binding protein, plasma, bovine serum albumin, interference, phosphodiesterase inhibitor, metal ions, U.K., non-specific, concentrations, cyclic AMP, ATP, cyclic GMP, assay, BSA, cyclic adenosine monophosphate, BIOLOGICAL SAMPLES, assay conditions, extraction procedure, Elsevier Scientific Publishing Company, Clinica Chimica Acta, assay results, separation methods, binding proteins, protein concentration, separation method, protein kinase, increase, association constant, EDTA, binding sites
Clinica Chimica Acta, 56 (1974) 221-234 @ Elsevier Scientific Publishing Company
, Amsterdam - Printed in The Netherlands CGA 6634 A SIMPLE DIRECT ASSAY FOR CYCLIC AMP IN PLASMA AND OTHER BIOLOGICAL SAMPLES USING AN IMPROVED COMPETITIVE PROTEIN BINDING TECHNIQUE
K.C. TOVEY, K.G. OLDHAM and J.A.M. WHELAN The Radiochemical Centre, Amersham, Buckinghamshire (Received April 8,1974)
Summary An improved cyclic AMP assay using a purified and activated binding protein
from bovine skeletal muscle is described. Activation of the purified binding protein by bovine serum albumin
was investigated and shown to increase the assocation constant from 0.4 * lo9 M-' to 1.0 . lo9 M-' . The fully activated binding protein was not affected by the presence of large amounts of protein in biological samples such as plasma. Using the fully activated binding protein and with optimised assay conditions it has been demonstrated that the assay has a remarkable freedom from non-specific interference by materials normally found in crude biological extracts. This freedom from interference, coupled with a simple method of preparing plasma, permits the hitherto unreported rapid direct assay of plasma cyclic AMP without the need for extraction procedure
s or the use of cyclic AMP-free plasma controls, thus making the method more reliable and suitable for use in the clinical laboratory. Development of an assay markedly insensitive to non-specific interference increases the number of potential applications of the assay.
Introduction Although a wide variety of methods [l] have been used for the assay of cyclic adenosine monophosphate (cyclic AMP) in tissues, urine and plasma, and in the assay of adenylate cyclase activity, saturation analysis methods
have become the method of choice . Specificity is an important requirement for any assay; that is, the assay for the compound being measured should be free from interference from other substances . Such interference can result not only from compounds similar
to that under assay but also from other materials (salts, ~)rotem~, neural IOII~, enzyme inhibitors, etc.) likely to be present. in biological extracts. Biological fluids, such as urine, which cotltain high levels of cyclic* AMP can be greatly diluted before assay. This dilution markedly decreases the passibilit.y of interference in most, \)ut not all assays [ 31 . Plasma r,ontaills ;I muc*h lower level of cyclic IV14P and interference cannot usually bt avoided 1j.vdilution. Current estimations of plasma cyclic AMP levels reqWY either tedious blank corrections using cyclic AMP-free plasma [5,6] , or rstraction methods [5,8] which require correction both for recovery and for furthtrr iilt~~r~ere~lce by substances isolated during the extraction procedure [5 ~lO.lZ] . No technique yet described has been sufficiently free from non-specific, interft>rrnce to permit a simple, direct assay of plasma cyclic rIMP. Differences in the composition of standards and unknowns cause environment.al changes which can result in gross interference with the primary binding reaction, with the separation procedure, or with both. One method of minimising such interferencr is to ensure that incubation mixtures containing unknowns and standards should 1~ as near identical as possible. This requires calibration with standard solution
s containing reagent blanks prepared from identical biological media which have been treated to remove the compound being assayed. Such procedures are laborious, particularly when different plasmas exert different. non-specific effects  thus requiring a separate reagent blank and calibration curve for each sample. Additionally, because cyclic AMP assays art: predominantly a research interest, t.hey are carried out on samples obtained from widely differing sources of Biological material
which necessitates the preparation of separate calibration curves for each source. ililternative methods used to minim&e non-specific int.erferenoe in cyclic AMP assays have involved the chromatographic isolation of the pure nucleotide, the use of complex intclrnal standardization methods [lo]. 01` the development of more sensitive assays. These methods are complicated and tedious. Thus an assay which is subject to minimal non-specific interference from substances likely to he found in biological extracts will. have many ~~c~vantag~~osver current methods. We have developed such an assay by selecting assay conditions to minimise interference, by the use of a purified binding protein having a high specificity and affinity for cyclic AMP, by using an improved charcoal separation technique, and by using an incubation mixture containing EDTA. The assay has been shown to permit the rapid and simple direct. assay of cyclic: AMP in plasma and other biological samples, and has a high degrtbe of freedom from non-specific interference.
Tritiated cyclic AMP, 1%' H] adcnosinc, 3',5'-monophosphate
with aspecific activity
of 29 Cj/mmole was obtained from The Radiochemica~ Centre
(Amersham, U.K). Cyclic AMP, cyclic GMP and ATP were obtained from the
Sigma (London) Chemical Co. (London, U.K.). Norit GSX charcoal was ob-
tained from Norit Clydesdale Co. (Glasgow, U.K.) and theophylline from
British Drug Houses (Poole, U.K. ).
Iteagents for liquid s~!intillation counting were obtained as follows: PCS
(Amersham/Searle) hquid scintillation cocktail for aqueous samples, from
Hopkin and Williams (Chadwell Heath, U.K.); Triton X-100 detergent and
toluene (sulphur-free) from British Drug Houses (Poole, U.K.); 2,5-diphenyl-
oxazole, (PPO) and p-b&(2-(5-phenyloxazolyl)-benzene),
Nuclear Enterprises (Edinburgh, U.K.).
Fresh bovine muscle was obtained from a local abattoir.
Methods Preparation of assay reagents (1) Binding protein was isolated from bovine skeletal muscle and purified essentially as described by Gilman [l,ll] but omitting the acid precipitation step. 400 g of muscle were homogenised in 500 ml of assay buffer, then centrifuged for 30 min at 10 000 X g. The supernatant was fractionated with ammonium sulphate. The 25-45% saturation fraction was dissolved in 0.005 M potassium phosphate buffer pH 7.5 containing 0.5 mM EDTA (buffer A) and dialysed overnight against 4 1 of the same buffer. The dialysed protein was then applied to a DEAE column (previously equilibrated with buffer A) and eluted with buffer A. Cyclic AMP-binding activity was eluted from the DEAE-cellulose column in two peaks. Binding protein from the first peak contained 50% of the initial binding activity and after dialysis, Tris-HCl buffer pH 7.5, EDTA and bovine serum albumin (BSA) were added to give the following final concentrations: Tris-HCl buffer pH 7.5, 50 mM; EDTA, 4 mM and BSA, 0.1%. The concentration of binding protein was adjusted so that 100 ~1 of the solution would bind 55% of the activity (zero-dose binding) when incubated with 0.9 pmoles of tritiated cyclic AMP in a final volume of 200 ~1. The specific activity of this preparation was 0.1 pmole cyclic AMP bound/pg binding protein and when stored at -20" was stable for several months. (2) Assay buffer. 50 mM Tris-HCl buffer pH 7.5 containing 4 mM EDTA. (3) Tritiated cyclic AMP reagent. 180 pmoles (ca 5 /..&i) in 10 ml of assay buffer (equivalent to 0.9 pmoles/50 1.11sample). (4) Standard cyclic AMP reagent. 1600 pmoles non-radioactive cyclic AMP in 5 ml of assay buffer (equivalent to 16 pmoles/50 ~1 sample). Four serial dilutions of this reagent with equal volumes of assay buffer provided additional calibration standards containing 8, 4, 2 and 1 pmole/50 ~1 sample. (5) Charcoal adsorbent. 520 mg of Norit GSX charcoal suspended in 20 ml of assay buffer containing 2% BSA. Preparation of toluene-Triton X-l 00 scintillation cocktail 4.4 g of PPO and 0.1 g of POPOP were dissolved in 1 litre of toluene; 500 ml of Triton X-100 were added, and the solution was well mixed and stored in the dark. Before adding radioactive samples for counting, vials were prepared by adding 10 ml aliquots of the scintillant cocktail and 1 ml of distilled water
to each vial and shaking to obtain a clear solution. Standard assay procedure To each tube the components of the reaction mixture were added in the
Fig. 1. A typical standard curve
for the assay of cyclic AMP. Assays were carried out as described in the text. following order: 50 ,ul of tritiated cyclic AMP reagent, 50 ~1 of unlabelled cyclic AMP (calibration standard or unknown), and 100 ~1 of binding protein solution. After mixing the contents, the tubes were incubated for 2 h (% 30 min) in an ice-bath. 100 ~1 of charcoal suspension were added to each tube which was then vortex mixed for 10 s. After ~entrifugation (ca 10 000 X g for 1 min at room temperature
) 200 ~1 samples of supernatants (bound fraction) were removed and their radioactivity determined by liquid scintillation counting in either PCS or a toluene-Triton X-100 scintillation cocktail. A blank reading was determined by measuring the cpm in 200 ~1 of supernatallt from a tube containing 50 1.11tritiated cyclic AMP, 150 ~1 of assay buffer, and 100 ~1 of charcoal suspension. This blank reading was subtracted from all other results. The cpm bound in the absence of unlabelled cyclic AMP (i.e. the zero-dose binding) was designated C, and the cpm bound in the presence of unlabelled cyclic AMP (calibration standard or unknown) was designated C,. The ratio Co/C, was calculated for each level of cyclic AMP. A straight line calibration curve was obtained by plotting Co/C, against pmoles of standard cyclic AMP on linear graph paper (see Fig. 1). The amount of cyclic AMP in the unknowns was then read off this line using the measured C, /C, values. Ox cerebral cortex
was homogenised at a concentration of 100 mg of tissue/ml of 0.1 M HCl and the homogenate was heated at 100" for 2 min as described elsewhere [lo] . After cooling, the suspension was centrifuged, the supernatant taken, freeze-dried and dissolved in water. Two ml of this extract* *A kind gift of Dr M. Weller, Institute of Psychiatry, London, U.K.
225 were adjusted to pH 7.5 with 1 M NaOH and diluted appropriately with assay buffer before assay. Preparation of plasma A freshly withdrawn sample of blood was added immediately to 1% of its volume of 0.5 M EDTA pH 7.5 contained in a cooled centrifuge tube and the contents were rapidly mixed. The EDTA-treated blood was centrifuged for 5 min at 4000 X g at room temperature. The plasma was then removed and assayed the same day or stored at -20" for assay later. Results and Discussion (A) Design of the assay Four major factor
s must be considered in attempts to design a competitive protein binding assay which will suffer minimal interference. These are: the sensitivity of the assay; the composition of the incubation mixture; and the influence of potentially interfering factors on both the binding reaction and the separation procedure. (1) The sensitivity of the assay Non-specific interference can be avoided by designing an assay of very high sensitivity so that assay samples can be diluted to an extent which reduces interfering factors to insignificant levels
. Maximum sensitivity, however, is not the only requirement in an assay. For optimum performance the assay should be simple, precise and rapid, and have an adequate range. Table I and Fig. 1 show that the assay technique described fulfils these requirements: counting times of 4 min are adequate; the detection limit
(2 standard deviation
s at zero-dose) is 0.05 pmoles/50 ~1 sample; the simple nonlogarithmic linear calibration curve has a mid-range (i.e. the dose required to reduce the response to half its initial value
) of 1.20-1.30 pmoles/50 ~1 sample; can be used with a within-assay coefficient of variation
of less than 11% over
Assays were carried out as described in the text. All samples were counted 6935 cpm; blank value was 64 cpm. Results are mean values * S.D. (n = 6).
for 4 min; total activity was
Dose (pmoles cyclic AMP)
Coefficient of variation (%)
Standard deviation (pmoles cyclic AMP)
0 0.25 0.50 1.0 2.0 4.0 8.0 16.0
62.5 51.5 43.1 34.4 23.0 14.6 7.9 4.9
+ 1.3 + 0.9 + 0.8 ? 0.9 + 0.8 ? 0.6 +_ 0.3 2 0.5
2.1 1.8 1.9 2.7 3.6 4.2 4.3 10.9
1.00 1.21 1.45 1.82 2.75 4.27 7.95 12.83
* 0.02 * 0.02 ?z 0.03 ? 0.05 * 0.10 i 0.18 + 0.34 + 1.40
0.021 0.027 0.035 0.061 0.12 0.23 0.43 1.75
10.8 7.0 6.1 6.0 5.1 5.3 10.9
the range 0.2516 pmoles/50 ~1 sample, and with maximum precision (coefficient of variation < 7%) over the range 0.5-8 pmoles/50 ~1 sample. Betweenassay reproducibility is excellent (see Table II). The sensitivity achieved is adequate for the vast majority of cyclic AMP assays but, because of the high affinity of the binding protein for cyclic AMP (K;, = 10" M-' ), it could readily be increased further if higher sensitivity is required. This could be done by reducing the concentrations of labelled cyclic AMP and binding protein, or by staggered addition of labelled cyclic AMP. These modifications would, however, make the assay less simple. (2) Composition of the incubation mixture The composition of the incubation mixture can have a significant effect on the susceptibility of an assay to non-specific interference. Assays were carried out at pH 7.5 rather than at pH 4 as in the Gilman assay [ 1,111, for several reasons: the binding protein's capacity was optimal at pH 7.5. A pH of 7.5 avoids the need to adjust the pH of most biological samples before assay. Many proteins precipitate at pH 4 and such precipitation would prevent simple direct assays on protein-rich samples like plasma. EDTA was included in incubation mixtures to chelate bivalent metal ions and t,hus simultaneously prevent enzymatic degradation of cyclic AMP and reduce the possibility of metal ion interference in the binding reaction [ 131 or separation procedure [ 141. The use of EDTA proved to have further marked advantages in the preparation of plasma samples (see section C, part 2). We therefore used an incubation mixture containing 50 mM Tris buffer pH 7.5 and 4 mM EDTA. Fig. 2 shows that the assay is insensitive to large increases in the concentrations of buffer and EDTA. (3) The binding reaction It is advantageous to use a binding protein with a high affinity for cyclic AMP, with minimal cross-reactivity with other nucleotides, and subject to minimal interference from other proteins present in incubation mixtures. Gilman [ 111 reported that a purified binding protein from bovine skeletal muscle had a high affinity for cyclic AMP (K, = 3.3-5 . lo8 M-' ) and a very low cross-
c] = normaf
R: 31r normai
q a:lOx normal H=ZOx normal
Fig. 2. The effect of increasing the buffer concentration on the binding of tritiated cvclic AMP was determined in reaction mixtures containing 100 1.t1 binding protein, 50 ~1 of tritiated cyclic AMP, and 50 ~1 of buffer whose concentration was increased from the normal value of 50 mM Tris-4 mM EDTA. pH 7.5 to a level 20 times the normal concentration. Co, cpm bound at zero-dose cyclic AMP in normal buffer concentration. CB, cpm bound at zero-dose cyclic AMP in higher buffer concentration. Broken lines indicate 95% confidence limits for zero-dose binding in normal buffer.
reactivity with ATP and cyclic GMP. The addition, to the purified binding protein, of a protein kinase
inhibitor fraction isolated from bovine muscle increased both the number of binding sites and the affinity for cyclic AMP (R, = 10' M-' ). Later Gilman [l] reported that the increase in total binding capacity was a non-specific effect and could be duplicated with albumin. The increase in affinity caused by the inhibitor preparations was presumed to be specific but the possible effect of albumin on the association constant was not investigated. We have shown that the increase in binding capacity caused by the addition of bovine serum albumin (BSA) is dependent upon the concentration of albumin present. M~irn~ activation was achieved with a BSA concentration in the incubation mixture of 0.05%. No further effect was observed on increasing the concentration of BSA from 0.05% to 2%. Double reciprocal plots of bound cyclic AMP against free cyclic AMP were used to calculate the association constants and the numbers of available binding sites both in the absence and in the presence of BSA. Straight line plots were obtained and the results, shown in Fig. 3, indicate that maximal activation by 0.05% BSA resulted both in an increase in the association constant from 0.4 . lo9 M-' to 1.0 *10' M-',and in a 60% increase in the number of available binding sites. Thus we have shown that the addition of BSA gave an increase in the association constant identical to that obtained with the protein kinase inhibitor. This increase, previously presumed to be a specific effect [l] can, therefore, also be duplicated nonspecifically with BSA. There have been some indications, although never investigated in detail, that similar effects are possible with other purified binding proteins [ 151.
Fig. 3. Determination protein concentration
I/nM Free [JHjc~cllcAMP of the association constant for cyclic AMP at pH 7.5. , without BSA; binding 15 pg/200 ~1. 0, with 0.05% BSA; binding protein concentration 7.5 !.&200 ~1.
Our results show that the fully activated binding is insensitive to large increases in the protein content of incubation mixtures. Failure to use fully activated binding proteins could contribute markedly to the interference noted in other assays [ 5-7,10] by proteins present in assay samples.
(4) Choice of optimum method of separation of bound and free cyclic AMP
An ideal method should give a clean separation and be free from inter-
ference by materials likely to be present in extracts. It should also be simple,
reproducible, rapid and cheap. Because of the difficulty of accurately counting
tritium-labelled compounds in heterogeneous solution, it is desirable to use a
technique which gives a soluble fraction for counting to avoid the need for
tedious or expensive solubilization procedures.
These requirements ruled out many separation methods: membrane filtra-
tion methods are tedious and expensive, require solubilization before counting
and are reported to be subject to interference by EDTA [ 7,121 and by proteins
[ 161 . Salt precipitation methods are greatly influenced by protein concentra-
tion, and the presence of salts causes counting problems. organic solvent
cipitation methods suffer similar defects and, in our hands, caused excessive
dissociation of the protein-cyclic AMP complex. Ion-exchange methods are
affected by salts, give incomplete uptake of cyclic AMP, and suffer interference
from plasma extracts [ 71 .
Only charcoal (centrifugation)
and Sephadex (column) separation
methods satisfied most requirements. A charcoal separation method was chosen
because it allowed the simpler centrifugation separation step.
Charcoal separation methods are very effective for the separation of low molecular weight ligands such as cyclic AMP. It is, however, important to use the correct amount and grade of charcoal [1,17] ; failure to do so will result in poor performance [7,8]. Optimal conditions must be determined for maximal adsorption of free ligand and minimal adsorption of bound ligand by the adsorbent. With charcoal adsorbents it is relatively easy to get complete adsorption of low molecular weight ligands but it is less easy to get 100% exclusion of bound ligand. Adsorption of bound ligand can be reduced, but not eliminated by adding dextran or protein to the separation mixture [ 181. Our results have shown that the optimum amounts of charcoal and bovine serum albumin used in the assay (see Methods) resulted in the adsorption of more than 99% of the free cyclic AMP and the exclusion of 90% of the bound cyclic AMP. The comparative freedom of the assay from non-specific interference by a wide range of substances (see section B) indicates that the efficiency of separation of bound and free cyclic AMP is constant over a wide range of environmental conditions
(B) Specificity of the assay
Specificity has been defined  as the extent of freedom from interference by substances other than the one intended to be measured. Interference in cyclic AMP assays can be expected from: other nucleotides which cross-react with the binding protein; other proteins present in the biological extract; phosphodiesterase inhibitors; bivalent metal ions added in some extraction procedures and to activate adenylate cyclase. The effects on the cyclic AMP assay of large variations in the concentrations of these potentially interfering factors have been determined:
(1) Interference by other nucleotides The naturally occurring nucleotides most likely to interfere in cyclic AMP assays are cyclic GMP and ATP. Fig. 4 shows the cross-reactivity of cyclic GMP and ATP in the assay. Approximately 200 times as much cyclic GMP as cyclic AMP is needed to give an equivalent reduction in the binding of tritiated cyclic AMP. Consequently cyclic GMP is unlikely to interfere in assays at concentrations reported in normal tissues. ATP does not interfere in this assay at concentrations up to 5 mM which represents an approximately million-fold excess over the tritiated cyclic AMP. This freedom from interference from ATP is significantly greater than that reported in other assays [1,9-111. Gilman [l,ll] , for example, reported 50% inhibition of binding by 1 mM ATP.
(2) Interference by other proteins Interference by proteins and unidentified present in biological extracts is a Common Source
non-dialysable substances of trouble in cyclic AMP
2:;o A A
hucleot idc conce~+ro?`or, Fig. 4. Cross-reactivity of various nurleotides 4.5 lo-`) M tritiated cvclic AMP.
n-0 ar' with cyclic AMP binding protem in the prcsrnce of
The results discussed previously (section A, part 3) indicate that provided sufficient BSA is added to fully activate the binding protein, neither the binding of cyclic AMP nor the separation procedure are affected by large increases in BSA concentration. Further evidence of the freedom from interference by proteins and other non-dialysable substances is given in Tables III and IV.
THE VALIDITY OF DIRECT MEASUREMENTS OF CYCLIC AMP IN PLASMA
The results are means ? S.D. (II = 3) _____-
pmoles cvclic AMP
Plasma * '
plasma* * *
1.06 (r 0.12) 1.oo (1 0.09) 0.45 (i 0.05) 1.00 (2 0.12) < 0.03
21.2 20.0 1X.0 20.0 i 0.6
* Prepared from blood containing as phosphodiesterase inhibitor. ** Prepared from blood containing *** Prepared by incubating plasma, 30
5 units of heparin/ml as anti-coagulant and 2 mg theophylline/ml EDTA as anti-coagulant and phosphodiesterasr Inhibitor. prepared from blood containing 5 units of heparin/ml, for 20 11at
TABLE IV THE VALIDITY OF DETERMINATIONS The results are means r S.D. (n = 8).
OF CYCLIC AMP IN OX-BRAIN EXTRACT
pmoles cyclic AMP
Per ml original extract
5.6-fold diluted extract
9.25 ? 0.93 4.57 f 0.25 6.20 + 0.60
0.066 f 0.002 -
1036 * 104 1024 + 56 - 7.4 + 0.2
* Prepared by treating 1 ml of brain extract with 10 mg of Norit GSX charcoal.
(3) Interference from phosphodiesterase inhibitors EDTA and theophylline are commonly used as phosphodiester~e inhibi- tors. Fig. 2 shows that there is no interference in assays from EDTA at concentrations up to 40 mM. Other experiments have shown that there is no interference from theophylline at concentrations up to 32 mM in the assay samples. These concentrations are much larger than those which might be expected in normal assay samples. (4) Interference from biualent metal ions Bivalent metal ions, such as Ca2+, Zn" and Mg2+, cause interference in some cyclic AMP assays . The effects of these cations on the binding of cyclic AMP are shown in Fig. 5. Concentrations of up to 80 mM Ca*` in the 50 ,ul assay sample (added to a 200 ~1 final volume incubation mixture) caused no decrease in binding. Similarly concentrations of up to 20 mM Zn2* caused no decrease in binding. Concentrations of Zn2+ `greater than 20 mM precipitated in the assay buffer and their effect could not be studied. Interference from Mg2+ became significant (> 2 standard deviations
) at 10 mM and increased steadily with increasing concentration causing 25% inhibition at 80 mM. Magnesium ions are most likely to be present when the assay is used to measure adenylate cyclase activity [ 10,191 and it is unlikely that Mg" concentrations as high as 10 mM will be present in such samples which are normally diluted considerably before cyclic AMP assay. Interference by bivalent cations in other assays has been shown to occur in the separation step . The marked freedom from interference in our assay results in part from the presence of EDTA in the reaction mixture and in part from the use of optimal proportions of charcoal and BSA in the separation procedure. In addition we noted that calcium and zinc ions form sparingly soluble complexes with the assay buffer and this might help to minimise metal ion interference in the assay.
~,i:=rOmM il?D=oO,M !% = BOr&
Fig. 5. The effect of bivaient metaf ians on the binding of tritiated cyclic AMP at zero-dose WLISdetermined in reaction mixtures containing 100 ill binding protein. 50 ~1 tritiated cyclic AMP and 50 1-1o1f solutions of the various metal ions, at the concentrations indicated, in assay buffer. Cm, cpm bound in the presence of metal ions: CO, cam bound in the absence of metal ions. Broken lines indicate the 95% confidence limits for the binding in the absence of metal ions.
(5) Interference from high salt and high buffer concentrations High salt and buffer concentrations usually result when samples are concentrated before assay. Fig. G shows that the assay is insensitive to very Large increases in salt and buffer concentrations.
Zero dose bmdiog
Bindrng I" presence of one ptcomolo cyclicAMP
Fig . C;. The effect of NaCl concentration on cyclic AMP binding: measured at zero-dose and in the presenceof 1 pmole CYCIIC AMP. The effect of increasing the NaCl concentration at zero-dose cyclic AMP was determined by measuring the percentage of cyclic AMP bound in reaction mixtures containing 300 ~1 binding protein, 50 ~1 tritiated eyciic AMP, and 50~1 of NaCl at the concentration indicated. For experiments in the presence of 1 pmoie crclic AMP the 50 i*i aliquots of NaCl solutions were supplemented with 1 pmole unlabelled cyclic AMP. Broken lines indicate 95% confidence limits for binding in the absence of NaCI.
(6) Interference by fluoride ion Many extracts contain fluoride ions. Potassium at concentrations up to 80 mM causes no significant change in zero-dose binding in the cyclic AMP assay.
fluoride in assay samples (> 2 standard deviations)
(C) the assay of cyclic AMP in urine, plasma and tissue extracts
(1) Cyclic AMP in urine Table II shows that, in assays of cyclic AMP in urine carried out on the same sample on six separate occasions over a period of seven weeks, a 1 : 50 dilution gave a value of 7.00 (? 0.52) pmoles/l and a 1 : 100 dilution gave a value of 6.80 (? 0.28) pmoles/l. Treatment of an undiluted urine sample
with charcoal to remove cyclic AMP reduced the cyclic AMP content of 50 ~1 of the undiluted sample to zero (within the standard deviation of the zero-dose binding).
(2) Cyclic AMP in plasma No previously published assay has permitted the direct assay of plasma cyclic AMP levels. The high specificity of the present assay has permitted a simple and rapid direct assay of plasma cyclic AMP. Plasma contains a high level of phosphodiesterase activity which must be completely inhibited immediately on sampling to prevent significant degradation of cyclic AMP. Plasma is usually collected in heparinized tubes to which the poorly soluble phosphodiesterase inhibitor theophylline is added. This procedure has been simplified by using EDTA which acts both as an anti-coagulant and as a phosphodiesterase inhibitor. Freshly collected blood is added to a tube containing 1% of its volume of 0.5 M EDTA pH 7.5, mixed and centrifuged. Table III shows that cyclic AMP levels in plasma treated in this way are identical to those in heparinized plasma containing theophylline. The high activity of phosphodiesterase in plasma was demonstrated by incubation of heparinized plasma containing no inhibitor for 20 h at 30". This resulted in a decrease in measured cyclic AMP from 20 pmoles/ml to < 0.6 pmoles/ml. Incubation of the EDTA/plasma for 20 h at 30" showed no decrease in cyclic AMP content indicating that the added EDTA was sufficient to cause complete inhibition of phosphodiesterase. The validity of the direct assay of plasma cyclic AMP levels was demonstrated by showing that the result was independent of sample size
; by using internal standards; and by showing that the addition of cyclic AMP-free plasma had no effect on the assay. The results are given in Table III. It is of interest to note that this simple procedure for the preparation of plasma cannot be used in many cyclic AMP assays because EDTA interferes both with membrane filtration
techniques [7,12] and with the binding of antibody to cyclic AMP in radioimmunoassays of cyclic AMP [ 131.
(3) Cyclic AMP in extracts of ox brain Many tissue extracts have been reported to contain unknown materials which interfere with saturation analyses. Weller et al. [lo] have reported that extracts of ox brain contained material which enhanced cyclic AMP binding at
low concentrations of extract and which caused inhibition at higher concrntrations. We have used such an extract (a kind gift from Dr M. Weller. Institute 01' Psychiatry, London) and shown that a simple direct assay is possible with the present assay (see Table IV). The increase in binding at low ~~il~entratioi~s of extract noted by Weller et al. [lo] may be explained as being clue to activation of the binding protein by non-dialysable material in the crude extract; the apparent inhibition of binding at high concentrations of extract may be explained as being due to interference by high molecular weight, impurities in the membrane filtration step [ 161 . Failure of these extracts to cause interference in the present assay would then be explained by the use of a fully activated binding protein and a carefully designed charcoal separation procedure which, unlike membrane filtration methods, is insensitive to large changes in protein concentration. Acknowledgements
The authors thank Dr W.P. Grove (MANAGING DIRECTOR
of The Radiochemical Centre) for permission t.o publish this paper, and Dr E.A. Evans for his interest and encouragement. Many helpful discussions with Dr B.L. Brown and Miss J.D.M. Albano (Institute of Nuclear Medicine
, Middlesex Hospital, London) are acknowledged with thanks. References
1 P. Greengard and G.A. Robison (Eds.), Advances in Cyclic Nucleotide Research, Vol. 2, Raven Press,
New York. 1972
2 M. Rodbell, Acta Endocrinol.
Suppl., 153 (1971) 337
3 R.P. Ekins, in In Vitro Procedures With Radioisotopes
in Medicine, International
Agency. Vienna, 1970, pp. 325-353
4 G.P. August and W. Hung. J. Clin. Endocrinol.
Metab., 37 (1973) 476
5 A.L. Latner and K. Prudhoe. Clin. Chim. Acta, 48 (1973) 353
6 P.M. Barling, J.D.M. Albano. S. Tomlinson,
B.L. Brown and J.L.H. O'Riordan,
Biochem. SOC. Trans.,
2 (1974) 75 7 C.P.W. Tsang, D.C. Lehotay
and B.E.P. Murphy.
J. Clin. Endocrinol.
Metab.. 35 (1972) 809
8 B. Rabinowitz and J. Katz. Clin. Chem., 19 (1973) 312
9 B.L. Brown, R.P. Ekins and J.D.M. AIbano,in ref. 1 (19'72) PP. 25-40
10 M. Weller. R. Rodnight and U. Carrera, Biochem. J., 129 (1972) 113
II A.G. Gilman, Proc. Nati. Acad. Sci. U.S., 67 (1970) 305
12 B.h?. Sanborn. R.C. Bhalia and S.G. Korenman.
92 (1973) 494
13 C.W. Parker, in Stefanini. M. (Ed.). Progress in Clinical Pathology, Vol. IV. Grune and Stratton. New
York and London, 1972, p. 119
14 J.D.M. Albano, G.D. Barnes, 1j.V. Maudsley, B.L. Brown and R.P. Ekins, Anal. Biochem., 60 (1974)
15 B.E.P. Murphy, in W.D. Ode11 and W.H. Daughaday (Eds.), Principles of Competitive
Assays, J.P. Lippincott,
1971, p. 130
16 M. Yarus and P. Berg, Anal. Biochem., 35 (1970) 450 17 M.A. Binoux and W.D. Ode& J. Gin. Endocrinol. Metab.. 36 (1973) 303 18 W.D. Ode% in W.D. OdeU and W.H. Daughaday (Ed%), Principles of Competitive Protein Assays:, J.P. Lippincott. 1971, p. 323 19 B.L. Brown, J.D.M. Albano, G.D. Barnes and R.P. Ekins, Biochem. Sot. Trans., 2 (1974) 10
KC Tovey, KG Oldham, JAM Whelan