Multivalent Regulation of Glutamic Dehydrogenases from Fungi

Guanylates (GTP, GDP, and GMP), short chain acyl-CoA derivatives; ATP, and ADP, were found to be allosteric activators of NAD-specific glutamic dehydrogenases isolated from a unique group of fungi, the Oomycetes. Long chain acyl-CoA derivatives and AMP were very strong inhibitors of the enzyme. Pythium glutamic dehydrogenase, in particular, was shown to be sensitive to control by energy charge. A total of five activators have now been discovered. These are NADP+, P-enolpyruvate, short chain acyl-CoA derivatives, GTP, and ATP. These activators functioned cumulatively to antagonize the effects of the inhibitors, citrate, AMP, and long chain acyl-CoA derivatives. The activators acted as unidirectional stimulants of the biosynthetic reaction, while some of them inhibited the catabolic reaction unidirectionally. These glutamic dehydrogenases are genetically controlled. They are subjected to marked catabolic repression and inducible by glutamate. The substrates (ammonia and a-ketoglutarate) of the biosynthetic reaction are allosteric inhibitors. The activators modulated the enzyme against allosteric inhibition by its substrates. This multivalent control has been explained as a mechanism by which the enzyme effects its biosynthetic and catabolic roles in amphibolic reactions of the citric acid cycle. A remarkable correlation has been made between lysine biosynthesis and allosteric control mechanisms among members of the Phycomycetes. The distribution of the two pathways of lysine biosynthesis among these simple fungi parallels the only two different forms of allosteric controls of NADspecific glutamic dehydrogenases elucidated for all major orders of the Phycomycetes studied to date.

Pythium glutamic dehydrogenase, in particular, was shown to be sensitive to control by energy charge.
A total of five activators have now been discovered. These are NADP+, P-enolpyruvate, short chain acyl-CoA derivatives, GTP, and ATP.
These activators functioned cumulatively to antagonize the effects of the inhibitors, citrate, AMP, and long chain acyl-CoA derivatives. The activators acted as unidirectional stimulants of the biosynthetic reaction, while some of them inhibited the catabolic reaction unidirectionally. These glutamic dehydrogenases are genetically controlled.
They are subjected to marked catabolic repression and inducible by glutamate. The substrates (ammonia and a-ketoglutarate) of the biosynthetic reaction are allosteric inhibitors.
The activators modulated the enzyme against allosteric inhibition by its substrates. This multivalent control has been explained as a mechanism by which the enzyme effects its biosynthetic and catabolic roles in amphibolic reactions of the citric acid cycle. A remarkable correlation has been made between lysine biosynthesis and allosteric control mechanisms among members of the Phycomycetes.
The distribution of the two pathways of lysine biosynthesis among these simple fungi parallels the only two different forms of allosteric controls of NADspecific glutamic dehydrogenases elucidated for all major orders of the Phycomycetes studied to date.  has suggested that at metabolic loci connecting energy-regenerating and energy-utilizing processes, the delicate balance between adenine nucleotides may be regarded as a critical factor in the regulation of enzymes involved in the reactions.
* This work was supported by a grant from the National Research Council of Canada.
$ To whom all correspondence should be addressed. 0 National Research Council Scholarship holder and predoctoral fellow.
7 Technical research assistant.
One such metabolic link is the oxidation-reduction reaction catalyzed by glutamic dehydrogenase.
Although NAD-specific glutamic dehydrogenase has been regarded as a catabolic enzyme, there would be times when it must perform a biosynthetic role; e.g. when the NADP-specific variety is absent in an organism.
This enzyme should therefore respond to fluctuations in adenylate concentrations (energy charge) as defined by Atkinson (3).
In this communication, we report on the effects of adenylates, short and long chain fatty acid-CoA esters, and guanylates (GTP, GDP, and GMP) as modulators of the catalytic activity of NAD-specific glutamic dehydrogenases obtained from a specialized group of fungi, Oomycetes.
Preliminary findings have shown that the enzyme is controlled both at the genetic and enzymatic levels (4). Synthesis of the enzyme is repressed by catabolites. NADP+, NADPH, and P-enolpyruvate activate the enzyme in an allosteric manner.
The activators act cumulatively to antagonize the inhibitory effects of citrate, AMP, and long chain fatty acid-CoA esters.
Among regulatory enzymes, glutamic dehydrogenase is unique. Glutamic dehydrogenases isolated from microorganisms and higher animals display different allosteric effects (4-S).
Evidently, this protein has been a sensitive target for the evolution of regulatory mechanisms.

MATERIALS AND METHODS
Methods used in the treatment of all the fungi for enzyme isolation have been reported elsewhere (4). Kinetic assays were carried out with a Gilford model 2400 recording spectrophotometer, equipped with temperature control unit, at 25". Chemicals-CoA, acetyl-CoA, malonyl-CoA, acetoacetyl-CoA, succinyl-CoA, dephospho-CoA, palmitoyl-Cob, and oleyl-CoA were purchased from P-L Biochemicals.
All other chemicals were obtained from Sigma.

Response to "Energy
Charge"-The organisms Achlya sp. The reductive amination (biosynthetic) reaction is unaffected by similar concentrations of AMP and ATP (Fig. lb). On the other hand, glutamic dehydrogenase from Pythium was found to be extremely sensitive to the influence of adenylates. ATP functioned as a mild activator and AMP as a strong inhibitor of the biosynthetic and catabolic reactions (Fig. 2, a and  b). In these experiments, addition of adenylate kinase was omitted.
According to the formulations of Atkinson (3), the response of Pythium glutamic dehydrogenase to varied proportions of AMP, ADP and ATP in catalyzing the catabolic and biosynthetic reactions was of the energy-utilizing (U-) type. In the absence of ADP, as in these experiments, absolute energy charge values were not obtained.
Our interest was to observe the general pattern of response of the enzyme to adenylates. Although ADP, like ATP, activates the enzyme slightly, the "energy charge" values given in Fig. 2, a and b, are only apparent. (b) One point of significance is the disproportionately large difference in adenylate concentrations required to control the biosynthetic and catabolic reactions.
The catabolic reaction was completely inhibited by 1 mM AMP whereas the biosynthetic reaction required 6 mM AMP for effective control.
Therefore, at other than equilibrium conditions, the adenylate control system would be operative unidirectionally as defined previously (9). Although the catabolic reaction may be completely inhibited at adenylic acid levels above 1 IYIM, the biosynthetic reaction would be operative, albeit, at a reduced efficiency. With ATP present, the difference would be more marked because the adenylate effect is a nonlinear function of the total adenylates, not of the single components. nia are among a few organisms shown to possess an NAD-specific glutamic dehydrogenase that is activated by NADP+, NADPH, and P-enolpyruvate (4). A linked metabolic reaction is catalyzed by an NADP-specific isocitric dehydrogenase. We speculated that these two enzymes may act cooperatively and function as a transhydrogenase system. Some transhydrogenases are energy-linked (10). It was of signal interest to find that Pythium glutamic dehydrogenase is particularly susceptible to adenylate control.
Because a multitude of activators have now been found (see later in this report), all activators, singly and in diverse combinations, were analyzed for their capability to antagonize or interact with the adenine nucleotides.
The results presented in Fig. 2a show that P-enolpyruvate and NADPf are the only ligands that can, independently, release the enzyme from AMP inhibition.
The other activators, GTP, CoA, and derivatives, were incapable of doing this even when combined.

Sigmoid
Inhibition-An enzyme that displays sigmoidal kinetics will satisfy the Michaelis-Menten equation with the following modification.
Koshland, Nemethy, and Filmer (11) have proposed a little used device that can discriminate between cooperative and noncooperative protein-ligand interaction. They suggested that by det.ermining the cooperativity index (R,), which is the ratio of substrate concentration necessary to give 90% saturation to that concentration which gives 10% saturation, any deviation from the Michaelis-Menten hyperbolic relationship can easily be determined. R, should be 81 for all cases that follow Michaelian kinetics.
At (S)sO and (X),0, Equation 1 can be written in the form of two simultaneous equations as follows. and NADP+ against oleyl-CoA inhibition. I;nset, an enlargement of the curves at low oleyl-CoA concentrations. The concentrations of activators used are given in Fig. 2 legend. The expression "All Activators" represents the four activators given above. between ligand and protein and evaluate the interaction coefficient, n, which is only an approximate indication of the number of substrate-binding sites involved in the interaction. When a substrate acts as an allosteric inhibitor, theoretically, it should be possible to analyze the saturation curve on the same basis by using Equation 5.
It must be emphasized that most substrate inhibition curves would give n. values greater than 1 when Equation 5 is used. Other diagnostic procedures would be required to confirm whether or not true cooperativity is involved. This approach has some value only when modulators activate an enzyme at high substrate concentration ranges. The computational procedure we followed was to select as zero inhibitor concentration the concentration of substrate at which there is neither an increase nor a decrease in the reaction rate when the substrate level was increased further by at least 10%. The concentration of substrate required to reduce this optimal rate by 10% was taken to represent (S&, and the concentration of substrate that reduced the optimal rate by 90% taken as (S1)90. (1)0.6 M 1.3 x 10-d 1.65 X lo+ 6.5 X 1O-4 3.3 x 10-d 3.0 x 10-d 4.5 x 10-S Extremely high 7.0 x 10-4 3.75 x 10-a 1.55 X 10-Z Extremely high 6.0 X 1O-6 9.1 x 10-e 6.1 X lo+ 6.1 X lo+ 6.2 X 1O-6 1.65 X lo-& 2.0 x 10-s Analysis of Fig. 3a in which ammonia acted as a substrate and an inhibitor of the reductive amination reaction of Pythium NAD-specific glutamic dehydrogenase, based on these suggestions, showed that the interaction of the inhibitor with the enzyme may be positive cooperative.
(X1)90/(Sl)10 was estimated as 3, and n value computed as 4. A similar computation done for a-ketoglutarate as an inhibitor and substrate (Fig. 3b) gave an (XI),~/(SI)n, value of 3.72 and n of 3.3.
ATP, as an activator, has been referred to above under adenylate control. Activation by all effecters was more pronounced on the Pythium catalyst than on AchZya glutamic dehydrogenase ( Table I). The activators were tested for their efficacy in the reductive amination and oxidative deamination reactions of both enzymes. Cyclic nucleotides, 3',5'-AMP and 3',5'-GMP, had slight stimulatory influence on the biosynthetic reaction but inhibited the catabolic process. Besides quantitative differences shown in the influence of the activators on the two enzymes, the effects of adenylates and P-enolpyruvate on these glutamic dehydrogenases were markedly different.
Pytlzium glutamic dehydrogenase was inhibited by P-enolpyruvate when the catabolic reaction was catalyzed whereas Achlya glutamic dehydrogenase was activated. ATP inhibited the catabolic reaction of Achlyu glutamic dehydrogenase but not that of Pythium.
Other than these subtle modifica- Allosteric Inhibiters-In addition to citrate and AMP, long chain fatty acid-CoA esters were found to be inhibitors of these glutamio dehydrogenases (Fig. 4). The two esters studied in some detail were palmitoyl-Coil and oleyl-CoA.
(I)0.5 values' for these esters are given in Table II. At very low concentrations, the esters activated the enzyme slightly. This may be caused by the presence of traces of free CoA that are present in the commercial preparation of the esters. Alternatively, the ester may be acting as a competitive inhibitor with one of the substrates that inhibit, the enzyme. At higher concentrations, the ester would bind at its own inhibitor site as well. @I) 0.5~ and Activation-We have mentioned elsewhere (4) that the activation mechanism operative on the glutamic dehydro-1 Following a suggestion of Koshland et al., (II), discussed in detail by Atkinson (12), we will restrict our use of the term (X),,., to mean the concentration of ligand required to fill half of the sites of a given type on the enzyme.
When the ligand is a substrate, we will use (S) in place of (X) ; when an inhibitor, genases is geared toward modulation of the enzymatic activity at high substrate concentrations where substrates appear to inhibit allosterically.
By studying the influence of effectorson the enzyme at (X1)o.b of a substrate, a valuable estimate of the extent of ligand activation or inhibition can be made. This approach was taken to evaluate all activators of these glutamic dehydrogenases.
The data given here are predominantly for the Pythium catalyst.
Ammonia was held fixed at @I),., and the other substrates kept at optimal levels. c-r-Ketoglutarate, being also an allosteric substrate inhibitor, was not used at saturating concentrations.
An activation curve for GTP is presented in Fig. 5 and the activation curves for four of the nine short chain acyl-CoA derivatives in Fig. 6. The results are presented as plots of (0, -a,) against the concentration of the activators. The notation ve is the reaction rate with activator and v0 is the rate without activator.
The same data were analyzed in Line-weaverBurk double reciprocal form. This method permitted us to determine the K, (activation constant) values for the activators and also evaluate the nature of ligand binding.
Although the activation of all of the short chain acyl-CoA derivatives were    analyzed in this manner, the plots given are the curves for CoA, acetyl-CoA, butyryl-CoA, and succinyl-CoA. A curve for dephospho-CoA, which neither activated nor inhibited the enzyme, has been included but the double reciprocal form omitted.
There were relatively little differences in the degree of activation caused by the acyl-CoA derivatives.
The activation constants for GTP and the acyl-CoA derivatives were approximately the same at about 4 X lop5 M. Since acyl-CoA derivatives and free CoA functioned as activators, the sulfhydryl group of CoA does not appear to be essential for activation.
The 3'-phosphate group of CoA, however, is essential (Table III).
All of the activators displayed what appears to be a linear function in kinetic binding of the activators to the enzyme.
But on replotting the data as log (v, -v,)/(V,,, -v,) against log of activator concentration, the log plots were biphasic with one slope of n value equaling 1 and a second slope with n values varying between 2 and 4. Only the data for GTP and butyryl-CoA are presented as they represent the extreme cases of the various biphasic log plots obtained.
Presumably, some form of cooperative binding of the activators does occur.
Although the activators of these glutamic dehydrogenases are phosphorylated compounds, their structures are so different that a clear picture cannot be obtained from the structures about important functional groups.
Phosphate is unlikely to be the single important factor because AMP is an inhibitor and several other phosphorylated compounds tested have no influence on the enzyme.
An alternative interpretation is that multiple sites may be involved in the activation process. Tests were done to see if there was any cumulative property of the activators in their interaction with the enzyme.

Multivalency
The ability of the activators to antagonize the inhibitory action of AMP, citrate, and long chain fatty acid-CoA esters was used as the test model.
The reason for this approach rested on an early observation that all of the activators, except ATP, can easily reverse the allosteric inhibition by the substrates, ammonia, and cY-ketoglutarate. Nonsubstrate inhibitors were more toxic and resisted antagonism by single activators. AMP InhiKttin-When the activators were used singly and in combination to antagonize AMP effects on Pythium glutamic dehydrogenases, only NADP+ and P-enolpyruvate could, independently, overcome most of the inhibition (Table IV). GTP, CoA (or acetyl-CoA) did not antagonize AMP inhibition significantly.
Combined, GTP and CoA were no better than CoA alone.
On the contrary, NADPf and P-enolpyruvate acted cumulatively and antagonized AMP completely. Although P-enolpyruvate was cumulative with either GTP or CoA, NADP+ did not show any cumulative property with these two compounds.
It would appear from these results that NADP+, GTP, and CoA were interacting at the same or closely related sites. P-Enolpyruvate, definitely, has a distinct site. However, because of the small relative differences in percentage deinhibition elicited when GTP and CoA were used, it became necessary to estimate the number of activator sites by using other inhibitors. Citrate Inhibition-All of the activators were tested for their ability to release the enzyme from citrate inhibition.
A previous study had shown that NADP+ and P-enolpyruvate acted cumulatively against citrate (4). Similar studies were done here with GTP, ATP, and acyl-CoA derivatives and then compared against NADP+ and P-enolpyruvate antagonism. Although ATP is an activator (based on the adenylate control hypothesis) it failed to antagonize citrate (Fig. 7). GTP, COA, and P-enolpyruvate were weakly antagonistic to citrate. Only NADP+ showed a significant antagonistic property.
GTP and CoA had some cumulative ability, but they were considerably less effective than NADP+ alone. When P-enolpyruvate and by guest on March 24, 2020 http://www.jbc.org/ Downloaded from either CoA or GTP were used, their cumulative effect did not quite match that of NADP+ as an antagonist (see Table II for a record of (Z)0.5 of citrate in the presence of various activators). P-Enolpyruvate, GTP, and CoA combined were cumulatively as effective as NADP+ alone. When all four activators (NADP+, P-enolpyruvate, GTP, and CoA) were used, complete antagonism of citrate inhibition occurred (Fig. 7). From these results, it appears that all of the activators have separate binding sites.
Okyl-CoA Inhibition-Long chain fatty acid-CoA derivatives have been studied in detail by Taketa and Pogell (13) with regard to their inhibitory effect on enzymes not related to fatty acid biosynthesis.
They concluded that palmitoyl-CoA may act as a detergent during inhibition. Normally, inhibition of this type would lead to an irreversible denaturation of the protein. Zahler,Barden,and Cleland (14) have estimated that the mixed micellar concentration of palmitoyl-Coil is 2 to 4 pM. Dorsey and Porter (15) showed that the critical mixed micellar concentration of palmitoyl-CoA and fatty acid synthetase is 5 PM. Above this concentration, the enzyme was inhibited irreversibly. We do not know what the critical mixed micellar level for Pylthium glutamic dehydrogenase and palmitoyl or oleyl-CoA may be. But Pythium glutamic dehydrogenase was markedly inhibited by oleyl-and palmitoyl-Coil at concentrations above 1.5 PM (Fig. 4) Activators used were at the concentrations indicated for Fig. 2. Reactants as outlined in Fig. 3b with 4 pg of enzyme. The values for ATP (3 mM) as antagonist were nearly coincident with the control plot and so have been omitted. PEP, phosphoenolpyruvate.
protective effect. As shown in Table II and Fig. 4 inset, significant protection against oleyl-CoA inhibition occurred when all the activators were present.
The enzyme was not inhibited by oleyl-CoA at concentrations below 6 PM. At 10 pM oleyl-CoA concentration, the enzyme was 80% inhibited in the absence of the activators, but only 20% inhibited when they were added.
Significant inhibition always occurred at oleyl-CoA levels above 10 PM whether activators were present or not. It is possible that the coenzyme may inhibit in this range by detergent action.
A more critical study of molar ratio of coenzyme and protein that leads to inhibition and how activators alter this ratio would be required to resolve this problem.

Physiological Action of Modulators
At this stage of our studies, it is not possible to give a complete synthesis of the physiological reasons that would account for the multivalent control of glutamic dehydrogenases from these fungi (Oomycetes). NADP+ and P-Enolpyruvate-Conclusions drawn from data presented previously (4) on the physiological basis of NADPi and P-enolpyruvate activation of Pythium, Achlya, and Sapro-Zegnia glutamic dehydrogenases are not contradicted. The NADP-specific isocitric and NAD-specific glutamic dehydrogenases act cooperatively to maintain a balance of pyridine nucleotides as follows.
Isocitrate + NH,+ + NADP+ + NADH $ NAD+ + NADPH + glutamate + CO, P-Enolpyruvate may not be involved in this reaction since the substrates can provide sufficient energy to effect the conversion.
Under energy-rich conditions when P-enolpyruvate accumulates, GTP and P-enolpyruvate activation of the biosynthetic reaction of glutamic dehydrogenase is a reasonable effect because the citric acid cycle would be operating as a biosynthetic unit supplying its intermediates for amino acids and nucleotides. The questions, why do these enzymes display marked sensitivity to inhibition by their substrates, ammonia and cY-ketoglutarate; and why do effecters modulate the enzyme only at these toxic levels of substrate, remain unanswered.
But two possible explanations can be offered. First, if cu-ketoglutarate is liable to accumulate during its production from carbohydrates and transamination reactions of glutamate, then this metabolite may act directly as a substrate and indirectly as an end product (transaminase-glutamic dehydrogenase couple) feedback effector of the glutamic dehydrogenase.
One role of the activators would be to relieve the enzyme from allosteric inhibition by substrates and permit continued biosynthesis. The second and more attractive proposal is that substrate control is related to catabolism.
The enzyme must have evolved as a catabolic catalyst, because it is subjected to catabolite repression.
During deamination reactions, ar-ketoglutarate and ammonia may accumulate.
This may have led to the development of a product inhibition control mechanism which is observed as substrate inhibition in these studies The multivalent control by activators may have evolved subsequently to overcome this inhibition when the enzyme has to function in a biosynthetic capacity.
Acyl-CoA Derivatives-According to the transhydrogenase hypothesis, NADPH is produced at the expense of NADH.
One use to which the NADPH may be put is to synthesize fatty acids. Activation of the enzyme by short chain intermediates of fatty acid synthesis in a manner that favors NADH utilization fits this concept very well. Palmitoyl-and oleyl-CoA, which may be considered "end products" of fatty acid biosynthesis feedback, inhibit the enzyme. This is a common method of control of the first enzyme involved in most biosynthetic sequences studied.
Although citrate is an inhibitor, its effect is qualitatively unidirectional on the catabolic reaction of the enzyme (4). Therefore, the anomaly of citrate inhibition would not contradict acetyl-CoA activation of the enzyme.
In fact, it does support the concept that the enzyme has catabolic function because citrate can be looked on as an end product.
GTP-A reasonable explanation for the unidirectional stimulation of the enzyme by GTP is difficult because many interrelated reactions utilize GTP. For example, substrate level oxidative phosphorylation of succinyl-Coil produces GTP; a GTP-linked fatty acid activation enzyme (acyl-CoA synthetase) is present in mitochondria and closely linked to the oxidation of a-ketoglutarate; GTP is utilized during synthesis of P-enolpyruvate from oxalacetate via P-enolpyruvate carboxykinase action. It is unlikely that GTP has the same function as ATP because the latter compound responds very differently in multivalent studies. The only definite conclusion we can draw is that GTP, like the other activators, activates the enzyme for active biosynthesis.
AdenyZutesThere are two suggestions that satisfactorily account for the effect of adenylates on the enzyme.
(a) The adenylate control may reflect the physiological response of the enzyme to the energy state of the cell. Under energy-rich conditions, there would be an ample supply of ATP that could be utilized for biosynthesis.
The biosynthetic reaction of the glutamic dehydrogenase, consequently, would be encouraged. (b) ATP may be involved in an NAD-kinase reaction of the type ATP + NAD+ ti NADP+ + ADP (7) which would assure a continued production of NADP+ for the transhydrogenase couple of isocitric and glutamic dehydrogenases.
Table V summarizes those physiological reactions that may be directly influenced by the activity of the glutamic dehydrogenase. The list is not exhaustive.
We have only presented those interrelations pertinent to the data presented in this communication.

Envoy
Two different mechanisms of allosteric control of NADspecific glutamic dehydrogenases have been observed among members of the simple fungi, Phycomycetes (4, 7). A broad sampling covering some 30 different species and all the major orders of this taxonomic group has been made.3 By a remarkable coincidence, the two complex enzyme control systems are distributed among the fungi in an identical fashion as the only two known pathways of lysine biosynthesis (16). Fungi are unique organisms in that they alone appear to possess these two biosynthetic pathways.
A biochemical rationale for this correlation is not yet evident.
It is hoped that more work on these and related enzymatic reactions of amino acid biosynthesis in these fungi may throw some light on the evolutionary aspect of our observations.