Products of the photophosphorylation reaction.

Abstract Most determinations of the stoichiometric relationship between electron transport and phosphorylation have employed radioactive orthophosphate and have been based on the assumption that ADP-dependent incorporation of radioactivity can be safely equated to ATP synthesis. The validity of using this procedure to measure photophosphorylation in isolated chloroplasts has been challenged recently. However, the studies reported here show that estimates of photophosphorylation based on label incorporation are probably correct. 1. The amount of labeled orthophosphate incorporated does not differ significantly from the amount of orthophosphate which disappears from the medium. Thus, there is no evidence of an exchange reaction between orthophosphate and ADP or other unlabeled, phosphate-containing substances. 2. Nearly all of the incorporated 32P is found in ATP but a significant amount of label is sometimes found in ADP. However, the 32P in ADP is all or nearly all derived from AT32P, presumably through the action of the ATP:AMP phosphotransferase (EC 2.7.4.3) known to be present in chloroplast preparations. Therefore, the sum of the radioactivities of ATP and ADP should be used in estimating the extent of the phosphorylation reaction.

2. Nearly all of the incorporated a2P is found in ATP but a significant amount of label is sometimes found in ADP. However, the 9 in ADP is all or nearly all derived from ATa2P, presumably through the action of the ATP:AMP phosphotransferase (EC 2.7.4.3) known to be present in chloroplast preparations. Therefore, the sum of the radioactivities of ATP and ADP should be used in estimating the extent of the phosphorylation reaction.
It has been widely accepted that 1 and only 1 ATP molecule is formed by illuminated chloroplasts during the transfer of a pair of electrons from water to ferricyanide or to other acceptors (1,2). However, Winget, Izawa, and Good (3) routinely observed P/e2 ratios between 1.1 and 1.3. Very rarely, even in a single determination, did they obtain values as low as 1 3. When ATP formation is measured as the incorporation of radioactive orthophosphate, the apparent value of the ratio P/e2 increases with decreasing amounts of electron transport. This observation they attribute to the fact that the alleged exchange reaction becomes a more significant proportion of the total incorporation as the incorporation through phosphorylation becomes less.
The experiments described in this paper were undertaken in an attempt to reconcile the observations of Winget et al. (3) with the observations of other workers. Particular attention has been given to checking the reliability of using 32P incorporation as an assay for ATP formation.
In the course of this work many of the experiments of Del Campo et al. (4) were repeated and extended.

Preparation
of Chloroplasts-Chloroplasts were isolated from market Spinach (Spina& olerucia L.) by the following procedures.
Petioles and the greater part of the midvein were removed from washed leaves which were then ground for about 10 set in a Waring Blendor with 0.3 M NaCl, 0.05 M tricine-NaOH, and 0.003 M MgCl (pH 7.5). The homogenate was squeezed through eight layers of cheesecloth and then centrifuged at about 2000 x g for 5 min.
The suspension was centrifuged briefly (about 15 set) at 2000 x g to remove cell debris and intact cells, then centrifuged again at about 2000 x g for 5 min. The pellet was washed once more in the same medium and was finally taken up in a small amount of the sucrose-tricine medium. Chlorophyll was estimated by the method of Arnon (5 containing 0.2% copper sulfate solution in water which acted as a combined heat filter and a condenser lens. The light intensity used was saturating. Estimation of Labeled Organic Phosphate-Organic-32P was estimated by a method adapted from the methods of Avron (6) and of Nielsen and Lehninger (7). The details of the procedure are as follows.
Acetone (1.2 ml), 1.0 ml of 10% ammonium molybdate, and 10.0 ml of butanol-benzene (1: 1, v/v, saturated with 10% perchloric acid) were added to a l.O-ml aliquot of the reaction mixture which had been diluted with 10 ml of 10% perchloric acid (saturated with butanol-benzene). The twophase mixture was stirred thoroughly by an up and down movement of a glass rod flattened at the end. After separation of the layers, the upper organic layer containing the phosphomolybdate was carefully sucked out of the test tube through a Pasteur pipette connected to a vacuum line with two traps.
The aqueous solution was then filtered into another test tube through a Whatman No. 4 filter paper soaked immediately before use with 0.5 ml of distilled water.
The wet filter paper held back chloroplast and any minute droplets of organic phase remaining suspended in the aqueous phase.
To the clear filtrate were added  In the experiment illustrated by p, 10-a M ADP was added, whereas in the experiments illustrated by ---, lO+ M ATP, 1e2 M glucose, and 1 mg of hexokinase were added. As soon as the ferricyanide was reduced, the light was turned off and the sample wm incubated in the dark for the stated length of time at 10". The products of the reaction, ATP and ADP or glucose 6-phosphate (G-6-P), were separated by paper chromatography and their radioactivities determined.
pearing cannot be very precise unless a considerable proportion of the total orthophosphate is consumed. Consequently, very low (decidedly suboptimal) concentrations of phosphate were used in these experiments.
For measurement of the incorporation of radioactive orthophosphate into other substances an excess of molybdate is again added, but in this case the phosphomolybdic acid is quantitatively removed by organic solvents and discarded.
Phosphate incorporation is measured as the residual radioactivity. The great, advantage of such determinations is that they can be very accurate even if huge amounts of unused orthophosphate are still present after the reaction.
For this reason the method has been used in most, studies of photophosphorylation.
Therefore, the report of Del Campo et al. (4) throws doubt on many data in the literature.
Fortunately a careful comparison of the two methods carried out, in this laboratory provides considerable support for the radioisotope method. During the reduction of 1.25 pmoles of ferricyanide, an average of 0.628 pmole of orthophosphate disappeared, while an average of 0.629 pmole of labeled phosphate appeared in the non-orthophosphate fraction. It is, therefore, probably safe to assume that there was no significant exchange of labeled orthophosphate with unlabeled phosphate compounds such as ADP.
It should be noted that the ratio P/e2 in these experiments was almost precisely 1.0. However, it should also be noted that in order to obtain reliable estimates of orthophosphate disappearance the concentration of phosphate used was only about 3% of the concentration which is optimal for phosphorylation. From other studies of the dependence of phosphorylation on phosphate concentration, it can be calculated that with these chloroplasts P/e ratios of about 1.2 would have been possible.
Di.stribution of 32P in Products of Phosphoyldion Reaction-Chromatographic analysis of the reaction mixture also showed that there was no significant ADPe3'Pi exchange during phosphorylation.
It is true that a small amount of AD32P was formed (see Table II) but this minor reaction depends on the continued presence of labeled ATP. Thus, when hexokinase and glucose kept the level of ATP very low, much less AD32P was formed (Table III).
Conversely, when hexokinase and glucose were absent and the reaction mixture was incubated in Reaction condition was as in Fig. 1 (with ATP, glucose, and hexokinase) except that ferricyanide was at the stated concentrations. Illumination was discontinued either as soon as the ferricyanide had been reduced (Zozuer line) or aft,er 4 min (upper line).
Because there is appreciable phosphorylation when chloroplasts are illuminated in the absence of ferricyanide, the ratio of organic phosphate formed to ferricyanide reduced becomes infinite when no ferricyanide is added. the dark after phosphorylation had ceased, there was an increase in the amount of labeled ADP at the expense of labeled ATP (see Fig. 1). It seems clear that most if not all of the trivial amount of label in ADP is derived from ATP through the action of the ATP :AMP phosphotransferase (adenylate kinase, myokinase) known to be present in chloroplast preparations ATazP + AMP = ADP + AD32P and, because the label is derived from ATP, this form of organic 32P must also be included in estimating the extent of the phosphorylation reaction.
Apparent P/e, as Function of Amount of Ferricyanide Reduced-As already pointed out, Del Campo et al. (4) have observed that the use of low ferricyanide concentrations results in apparent P/e2 values higher than 1.0 and the lower the ferricyanide concentration the higher the apparent P/et. However, it would seem from the published description of their experiments that they illuminated the chloroplasts for a constant period of time (4 min), which was greater by varying degrees than the time actually required for complete reduction of the amount of ferricyanide added. Therefore, it is reasonable to suspect that the absence of an exogenous electron acceptor permitted some cyclic electron transport catalyzed by endogenous carriers in the manner described by Mehler (9). Maximum "endogenous" phosphorylation would therefore occur with minimum ferricyanide and the computed ratio of P/e2 would rise, because an increasing proportion of the electron transport would be unrecognized and unrecorded. We find that the ratio of ATP formed to ferricyanide reduced is independent of the amount of ferricyanide reduced when the light is turned off as soon as the ferricyanide is reduced. Only when there is a constant period of illumination does the ratio rise with decreasing amounts of ferricyanide (Fig. 2). they must have either overestimated ATP formation or underestimated electron transport. This study shows that their measurements of phosphorylation are very probably correct. We cannot be so sure about the measurements of electron transport because light-driven electron movements can never be measured unambiguously; electrons raised to a higher potential by a photochemical reaction can always return to the lower level unnoticed. Therefore, there is no way of eliminating the possibility of some unmeasured cyclic electron flow. However, it should be noted that phosphorylation supported by cyclic electron flow is usually quite slow (although probably never negligible).
This low rate can be subtracted from the rate observed in the presence of ferricyanide and still leave a ratio of P/e* well over 1.0. Therefore, unless the presence of ferricyanide or other electron acceptors actually increases the cyclic flow of electrons (which seems improbable), the true value of P/e2 must be greater than 1.0. Indeed, for other reasons Izawa and Good (10) have suggested that the phosphorylating part of the electron transport may make 2 ATP molecules for each pair of electrons transferred.