THE JOUR~VAL OF RIOL~GICAL CHEMISTRY Vol. 24i, No. 22, Issue of November 25, pp. 7430-7438, 1972

Printed in U.S.A.

Biochemistry of Deoxyribonucleic Acid-defective Amber

Mutants of Bacteriophage TP

III. NUCLEOTIDE POOLS*

CHRISTOPHER I<. MATHEWS

(Received for publication, June 29, 1972)

From the Department of Biochemistry, University of Arizona, College of Medicine, Tucson, Arizona 85724

SUMMARY

Nucleotide pools were analyzed under different conditions of infection by bacteriophage T4, for three reasons: (a) as an approach to identifying the defective functions associated with conditional lethal phage mutants bearing uncharacterized lesions in DNA replication; (b) to gain information about the control of nucleotide synthesis in phage-infected Escherichia coli; and (c) as an aid to interpretation of earlier data (MA- THEWS, C. K. (1968) J. Bid. Chem. 243,561O) which suggested that the rate of total RNA synthesis decreases when DNA replication is blocked. In infection by wild type T4D there are no significant changes in ribo- or deoxyribonucleoside triphosphate pool sizes, except for replacement of dCTP by S-hydroxymethyl-dCTP. By contrast, infection by DNA- negative mutants causes up to do-fold pool expansions of dATP and dTTP, with hydroxymethyl-dCTP accumulating to similar extents. The dGTP pool, however, does not expand significantly, suggesting that the synthesis of this nucleotide is regulated more closely than those of the others. Ribonucleo- side triphosphate pools do not expand dramatically in infec- tion by mutants in genes 42, 44, or 62, but up to S-fold pool expansion is seen in infection by mutants in genes 41 and 45. These studies indicate that the products of genes 41, 44, 45, and 62 do not directly affect the synthesis of DNA precursors. Moreover, comparative studies on the rates of labeling of RNA precursors late in infection support the earlier reported conclusion that the rate of RNA synthesis decreases consid- erably when DNA synthesis is blocked. Finally, the data allow calculation of the adenylate energy charge and of ap- proximate concentrations of nucleotides in exponentially growing E. coli. The energy charge is about 0.9, in agree- ment with published work, and this value does not change significantly under the conditions of infection examined. The intracellular concentrations of nucleoside triphosphates range from 0.1 mu for dGTP to 2.7 mu for ATP. These values allow one to ask which of various enzyme feedback effects observed in uifro might reasonably be assumed to be operating in vivo.

* This research was supported by Grant AI-08230 from the National Institute of Allergy and Infectious Diseases, and Grant 70 1013 from the American T-Ieart Association. Paper II in this series is Reference 6.

The genome of bacteriophage T4 contains some 30 genes known to be involved, either directly or indirectly, in viral DNA synthesis (l-3). Some of these genes code for enzymes of DNA precursor synthesis, such as thymidylate synthetase or ribo- nucl.eoside diphosphate reductase. Others code for enzymes of DNA metabolism at the macromolecular level. such as DNA ligase or 5-hydroxymethylcytosine-glucosyltransferases. At least one gene, gene 32, codes for a nonenzymatic protein es- sential to DNA replication and recombination (4). Interest in this laboratory (5, 6) is focused upon those genes, particularly 41, 44, 45, and 62, whose products have not been identified either in terms of reaction catalyzed or specific function in DNA replication. Amber mutants in genes 44 or 45 are classified as DNA-negative, or DO, because phage DNA synthesis is virtu- ally undetectable under nonpermissivc conditions of infection (2, 5). Mutants in genes 41 and 62 have been classified by Warner and Hobbs as “DS” (2) because a small amount of DNA synthesis does occur after infection. None of the products of these four genes have been characterized, although it has been suggested that the gene 44 product is involved in initiation of replication (6), and that the gene 41 product participates along with DNA ligase (gene 30) in the conversion of nascent DNA fragments to high molecular weight DNA (7, 8).

By way of asking whether genes 41, 44, 45, or 62 control hitherto undiscovered steps in DNA precursor synthesis, I have determined nucleotide pool sizes following infection of Escke- richia coli B with T4D (wild type) and appropriate conditional lethal mutants, as reported in this paper. Previous workers, notably Warner and Hobbs (9) and Elan1 and Koerner (lo), have studied nucleotide pools in T-even phage-infected cells. However, Elam and Koerner studied only the nucleoside mono- phosphate pools and they confined their studies to wild type T2 phage, while Warner and Hobbs converted all nucleosidc polyphosphates to monophosphates prior to separation and analysis. Thus, although they did observe deoxynucleotide pool changes in infection by various T4 mutants, they could not tell whether the changes were occurring primarily in pools of mono-, di-, or triphosphates. Moreover, because they used [r4C]uracil to label nucleotides being analyzed, their data pro- vided information only about t,he pyrimidine nucleotides. In the present study I have used Y’ to label all nucleotides and have used thin Iayer chromatography on PEI-cellulose (ll), t,o obtain data primarily on nucleoside triphosphate pools, al-

7430

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

de nova -~ synthms

I +

U-UMP+UDP.UTP

4,

dUTrdUPP dCfP.dCTP

/ \,+

dUMP &-dCtk A

dTMP dHMP

i 4 dTDP dHDP

Strain Mutant type Mutant gene product

T4D Wild type N81 am41 Unknown N122 am42 Deoxycytidylate hydroxymethylase L13 is42 Deoxycytidylate hydroxymethylase N82 am44 Unknown El0 am45 Unknown El140 am62 Unknown

-

7431

TABLE I

Phage siTaim used in this paper

i \1 dTTP dHTP

J J DNA

Icro. 1. Pathways of pyrimidine nucleotide metabolism in T4 phnge-infected Escherichia coli. In the upper left-hand corner U represents exogenous umcil. Reactions marked with an asterisk are those known to be regulated by nucleoside triphosphates in vitro.

though some data are presented as well on ribonucleoside mono- and diphosphate pools.

Two additional reasons for our interest in nucleotide pool analyses arc as follows: (a) Several phage-coded enzymes of nucleotide metabolism, notably ribonucleoside disphosphate reductase (12, 13), deoxycytidylate deaminnse (14, 15), and possibly thymidine kinase (IQ, exhibit complex feedback activa- tion and inhibition effects in vitro in the presence of various nucleotides. By learning the intracellular concentrations of these nucleotides, and their fluctuations under various conditions of infection, one might ask whether effects observed in vitro might actually be physiologically significant; (b) Several years ago, on the basis of pulse-labeling data with [14C]uracil, I re- ported that the rate of RNA synthesis decreases at least IO-fold late in infection by DO mutants when compared with infection by wild type T4. However, these data could be explained just as easily by proposing that blockage of DNA synthesis leads to accumulation of both deoxyribonucleotides and ribonucleotides (see Fig. 1). Thus, in a pulse-labeling experiment the intracel- lular RNA precursor pools would be of lower specific activity in a DO infection than under normal conditions, and radioactive label incorporated into RNA would be decreased even if the actual rates of RNA synthesis were identical. Since the puta- tivc RNA-DNA coupling is a phenomenon of some interest, it secmed desirable to determine unambiguously whether my earlier conclusion was correct. Data in the present paper in- dicate that it was.

ISXPJ’RIA~ENTAL PROCEDURE

Bscherichia coli CR63 was used as a permissive host for propa- gation and assay of amber mutants of T4, and i’?. coli B was used as nonpermissive host for all experiments described in this paper. T4D and all phage mutants used came originally from the lab- oratory of Dr. R. S. Edgar. Table I lists all phage strains used, both by strain number and mutant type. For example, a strain of mutant type am44 refers to a phage bearing an amber muta- tion in gene 44, while ts42 refers to a phage bearing a tempera- ture-sensitive mutation in gene 42, and so forth. This nomen- clature will be used throughout the paper to identify strains used in each experiment both by mutant gene and type of mu-

tation. All strains used were purified genetically in this labora- tory by backcrossing at least twice against T4D.

The culture medium used for all experiments was Medium A, a low-phosphate Tris-glycerol casamino acids medium de- scribed by Kaempfer and Magasanik (17). The only phos- phorus in this medium is the phosphate present in the casa- mino acids. The actual phosphate concentration of each batch of medium was determined by the method of Lowry et al. (18) and was usually about 0.1 mM. Unless otherwise specified, cells were grown and infected at 37”, under which conditions E. coli B has a mass doubling time of about 50 min. Phage lysates were prepared by growth and infection of cells in phosphate-buffered glycerol-casamino acids medium (Medium B) (19). Phage were purified by differential centrifugation and treatment with DNase and RNase. and the final phage pellets were resuspended in Medium A.

All nucleotides used as chromatographic markers were pur- chased from P-L Biochemicals, except for dHTP,l which was generously supplied by Dr. A. Kornberg. H332P04, [2-YJuracil, and [5-3H]uracil were purchased from New England Nuclear. Thin layer chromatography was carried out on 20 x 20 cm sheets of PEI-cellulose, O.l-mm layers backed by heavy alumi- num foil, supplied by Brinkmann Instruments.

Labeling and extraction of nucleotides was carried out as follows. Bacberial cultures were grown with forced aeration for at least 2 36 generations, to about 3 x lo8 ml-l, in Medium A containing 32Pi at 10 &i per ml. The specific activity of phos- phorus in the medium was about 100 &i per pmole, a level previously shown (20), and confirmed in these studies, not to affect the growth rate of the bacteria. The reason for the rel- atively long growth period in radioactive medium is the fact that a considerable fraction of the deoxyribonucleotide pool in phage-infected cells is derived from the breakdown of bacterial DNA (lo), and it was necessary to bring the specific activity of this DNA as nearly as possible to the corresponding value for the acid-soluble pool. The actual specific activity of the medium in each experiment was determined from its phosphate con- centration and its radioactivity, determined under counting conditions identical to those used for radioactive nucleotides.

Cells were sampled for viable count, then infected with phage at an average multiplicity of five to six, and at the indicated time intervals 5-ml aliquots were removed for extraction of nucleotjdes as follows. Each sample was passed rapidly through a 0.45.pm Millipore filter, which was immediately placed in 2.0 ml of 0.4 M ammonium format’e buffer, pH 3.0, in an ice bath

1 The abbreviations used are: dHMP, dHDP, and dHTP, 5- hgdroxymethgldeoxycytidine 5’.mono-, di-, and triphosphates, r&pectively; -rNMP, rNDP, and rNTP, ribonucleoside mono-, di-, and triphosphate, respectively; dNTP, deoxyribonucleoside triphosphate; HMC, 5-hydroxymet.hylcytosine.

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

7432

The cell layer was scraped from the filter and resuspended with a flat-ended spatula, and each mixture was allowed to sit at ice temperature, with occasional gentle agitation, for 1 hour, fol- lowing which it was centrifuged at 4” for 10 min at 3000 x 9 and the precipitate discarded.

In some early experiments the ammonium formate buffer used for extraction was at pH 3.4. While this work was in progress Bagnara and Finch (21) reported that the extractability of nucleotides from E. coli by sodium formate buffers decreases sharply above pH 3.0. I found no significant differences in nucleoside triphosphate recoveries between ammonium formate buffers at pH 3.0 or pH 3.4. However, both ammonium formate extraction techniques gave considerably higher recoveries, in my hands, than extraction with perchloric acid, either by the recommended technique of Bagnara and Finch (21) or by the method of Neuhard and Thomassen (22).

Nucleotides in each extract were separated by two-dimensional thin layer chromatography on PEI-cellulose sheets (11). Ali- quots of each extract were mixed with aqueous solutions of marker nucleotides at 0.01 M each. Each mixture was applied as a small spot at a location 2 cm in from each edge, near a corner, such that the final application was 20 ~1 of extract and 0.05 pmole of each marker. The sheets were washed in methanol (11) and dried. For chromatography of di- and triphosphates a Whatman No. 3MM filter paper wick was stapled to one edge of each sheet, and the solvent for the first dimension, 1.0 M Licl saturated with boric acid and adjusted to pH 7.0, was allowed to run up the sheet and about 3 cm onto the wick. The sheets were removed and dried, each wick was cut off and discarded, and the sheets were washed in methanol and dried again. Chro- matography, in the second dimension employed the ammonium sulfate system of Randerath and Randerath (II), or, when it was desired to quantitate diphosphates as well as triphosphates, the stepwise formate buffer elution system (11). Nucleoside mono- phosphates were resolved by Procedure 2 of Randerath and

Randerath (11). In general G;1IP and GDP were poorly re- solved in the solvent systems used, and data for these nucleotides are not shown.

After chromatography, spots of marker nuclcotides were located under ultraviolet light. Correspondence between radioactive material and ultraviolet light-absorbing material was verified by radioautography. Each chromatogram was placed against 3-1M type R x-ray film for at least 16 hours. Pro- vided that radioactive areas coincided with marker nucleotide areas, each spot containing a nucleotide of interest was cut out with scissors, placed in a counting vial along with 5 ml of a toluene-based phosphor (5), and counted in a Beckman LS-250 liquid scintillation spectrometer. It was found that recovery of radioactive material is about 20% higher when radioactive spots were cut out than when the adsorbent containing radioactive material was scraped off the surface of a sheet and placed in a counting vial.

From the radioactivity of each nucleotide, the specific act,ivity of phosphorus in the medium, and the ceI1 concentration at the time of infection, it was possible to calculate each nucleotide pool size as nanomoles per lOlo cells. For presentation in this paper, these values have been converted to molecules of nucleotide per cell.

RESULTS

Ttiphosphate Pools-Fig. 2 shows the ribo- and dcoxyribonu- cleoside triphosphate pools as a function of time after infection by T4D and mutants N82 (am44) and N122 (am42). Except for the replacement of dCT1’ by dHTP, there are no significant changes in triphosphate pool sizes after T4D infection. The failure of deoxyribonucleoside triphosphate pools to expand is of interest in light of the much higher rate of DNA synthesis in phage-infected, as compared with uninfected, cells (5, 23). Pre- sumably the dNTP pools are maintained constant after infection by a much higher rate of turnover, although this has not, yet been

24.

u T4D rNTP’s 382 rNTP’s N122 rNTP’s k R 32-

ATP

/ .y. ATP

l 1 o-• ;;/*

-.a ,e-+TP

eAO----l -,

‘*

,aJJJP ,= I /----a-----m. .GTP

-.----.CTP I

0 10 20 0 IO 20 0 ‘I 0 20

Minutes Af tw Infection

FIG. 2. Nucleoside triphosphate pools following infection of Es&~-i&a coli B by T4D (Z&-hand panels), N82 (anl4J; cenler panels), or N122 (am42; right-hand panels).

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

7433

established directly. Such a determination would be of value in supporting the idea that 5’ dNTPs are the immediate precursors for DNA replication, an idea which has recently been challenged (24).

By contrast, infection with DNA-negative mutants leads to dramatic expansion of dNTP pools, with relatively little change in the rNTP pools. In the N82 experiment depicted in Fig. 2, the dATP and dTTP pools expanded some 30.fold, while the dHTP pool rose from zero to a similar level. However, dGTP did not a accumulate to a significant extent. The failure of dGTP to accumulate is not a specific consequence of the gene 44 mutation, for the same phenomenon was observed in infection by all DNA-defective mutants studied. This observation is reminiscent of the finding of Neuhard (25) that when DNA synthesis is blocked in uninfected E. coli by thymine starvation, the pools of dXTP and dCTP expand considerably, but that of dGTP does not change. Thus, it seems that in both infected and uninfect,ed cells the synthesis of dGTP is regulated more stringently thau synthesis of the other dNTPs, although it has not yet been ruled out for infected cells that failure of dGTP to accumulate might result from increased turnover rather than decreased synthesis. In the case of uninfected E. coli, Neuhard and Thomassen (22) have demonstrated that dGTP turns over actually less rapidly than dCTP and dATP while the latter are accumulating.

Iu infection by N122 no detectable dHTP is fornled, as ex- pected from the mutation in gene 42, which leads to failure of these cells to form active dCMP hydroxymcthylase. However, in other regards the pattern seen is similar to that observed with N82.

From the experimentally determined pool sizes, the number of nucleotides per T4 DNA molecule (approximately 4 X 105), and the nucleotide composition of ‘1’4 DNA (26), one can esti- mate the amount of each dNTP accumulating, in terms of phage-equivalent units per cell. For the 21.min sample in the N82 experiment of Fig. 1, these values are, for dATP, dTTP, dHTP, and dGTP, respectively, 27, 18, 50, and 1. Since N82 synthesizes far less than 1 phage-equivalent unit per cell of l)NA (2, 5, 27, 28), it is apparent that the failure of dGTP to accumulate is not directly responsible for the absence of detect- able DNX replication. Further insight into this point can bc gained if we consider pool sizes in terms of millimolar concen- trations. Freedman and Iirisch (29) have presented data from a Coulter counter which indicate that the mean cell volume of E’. coli Is/r, grown under conditions similar to those of this study, is about 0.9 pm3. If we assume that cells of E. coli 13 are of similar size and that triphosphates are not compartmented in the cell, then we can calculate approximate intracellular con- centrations, as showu in Table II. Note that the least abundant nucleotide, namely dGTP, is present at approximately 0.1 mM. It is probable that this value is within a range of optimal con- centrations for the T4 DNA replication machinery. For ex- ample, Goulian et al. (30) used a routine assay for T4 DNA polymerase (possibly the true “replicase”) in which each tri- phosphate was present at 0.033 InM, presumably within the range of optimal concentrations. Thus, the low level of dGTP following DO infections does not appear to contribute directly to the limitation of DNA synthesis.

In comparing the values determined in Table II with prc- viously published data on triphosphate pools in E. coli, one is struck by the fact that most workers have reported pool sizes in terms which cannot be directly compared with our units of mass per cell, terms such as mass per unit of dry weight of cells (22, 25)

TABLE II Nucleoside lriphosphate pools in E. coli B

These values were obtained as the average of the values for the three zero time samples in the experiments of Fig. 1.

Nucleotide I

Pool size Concentration

ATP GTP..... CTP UTP..... dATP dGTP.... dCTP dTTP .

. .

molecules/cell x 106

1.47 0.57 0.37 0.75 0.16 0.07 0.11 0.12

??zM

2.7 1.1 0.7 1.4 0.3 0.1 0.2 0.2

/

mass per ml of culture (31,32), mass per optical density unit (32), or relative counts per min (33). Thus, although the relative values determined in this study are in reasonable agreement with the relative values in the above cited studies, absolute values cannot be compared. However, triphosphate pool sizes have been reported on a per cell basis by Edlin and Stent (34) and by Huzyk and Clark (20)) and the data of Table II are in good agree- ment with these values.

Yeh and Tessman (13) have recently reported that the ADP reductase activity of the T4 ribonucleotide reductase system is stimulated by dGTP. The intracellular dGTP concentration, 0.1 mM, is within the range where this activity is most sensitive to changes in dGTP concentration, as might be expected if this effect is physiologically significant. However, from the mag- nitude of nucleotide accumulations seen in DO infections, it would appear that neither the E. coli nor the T4 reductase, both of which are preseut in infected cells, is efficiently regulated in vivo by feedback inhibition.

Ribonucleoside Mow- and Diphosphates-The above results show that dNTP pool expansion can occur without concomitant accumulation of rNTPs, and they support the idea that pulse labeling with uracil at different times in infection does lead to valid estimates of rates of RN,4 synthesis (5). However, to draw this conclusion unambiguously, one needs to establish that rNMP and rNDP pools do not expand either (see Fig. 1). As shown in Fig. 3, this seems to be the case. Although there are some fluctuations in the levels of mono- and diphosphates, the pools of total uridine nucleotides and total cytidinc nucleotides are similar over the 21-min time period studied. However, in our earlier study (5)) the rate of uracil labeling of RNA was decreased by some lo-fold by 20 min after infection. The present data support the idea that this represents a true decrease in the rate of RNA synthesis.

Uracil Uptake-One could reasonably argue that pulse-labeling data do not reflect true rates of RNA synthesis even in the absence of ribonucleotide pool expansion because of the possible development of a barrier to the uptake of labeled uracil into cells. I f such a barrier developed in DO infections but not in normal infection, then pulse-labeling data late in infection would underestimate the rate of RNA synthesis in DO infections. Accordingly, I rneasured rates of [2-W]uracil uptake into cells early and late after infection by T4D, N82 (am44), and El0 (am45). As shown in Table III, a barrier to the uptake of uracil does not develop late in infection by the DNA-negative mutants.

Rates of Labeling of RNA Precursors-A direct demonstration of the validity of pulse-labeling data can be obtained if we can

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

7434

1 I

u-l p

T4D __ N82 m-N122

r Y X

I

&OeT4D __ N82 __ N122

:

f 5”~P l ---- AMP

(I&=-===. .CMP 1 L-----T :.

0 10 20 0 10 20 0 10 20

M inu t Qs Af tQr Infection FIG. 3. Ribonucleoside mono- and diphosphate pools following infection of Escherichia coli B by T4D (left-hand panels), N82 (center

- panels), or N12’ (right-hand panels).

TABLE III

[W]Uracil uptalce

E. coli B was grown in M9 medium (5) plus 20 pg per ml of ura- cil, to about 3 x 108 ml-l. The cells were centrifuged and re- suspended in M9 cont,aining 5 pg per ml of uracil, plus 20 pg per ml of L-tryptophan 8s adsorption cofactor. The cells were chilled, and 20-ml aliquots were infected at ice temperature with T4D, T4 amN82, or T4 amEl (multiplicity of infection = 5). After a 5-min adsorption period the cultures were t,ransferred to a 37” water bath (time zero), and aeration was continued. At 2 min and at 22 min 5-ml aliquots were transferred to culture tubes containing 5 pCi each of [2-‘%]uracil. One-milliliter aliquots were removed at l-min intervals and passed through Milliporc filters, which were immediately washed with 2 ml of cold 0.1 M

potassium phosphate buffer, pH 7.0, and subsequently dried and counted. Uptake rates were linear over the time period examined (3 min).

-

Phage Rate of wad uptake

Z-5 min I

22-25 min

cfm/ml/min

T4D 645 852 N82 (~244). . . . 383 715 El0 (arn45) 511 710

show that immediate RNA precursors, namely the ribonucleoside triphosphates, become labeled during a pulse at equal rates in

normal and DO infections. As many workers have pointed out (cf. References 35-37), the relationship between rate of incor- poration of isotope into RNA and the rate of RNA synthesis depends critically upon the specific activity of labeled RNA precursors. Our earlier experiments were carried out under conditions where cells were grown before infection at a sufficiently high uracil concentration to repress endogenous pyrimidine synthesis. Hence, the infected cells were dependent upon exogenous uracil for synthesis of all pyrimidine nucleotides. Therefore, in order to conclude directly that pulse-labeling with urscil provides a valid index of rate of RNA synthesis, we must show that the rates of labeling of CTP and UTP pools are equal

I

0.4- T4D -- N82

0 30 60 0 30 60 SQconds AftQr I-i3-Uracil Addition

FIG. 4. Rate of labeling of UTP and CTP pools. Escherichia coli B was grown in Medium A plus 0.6 PCi per ml of H322P04 plus 20 pg per ml of unlabeled uracil for 2% generations, to a viable cell concentration of about 3 X 108 ml-l. Infection with T4D OF N82 (an244), both at a multiplicity of 6 phage per cell, was carried out in the same medium, At 20 min both cultures were rapidly chilled, centrifuged, washed, and resuspended in Medium A at the same specific activity of Pi but with uracil present at 1 rg per ml instead of 20. Both cultures were incubated for a further 5 min, at which time [5-3H]uracil was added to each culture at 8 &i per ml. Each culture was sampled at the indicated times, and the isolated pyrimidine NTPs were counted in a liquid &I>- tillation spectrometer set for dual isotope counting of 3H and 32P. Spillover corrections were determined from data obtained with the 3Wlabeled purine rNTPs, isolated and prepared for counting in identical fashion to that used for the pyrimidine rNTPs.

when one compares normal 2nd DO infections late in infection, when the difference in rates of RNA labeling is greatest.

An experiment demonstrating the above point is sho\vn in Fig. 4. In this experiment cells were labeled continuously with 32P, as described for the experiments in which absolute pool sizes were measured, and unlabeled uracil was present continuously as in our pulse-labeling experiments. At 25 min after infection by T4D or N82 (um44), [5-3HJuracil was added, and the cultures were sampled at 30 and 60 s after labeling. Under these con- ditions the 3H :32P ratio provides a valirl index of specific activity,

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

7435

since Y radioactivity is directly related to total nuclcotide concentration. As shown in Fig. 4, the rates of UTP labeling and of CTP labeling are virtually identical between the two cultures.

Riva et al. (37) have performed a very similar experiment to show that the absolute rate of RNA synthesis decreases following a temperature upshift in a ts43 1‘4 mutant. However, as these authors pointed out, their thin layer chromatographic system did not resolve ribonucleotides from deoxyribonucleotidcs, so that dNTP pool expansion occurring after a temperature upshift would tend to increase the 32P content of recovered UTP and CTP, and hence to cause underestimation of the true specific activities of pyrimidine rNTPs. However, both my experiment and that of Riva et al. lead unambiguously to the same conclusion that inhibition of T4 DNA synthesis leads to an inhibition of RNA synthesis. Thus, not only are replication and late gene transcription coupled (36), but replication and total tran-

RNA synthesis, although I have not yet tested this directly by an experiment of the type shown in Fig. 4.

The expansion of ribonucleotide pools following El0 infection occurs primarily among the triphosphates, although, as shown in Fig. 6, limited accumulations of mono- and diphosphates can be seen.

Fig. 7 presents the triphosphate pool patterns for N81 (~241) and El140 (am62). Both show the same type of dNTP pattern as previously described. Regarding the rNTP patterns, El140 resembles N122 and N82 in that no dramatic changes occur after infection, while N81 shows large rNTP accumulations, just as seen with ElO.

On “Dual Function” of dCMP H&rozymethylase-Chiu and Greenberg (39) have suggested that the T4 gene 42 product, deoxycytidylate hydroxymethylase, plays two roles in the syn- thesis of phage DNA, first, the already established role of forming 5-hydroxymethyldeoxycytidylate from deoxycytidylate and

scription are coupled as well. Adenylate Energy Charge-Knowledge of the AMP, ADP, and

ATP pools, as shown in Figs. 2 and 3, allows one to calculate the adenylate energy charge, defined by Atkinson and coworkers (38) as ([ATP] + 36 [ADP])/([ATP] + [A4DP] + [AMP]). For uninfected E. coli this value is about 0.90, only slightly higher than the value reported by Chapman et ~1. (38), and this value does not change significantly over the 21-min time period covered in these experiments.

Other DO Mutants-In Fig. 5 we see that extending the period

IO- N82 El0 .ADP

‘: 0 ; 5 / S -.~:--.CDP ,.UDP 1 0

of infection with N82 to 30 min allows one to detect a limited 2

rNTP pool expansion. Considerably greater rNTP accumu- 05 .X

lation, however, up to 5-fold, is seen in infection with El0 (~~45). CMP ,,$~&-P

This is interesting in light of our earlier observation (5) that Oo-?-==~=~~UMP) l UMP

10 20 30 0 10 20 30 uracil incorporation into RNA is even more severely depressed in MinUtQs After InfQction

infection by El0 than in infection by other DO mutants. It would appear that this extra inhibition is due to this rNTP pool

FIG. 6. Ribonucleoside mono- and diphosphate pools following infection of Escherichia coli B by NS2 (am44; left-hand panels)

expansion rather than an even more severely depressed rate of or El0 (am45; right-hand panels).

24 I

16 1

In b 81 dHTP dGTP

Y i .TTPi riATP /

T

“, ,, II a----- 8 ! -i ‘.‘&- ,dCTp~~sd-.- 3

-c- T4D rNTP’s N82 rNTP’s 1 El0 rNTP’s ._-

N82 dNTP’s l -. dTTP/

El0 dNTP’s

TP

---.GTP

-*----.CTP

Minutes After Infection

0 IO 20 30

FIG. 5. Nucleoside triphosphate pools following infection of Escherichia coli B by T4D (left-hand panels), N82 (am44; center panels), and El0 (am45; right-hand panels).

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

El140 dNTP’s 25

$50 N81 rNTP’s t

El140 rNTP’s _rTn

.-----•UTP

0 ? 10 20 30 0 10 20 30

M,nules After lnloctlon

FIG. 7. Nucleoside triphosphate pools following infection of Escherichia coli B by N81 (am41; left-hand panels) or El140 (am62; rioht-hand uanels). Not,e the difference in scales for the N81 and l&40 data: ’

hence generating a pool of hydroxymethylcytosine nucleotides, and second, an as yet uncharacterized role as an element of the “DNA replication complex.” This suggestion stemmed from the observation that an active dCMP hydroxymethylating activity seemed to be present following infection by T4 tsL13, a ts gene 42 mutant, at a nonpermissive temperature, but that no DNA is synthesized. Chiu and Greenberg concluded from this that the replication complex function is more heat-labile in the mutant gene 42 protein than is the hydroxymethylase activity. However, they did not verify their conclusion that the mutant hydroxymethylase is active in vioo by demonstrating that a pool of hydroxymethylcytosine nucleotides is formed at 42”. Pre- sumably the HMC nucleotide accumulating would be dHTP, since an active gene 1 protein (deoxynucleotide kinase) is present. I have asked whether any dHTP is detectable following infection by L13 at 42”. As shown in Pig. 8, no detectable dHTP is formed, even though the dATP and dTTP pools expand signifi- cantly. This suggests that the mutant gene 42 protein is not active as a dCMP hydroxymethylase in tivo at 42”. In turn this makes somewhat less compelling the postulate that this protein plays a dual role, since the failure to generate HMC nucleotides can alone account for the inability of fs42 mutants to synthesize DNA at restrictive temperatures.

DISCUSSION

Three major questions emerge from the present work. (a) What is the nature of replication-total transcription coupling? (5) Why do ribonucleoside triphosphates accumulate in infection by some mutants (am41, am45) but not others (am42, am44, am62)? (c) Why do dGTP pools not expand concomitantly with those of dATP, dTTP, and dHTP?

Regarding the relationship between DNA synthesis and total RNA synthesis, Riva et al. (37, 40) have argued plausibly that replication-late transcription coupling is related to the formation of single strand interruptions as a consequence of discontinuous replication of DNA. Such gaps or nicks, presumably, must be present for binding of RNA polymerase and initiation of tran- scription or both. In nonreplicating DNA nicks are repaired by

I 1

107 dATP

~------i~

/

01 0 10 20 30

Minutes After Infection

FIG. 8. Nucleoside triphosphate pools following infection of Escherichia coli B by L13 (ts42) at 42”. Bacteria were grown at 37” for two generations, shifted to 42” at one-half generation time before infection, and maintained at that temperature throughout the remainder of the experiment.

DNA ligase essentially as rapidly as they are formed, such that initiation of transcription cannot occur. In this context it would be of interest to examine the integrity of parental DNA strands late in infection by DNAnegative mutants. If coupling is causally related to the presence of nicks, we would expect these strands to be relatively intact, whereas a considerable amount of short-stranded phage DNA can be detected at late times in normal infection (41). Of course, interpretation of the results of such an experiment might be complicated by repair or recom- bination effects.

A question which must be asked about the different patterns of rNTP accumulations is whether the observed phenomena are gene-specific or mutant-specific. So far all of our pool studies have been carried out with only one amber mutant in each gene. If we are to conclude that rNTP accumulation or nonaccumu- lation is related to the function of a particular gene product, we must be sure that all nonleaky amber mutants in a given gene show the same pattern, in other words, that the pattern ob- served does not result from some peculiarity of one mutant within a given gene. If the effects are found to be gene-specific, then we would want to ask whether rNTP accumulation under certain conditions is due to increased synthesis or decreased turnover. I f it is the latter, then this would suggest that total RNA syn- thesis rates are even more severely blocked in those mutants showing accumulation than in those in which the rNTP pools do not expand, and this may ultimately provide a clue to the mecha- nism of coupling, particularly in conjunction with information on DNB strand integrity.

Presently available data shed little light on the question of why dGTP pools do not expand to the same extent seen with other dNTPs. In uninfected cells the failure of dGTP to accumulate when DNA synthesis is blocked results from decreased synthesis rather than increased turnover (22), and it seems reasonable to

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

UDP

host x 0 * T4 red uctase reductase

dUTP\IIrpydCMP

/ \

syXetase/ ’ T4

i rhetase T TdR * > dTMP--+dTDP+dTTP-DNA

* TdR’

Ii cell

surface

I host DNA

FIG. 9. Pathways of thymidylate metabolism in T4 phage- infected Escherichia coli. T and TdR represent thymine and thymidine, respectively. Reactions marked with an asterisk are those known to be regulated in vitro by dTTP.

assume that the same is true in infected cells. One might have expected that ribonucleotide reductase would be the key step at which deoxyribonucleotide synthesis is regulated. However, (a) the extensive accumulations of dATP, dTTP, and dHTP which occur suggest that feedback inhibition of this enzyme is quite ineffective in vivo; and (b) on the basis of in vitro studies with this enzyme,2 no conditions have yet been found in which ADP reduction is stimulated while GDP reduction is inhibited. Therefore, the critical step in regulating the rate of synthesis of dGTP probably lies elsewhere, and its ultimate identification will require a thorough analysis of all purine ribo- and deoxyri- bonucleotide pools.

Finally, some cautionary remarks should be made in light of the present work, about the widespread use of pulse labeling with radioactive thymine or thymidine to estimate rates of DNA synthesis in T-even phage-infected cells. As shown in Fig. 9, there are three metabolic routes for formation of thymidylate in the phage-infected cell; the action of either the host or phage- coded thymidylate synthetase and the breakdown of host DNA (27). In order for pulse-labeling data to accurately reflect rates of DNA synthesis, one must show that the relative exogenous contribution of thymine nucleotides to DNA is the same under all conditions studied. This is almost certainly not the case when one compares infected with uninfected cells, for in unin- fected cells two of the three endogenous pathways are not present. Moreover, the activity of these pathways changes with time after infection, with the viral synthetasc beginning to be formed at about 5 min and thyminc nucleotide release from host DNA beginning at 10 min (27). Thus, it seems quite unlikely that the esogenous contribution would remain constant during in- fection.

Especially vulnerable to misinterpretation is the use of thy- midine labeling to compare rates of DNA synthesis before and after temperature shiftup in infection by a tsDNA-defective mutant (cf. 37, 39, 42). This procedure is quantitatively in- accurate, for, as shown in this paper, the blockage of DNA resulting from a shiftup expands the dTTP pool, with a resultant decrease in its specific activity, and hence of the exogenous con- tribution. In addition, since several steps of thymidylate metabolism are feedback-regulated by dTTP (see reactions

2 0. Berglund, personal communication.

7437

marked with asterisk in Fig. 9), there are additional, and un- predictable, changes in the relative activities of the various pathways which would result from dTTP pool expansion. To be sure, the quantitative errors introduced by these phenomena may be relatively small in experiments of short duration; I have not yet measured the precise kinetics of dNTP pool expansion after a shiftup. However, since all three endogenous routes to dTMP can be blocked by appropriate mutations (27, 43), with resultant total dependence upon exogenous thymine or thymidine for DNA4 synthesis, it is suggested that investigations on rates of DNA synthesis use appropriate mutant phage and bacterial strains.

Acknowledgments-1 thank Mrs. Diana Spatola and Miss Kathleen Scott for capable technical assistance.

REFERENCES

1.

2. 3.

4. 5. 6.

7. 8. 9.

10.

11.

12.

13.

14.

15.

1G.

17.

18.

19.

20. 21.

22.

23.

24. 25. 26.

27. 28.

29. 30.

31.

32.

EPSTEIN, R. H., BOLLI”, A., STEINBERG, C. M., KIGLLUN- BERGER, E., BOY DE LA TOUR, E., CHI<;VALL*>Y, R., EDGAR, R. S., SUSMAN. M., DENHARDT, G. H., AND LIELAUSIS, A. (1963) Cold Spkng harbor Symp. Quant: Biol. 28, 375

WARNER, H. I~., AND HOBBS, M. D. (1967) Virology 33, 376 MATHEWS, C. K. (1971) Bacteriophage Biochemistry, Van

Nostrand Reinhold Co., New York ALBERTS. B. M.. AND FRIGY. L. (1970) iliature 227. 1313-1318 MATHEWS, C. K'. (1968) J. bioZ.‘Chek. 243, 5610-5615 MURRAY, R. E., AND MATHINS, C. K. (1969) J. IMoZ. Biol.

44,249-262 OISHI, M. (1968) Proc. Nat. Acad. Sci. U. S. A. 60,lOOO SCOTTI, P. D. (1969) Proc. Nut. Acad. Sci. U. S. A. 62, 1093 WARNPR, H. R., AND HOBBS, AI. D. (1968) Virology 36,527-537 ELAM, J. S., AND KOERNER, J. F. (1970) J. Viol. Chem. 246,

1012-1019 RANDERATH, K., AND RANDERATH, E. (1967) Methods Enzymol.

12, 323 BERGLUND, O., K~RXTR~M, O., -[ND REICHARD, P. (1969)

hoc. Nat. Acad. Xci. U. S. A. 62, 829 YEH, Y.-C., AND TESSMAN, I. (1972) J. Biol. Chem. 247, 3252-

3254 Scocca, J. J., PAXNY, S. R., AND BESSMAN, M. J. (19G9) J.

Biol. Chem. 244. 3698-3706 MALEY, G. F., ~UARINO, 1). U., 9ND MALEY, F. (1967) J.

Biol. Chem. 242, 3517-3524 HIRAGA, S., IGAR.~SHI, K., AND YURA, T. (1967) Biochirn.

Biophys. Acta 146, 41 KBI~;MPFER, R. 0. R., AND MAG.ISANIIi, B. (1967) J. itfOl. Biol.

27, 453-468 LOWRY, 0. H., ROB~XTS, N. R., LEIiXIcR, K. Y., WV, M.-L.,

AND FARR, A. L. (1954) J. Biol. Chem. 207, 1 M.\THPWS, C. K., .~ND HEWLETT, M. J. (1971) J. vi’irol. 8,

275-285 HUZYI~ I,., AND CLARIC, D. J. (1971) 1. Bacterial. 108, 74-81 BAGNARA, A. S., AND FINCH, L. 11. (1972) Anal. Biochem. 46,

24-34 NEUHARD, J., AND THOM~SSEN, 1:. (1971) Eur. J. Biochcm.

20, 36 COHEN, S. S. (1947) Cold Spring Harbor Symp. Quant. Hiol.

12, 35 WERNER, R. (1971) i\Tature n’eu, Biol. 233, 99 NEUHARD. J. (1966) Biochim. Biophus. Acta 129, 104-115 SINSHEIM~R, it. L: (1969) in 2'hk kucleic Acids (CHARGAFF,

E., AND DAVIDSON, J. N., eds) Vol. III, p. 187, Academic Press, New York

MATHEWS, C. K. (1966) Biochemistry 6, 2092-2100 KOZINSI<I, A. W., AND FELGENHAUER, Z. Z. (1967) J. Viral. 1,

1193-1202 FREEDMAN,M. L., .IND KRISCH, R. E. (1971)J. viral. 8,87-94 GOULIAN, M., LUCAS, Z. J., AND KORNBERG, A. (1968) J. Biol.

Chem. 243, 627-638 RUFF, W., KIRBY, E. P., AND GOLDTHw.IIT, D. A. (1971) J.

Bacterial. 106, 994-1004 GALLBNT, J., AND HARADA, B. (lYG9) J. Biol. Chem. 244, 3125-

3132

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

7438

33. EDLIN, 34. EDLIN.

G., AND N~UHARD, J. (1967) J. Mol. Biol. 24, 225-230 G., AND STF,NT, G. S. (1969) Proc. Nat. Acad. Sci.

U. i A. 62, 475-482 35. NIIIRLICH, D. P. (1967) Science 168, 1186-1188 36. SALSIZR, W., JANIN, J., AND LEVINTHAL, C. (1968) J. Mol.

Biol. 31, 237-266 37. RIVA, S., CASCINO, A., AND GEIDUSCHEK, E. P. (1970) J. Mol.

Biol. 64, 85-102 38. CHAPMAN, A. G., FALL, L., AND ATKINSON, D. E. (1971) J.

Bacterial. 108, 1072-1086

39. CHIU, C.-S., AND GREENBERG, G. R. (1968) Cold Spring Harbor Symp. Quant. Biol. 33, 351

40. RIVA, S., CASCINO, A., AND GSIDUSCHEK, E. P. (1970) J. Mol. Biol. 64, 103-119

41. FRANKEL, F. R. (1968) Cold Spring Harbor Symp. Quant. BioE. 33,485493

42. SAUERBIER, W., AND BR~UTIGAM, A. R. (1970) J. Viral. 6, 179-187

43. WARNER, H. R., SNUSTAD, D. P., JORGENSEN, S. E., AND KOERNER, J. F. (1970) J. Viral. 6, 700

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from

Christopher K. MathewsT4: III. NUCLEOTIDE POOLS

Biochemistry of Deoxyribonucleic Acid-defective Amber Mutants of Bacteriophage

1972, 247:7430-7438.J. Biol. Chem. 

  http://www.jbc.org/content/247/22/7430Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/247/22/7430.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on July 8, 2020http://w

ww

.jbc.org/D

ownloaded from