A novel nucleotide implicated in the response of E. coli to energy source downshift

Cell, Vol. 7, 75-84, January 1978, Copyright 0 1976 by MIT

A Novel Nucleotide Implicated in the Response of E. coli to Energy Source Downshift

Jonathan Gallant, Linda Shell, and Rex Blttner Department of Genetics University of Washington Seattle, Washington 98195

Summary

When E. toll cells are subJected to energy source downshift, the accumulation of RNA (and overall ceil growth) is drastically restricted within 1 to 2 min. However, the Identity of the prlmary metabolic slgnal for this adjustment Is a mystery. Earller studies, and further evidence presented here, show that there Is no satisfactory correlation between the sudden adjustment of RNA accumulation and the klnetlcs of changes In the levels of prospective slg- nailing compounds, such as glycolytic Interme- diates, ppGpp, ATP, or the three adenyiate nucleotides. We have discovered an unusual nucieotide, which we call the phantom spot, whose level decreases dramatlcally wlthin a mlnute of downshift, correlating well with the adjustment of RNA accumulation. Preilminary characterization of the phantom spot indicates that it Is a triphosphate derived from the guanylate pathway, and suggests that it is a form of GtP with a modification of the lmidazoie portlon of the purlne rlng. We postulate that this nucieotlde serves as a regulatory facsimile of ATP, linking the rate of RNA accumulation and other anaboilc processes to the overall rate of phosphoryiation.

Introduction

When bacterial cells shift down from growth on a rich carbon and energy source to growth on a poorer one, they exhibit a rapid adjustment whose most familiar sign is a shut-off of RNA accumulation (Neidhardt, 1963; Maalbe and Kjeidgaard, 1966). The regulatory process which effects this (and other) adjustments during downshift is as yet poorly understood. The unusual nucleotide ppGpp is probably one aspect of the process. There is consider- able evidence that ppGpp inhibits the synthesis of stable species of RNA (Travers, Kamen, and Cashel, 1970; Van Ooyen et al., 1975; Block, 1975; Reiness et al., 1975) and downshift does trigger a rapid accumulation of the nucleotide in certain strains (Laz- zarini, Cashel, and Gallant, 1971; Winslow, 1971; Hansen et al., 1975).

On the other hand, it is equally clear that effects of ppGpp cannot be the whole story. This conclu- sion follows from instances where the rate of RNA accumulation and the level of ppGpp are not corre- lated at all. For example, the data of Winslow (1971)

show that ppGpp rises to a high level immediately following downshift in strain CP78 (re/+) and RNA accumulation stops at the same time: however, ppGpp subsequently falls back to its basal level well before RNA accumulation resumes. In this case, therefore, something other than the level of ppGpp must restrict the accumulation of RNA during the later stage of downshift. Hansen et al. (1975) have recently reported a symmetrical case where RNA accumulation resumes, upon reversal of downshift, before any reduction in the level of ppGpp has occurred.

Still another instance is represented by the behavior of certain re/A- mutants. In these mutants, the accumulation of ppGpp triggered by downshift occurs very slowly (Lazzarini et al., 1971; Hansen et al., 1975) much too slowly to account for the prompt cessation of net RNA synthesis. Figure 1

.7 i

.6

$ .5

8 t/ .4-

30,000 .

20,000 a 8

10.000

LL

.

0

8 0.10

8\ b % T$ ‘0

F 0.05

I ** +-x---

c

f20 +40

Figure 1. Kinetics of Net RNA Synthesis and ppGpp during a Glu- cose to Succinate Downshift

Strain NF162 (argA- metS- spoT- W/A-) was cultivated in excess succinate and limiting glucose. At the point indicated by the arrow, an aliquot of the culture was labeled with X-uracil (10 j&i/O.04 pmole/ml) in the presence of 100 pg/ml cytosine, and 100 PL aliquots were removed at intervals to tubes containing 1 ml of 10% trichoroacetic acid (TCA), and the precipitate, separated by filtra- tion, was counted in a liquid scintillation counter; the data are given in cpm/lOO pL sample in the middle panel. (Under these conditions, radioactivity in DNA represents only 4-6% of the total and was not corrected for.) A parallel culture was labeled with ‘2P0, for measurement of ppGpp. Time is specified relative to the sharp break in the OD curve.

Cell 76

is illustrative. In this re/A- mutant, the incorporation of C’4-uracil into RNA stops within 1 min of the sharp break in the growth curve characteristic of downshift. (Similar results have been obtained with 3*PO4 labeling, as in Figure 3, Confirming that UraCil labeling reflects net synthesis of RNA.) Yet the level of ppGpp has scarcely begun to rise at this point (Figure 1). In fact, by 10 min, well after RNA accumulation has stopped, the level of ppGpp has only just reached 0.1 nmole/OD, which is the normal basal level of the nucleotide in actively growing cells of a relA+ strain isogenic with the one used here (Gallant et al., 1970).

The same is true of another kind of energy source downshift illustrated in Figure 2. Here we have subjected the same re/A- strain to a downshift from aerobic to anaerobic growth on glucose. Once again, RNA accumulation halts within a minute or two, while the level of ppGpp rises even more sluggishly, taking 30 min to reach 0.1 nmole/OD. In both cases, it is clear that the rate of RNA accumulation must respond to some signal other than the level of ppGpp. [We are not concerned here with the question of whether the rate of RNA accumulation, in the conditions under discussion, is con-

20,000 1 /$--=--

-20 6 +20 f40 Time I” minufes

Figure 2. Kinetics of Net RNA Synthesis and ppGpp during an Anaerobic Downshift

Strain NF162 was cultivated in excess glucose, and aeration was replaced by Nz:C02 (95%:5%) bubbling at time 0. Methods were as in Figure 1.

trolled at the level of synthesis or turnover. There is some evidence that the adjustments we are consid- ering, where ppGpp does not increase very much, reflect turnover of newly made ribosomal RNA rather than reduced synthesis. This evidence is discussed elsewhere (Gallant and Lazzarini, 1976).1

What might this other signal be? The rapidity with which RNA accumulation responds to downshift and the fact that downshift is in essence a change in metabolic activity both imply that the signal too is metabolic, that is, a change in the level of a small molecule. However, the equivalent response to glucose exhaustion (Figure 1) and anaerobiosis (Fig- ure 2) make it most improbable that the signal is the level of glucose itself or any glycolytic intermediate. Under anaerobic conditions, we would expect glucose transport and the concentrations of glycolytic intermediates to increase rather than decrease as a consequence of the Pasteur effect. Indeed, we have confirmed that this is the case for one representative glycolytic intermediate, fructose diphosphate (Table 1).

Both of the downshifts illustrated above entail a sudden decrease in the rate of high energy phosphate regeneration. Naively, therefore, one might expect the ATP pool size to serve as the signal. However, numerous measurements have shown that the ATP pool is peculiarly stable to metabolic shifts. For example, in the aforementioned study by Winslow, ATP dropped very sluggishly in response to a glucose to lactate shift: a decrease of 10% at most had occurred at the time RNA accumulation stopped (see Winslow, 1971, figure 4). Friesen, Fiil, and von Meyenburg (1975) report a similarly small increase in ATP following upshift from acetate, a poor growth substrate, to rich medium. Greene and Magasanik (1967) found virtually no change in the

Table 1. Fructose Diphosphate Levels during Two Kinds of Downshift

Downshift

(A) Glucose to Succinate

(8) Aerobic to Anaerobic

Time Fructose Diphosphate

- 25 min 6.5 - 5 min 5.6 + 5 min 0.64 +15 min 0.73

- 2 min 6.8 + 1 min 10.0 + 2 min 17.0 + 5 mln 18.0 + 8 min 30.0

Cultures of strain NF162 were cultivated in 3ZPOd-labeled medium in limiting glucose and excess succinate for downshift (A), or containing excess glucose in downshift (B). For downshift (A), time is tabulated in minutes relative to the sharp break in the growth curve characteristic of downshift. For downshift(B), time is tabulated relative to the moment when the cultures’ aeration was replaced by 95% Nz:5% COz. Fructose diphosphate was resolved by method 1. and is reoorted in nmoles/OD unit of cells.

Energy Downshift and the Phantom Spot 77

ATP pool when cells growing on succinate were subjected to anaerobiosis or to azide inhibition. These workers also measured the levels of ADP and

Downshift

s I.01 ( , ,

40 60 80 100 Time in Minutes

Figure 3. Kinetics of Net RNA Synthesis and ATP during a Glucose to Succinate Downshift

Strain NF162 was cultivated in excess succinate and limiting glucose, as in Figure 1. Total PO, in RNA was measured in the labeled culture by correcting total TCA-precipitable radioactivity for radioactivity in DNA, as described by Gallant and Harada (1969).

AMP under both conditions, and found only small changes in the levels of these adenine nucleotides; examination of their data reveals that neither the overall adenylate energy charge (Atkinson, 1970) nor the ratio of ADP to ATP, nor the ratio of AMP to ATP shows more than a slight systematic variation in response to downshift. Similarly, Chapman, Fall, and Atkinson (1971) report only slight changes in adenylate energy charge in the course of several kinds of downshift. Our own measurements, reported in Results, are in close accord with these earlier indications of essential stability in the adenine nucleotide pools.

It is most peculiar. When high energy phosphate generation is restricted, there is a sudden drastic restriction of RNA accumulation and of overall growth rate as well. These adjustments at the mac- romolecular level must be remote effects of a signal provided by metabolites involved in high energy phosphate capture. Yet the signal provided by the adenine nucleotides, the very metabolites we sup- pose to be involved in high energy phosphate capture, is feeble. This discrepancy at least raises the possibility that the metabolic signal of downshift re- mains to be discovered. We have therefore enter- tained the hypothesis that some hitherto unknown nucleotide, produced in reactions paralleling those of ATP regeneration, serves as the signal. (We elab- orate further on this hypothesis in the Discussion). In this paper, we report the discovery of an unusual nucleotide whose level shows large, rapid, and systematic changes in response to several sorts of energy source downshift, a response kinetically’appropriate to the role of the primary downshift signal.

Results

The Response of Adenlne Nucleotldes to Downshift Figure 3 shows RNA accumulation and ATP pool size in cells traversing a glucose to succinate downshift. The results are in close agreement with those previously reported by Winslow (1971) with a different strain and a different metabolic regime. The

Table 2. Adenine Nucleotide Levels during Anaerobic Downshift

Time ATP ADP AMP Energy Charge

- 1 min 7.3 1.2 0.49 0.87

+ 2 min 4.9 1.1 0.36 0.86

+ 4 min 4.0 1.1

+ 8 min 4.8 1.5

+12 min 3.8 0.97 0.41 0.83

+ 20 min 4.8 1.2

A culture of NF162 was grown in excess glucose with 32PQ. At time 0, air was replaced by 95% Nz : 5% CO2 bubbling. Data are reported in nmoles/OD unit of cells.

Cell 78

ATP pool declines rather sluggishly following the break in the growth curve, and a drop of at most 20% has occurred at the point where RNA accumulation stops dead. It is worth emphasizing that even this modest decrease is the largest we have detected in similar measurements on a number of different strains. Given the parade of data cited in the Intro- duction, the point need not be belabored further.

In this kind of downshift, the exact moment of transition is defined somewhat imprecisely. We have therefore also reexamined the downshift from aerobic to anaerobic growth conditions, where the transition in cultural conditions can be defined ex-

Figure 4. Some Nucleotides before and after Glucose to Succinate Downshift

Strain NF162 was cultivated in excess succinate and limiting glucose in W04-labeled medium. The upper panel shows a radioautogram of 20 AL of a sample taken 15 min before the OD break; the lower panel shows a corresponding radioautogram of 20 pL of a sample taken 5 min after the OD break. The nucleotides were resolved by method 2 and exposed to film for 3 days, The dotted circles locate the position of marker adenosine-5’-tetraphosphate, identified by ultraviolet absorption. The heavily labeled spot directly above it is GTP. The spot below and to the left of GTP is ppGpp. The phantom spot is marked P.

actly and more detailed kinetic data are easier to obtain. The results (Table 2) are much the same: ATP drops by a third or so, and the other adenylate nucleotides show little change. In short, there is little signal from the adenine nucleotides.

A Nucleotlde Which Does Respond to Downshift Figure 4 shows two autoradiograms of 3*P04- labeled small molecules extracted from a bacterial culture during the course of a glucose to succinate downshift. The chromatographic separation used moves nearly everything into a vast blob in the upper right hand corner of the plate, resolving only GTP and a few slower moving, super-phosphorylated nucleotides. In the upper autoradiogram, which displays an extract made 15 min before downshift, a compound labeled “P” can be seen below and slightly to the right of GTP. In the lower autoradiogram, which displays an extract made 5 min after the downshift, this compound has disappeared. We call this material the phantom spot (PS). The first set of measurements in Table 3 shows the relative

Table 3. Relative Levels of PS during Three Kinds of Downshift

Downshift Time Relative Level PS

(A) Glucose to Succinate - 10 min 1 .o

- 5 min 1 .o

0 min 0.29

+ 5 min <O.lO

+lO min to.10

(B) Aerobic to Anaerobic - 1 min 1 .o

+ 1 min 0.5

+ 2 min 0.22

+ 4 min 0.16

+ a min 0.20

+12 min 0.15

- 1 min 1 .o

+ 1 min 0.36

+ 2 min 0.16

+ 3 min 0.18 + 5 min 0.17

Strain NF162 was cultivated in 32POJabeled medium under the following conditions: in downshift (A), limiting glucose and excess succinate were present as growth substrates, and time is specified relative to the sharp break in the OD curve; in downshift(B), excess glucose was present throughout, and time is specified relative to the moment when aeration was replaced by 95% NI: 5% CO?; in downshift (C), excess glucose was present throughout, and time is specified relative to the moment of addition of the uncoupling agent carbonyl cyanide metachlorophenylhydrazone (20 pg/ml). The levels of PS are normalized to those measured before downshift: these varied between 0.07 and 0.14 nmoles of PO, per OD unit of cells in different experiments, perhaps reflecting differences in growth rate. Each kind of downshift was performed at least 3 times. with closelv similar results.


levels of PS during the course of a glucose to succinate downshift. It is worth noting that the decline in PS is already evident in a sample taken at the moment of the break in the growth curve.

The level of PS in growing cells is very low, about 40 times less than that of GTP. Its chromatographic behavior in most one- and two-dimensional separations is so similar to that of GTP that it is lost in the GTP spot. The method used in Figure 4, however, resolves PS reasonably well.

Figure 5 shows similar autoradiograms of extracts prepared 1 min before and 2 min after a downshift from aerobic to anaerobic growth on excess glucose. Once again, PS virtually disappears. The sec-

Figure 5. Some Nucleotides before and after Anaeroblc Downshift

Strain NF162 was cultivated in excess glucose in 32P04-labeled medium. Anaerobic downshift was produced by bubbling the culture with N? : CO2 (95% : 5%) in place of air. Chromatography and radioautography were as in Figure 4. Upper panel: 1 min before anaerobic downshift; lower panel: 2 min after anaerobic downshift.

ond set of measurements in Table 3 shows the kinetics of the drop in PS during anaerobic downshift-a 5 decrease within 2 min. It will be re- called that anaerobic downshift results in the accumulation of fructose diphosphate (Table 1) and pre- sumably of other glycolytic intermediates. Thus the level of PS does not seem to reflect the activity of the glycolytic pathway.

Finally, we examined a downshift in which energy generation is specifically uncoupled from oxidative metabolism without altering the availability of either the growth substrate or oxygen. This can be done by adding an uncoupling agent to cells growing aerobically. Figure 6 and the third set of entries in Table 3 show that PS drops as rapidly after uncoupling as it does after anaerobic downshift.

Speclflcity of the PS Response to Downshift Each of the downshift conditions we have explored thus far entails a large reduction in the rate of protein synthesis. It is conceivable, therefore, that the

Figure 6. Some Nucleotides before and after Uncoupling Downshift

Strain NF162 was cultivated in excess glucose in ‘2P01-labeled medium. Oxidative phosphorylation was uncoupled through the addition of carbonyl cyanide metachlorophenylhydrazone (20 pg/ml). Chromatography and radioautography were as in Figure 4. Upper panel: 1 min before uncoupling; lower panel: 2 min after uncoupling.

Cell 80

observed decrease in the level of PS is primarily related to the rate of protein synthesis and is only a secondary reflection of energy source downshift. This interpretation is ruled out by the data presented in Table 4. Here we have blocked protein synthesis by means of inhibitors of serine or tryptophan activation, and it can be seen that the level of PS did not decrease.

Since downshift curtails the accumulation of RNA (by whatever means), another possible interpretation of the drop in PS is that it is a primary consequence of blocked RNA accumulation and again only secondarily related to downshift. The data of Table 4 also argue against this interpretation, for a re/+ strain was used in these experiments to trigger the stringent response. Thus RNA accumulation was also blocked in response to the inhibitors of amino acid activation but, as noted above, no decrease in the level of PS occurred. Similarly, direct inhibition of RNA synthesis by the RNA polymerase inhibitor rifampicin also failed to elicit any decrease in the level of PS (Table 4, column 4).

These experiments show that the dramatic response of PS to energy source downshift is un- likely to occur via reduction in the rate of either protein or RNA accumulation. Rather, they support the view that it is energy input itself which affects the level of the nucleotide.

Most of our initial studies were conducted with NF162, a strain carrying two mutations, relA- and

SPOT-, which affect the metabolism of the MS nucleotides (Cashel and Gallant, 1969; Laffler and Gallant, 1974). However, neither the presence of PS in the acid-soluble pool nor its drastic response to downshift are specifically associated with these genetic lesions, The strain used in the experiments of Table 4, which carries the wild-type alleles of relA and SPOT, produces normal levels of PS during exponential growth. In this strain, energy source downshift also elicits the same large reduction in PS level as we observed in strain NF162. Similar results have been obtained in other strains of relA- SPOT+ and relA+ SPOT- genotypes.

To summarize, PS appears to be a normal com- ponent of the acid-soluble pool in actively growing E. coli; its level responds in a specific manner to energy source downshift, but shows little response to direct inhibition of protein or RNA synthesis; and gene products which affect the metabolism of the MS nucleotides are seemingly not involved in the metabolism of PS. However, we will demonstrate below that the formation of PS is affected by a muta- tion in the guanylate pathway, which provides an important clue to the nucleotide’s chemical com- position.

Preliminary Characterization of PS PS is quantitatively adsorbed by activated charcoal, showing that it contains an aromatic ring constitu- ent; since it also contains phosphate, it is therefore

Table 4. Effect of Inhibitors on Levels of PS

Inhibitor

PS (nmoles PO, per OD unit)

None

0.10

Serine 5-Methyl- Hydroxamate Tryptophan (250 pg/ml) (100 pg/ml)

0.13 0.099

Rifampicin (200 pg/ml)

0.145

A culture of strain Wl was grown in 3zPOJabeled medium. The inhibitors indicated were added to aliquots of the culture, and samples removed for measurement of PS after 5 min of inhibition in the case of the inhibitors of protein synthesis, and after 13 min in the case of rifampicin.

Table 5. Resoonse of PS and Other Guanine Nucleotides to Guanosine Starvation

Growth Condition

nmoles of ‘2PO., per Optical Density Unit in:

PS GTP PPGPP

(A) Exponential growth

(B) Guanosine starved for 30 min

(C) 20 min after auanosine restoration

0.13 2.8 0.38

‘zo.01 0.36 0.012

0.28 2.0 0.32

A culture of CP78GB2, a guaS mutant, was cultivated in ‘zPO,-labeled medium containing limiting guanosine (0.03 pmoles/ml). Samples were removed for analysis during exponential growth before the guanosine had been exhausted (A); after 30 min of guanosine starvation (B); and 20 min after resupplementing the culture with excess (0.1 amole/ml) guanosine (C). Guanosine exhaustion was signalled by a sharp break in the growth curve.


probably a nucleotide. The level of PS decreases, along with that of representative guanine nucleotides, when a guaB mutant is starved of guanosine, and rises with them upon restoration of guanosine (Table 5).

This result suggests that PS is derived from the guanylate pathway, an inference confirmed by the

Figure 7. Some Nucleotides Labeled by W-Guanosine

Strain CP76GB2 was cultivated in excess glucose and limiting guanosine (0.03 pmole/ml). Exhaustion of guanosine was signalled by a sharp change in the OD curve from exponential growth to linear growth at a 4 fold reduced rate, as reported by Gallant and Harada (1969). After about 60 min of guanosine starvation, an aliquot of the culture was labeled with r*C-guanosine (50 pWO.1 pmole/ml). Resumption of exponential growth at the normal rate was confirmed by OD measurements. After 27 min, 2 ml were acidi- fied with formic acid in the usual way, neutralized with 10% NHIOH after 30 min in the cold, and centrifuged. The supernatant was concentrated 10 fold by lyophilization, and 40 pL of the concentrated extract were chromatographed by method 2. The figure shows the radioautogram produced by 11 days exposure to film. (Similar results, showing the labeling of PS by r4C-guanosine, have been obtained in five separate experiments.) Label in GTP, ppGpp, and pppGpp is also evident in the radioautogram.

following observations. First, label from X-guanosine finds its way into PS (Figure 7). Second, the entry of ‘4C-guanosine into PS shows the competi- tion characteristics of a compound made from the guanylate rather than the adenylate pathway. When duplicate cultures were labeled with ‘4C-guanosine in the presence and absence of an excess of cold adenosine, the competitor reduced the labeling of ATP and GTP by factors of 4.65 and 1.3, respectively. The labeling of PS was reduced by a factor of 1.3, in exact agreement with that for GTP.

When duplicate aliquots were labeled with ‘4C- guanosine and 32P04, the ratio of r4C to 32P found in PS was 0.95 of that found in GTP, 0.77 of that found in ppGpp, and 0.65 of that found in pppGpp. These values are in good agreement with those expected (1 .O, 0.75, and 0.60, respectively) for a triphosphate. Thus labeling data indicate that PS is a triphosphate of a nucleoside derived from guanosine.

To our surprise, we were unable to detect signifi- cant tritium labeling of PS by either 8-JH-guanosine or 8-3H-adenosine, where abundant labeling of GTP was detected in the same experiments. Given the unequivocal evidence that carbon atoms of the guanosine ring enter PS, this can only mean that the 8 position is modified in PS or else that substitution elsewhere in the ring renders the 8 position labile to exchange.

We have prepared small quantities of 32P-labeled PS in pure form by eluting it from thin layer chromatograms in 10% NH40H. The compound is sensitive to hydrolysis by alkaline phosphatase and by snake venom phosphodiesterase. After prolonged incuba- tion with venom phosphodiesterase, the 32P label is found mainly in PO4 and in another compound with the chromatographic properties of a nucleoside monophosphate, although it does not comi- grate with either AMP or GMP. These properties

Table 6. Rr Values of PS and Other Purine Nucleotides

Separation System

Nucleotide A 0 C D

PPPPA 37 51

PPPA 17 9 58 46

PPA 36 39 39

PA 52 a3 32

PPPPG 20 76

PPPG 13 3 41 69

PPG 24 15 67 60

PG 42 52 77 54

PS 7 3 49 65

Purified ‘ZPQ-labeled PS was chromatographed together with marker nucleotides in the separation systems indicated. PS was located by radioautography, the marker nucleotides by ultraviolet absorption. Distances migrated were measured from the middle of each spot, and Rr values are given as percentages. The separation systems are as follows: (A) the first dimension of method 1, on PEI-cellulose; (6) the second dimension of method 1, on PEI-cellulose; (C) the second dimension of method 2, on PEI-cellulose; (D) the ammonium sulphate-sodium acetate-isopropanol solvent described by Randerath (1962), on cellulose.

Cell a2

strongly suggest that PS is a nucleoside polyphos- phate. Moreover, the sensitivity to venom phoso- phodiesterase suggests that the 3’ position is not esterified. In agreement, we find that periodate oxi- dation alters the chromatographic behavior of PS.

Chromatography of purified PS in a variety of solvent systems provides further information as to its chemistry. On layers of the ion exchanger PEI- cellulose, the Rt values of nucleotides vary inversely with the number of phosphate groups within each homologous series, and adenine nucleotides migrate faster than corresponding guanine nucleotides. Table 6 shows that PS migrates most like GTP on PEI-cellulose in several solvent systems. Another separation method, on cellulose layers, involves different chemical specificities, since the rules are re- versed: Rt values increase with the number of phosphate groups, adenine nucleotides migrate slower than corresponding guanine nucleotides (Ran- derath, 1962). In this separation system, PS again migrates most like GTP (Table 6).

The chromatographic properties of PS are thus similar but not quite identical to those of GTP. Taken together, our results suggest that PS is a modified form of GTP. Its sensitivity to venom phosphodiesterase and to periodate rules out the possibility that it is a 3’ phosphorylated guanylate like the MS nucleotides. The failure of 84ritiated adenosine or guanosine to label PS indicates some funny busi- ness in the purine ring at the 8 position or perhaps nearby. These properties are suggestive of a form of GTP with a substituent on the imidazole part of the purine ring.

Discussion

We began by pointing out that the control of RNA accumulation during downshift could not be fully accounted for by variation in the level of ppGpp. Further consideration indicated that a metabolite must be employed as the intracellular signal of downshift, but that no known metabolite seemed to vary with appropriate kinetics. In particular, neither the concentration of ATP nor the relative levels of the three adenylate nucleotides show more than quite slight changes during various kinds of downshift. An electronics engineer could no doubt devise an amplification system by which a 10 to 30 percent drop in the ATP pool might serve as the signal for a 10 to 30 fold change in the rate of controlled processes. But there is little warrant for assuming that intracellular control systems involve such amplification. On the contrary, in certain control systems where the signals and their effects are known to a degree, there is no signal amplification: the levels of CAMP and of ppGpp, for example, show varia-

tions approximately commensurate with the rate changes they bring about. Yet no metabolic signal of downshift has hitherto been observed which is remotely commensurate with the large observed change in the rate of RNA accumulation.

The relative constancy of the ATP pool, in the face of restricted ATP regeneration during downshift, is itself another instance of the problem. Clearly, the rate of ATP expenditure must be prompt- ly adjusted to decreases in the rate of ATP regeneration, for otherwise much larger fluctuations in the adenylate pool parameters would be observed during such stresses as downshift. But what is the signal to which this adjustment responds? A conven- tional feedback design leads to the same sort of dilemma as that noted above: the magnitude of the adjustment is grossly out of line with the magnitude of the feeble signal provided by the adenylate pool parameters themselves.

A simple way out of the dilemma could be provided by feed-forward control of the kind discussed, for example, by Milsum (1966). By feed-forward we mean a control element which responds to the rate of phosphorylation itself, rather than to the level of ATP, which is the output of phosphorylation. Figure 8 shows a biochemical model for such a feed-forward design.

The model postulates that a portion of the cell’s phosphorylating capacity is tapped off to generate a regulatory facsimile of ATP, designated pppX in the figure. The level of this nucleotide (perhaps in relation to its precursor ppX) then regulates a range of anabolic processes, RNA accumulation among them, which consume ATP. We postulate that pppX acts as a positive effector of these processes. Thus

LpppA/ ‘t

Regulbtion I

L i : pppx ---4’

Figure 8. A Feed-Forward Model for the Control of ATP Expenditure

The model postulates that phosphorylating capacity, symbolized as zP at left, is used in two parallel sets of reactions. The major set, symbolized by the upper heavy line, produces the familiar phosphorylation of ppA to pppA. A second set, symbolized by the lower thin line, produces the phosphorylation of a hypothetical nucleoside diphosphate ppX to its corresponding triphosphate pppX. The level of pppX then regulates some of the anabolic reactions driven by the hydrolysis of pppA.

Energy Downshift and the Phantom Spot a3

when the rate of phosphorylation falls, the decline in the level of pppX brings about a reduction in the expenditure of ATP and thus serves to stabilize the relative levels of ATP and ADP.

The advantage of this design is that it would per- mit the cell to link anabolic processes to the rate of energy input without suffering the possible disad- vantages of large fluctuations in the ATP pool or the adenylate energy charge. There are some prec- edents for such a design. Tomkins (1975) recently pointed out that certain broadly integrative intracellular control systems operate through effector molecules (cyclic AMP and ppGpp) which are not themselves essential metabolites, but rather function as “symbols” of environmental sufficiency. In effect, our model is simply a translation of Atkinson’s con- cept of energy charge into Tomkins’ language of regulatory “symbols”.

Departing from these considerations, we set out to search for a regulatory symbol of energy input. We studied three metabolically distinctive forms of downshift to screen out of consideration compounds which were trivially related to patterns of carbon source catabolism. For example, compounds related to the glycolytic pathway would be expected to decrease during a glucose to succinate downshift, but increase during an anaerobic downshift, just as fructose-diphosphate does (Table 1). A compound related to the oxidative shunt might decrease under both of these circumstances, but should not decrease in response to uncoupling, since oxidative metabolism proceeds. In contrast, a symbol of energy input should decrease during each of the three forms of downshift, just as PS does.

The results presented in Table 4 argue that the level of PS does not respond to the rates of protein or RNA accumulation as primary determinants, but rather to some more elementary input from energy metabolism, again in agreement with our hypothesis. Finally, the evidence that PS is a purine ribonucleoside triphosphate conforms to the hypothesis in strict chemical terms.

Further tests may, of course, falsify elements of the model presented in Figure 8; that is what models are for. From a more general point of view, however, the specific response of PS to energy source downshift suggests that this nucleotide represents a new clue to the mechanism of the downshift adjustment.

The ideas presented here are closely related to the more comprehensive synthesis advanced by Tomkins (1975). Tomkins drew attention to the roles of cyclic AMP and of ppGpp as “symbols” of particular kinds of environmental stress. We would add that the environmental conditions they symbolize occur in subsystems of cellular activity distinct from the “domains” of cellular activity which they con-

trol. Thus, for example, ppGpp is made in a ribosomal reaction in response to the presence of un- charged tRNA (Block and Haseltine, 1974) but its “domain” includes a variety of reactions outside of the protein synthesis subsystem (Cashel and Gal- lant, 1974). Thus Tomkins’ “symbols” can also be understood as metabolic devices for channeling information from one intracellular subsystem to others. In this context, our model postulates that PS serves to channel information from the subsystem of energy metabolism to that of RNA accumulation, among others.

Both cyclic AMP and ppGpp are modified nucleotides, so are cyclic GMP and pppGpp; so is the phantom spot. Is it mere coincidence that all of these compounds are modified nucleotides? We wonder. The biochemical literature includes scat- tered reports of other funny nucleotides (Su and Hassid, 1982; Finamore and Warner, 1983; Hepner and Smith, 1967; Zamecnik and Stephenson, 1969; Gallant and Margason, 1972; Rhaese, Dich- telmiiller, and Grade, 1975; Le John et al., 1975; Hamagishi, Nishino, and Murao, 1975), and we have detected still others in the course of the present studies. Perhaps it is time to inquire into the roles of these eccentric compounds in the regulatory net- work of the cell.

Experimental Procedures

The following strains of E. coli were used in these studies: NF162 (arg, met, relA-, spoT-); WI (arg, his, pro, thr, leu, fho; and CP76GB2 (arg, his, fhr, leu, fhi, gua). In other experiments alluded to in the text, although not reported in detail, we used NF161, a re/A+ strain isogenic with NF162, and W2, a SPOT- strain isogenic with Wl

Bacteria were cultivated under forced aeration in a Tris-minimal medium containing 0.5 mM phosphate. Glucose to succinate downshifts were performed by growing the bacteria in the presence of growth-limiting concentrations of glucose (0.03-0.06% in different experiments) and excess succinate (0.3%), as described by Gaf- lant, Margason, and Finch (1972). Growth was monitored by mea- suring optical density at 720 nm in a Beckman DB spectrophoto- meter: an OD of 1 corresponds to about 109 cells per ml or about 250 ug/ml protein.

RNA accumulation was measured either by ‘ZPO, incorporation or by I’C-uracil incorporation, as described in the figure legends.

For measurement of nucleotides and other small molecules, the cells were equilibrated with JzPO4 for at least one doubling. Extracts were prepared by adding formic acid to a final concentration of 0.1 N. After about an hour in the cold, the extracts were neutralized by addition of excess Tris buffer or, in some cases, 10% NH40H, centrifuged, and the supernatants stored at -2O’C for analysis. In experiments which involved labeling with ‘4C-guanosine, the extracts were concentrated 10 fold by lyophilization.

Nucleotides and other small molecules were resolved by thin layer chromatography on layers of PEI-cellulose (Randerath, 1966). The following solvent systems were used. Method 1: three steps of LiCl in the first dimension, followed by three steps of sodium formate in the second dimension, as described by Randerath (1966, Figure 66). Method 2: 2 M formic acid containing 1.5 M LiCf in the first dimension, followed by 1.5 M KHzPOe in the second dimension. Method 3: 4 M sodium formate (pH 3.4) containing 0.5 M

Cell 84

LiCl in the first dimension and 1.5 M KHlPOa in the second dimension

Method 1 gives good resolution of most of the ribonucleoside mono-, di-, and triphosphates, as well as of several sugar phos- phates and other compounds (see Randerath. 1966; Cashel and Gallant, 1969); we used this method to resolve ATP, ADP, AMP, and FDP. Methods 2 and 3 resolve only the slowest moving species, such as GTP, the MS nucleotides, and PS. The only difference between the two systems is that the first dimension Rr values are somewhat greater in method 2, with the result that PS is better resolved from GTP, but the GTP and ppGpp spots are somewhat more diffuse. We invariably used method 2 to measure PS, in some cases checking the results with duplicate separations by method 3. GTP and ppGpp were invariably resolved by method 3, sometimes checked as well by method 2.

Labeled compounds were located by radioautography and identified (where known) by comigration with authentic standards. All data were corrected for nonspecific radioactive trailing and counting background either by counting blank regions of the chromato- gram near the compound of interest, or else by chromatographing the supernatants of Norite-adsorbed control extracts and counting appropriate regions of these control chromatograms. These cor- rections are negligible for major compounds but can be quite signif- icant for rare nucleotides such as PS. In the latter case, the blank correction typically constituted 1 O-30% of the uncorrected radioactivity in samples taken from actively growing cells. Obviously, the level of PS can be determined with reasonable accuracy in growing cells, but not in cells where its level has declined by a large amount. For this reason, we have illustrated the decline in PS with the visual evidence of autoradiograms as well as with numbers.

Acknowledgments We are indebted to Janet Walsh and Gary Parker for expert techni- cal assistance in several of the experiments, This work was sup- ported by grants from the NIH and ACS.

A preliminary account of some of this work was presented at the Alfred Benzon Symposium IX, “Control of Ribosome Synthe- sis,” to be published by Munksgaard (Copenhagen) under the edi- torship of N.-O. Kjeldgaard and 0. Maalbe.

Received October 17, 1975

References

Atkinson, D. E. (1970). In The Enzymes, I, third edition, P. Bayer, ed. (New York: Academic Press), pp. 461-489.

Block, R. (1975). In Alfred Benson Symposium IX, Control of Ribo- some Synthesis, N.-A. Kjeldgaard and 0. Maalbe, eds. (Copenha- gen: Munksgaard), in press.

Block, R., and Haseltine, W. A. (1974). Ribosomes (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory), p. 747.

Cashel. M., and Gallant, J. (1969). Nature 227, 838-841.

Cashel, M., and Gallant, J. (1974). Ribosomes(Cold Spring Harbor, New York: Cold Spring Harbor Laboratory), p. 733.

Chapman, A., Fall, L., and Atkinson, D. E. (1971). J. Bacterial. 108, 1072.

Finamore, F. J., and Warner, A. H. (1963). J. Biol. Chem. 238, 344.

Friesen, J., Fiil, N., and von Meyenburg. K. (1975). J. Biol. Chem. 250, 304.

Gallant, J. and Harada, 8. (1969). J. Biol. Chem. 244, 3125.

Gallant, J. and Lazzarini. R. (1976). Protein Synthesis: A Series of Advances, II, E. McConkey, ed. (New York: Marcel Decker).

Gallant, J., and Margason, G. (1972). J. Biol. Chem. 247, 2289.

Gallant, J., Margason, G., and Finch, B. (1972). J. Biol. Chem. 247. 6055-6058.

Gallant, M., Erlich, H., Hall, B., and Laffler, T. (1970). Cold Spring Harbor Symp. Quant. Biol. 35, 397-405.

Greene, R., and Magasanik, B. (1967). Mol. Pharmacol. 3, 453.

Hamagishi, Y., Nishino, T., and Murao, S. (1975). Agric. Biol. Chem. 39, 1015.

Hansen, M., Pato, M. L., Molin, S., Fiil, N. P., and von Meyenburg, K. (1975). J. Bacterial. 722, 585.

Hepner. A. S., and Smith, M. (1967). Biochem. Biophys. Res. Com- mun. 26. 584.

Laffler, T., and Gallant, J. (1974). Cell 3, 47-49.

Lazzarini, R., Cashel, M., and Gallant, J. (1971). J. Biol. Chem. 246, 4381.

Le John, H. B., Cameron, L. E., McNaughton, D. R., and Klassen, G. R. (1975). Biochem. Biophys. Res. Commun. 66, 460.

Maalbe, O., and Kjeldgaard, N.-O. (1966). Control of Macromolecu- lar Synthesis (New York: Benjamin).

Milsum. J. H. (1966). Biological Control Systems Analysis (New York: McGraw-Hill).

Neidhardt, F. C. (1963). Biochem. Biophys. Acta 68, 365.

Randerath, K. (1962). Biochem. Biophys. Res. Commun. 6, 452.

Randerath, K. (1966). Thin-Layer Chromatography (New York and London: Academic Press).

Reiness, G., Yanz, H.-L., Zubay, G., and Cashel, M. (1975). Proc. Nat. Acad. Sci. USA, in press.

Rhaese, H.-J., Dichtemuller, H., and Grade, R. (1975). Eur. J. Bio- them. 56, 385.

Su, J.-C., and Hassid, W. Z. (1962). Biochemistry 1, 474.

Tomkins, G. M. (1975). Science 189, 760.

Travers, A., Kamen. R., and Cashel, M. (1970). Cold Spring Harbor Symp. Quant. Biol. 35, 415.

Van Ooyen, A. J. J., De Boer, H. A., Ab. G., and Gruber, M. (1975). Nature 254, 530.

Winslow, R. (1971). J. Biol. Chem. 246, 4872.

Zamecnik, P. C., and Stephenson, M. L. (1969). In Alfred Benron Symposium I, The Role of Nucleotides for the Function and Confor- mation of Enzymes (Copenhagen: Munksgaard).

Note Added In Proof Ideas about the biological significance of abnormal nucleotides and related compounds have also been advanced by Bruce Ames (personal communication) and by Emil Fischer 11907, Untersu- chungen in der Puringruppe (Berlin: Springer Verlag)].

Date post:	31-Dec-2016
Category:	Documents
Upload:	rex
View:	214 times
Download:	2 times

A novel nucleotide implicated in the response of E. coli to energy source downshift

Documents