+Gln enzyme showed that it is devoid of co-operativity with respect to aspartate. Furthermore, its activity was inhibited by low concentrations of PALA at low and saturating concentrations of aspartate carbamyl phosphate (Ladjimi & Kantrowitz, 1988),
suggesting that the active sites of the Glu239+Gln enzyme are in the active R form, even a,t low concentrations of aspartate (with saturating carbamyl phosphate). Therefore, kinetically, the Glu239+Gln enzyme appears to be in an R-like state. The determination of the radius of gyration of the Glu239-rGln enzyme provides direct structural evidence to support this conclusion. The value of the radius of gyration of unligated Glu239--+Gln enzyme is close to the value obtained for the Glu239+Gln enzyme saturated with PALA. For the mutant enzyme, the binding of PALA seems t’o cause almost no structural transition. The scattering curves at larger angles show that the situation described above is incomplete. The Glu239-+Gln enzyme, in the absence of all ligands, is in neither the wild-type T nor R structure, since the scattering curve of the mutant is different from the two curves of the wild-type enzyme (seeFig. 3). How can this be reconciled with the radius of gyration data indicating that the mutant enzyme is almost entirely in the R structure? The contradiction is only apparent. Indeed, the position of the first subsidiary maximum is correlated with the distance between the two catalytic trimers along the 3-fold axis, on each side of the central cavity. This position on the curve of t’he unligated mutant is close to the position on the R curve, showing that the two catalytic trimers are almost as far apart as in the wild-type R structure. Since the catalytic subunits account for two-thirds of the molecular mass of the enzyme, their distance determines the value of the radius of gyration. Thus, the R, values are consistent with the scattering curves at larger angles. Changes in the scattering pattern of aspartat,e transcarbamylase at s-values lower than 0.05 A ’
Structural
Consequences of a Mutation
have been shown to be specifically sensitive to quaternary structure changes (Altman et al., 1982; Herve et al., 1985). The scattering curve of the unligated Glu239+Gln enzyme differs from the scattering curves of both the unligated (T) and PALA-saturated (R) structures of the wild-type enzyme. This suggests that either the solution of unligated Glu239+Gln enzyme contains a mixture of molecules in T and R structures or that the mutant enzyme exists in an entirely new structural state. The fact that this curve can not be fit to any linear combination of the wild-type T and R scattering curves strongly supports the second hypothesis. Furthermore, the radius of gyration of the mutant; enzyme along with the position of the first subsidiary maximum indicates that the structure of the unligated Glu239+Gln enzyme is more similar to the R structure than to the T structure of the wild-type enzyme with an elongation along the 3-fold axis. This structure will be hereinafter designated as the R’ structure. The fact that, in the absence of ligands, the Glu239+Gln enzyme does not exist in a wild-type like T structure, suggests that the Glu239 interactions are essential for the stabilit#y of the T structure. Other mutations have been made in the Cl-C4 interface, Tyrl65+Ser (Robey & Schachman, 1984) and Tyrl65--+Phe (Wales et al., 1988), which do not produce enzymes with R-like kinetic properties. The catalytic subunit of the Tyrl65-Ser enzyme has reduced activity and affinity, while the TyrlSEi-tPhe enzyme has a much lower affinity for aspartate. A plausible interpretation is that the Glu239 to Lysl64 intersubunit link is essential for the stability of the T state, together with the (~1~239 to Asp236 intrachain bond, while the Glu239 to Tyrl65 intrasubunit bond contributes to the stability of the R-state. The Glu239-+Gln enzyme is not the first case to be reported of a mutant version of aspartate transcarbamylase adopting a new quaternary structure in the absence of ligands. The pAR5 enzyme is a mutated form of aspartate transcarbamylase in which residues 145 to 153 at the C terminus of the R chain are replaced by a new sequence of six residues (Ladjimi et al., 1985) that contribute to the Rl-C4 interfa,ce. The quaternary structure of the unligated pAR5 enzyme has been shown to be different from both the T and R structures of the wild-type enzyme (Cherfils et al., 1987). It is also different from the structure of the unligated Glu239+Gln enzyme; more precisely, the scattering pattern of the unligated pAR5 enzyme is much closer to the wild-type T curve than is the curve of the Glu239+Gln enzyme. In the presence of saturating amounts of PALA, the scattering curves of the wild-type and of the Glu239+Gln enzyme are indistinguishable. So, upon substrate binding, the mutant enzyme undergoes the R’-+R structural transition. Are there other, intermediary quaternary structures accessible to the Glu239+Gln enzyme? All the scattering curves of the Glu239+Gln enzyme presented here
in Aspartate
Transcarbamylase
333
can be satisfactorily fit by a linear combination of the two extreme scattering patterns of R and R’. So, the structural behavior of the mutant enzyme under the variety of conditions investigated can be accounted for with only two quaternary structures, R and R’. Similar observations were made for two other mutant enzymes, Tyr240-+Phe enzyme (Cherfils et aE., 1989) and the pAR5 enzyme. The addition of saturating carbamyl phosphate does not cause a significant change in the scattering pattern of the wild-type enzyme, while it causes considerable modifications in the case of the Glu239+Gln enzyme. In fact, carbamyl phosphate alone can convert most of the Glu239-+Gln enzyme to the R structure. The binding of carbamyl phosphate is known to occur first and to cause tertiary conformational changes in the carbamyl phosphate domain (Collins & Stark, 1969; Griffin et al., 1972) that help to create the aspartate-binding site. For the wild-type enzyme, the intersubunit interactions involving Glu239 keep the enzyme in the T structure even when carbamyl phosphate binds. It is binding of aspartate after the binding of carbamyl phosphate that triggers the allosteric quaternary conformational change (Kantrowitz & Lipscomb, 1988). In the case of the Glu239+Gln enzyme, the (presumably) same conformational change in the carbamyl phosphate domain results in the conversion of most (approx. 90%) of the enzyme to the R structure. In other words, the value of the allosteric constant L, as defined in the Monod, Wyman & Changeux (Monod et aZ., 1965), of the enzyme-CP complex is about 61; this very low value explains the absence of co-operativity observed with aspartate (Ladjimi & Kantrowitz, 1988). A similar, though weaker, effect of CP was observed with the pAR5 enzyme. Neither ATP nor CTP alter the activity of the Glu239+Gln enzyme (see Fig. 8), although both are able to alter the structure of the mutant enzyme. In the absence of carbamyl phosphate and aspartate, CTP is unable to shift the structure towards the wild-type T state. However, in the presence of carbamyl phosphate, which itself has converted most of the mutant enzyme into the R structure, CTP is able to promote a return towards the R structure, bringing the curve from 907; R to about 45% R. This first proves that CTP does bind to the Glu239-+Gln enzyme; the absence of effect of CTP on the unligated Glu239+Gln enzyme can not thus be explained by the absence of binding. These data suggest also that carbamyl phosphate is not very effective in stabilizing the R structure, since CTP can, at least partially, reverse its effect. However, they can not be correlated directly with the activity measurements, since the latter were performed in the presence of a significant amount of aspartate, which could have stabilized the R structure. To test this possibility, experiments were performed at subsaturating concentrations of CP; they show neither inhibition by CTP nor activation by ATP (data not shown). This suggests that the structural conversion of the enzyme population from mostly R
P. Taut et al.
334
to half R half R’ by CTP is insufficient to inhibit the activity of the enzyme. A possible interpretation could be that the mutant enzyme in the R’ structure has the same affinity for aspartate as in the R structure. In fact, we have no direct indication of the affinity of the unligated Glu239+Gln enzyme for aspartate, since the only available data pertain to the ATCase-CP complex, which is mostly in the R structure. l?or the wild-type enzyme, subsaturating PALA has been used to shift the TtsR equilibrium towards the R structure. Under these conditions, ATP has no detectable influence on the scattering curve of the wild-type enzyme, while CTP has only a small effect (HervB et al., 1985). At subsaturating concentrations of PALA, CTP does not shift the scattering curve of the Glu239--+Gln mutant toward the R’ structure. This result suggests that PALA binding to the mutant enzyme stabilizes the R structure sufficiently to prevent the mutant enzyme from reverting back t)o the R’ structure upon the addition of CTP. On the other hand, ATP is able to shift the scattering curve, bringing a population about equally distributed between R’ structure and R structure mostly into the R structure. Together with the absence of effect observed with CTP under the same conditions, this supports the view that ATP and CTP operate in different ways. The role of Glu239 for co-operativity in aspartate transcarbamylase and in maintaining the enzyme in the T structure has been suggested: Cl-C4 intersubunit interactions with Lysl64 and Tyr165 (Kantrowitz & Lipscomb, 1988) and an intrachain hydrogen bond with Asp236 (E. Gouaux, personal communication). The scattering data reported here substantiate this interpretation. The mutant enzyme is unable to adopt the T structure. Furthermore, the Cl-C4 interface in the R’ structure that it adopts instead is less stable than in the T structure of the wild-type enzyme. In order to establish the importance of the R’ structural state observed in the Glu239-+Gln enzyme for co-operativity of aspartate transcarbamylase, additional functional and structural studies are in progress. After this paper was submitted, the crystallographic enzyme
structure
of the
unligated
Glu239-+Gln
was reported by Gouaux et al. (1989). The quaternary structure of the enzyme appears to be different from both the T and R structures. However, this structure seemsto be much too close to the T structure to be compatible with our results in solution. This discrepancy, which could be due either to the constraints from the crystal lattice or to the low pH (pH 58, close to t,he p1 of the enzyme) at which crystals are grown, will be further investigated by calculating scattering curves from crystallographic co-ordinates.
It is a pleasure to acknowledge the help from the technical staff from LURE-DC1 and from the computing center of LURE. We thank E. Gouaux for attracting our attention towards the role of Asp236, Wei Xu for help with some of the activity measurements, W. N. Lipscomb
for providing X-ray co-ordinates of the enzyme and Jo#l Janin for critical reading of’ the manuscript. E.R.K. acknowledges support from USPHS research grants GM-26237 and DK- 1429.
References Altman,
R. B., Ladner,
J. E. & Lipscomb, W. N. (1982). c’ommun. 108, 592-595. Bordas, cJ., Koch, M. H. J., Clout, P. N., Dorrington. K:.. Boulin. C. & Gabriel, A. (1980). J. Whys., E: 8ci. In&rum. 13. 938-944. Cherfils. J.. Vachette, P., Taut, P. dt Janin. J. (1987). Biochem.
EMBO
Biophys.
Res.
J. 6, 2843-2841.
Cherfils, J.. Sweet, R. M.. Middleton, S. A., Kantrowitz. E. R., Taut, P. & Vachette, P. (1989). PEBS Letters, 247,
361-366.
Collins, K. D. & Stark. U. R. (1969). J. Hiol. Chrvn. 244. 1869~-1877. Depautex, C., Desvignes, C.. Leboucher. P.. Lemonnier, 81.. Dagneaux, D.. Benoit, ,J. P. & Vachette. P. (1987). LURE Annual Report. Laboratoire pour I’Utilisation du RayonnementElectromagnBtique,Paris. Gerhart. ,J. C. (1970). CWT. Top. Cell. Reg. 2, 275~3%. Gerhart. ,J. C. & Pardee. A. H. (1962). J. Biol. C’hevn. 237, 891.-896. Gouaux, J. E., Stevens, R. C.. Ke, H. M. & Lipscomb. W. N. (1989). Proc. ,Vat. Acad. Ski., fT.8.A. 86. 8212-8216. Griffin, J. H., Rosenbusch. J. P.. Weber. K. K. b Blout. E. R. (1972). J. Biol. Chem. 247, 6482%6490. Guinier, -4. & Fournet, G. (1955). Srn.all-Angle Scattering of X-rays, Wiley. New York. HervB, U., Moody. M. F.. Taut, P., Vachette. 1’. Cyr. Jones. P. T. (1985). J. iuol. Riol. 185, 189-199. Honzatko. R. B., Crawford, J. L., Monaco, H. L., Ladner. J. E.. Edwards, B. F. P.. Evans. D. It.. Warren. 8. 0.. Wiley, D. (‘., Ladner, R. C. & Lipscomb. W. N. (1982). J. Mol. Biol. 160, 219-263. Hsuanyu, Y. & Wedler. F. (‘. (1987). Arch. Nioch.rm Biophys
259,
,Jacobson,
316--330.
G. R. & St,ark.
G. R. (1973).
Cnzyvnes
9.
225308.
Kantrowitz, E. R. bz Lipscomb, W. N. (1988). Science. 241, 669-674. Kantrowitz, E. R., Pastra-Landis. S. (‘. & Lipscomb. W. N. (1980a). Trend4 Biochem. Sci. 5. 124-128. Kantrowitz. E. R.; Pastra-Landis, S. C. & Lipscomb. W. N. (198Ob). 7’rendB B&hem. Sci. 5, 150--163. Kim. K. H.. Pan. Z., Honzatko, R. B.. Ke. H. M. & Lipscomb, W. N. (1987). ,/. Mol. Biol. 196, 853-875. Krause, K. I,.. Volz, K. W. 6t Lipscomb, W. N. (1987). J. Mol. Biol. 193, 527ki53. Ladjimi, M. & Kantrowitz. E:. R. (1988). Biochemistry, 27, 276-283.
Ladjimi, M.. Ghklis, (1.. Feller, A.. Cunin, R.. Ulansdorff, S.. PiBrard, A. & HervB. G. (1985). J. Mol. Riol. 186. 715. 724. Monod, ,J.. Wyman. ,J. & (Ihangeux,
