Organophosphates of the crystalline lens: a nuclear magnetic resonance spectroscopic study

Transcription

1 Organophosphates of the rystalline lens: a nulear magneti resonane spetrosopi study Jak V. Greiner, Stephen J. Kopp, Donald R. Sanders, and Thomas Glonek We quantitated the onentrations of the prinipal organophosphate metabolites present in the intat rystalline rabbit lens, measured the intralens ph, and evaluated dynami hanges during 24 hr inubations, using phosphorus-31 nulear magneti resonane (P-31 NMR) spetrosopy. Tissue perhlori aid extrats prepared from these same lenses were analyzed by this tehnique to verify metabolite identifiations and to quantitate the onentrations of the minor lens metabolites. Values for lens organophosphate onentrations, inluding three groups of previously unidentified phosphorus-ontaining substanes, were established for freshly exised lenses, 24 hr inubated lenses, and lenses inubated in gluose-defiient media. Lens metabolite levels were not adversely affeted by inubation in a medium previously shown to maintain lens larity and ion transport apabilities. Conversely, lens inubation in gluose-defiient media indued signifiant metaboli hanges haraterized by a time-dependent deline in ATP, orresponding inreases in ADP, inorgani phosphate, and phosphorylated hexoses. Catarat formation was noted after inubation in this medium. These findings support the hypothesis that alterations in the organophosphate levels of the lens atually preede hanges in the Na + and K + onentrations and therefore may be the "initiating fator" in formation of lens atarats. (INVEST OPHTHALMOL VIS SCI 21: , 1981.) Key words: rystalline lens, organophosphates, phosphorus-31 nulear magneti resonane spetrosopy, tissue ph The rystalline lens is not a tissue with high metaboli ativity, suh as musle or liver, but it does require hemial energy in the From the Department of Pathology (Dr. Greiner) and the Nulear Magneti Resonane Laboratory (Drs. Glonek and Kopp), Chiago College of Osteopathi Mediine, and the Department of Ophthalmology, University of Illinois Eye and Ear Infirmary (Drs. Greiner and Sanders), Chiago. This study was supported in part by a grant-in-aid from the National Soiety to Prevent Blindness (Dr. Greiner), and ore grant EY awarded to the Department of Ophthalmology, University of Illinois Eye and Ear Infirmary, by the National Eye Institute, National Institutes of Health. Submitted for publiation De. 16, Reprint requests: Jak V. Greiner, Ph.D., Chiago Osteopathi Hospital, 5200 South Ellis Ave., Chiago, form of adenosine triphosphate for ative transport proesses. Moreover, lens ATP is utilized in syntheti and degradative pathways that are required to maintain tissue growth, struture, and funtion. Despite this desriptive understanding of lens metaboli ativity, little is known about the atual dynami energetis of intat rystalline lens metabolism, partiularly the overall metaboli events that preede, and therefore predispose, the lens to initiate atarat formation. Tehnial limitations have previously preluded measurement of dynami metaboli ativity in intat tissues. Reent advanes in phosphorus-31 nulear magneti resonane (P-31 NMR) tehnology, however, have provided the neessary bioinstrumentation to analyze dynami metaboli events nondes /81/ $01.40/ Asso. for Res. in Vis. and Ophthal., In.

2 Volume 21 Number 5 NMR of lens organophosphates 701 trutively and quantitatively as they our in intat, funtioning mammalian tissues. '~ 6 P-31 NMR spetrosopy permits rapid, aurate quantitative and qualitative determinations of the prinipal organophosphates present in the intat tissue.' 3 6 ' Adenosine triphosphate, adenosine diphosphate, reatine phosphate, and the phosphorylated hexoses and trioses are among the phosphati metabolites that an be aurately quantitated in the intat tissue. With in vitro tissue inubation tehniques, quantitative time-dependent metaboli hanges an be monitored ontinuously for orrelation with funtional disturbanes during defined inrements in time in the same tissue. 7 ' 8 We suessfully studied the dynami measurements of organophosphate metabolite levels in the intat mammalian rystalline lens during ontrol and experimental inubations. We desribe the signifiane of these findings and the potential use of this analytial tehnique in the study of ataratogenesis. Methods Surgery. Albino rabbits, weighing 2 to 3 kg eah, were killed with sodium pentobarbital injetions, and both eyes were enuleated immediately. The lenses were obtained by opening the eyes at the posterior pole, gently separating the vitreous humor to the side, separating the zonules with urved blunt sissors, and removing the lens with a Parafilm-oated wire lens loupe. In vitro inubation. Lenses for inubation were freshly exised, isolated, weighed, and plaed together in a tared 12 mm NMR tube ontaining a volume of physiologi Earle's buffer (Difo Laboratories, Detroit, Mih.) at 37 C, ph 7.4. This buffer [NaCl gm/l (116.4 mmol); dextrose (5.6); KC (5.4); CaCl 2, anhydrous (1.8); MgSO (1.4); NaH 2 PO 4 H 2 O, (0.9); NaHCO 3, (26.4)], with an osmolarity of 295 mosm, has maintained lens larity and ion transport apabilities for at least 24 hr of in vitro inubation. 9 All inubated lenses were equilibrated in Earle's buffer for 2 hr at 37 C prior to initiation of eah experiment. At eah 1 hr point during the experimental inubation periods, used buffer was aspirated from the base of the NMR tube, the lenses were washed three times with fresh buffer, and fresh buffer was added to a onstant volume. This proedure was neessary to prevent gluose depletion of the buffer bathing the lenses in the NMR tube. For the temporal ontrol phase, 10 lenses were inubated for 24 hr in physiologi Earle's buffer. Individual P-31 NMR profiles were aquired during onseutive 1 hr intervals for the duration of the entire inubation period. Gluose-defiient inubations were onduted with gluose-defiient Earle's buffer, aording to the same analytial protool. The lenses in this phase of the study were inubated for a 12 hr period. After the experimental inubations, the lenses were weighed and frozen in liquid nitrogen and prepared for perhlori aid (PCA) extration. 10 Ten freshly exised lenses, whih served as noninubated ontrols, were also prepared for PCA extration by freezing in liquid nitrogen immediately after exision and weighing. Lens PCA extrats. The frozen lenses were pulverized to afinepowder with a stainless steel mortar and pestle hilled with liquid nitrogen. The mortar and pestle were maintained in a liquid nitrogen bath during the pulverization proedure. The tissue powder was added to a entrifuge tube ontaining 0.1 vv/v 60% PCA prefrozen in liquid nitrogen. The powder was stirred ontinuously while it was warmed to a paste onsisteny. The sample was then immediately entrifuged at 43,500 x g for 15 min at -5 C. This proedure made possible omplete oating of the tissue partiles with PCA and subsequent extration of the powder at a temperature of about 20 C. After entrifugation, the supernatant was immediately transferred to an equivalent volume of ION KOH, and the ph was adjusted to 10. The sample was entrifuged to remove preipitated KC1O 4. The final supernatant was passed through a potassium Chelex-100 olumn to remove alkaline earth and transition metal ations. The olumn effluent was lyophilized and redissolved in 1.0 ml of 20% D 2 O for NMR analysis. The preparation of the extrats and the P-31 NMR alibrations and analyses were performed aording to aepted and well-validated proedures previously desribed for P-31 NMR intat tissue and tissue extrat analysis. 3 ' 6 Subsequent to the sample NMR analysis, total phosphate determinations were performed on the tissue extrats and referened to the extrated protein onentrations to enable expression of the metabolite levels as onentrations in miromole per gram of lens protein. P-31 NMR spetrosopy. The NMR spetrometer used in this investigation was a Niolet NT-200 system equipped with deuterium stabilization, variable temperature, and Fourier-transform ap-

3 702 Greiner et al. Invest. Ophthalmol. Vis. Si. November 1981 abilities operating at MHz for P-31. This system is interfaed to a wide-bore Oxford superonduting magnet (4.7 T). Intat lenses were analyzed under nonspinning, proton-oupled onditions. Extrats prepared from these lenses were analyzed with and without proton deoupling, while the sample was being spun to enhane signal resolution. Twelve-millimeter sample tubes were used for the analysis of the intat lenses. In this manner usable data were obtained in as little as 5 min, although 1 hr was needed to gather the bulk of the P-31 NMR data presented here. For the analysis of the PCA extrats, the 1.0 ml samples were plaed in NMR miroell assemblies so that signal intensities ould be enhaned as muh as possible for the quantitation of minor lens metabolites. The NMR analyses were onduted at 37 C. Chemial shift data are reported relative to the usual standard of 85% inorgani orthophosphate (Pi)"; however, the primary internal standard was the natural glyerol 3-phosphorylholine (GPC) resonane of the lens. This ompound has a relatively onstant hemial shift for a phosphate ( 0.13 ppm), whih is not influened by variable physiologi ph, ioni strength, or ounteration onditions. l2 A variety of spetrometer onditions were used in this study; however, those desribed below, whih were used in the miroell analysis of lens PCA extrats, are typial: pulse sequene 1 pulse; pulse width 9 fxse (45 degree flip angle); aquisition delay 200 /use; yling delay 200 /xse; number of sans 144,100; number of data points per free-indution-deay 16,384; aquisition time 1.64 se; sweep width ± 2500 Hz. In addition, a omputer-generated filter time onstant introduing 0.6 Hz line broadening was applied. Data redutions, inluding peak area and hemial shift measurements, were effeted with the spetrometer's omputer. The hemial shifts reported here follow the International Union for Pure and Applied Chemistry onvention and are reported in the field-independent units of parts per million (ppm); oupling onstants (J values) are reported in units of hertz (Hz). Chemial identifiatioti of intat tissue resonane peaks. P-31 NMR spetrosopi analysis performed on the intat rabbit lens yields a spetrum in whih eah peak orresponds to a single phosphorus-ontaining funtional group having a disrete resonane shift position. The preise resonane shift position of eah peak is a harateristi physiohemial marker for individual organophosphate metabolites present in the tissue. The shift position is determined by the moleular and maromoleular environment of the phosphorus atom ontained in eah different metabolite; however, the maromoleular environment to whih the phosphorus atom of various metabolites is exposed annot be assumed to be uniform throughout the ell. Ioni strength, ph, and ationi and protein onentration differenes in ellular miroompartments of the lens may ompliate the hemial identifiation of a resonane peak reorded from intat tissue. For this reason, the hemial identity of eah resonane peak reorded from the intat lens is based on several physial and hemial riteria: (1) magneti resonane analysis of intat lenses and subsequent analysis of PCA extrats prepared from these same lenses to demonstrate a orresponding pattern of resonane peaks; (2) analysis of the tissue extrat under proton-oupled onditions to define peak multipliities of proton-deoupled resonane peaks and to determine the orresponding peak J values (spaing intervals between multiplets of the same deoupled peak, in Hz); (3) variable ph titration of the tissue extrat using P-31 NMR analysis to define the ph-dependent resonane shift urve for eah peak (the shift urve measures the mole ratio of an aidi anion and its orresponding onjugate base involved in protonation-deprotonation reations); (4) magneti resonane analysis of the extrat under proton-oupled and proton-deoupled onditions at various sample ph values alter addition of known phosphate ompounds to the sample to demonstrate that a speifi resonane peak enhanes, ompletely superimposes, and migrates in the ph spetrum as the added (known) ompound; (5) P-31 NMR ph titration analysis of aqueous mixtures of known organophosphate ompounds under idential extrat onditions (ounteration, ioni strength, and temperature) to demonstrate that the ph titration-resonane-shift-position urve derived from peaks ontained in the extrat superimpose with orresponding known ompounds present in the aqueous mixtures; and (6) onfirmation of peak assignments based on the known theoretial moleular hemistry that defines the magneti resonane patterns of speifi phosphate moleules. By a omparison of the ph titration urves derived from standard phosphates with those obtained from the lens PCA extrats, the ommon energy metabolites involved in intermediate metabolism were identified with a high degree of ertainty. These peak assignments were further reinfored by the demonstration that, under expanded

4 Volume 21 Number 5 NMR of lens organophosphates 703 INTACT RABBIT LENS ;ATP DN I ' ' ' -90 PPM Fig. 1. P-31 NMR spetrum of three rabbit lenses aumulated over a period of 6 hr to reveal the fine struture of the spetrum. The lenses were maintained during the experiment by Earle's buffer (with gluose) at ph 7.4 at 37. Arrow indiates a new phosphorus metabolite at 6 ppm; SP, other sugar phosphates; ADP and ATP are shown with respetive y-, a-, and /3-resonane signals; DA 7, dinuleotides; NS, nuleoside diphosphosugars and related speies. The hemial shift sale follows the shift onvention of the International Union for Pure and Applied Chemistry and is given relative to the resonane position of 85% phosphori aid. spetral onditions, known phosphates superimposed with speifi extrat resonane peaks throughout the ph titration range. In regions where a number of losely spaed resonanes were found, however (e.g., the hexose region), the peak assignments were reinfored by seletively shifting the resonane of interest with respet to the others by judiiously adjusting ioni strength 13 or by hanging the ounteration 14 in solution and subsequently adding known ompounds' to the sample. Mathematial analysis of dynami lens hanges in gluose-depleted medium. The time-ourse data obtained in this study were redued by a leastsquares regression analysis using a polynomial expression of the form y = Ax 1 + Bx 2 + Cx + D. The regression analysis was arried out until the hi-square was redued to 0.1%. The best fit values of the onstants are presented in tabular form along with their standard deviations. The thirddegree expression was neessary to adequately approximate the real time-ourse data obtained from the intat lens tissue. The loation of maxima and minima in the time ourse of hanging rate expressions was obtained by differentiation of the lous determined from the regression analysis. Results The P-31 NMR spetrum obtained over a period of 6 hr from intat rabbit lenses inubated in ontrol Earle's buffer is shown in Fig. 1. The 6 hr measuring time was neessary to easily observe the minor resonanes above the bakground noise. Metabolite onentrations derived from lens spetra are summarized in Tables I and II. The signal indiated by the arrow at 5.7 ppm arose from a phosphorus-ontaining moleule of unknown origin and hemial nature. The large signal immediately up-field arose from a- glyerophosphate (agp). This ompound is muh more abundant in lens than in any other intat tissue thus far examined by P-31 NMR. 15 The sugar phosphate band (SP) was omposed of two losely spaed groups of resonane signals arising from other triose

5 704 Greiner et al. Invest. Ophthalmol. Vis. Si. Noventier 1981 Table I. Lens P-31 NMR profiles Chemial shift (ppm*) Amount (as % of the total P in i he profile) Control PC A extrat Phosphati ompound Intat lens a, ; /3, ; y, a, ; /3, PC A extrat Intat lens Freshly exised 24 hr inubation Gluosedepleted Unknown 8 Unknown 13 Unknown 8 Trioses Hexoses PE PC Pi GPE GPC PCr Unknown 8 ATP ADP Dinuleotides Nuleoside diphosphosugars Unknown , 10.72, K 6.00 F 4.30 G 3.78 H ' , -5.50, E a, ; /3, ; y, -5.80* a, ; 0, L M i A Field-independent nulear magneti resonane units relative to the shift position of the 85% inorgani orthophosphori aid referene phosphate at 25. B Compound, as of this writing, is not identified with any known phosphorus-ontaining biomoleule. Minor omponent not detetable in the intat tissue. "Resonane band, omprising a group of separate resonane signals, is present. B Three separate resonane signals are deteted. F J(P-H[phosphorus-hydrogen]) = Hz. G Complex resonane band, the prinipal resonane signals of whih ome from the agp triplet; J(POCH) = 6.68 Hz. "Hexose and pentose phosphates, prinipally the resonane triplet of inosine monophosphate; J(POCH) = 3.81 Hz. 'Cannot be determined in the intat tissue as a separate resonane band but is ombined with the Pi signal. J J(POCH) for the triplet = 5.63 Hz. K J(POP, a/3) = Hz; J(POP), 0y) = Hz. 1 J(POP, a/3) = Hz. "Prinipal resonane signals of this band arise from the P,P'-diesterified pyrophosphate residues of NAD and NADH. "Complex resonane band omposed of three sets of overlapping, 31 p. 31 P ) ab NMR multiplets. phosphate sugars in one group and hexose and pentose phosphate sugars in the other. The next up-field resonanes were the signals from Pi and two small signals arising from the phosphodiesters, glyerol 3-phosphorylethanolamine (GPE) and GPC. These were followed by the ionized y hain-terminal phosphate of ATP and the (3 hain-terminal phosphate of ADP; the alpha-esterified phosphates of ATP and ADP; the P,P'-diesterified pyrophosphate residues of the dinuleotides, prinipally niotinamide adenine dinuleotide (NAD) (DN); a low-amplitude broad resonane of nuleoside diphosphosugars (NS); and finally the /3 hain-middle phosphorus of ATP. When the ontrol Earle's buffer that ontains gluose was hanged at regular 1 hr intervals, the intat lens NMR profile did not hange adversely at 37 C for at least 26 hr. The hemial shift position of eah peak as well as their respetive amplitudes remained idential within the preision of the NMR measurement, whih in this example was ±0.015 ppm. The stability of these intat lens preparations permitted quantitative data of the major lens metabolites to be obtained from a single lens. Four-hour signal averaging was required for single lens analysis. The peak assignments given for the intat lens spetrum orresponded to assignments reported for a large variety of mammalian

6 Volume 21 Number 5 NMR of lens organophosphates 705 INTACT LENS CONTROL PCA EXTRACT CONTROL PCA EXTRACT MINUS GLUCOSE Fig. 2. P-31 NMR spetra from intat rabbit lens and its orresponding PCA extrat and from an extrat of lens inubated in a gluose-defiient medium. Some of the signals are multiplets. DA 7, Dinuleotides; SP, sugar phosphate resonane group; NS, nuleoside diphosphosugars; PCr, phosphoreatine. Signals at 17.9, 10.7, and 28.0 ppm represent phosphati metabolites of unknown hemial nature. and other NMR. 2 3) ' living The tissues studied by P-31 resonane from inorgani orthophosphate was readily identified as a result of its aumulation to high onentrations after death of the tissue. The other resonane signals were haraterized by omparing the intat tissue spetra to similar spetra obtained from lassi PCA extrats of the same tissue. This is illustrated in Fig. 2, in whih an intat rabbit lens spetrum aumulated over a period of 1 hr is ompared to orresponding PCA extrat spetra.

7 706 Greiner et al. Invest. Ophthalmol. Vis. Si. November 1981 Table II. Conentrations of aid-extratable phosphates in rabbit lens (/xmol/100 gm of fresh tissue) Referene Total phosjihortis ATP ADP Pi PCr agp Hexoses Trioses NS Present paper van Heyningen and Pirie 34 LePage 27 Haushildt et al. 35 Frohman and Kinsey 28 Pirie et al. 3(i Klethi and MandeF Nordmann and Mandel 38 Palm 39 van Heyningen 40 Pirie C 663 E C 245 F D D A Summed onentrations of all unidentified speies. B AMP resonane not identified as suh in the phosphorus resonane spetrum. Pooled adenyl nuleotide fration. "Pooled phosphorylated sugar fration. 15 Pooled adenyl diphosphate and triphosphate frations. K Combined nuleoside triphosphates. The signals from the extrats, whih beause of the favorable physial properties of dilute aqueous solutions were highly resolved, were identified through their hemial shift position and other spetral properties as detailed in Methods. The favorable signal-to-noise ratio of the extrat spetrum yielded usable quantitative data on several minor phosphorus resonanes as well as on the prominent resonanes. The extrat spetrum showed two additional signal groupings at 17.9 and 10.7 ppm arising from ompounds that have not been identified as to the nature of their soure phosphati ompound. The hemial shift position of these two signals suggest that the soure materials were not phosphates or phosphoroamidates ommon to biologi systems, sine suh ompounds do not generate resonane signals at the loation observed for the unknowns. The 17.9 ppm resonane was in a harateristi position of phosphonates, i.e., ompounds ontaining the C P bond. 16 The 10.7 ppm resonane ould not be assigned with any degree of ertainty beause it lay in a region of the spetrum where only a few model ompounds are known, inluding anhydrides of phosphoni aids and their derivatives. 17 Between the 10.7 ppm resonane and the sugar phosphate grouping was a doublet orresponding to the unidentified resonane observed in Fig. 1. The down-field resonane position of this ompound is unique, 18 and it is unlikely that the substane is a rare sugar phosphate. The substanes giving rise to the 17.9 and 10.7 ppm signals were only sparingly soluble in the PCA extration solvent, suggesting that the soure moleules may be of relatively high moleular weight similar to the "presumptive phosphopeptides" isolated from bovine lens by Bettelheim 19 and Bettelheim and Wang. 20 These low field signals, however, annot arise from phosphoserine or phosphothreonine residues in peptides or proteins, sine their shifts were 3.7 2I> 22 and ppm, respetively, and varied only a few tenths of a ppm with the nature of the protein. The sugar phosphate resonane band in almost all tissues examined exhibited three harateristi groups of signals: a band orresponding to triose phosphates at 4.4 ppm, a band orresponding to hexose or pentose phosphates at 3.8 ppm, and a band orresponding to phosphoholine (PC) and phosphoethanolamine (PE) at 3.3 ppm. The prinipal triose phosphate in the lens is agp, whih exhibits a harateristi triplet due to the oupling to the neighboring CH 2 group. The hexose and pentose resonane band shows a omplex system of overlapping resonane multiplets that, depending on on-

8 Volume 21 Number 5 NMR of lens organophosphates 707 RABBIT LENS, MINUS GLUCOSE DN AMP GPC GPE PC PE Unknown^ ATP a B SP 1 ON NS 0 HR ditions, may show prominent signals from moleules suh as frutose-1, 6-diphosphate, inosine monophosphate, or adenosine monophosphate. Although the resonane from phosphoreatine was of very low intensity in the lens, it ould be identified at 3.1 ppm. Ionized polyphosphate end group resonanes from other nuleoside triphosphates, prinipally guanosine triphosphate, ould also be identified at 6 ppm. The ionized end groups of the nuleoside triphosphates were split into doublets though their interation with the middle group phosphate at 21.5 ppm; this phenomenon is well resolved in the middle spetrum of Fig. 2. Like the end group of ATP, the end group of ADP was also a doublet that, under the san onditions employed, lay lose to the ATP end group doublet, so that the high-field arm of the ATP doublet and the low-field arm of the ADP doublet nearly overlapped. The esterified phosphate end group region at 11 ppm was ompliated by the presene of numerous signals arising from the dinuleotide ofators of intermediate metabolism. In this region lay the resonanes of NAD, niotinamide adenine dinuleotide phosphate (NADP), flavin adenine dinuleotide, and oenzyme A, as well as resonanes from other monoesterified pyrophosphate groups. Up-field of this group at -13 ppm were a group of multiplets orresponding almost exlusively to nuleoside diphosphosugar resonanes. 23 The bottom spetrum of Fig. 2 shows a VyW 2 HR 5 HR PPM Fig. 3. P-31 NMR spetra from three time points during an inubation of lens in a gluose-defiient medium. PCA extrat of lenses deprived of gluose for 5 hr at 37 C. Signals from the same phosphorus groups ould be deteted, although their relative proportions were markedly altered. In addition, a new resonane appeared at

9 708 Greiner et al. Invest. Ophthalmol. Vis. Si. November 1981 Fig. 4. Time ourse for the hanges in the lens phosphate profile when the lens is inubated in a gluose-defiient medium. The ph urve is the intralens ph determined from the resonane shift of the Pi signal ppm. This resonane has not been identified, although its hemial shift suggests that it may be a redued phosphorus grouping, suh as a phosphine, 24 and not a phosphate. Intat lens spetra taken at various times while the lenses were being inubated in a gluose-free buffer are shown in Fig. 3. Under these onditions the phosphate profile showed the loss of ATP with the onomitant rise of Pi, so that after 5 hr the major resonane signal was from Pi and the resonane signals from ATP had virtually disappeared into the spetral noise. A time ourse plot orresponding to suh spetra is shown in Fig. 4. When gluose was removed from the buffering medium, the onentration of ATP fell to values orresponding to less than 10% of the total phosphorus while the onentration of Pi inreased orrespondingly. ADP also inreased, approximately doubling its value at the end of 6 hr, and the hexosephosphate band also inreased after about 4 hr. Examination of the PCA extrat shows that the prinipal phosphate ontribution to the hexose band inrease was adenosine monophosphate. The levels of the triose sugars, the nuleoside diphosphosugars, the phosphodiesters suh as GPC, and the unknown phosphorus-ontaining moleules were unhanged during this time ourse. Lens ph. Phosphate hemial groups bearing a dissoiable proton underwent hemial shift hanges as a funtion of ph in the range orresponding to the pka of the dissoiable proton. This is illustrated in Fig. 5, whih shows the P-31 spetrosopi ph titration urves at 37 of several phosphates of biologi origin whose resonane signals have been deteted in lens extrats. This property renders the hemial shift behavior of moleules suh as Pi or agp useful in determining the ph value of media in whih the ion is a solute. Note that in Fig. 5 the phosphorus resonane signals may be up-field or down-field of eah other, depending on the solvent ph; therefore the ontrol of ph is absolutely essential for the proper interpretation of phosphate hemial shifts. We and a number of other investigative groups 2 " 6 have used the hemial shift of Pi to measure intraellular ph values in a number of intat ellular systems. Applying these priniples to the lens and using the shift of both Pi and agp, we have determined that the average ph of the interior of the lens is 6.9. This value is surprisingly low for mammalian tissues; it is 0.5 ph unit lower than the average value of 7.4, whih has been reported as the probable ph of the lens. 25 ' 26 When gluose was removed from the buffering medium (ph 7.4), the intralentiular ph, as measured by the hemial shift of both phosphates, began to drift toward higher values as the ATP ontent of the lens was depleted (Fig. 4). At the end of 6 hr the ph had inreased more than 0.3 units to a ph of 7.2. This behavior is atypial for mammalian ells, whih usually beome aidi on loss of ATP, and it is diret evidene in the lens for an ative hydronium ion pump.

10 Volume 21 Number 5 NMR of lens organophosphates Fig. 5. Control rabbit lens PCA extrat ph titration data. P-31 NMR-pH titration urves in water (5% D 2 O) at 37 C with potassium as the ounteration, and a total salt onentration approximately 100 mm. F-l,6-diP, Frutose-l,6-diphosphate; Glu-l-P, gluose-1-phosphate. Table III. Coeffiients for the expression, y = Ax 3 + Bx 2 + Cx + D, obtained in the linear regression analysis of lens time-ourse data* Component of time ourse B Value of oeffiient D Adenosine triphosphate Adenosine diphosphate Pi Hexose phosphate ph ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± *The ordinate is the perent of total phosphorus deteted and the absissa is time in hours. Other intat tissues suh as hiken petoralis musle 3 ' 15 frequently show more than one intraellular pool of Pi; thus far, we have not observed this property in the lens. The data of Fig. 4 were subjeted to a linear-regression analysis, so that the rate of hange of any omponent ould be determined at any time. It was found that a thirddegree equation of the form y = Ax 3 + Bx 2 +Cx + D was neessary to fit the time ourse data adequately within the 6 hr period examined. Only ATP, ADP, Pi, the hexose phosphates, and the ph urves were analyzed, sine the other urves varied little over this period of time. The oeffiients of the above equation, whih were determined from the data of Fig. 4, are given in Table III along with their standard deviations. The fit to the above polynomial was reasonably preise, as indiated from the standard deviations in the value of the oeffiients and the fit of the regression-analysis-derived urve drawn among the data points in Fig. 4. The ATP urve exhibited a sigmoid shape,

11 710 Greiner et al. Invest. Ophthalmol. Vis. Si. November 1981 with a maximum rate of onsumption somewhere near 3 hr. The derivative of the fitted ATP polynomial was used to alulate the preise position of the maximum rate of ATP utilization. This was found to our at 2.9 ± 0.8 hr. Further analysis indiated that ADP aumulation reahed a minimum value at 3.7 ± 0.3 hr, Pi aumulation reahed a maximum at 3.8 ± 0.2 hr, and hexose aumulation obtained a maximum at 3.0 ± 0.2 hr. The minimum hange in the intralens ph ourred at the beginning of the time ourse. Disussion The mammalian lens, like numerous other intat tissues, produes a highly resolved P-31 NMR spetrum that exhibits resonane signals from the ommon low-moleularweight metabolites of intermediate metabolism. Compared to profiles of a variety of mammalian tissues, 3 ' 4 that of the intat rabbit lens exhibits several distintive features. The tissue is rih in glyolyti sugar phosphates and in the dinuleoside ofators required for the metabolism of these phosphates. The tissue also shows a surprising amount of low-moleular-weight phosphodiester and nuleoside diphosphosugars. Readily deteted were the resonanes from ATP, agp, Pi, various phosphorylated sugars, and several nuleotides. In addition, these metabolites ould be quantitated in reasonably preise agreement with hemial analyses performed through lassi biohemial proedures (Table II). P-31 NMR, however, made possible the simultaneous determination of phosphorus-ontaining ompounds on a single intat lens without the extensive preautions that are neessary to avoid breakdown of deliate high energy phosphates in destrutive proedures suh as homogenization. The simpliity, auray, and omprehensiveness of this method are extremely attrative. The intat lens phosphorus spetrum exhibits the highest resolution that has yet been obtained from an intat tissue. In addition to ATP and prominent sugar phosphates, the phosphodiesters GPE and GPC are readily observed above the bakground noise and quantitated; similarly, ADP, the dinuleotide signal, and the nuleoside diphosphosugar resonane an be preisely determined. With suh detailed spetrosopi data, we believe that we have, for the first time, deteted a resonane signal from an unknown phosphati moleule in the intat tissue, thus preluding the possibility that the new substane, also observed in the PCA extrat, was generated by the PCA preparative proedure. The ATP/inorgani phosphate ratio at 2.15 is similar to that exhibited by other tissues, whereas the ATP/ADP ratio at 6.44 is quite high. The lens does not ontain appreiable quantities of phosphoreatine (hemial shift 3.1 ppm), although this may be a speies variation. This finding is in ontrast to those of other reports, in whih the ourrene of phosphoreatine in appreiable quantities in rabbit lens has been reported. 27 ' 28 From the phosphorus profile, the phosphorylation potential 29 of the lens, ATP/ADP Pi, is alulated to be per mole. In the intat lens, the hemial shift of the ADP /3-group is 0.1 ppm higher than observed in other tissues, partiularly musle 3 and syntheti systems 3 30 ' involving ATP, ADP, and PCA extrats. This hemial shift differene does not appear to be a funtion of ph involving separate pools for ADP and ATP, sine the relative shift differene remains onstant throughout the time ourse of the death of the lens when the ph, monitored by Pi and agp is inreasing. This hemial shift differene is unusual and undoubtedly reflets a speifi property involving the miroenvironment of the nuleotides in the lens. The hemial shift of ATP at 5.62 ppm is nearly idential to that observed in various musles ( 5.60 ppm), 3 whih suggests that it is the miroenvironment of ADP that has speifi alterations in lens relative to that found in a number of other tissues. The relative y-group and /3-group ATP hemial shifts in the intat tissue support the suggestion that ATP exists as its monomagnesium omplex. This finding parallels that in intat musle 3 and liver mitohondria. 31 The ph of the intat lens measured spe-

12 Volume 21 NMRoflensorganophosphates 711 Number 5 trosopially is 6.9, whih is a full 0.5 ph unit lower than the ommonly aepted value for lens ph. 25> 26 This marked disrepany undoubtedly reflets the fat that previous determinations involved disruption of ellular integrity and not the inability of the NMR method to determine ph in lens. The inorgani phosphate signal in phosphorus spetra from intat tissues has been used by investigators as an intraellular indiator of intat tissue ph 32 ' 33 ; however, our determination involved the shift position of two intraellular metabolites, Pi and agp. The ph value given by eah resonane shift was the same to within 0.02 ph unit. This agreement supports the hypothesis that the resonane shift of Pi reflets the intraellular ph and is not the result of nonspeifi solvent effets. If solvent effets were operant, it would be neessary that they influene the alohol ester to the same degree and in the same sense that they influene the inorgani anion; suh a hemial anomaly is not likely. The PCA extrat spetra exhibit relatively intense signals from several phosphates of intermediate metabolism that have been identified through their harateristi spetrosopi properties and their phosphate titration data; these are agp, PE and PC, GPE and GPC, phosphoreatine, ADP and ATP, and NAD. In addition, three heretofore unidentified groups of phosphate moleules have been deteted. The resonane at 6 ppm is unique to lens tissue; those at 10 and 18 ppm have also been deteted in other tissues, in partiular, liver and kidney. The 18 ppm resonane is harateristi for phosphoni aids. The 10 ppm signal is unidentified exept that it annot arise from an orthophosphate or any of its ester or anhydride derivatives. None of these signals orresponds to known resonanes from a variety of phosphorylated proteins. Metaboli deay proesses. When the gluose buffer is replaed with a gluose-defiient buffer, the tissue deays metabolially before evidene of gross physial hanges, whih involve formation of a atarat-like opaity. The metaboli deay involves the loss of ATP, with the onomitant inrease of Pi, a sugar phosphate, and ADP. Other ompounds are not mobilized to sustain ATP levels as they are, for example, in the musle where energy soure-deprivation results in the atabolism of glyogen with the subsequent prodution of triose phosphates detetable in the P-31 NMR spetrum. 3 The atastrophi loss of ATP begins about 1 hr after the removal of gluose-ontaining media, with the lag-time probably refleting the size of the intralens gluose pool. 16> 17 The optimum rate of ATP onsumption ours at 2.9 hr, whih oinides with the optimum rate of hexose phosphate aumulation. The optimum rates of ADP and Pi aumulation lag behind this value by approximately 0.8 hr. The hemial destrution of ATP an be interpreted as a straightforward hydrolysis of ATP to ADP and Pi, with the adenylate kinase ativity, 2ADP. ATP + AMP, operating to generate ATP from ADP with the resultant prodution of AMP; AMP is the probable soure of the inreasing sugar phosphate resonane. The aumulated endproduts of the overall reation are AMP and Pi. Thus inteqdreted, the first event that ours after the loss of gluose from the sustaining buffer is the phosphorylation of intralens metabolites by ATP. Shortly thereafter, this pool of metabolites is exhausted, whereupon ATP annot be regenerated from ADP, and ADP begins to aumulate. The loss of energy ours as ATP ompromises the ability of the lens to pump hydronium ions (and, presumably, sodium ions as well). Under the onditions of our experiment, whih is parallel to standard linial praties, this loss results in an elevation of the intralens ph. The speifi ontribution of this more alkaline intraellular lens ph to ataratogenesis in laboratory models is unertain. The ATP onentration deteted in the intat lens and lens extrats by P-31 NMR analysis was within the range of those values previously reported (Table II). The Pi ontent is the lowest value reported. This finding is onsistent with the NMR method of analysis, whih requires minimal hemial treat-

13 712 Greiner et al. Invest. Ophthalmol. Vis. Si. November 1981 ment of the sample, thereby dereasing the opportunity for hydrolyti breakdown of labile ompounds. The phosphoreatine onentration is low, as is the agp, but in aord with previously reported values. 28 The values obtained for the holine and ethanolamine derivatives are also in aord with the same general onentration range as observed previously. 34 The ratio of the holine to the ethanolamine derivatives agrees rather well. The underlying metaboli hanges that result in the funtional and morphologi damage assoiated with atarat development are not well understood. Physiologi stressors have been impliated as etiologi fators that predispose the rystalline lens to atarat formation, and the funtional 42 and morphologial 43 ' 44 damage resulting from the effets of these stressors are desribed in animal models. The use of NMR spetrosopy on the intat rystalline lens allows omprehensive biohemial analysis and permits subsequent funtional (measurement of Na and K levels) and morphologial (examination by eletron mirosopy) analyses to be performed on the exat same lenses. Moreover, NMR spetrosopy is a totally nondestrutive tehnique, whih not only permits the measurement of metaboli hanges in live tissue as they are ourring and permits examination of tissue morphology by eletron mirosopy in the same tissue sample but also allows for the monitoring of tissue reovery subsequent to experimental therapeuti intervention. A tehnique suh as this an substantially enhane our understanding of lens ataratogeni mehanisms by providing a more omprehensive view of the multifatorial basis of this degenerative proess. REFERENCES 1. Moon RB and Rihards JH: Determination of intraellular ph by 31 P magneti resonane. J Biol Chem 248:7276, Hoult DI, Busby SJW, Gadian DK, Radda GK, Rihards RE, and Seeley PJ: Observation of tissue metabolites using 3I P nulear magneti resonane. Nature 252:285, Burt CT, Glonek T, and Barany M: Analysis of phosphate metabolites, the intraellular ph, and the state of adenosine triphosphate in intat musle by phosphorus nulear magneti resonane. J Biol Chem 251:2584, Glonek T: Appliations of 3I P NMR to biologial systems with emphasis on intat tissue determinations. In Phosphorus Chemistry Direted Towards Biology, Ste WJ, editor. New York, 1980, Pergamon Press, pp Navon G, Ogawa S, Shulman RG, and Yamane T: 3I P nulear magneti resonane studies of Ehrlih asites tumor ells. Pro Natl Aad Si USA 74:87, Gadian DG, Radda GK, Rihards RE, and Seeley PJ: 31 P NMR in living tissue. In Biologial Appliations of Magneti Resonane, Shulman RG, editor. New York, 1978, Aademi Press, In., pp Dawson MJ, Gadian DG, and VVilkie DR: Contration and reovery of living musles studied by 3I P nulear magneti resonane. J Physiol 267:703, Hollis DP, Nunnally RL, Jaobus WE, and Taylor GJ IV: Detetion of regional ishemia in perfused beating hearts by phosphorus nulear magneti resonane. Biohem Biophys Res Commun 75:1086, Sanders D, Peyman GA, MClellan K: Dextran 40- ontaining inubation media. Effet on lens eletrolyte and water balane. Arh Ophthalmol 97:156, Glonek T and Marotta SF: 31 P magneti resonane of intat endorine tissue: adrenal glands of logs. Horm Metab Res 10:420, Glonek T: Aqueous tetrahydroxyphosphonium perhlorate as a narrow-line 31-P NMR referene substane. J Magneti Reson 13:390, Burt CT, Glonek T, and Barany M: Phosphorus-31 nulear magneti resonane detetion of unexpeted phosphodiesters in musle. Biohemistry 15:4850, Costello AJR, Glonek T, and VanVVazer JR: Phosphorus-31 hemial shift variations with ounteration and ioni strength for the various ethyl phosphates. Inorg Chem 15:972, Glonek T, Kleps RA, Griffith EJ, and Myers TC: Phosphorus-31 nulear magneti resonane studies on ondensed phosphates. I. Some fators influening the phosphate middle group hemial shift. Phosphorus 5:157, Glonek T, Burt CT, and Barany M: 3I P nulear magneti resonane analysis of intat tissue inluding several examples of normal and diseased human musle determinations. In NMR Basi Priniples and Progress. Vol. 19. NMR in Mediine, Damadian R, editor. New York, 1981, Springer-Verlag (in press). 16. Hilderbrand RL, Henderson TO, Glonek T, and Myers TC: Isolation and haraterization of a phosphoni aid rih glyoprotein preparation from Metridium dianthus. Biohemistry 12:4756, Glonek T, VanWazer JR, and Myers TC: Full anhydrization of methylenediphosphoni aid and of

14 Volume 21 Number 5 NMR of lens organopnosphates 713 phosphori aids by a arbodiimide. Inorg Chem 14:1597, Navon G, Navon R, Shulman RG, and Yamane T: Phosphate metabolites in lymphoid, Friend erythroleukemia, and HeLa ells observed by highresolution 31 P nulear magneti resonane. Pro Natl Aad Si USA 75:891, Bettelheim FA: The nature of rystallinity of a-rystallin preparation of bovine lenses. Exp Eye Res 14:251, Bettelheim FA and Wang TJY: X-ray diffration studies of maromoleular aggregates of bovine lens. Exp Eye Res 25:613, Ho C, Magnuson JA, Wilson JB, Magnuson NS, and Kurland RJ: Phosphorus nulear magneti resonanes studies of phosphoproteins and phosphorylated moleules. II. Chemial nature of phosphorus atoms in a s -asein B and phosvitin. Biohemistry 8:2074, Lee SL, Veis A, and Glonek T: Dentin phosphoprotein: an extraellular alium-binding protein. Biohemistry 13:2971, Pettegrew JW, Glonek T, Baskin F, and Rosenberg RN: Phosphorus-31 NMR of neuroblastoma lonal lines. Effet of ell onflueny state and dibutyryl yli AMP. Neurohem Res 4:795, Cruthfield MM, Dungan CH, and VanWazer JR: The measurement and interpretation of high-resolution ; "P nulear magneti resonane spetra. Top Phosphorus Chem 5:1, Salit PW: The reation and buffer ativity of normal ox lenses. Am J Ophthalmol 22:413, Bellows JG: Catarat and Anomalies of the Lens. Growth, struture, omposition, metabolism, disorders and treatment of the rystallin lens. St. Louis, 1944, The C. V. Mosby Co., p LePage GA: Glyolysis in tumor homogenates. J. Biol Chem 176:1009, Frohman CE and Kinsey VE: The rystalline lens. V. Distribution of various phosphate-ontaining ompounds and its signifiane with respet to energetis. Arh Ophthalmol 48:12, Lehninger AL: Biohemistry, The Moleular Basis of Cell Struture and Funtion. New York, 1975, Worth Publishers, In., p Labotka RJ, Glonek T, and Myers TC: Phosphorus- 31 nulear magneti resonane studies on nuleoside phosphates in nonaqueous media. J Am Chem So 98:3769, Ogawa S, Rottenberg H, Brown TR, Shulman RG, Castillo CL, and Glynn P: High-resolution 31 P nulear magneti resonane study of rat liver mitohondria. Pro Natl Aad Si USA 75:1796, Seeley PJ, Busby SJW, Gadian DG, Radda GK, and Rihards RE: A new approah to metabolite ompartmentation in musle. Biohem So Trans 4:62, Busby SJW, Gadian DG, Radda GK, Rihards RE, and Seeley PJ: Phosphorus nulear-magnetiresonane studies of ompartmentation in musle. Biohem J 170:103, van Heyningen R and Pirie A: Aid-soluble phosphates in the lens. Biohem J 68:18, Haushildt JD, Harris JE, and Nordquist LT: Changes in the phosphate frations of the lens under various onditions whih influene ation transport. Am J Ophthalmol 39:155, Pirie A, van Heyningen R, and Flanders PH: Changes in lens during the formation of X-ray atarat in rabbits. 3. Phosphates. Biohem J 61:341, Klethi J and Mandel P: Eye lens nuleotides of different speies of vertebrates. Nature 205:1114, Nordmann J and Mandel P: Metabolism of gluides in the lens. I. Anaerobi glyolysis. Ann Oul 185:929, Palm E: On the phosphate exhange between the blood and the eye. Ata Ophthalmol (Suppl) 32:1, van Heyningen R: Some glyolyti enzymes and intermediates in the rabbit lens. Exp Eye Res 4:298, Pirie A: Metabolism of a-glyerophosphate in lens. Pro Biohem So 88:48, Kinoshita JH: Mehanisms initiating atarat formation. INVEST OPHTHALMOL 13:713, Kuwabara T, Kinoshita JH, and Cogan DG: Eletron mirosopi study of galatose-indued atarat. INVEST OPHTHALMOL 8:133, Greiner JV and Chylak LT Jr: The anatomy of the experimental "hypoglyemi" atarat in the rat lens. Ophthalmi Res 8:113, 1976.