1076(1991}37-48 © 1991ElsevierSciencePublishersB.V.(BiomedicalDivision)0167-4838/91/$03.50 ADONIS 01674538910Q0633 Biachirnica et Biophysica Acla,
37
BBAPRO33799
~H-NMR. study and structure determination of 4,4- and 4,6-dimers from electrochemical reduction of NADP ÷ Enzio Ragg 1, Leonardo Scaglioni t, g o s a n n a Mondelli 1 Vincenzo Carelli :, ltalo Carelli 3, Antonio Casini z, Alessandro Finazzi-Agrb 4 Felice Liberatore : and Silvano Tortorella 2 J Diparlimento d3 Scien~ Mnlecolari Agroahmeatari, Oaiversit~ di Milang Milano (Italy). ' Dipartimento di Studi di Chimiea Tecnolagia delle Sostanze glotogtcamente Attwe, UnJ~ersllhdi Roma "La Sapiep~za'. Roma (Italy). ~ Dipanimrnt~ .~.iChinffca, LC.M., Universitd de L'Aquda, L'Aquila (Italy) and ~ Dipartimento di Med~rina Sfc*i,~itntate e Scienze Blochirniche, Uni~ersitd di Roam Tot Vergata. Roma (Italy)
(Received9 August19~)
Keywords: NMR,1ft.-:Nucleolidestructure; Electrochemicalreduction;Nicolinamideadeninenucleotide The products arising from ono-electrun electrochemical reduction of the eoenzyme nicolinamide adenine dinucleollde phosphate (NADP +) have been studied by ItPLC chromatography and 1H-NMR spectroscopy. I-1PLC and NMR analyses have shown seven dimefie species, the most abundant of which (40%) has been isolated and has resulted to be an NADP 4,4~linked dimer. The other two diaster~isomeric 4,4-directs present for the 25% and 10%, respectively, have been detected in the crude reaction mixture, but have not been isolated. The 4,4-tetrahydrobipyridine s~ructme and the stereochemistry at the ring-ring junction [or these three isomers have been determined on the basis of their NMR parameters. Preparative HPLC chromatography also led to two fractions enriched in another four dimers, present in the crude mixlure, which turned out I~ have a 4,6-tetrahydrobipyridino slructure. All the chemleal shifts and the It,H coupling ~onstants of the 4,4- and 4,6-tetrahydrobipyridine systems have been ohlained for the seven COmlmunds. For the most abundant among the 4,4-directs Ihe NMR analysis also gave the coupling constant values of the ribose-diphosphate chain.
ln~eduetlen The electrochemical reduction of the coenzyme nientinamide adenine dinudeotide (NAD ÷) has been the subject of several investigations [1,2], The uptake of two electrons by the pyridinium cation present in the coenzyme gives rise to the formation of dihydropyridines, while the uptake of one electro.n,prc~duces a free radical species which after~vards dimerizes. The Facile back oxidation of (NAD) 2 obtained by chemical [3,4], electrochemical [5] and photochemical [61 methods proves the occurrence of an in vitro redox cycle NAD+/(NAD)2 analogous to the couple NAD+/ NADH. Furthermore, the oxidation of (NAD)z to NAD ÷ by molecular oxygen in the presence either of some proteins [7,8], or of the supematant of a standard
Correspondence: R. Mondelli,Dipartimcntodi Scienz¢Mol~r.olari Agroalimentar/,ViaCeloria2, 20133, I~lano, Italy.
liver mylochnndrial preparation [2], and the ability of (NAD) z to bind some NAD-dependent dchydrogenases [3]. might be of biological interest. However. dimerie species of NAD(P) have not so far been identified in viva. Other authors have also studied the electrochemical reduction of the other coenzyme, the nicotinamide adenine dinucleotide phosphate, (NADP+), and the binding of NADP dimers to NADP-dependent dehydrogenas~s [1,9,10], but none of these works deals with the isolation and structural characterization of the reduction products. In the present paper we report on the results of an NMR investigation, performed on the crude reaction mixture obtained from one-electron dectrochemical reduetion of NADP + and on fractions obtained by preparative HPI.C, which led to the structure determination of the 4A-linked and the 4,6-1inked dimers. The NMR parameters have also been used for the molecular eonformationa] study reported in the accompanying paper.
38 Materials and Methods ~-NADP +, free acid, 96% pure, Mr 797.4 (trihydrate) was purchased from Sigma (U.S.A.). Its molar absorption coefficient was found to be 3,2~0 - 18000 4400 M 1. cm (water). This was the starting product used for the preparation of the directs, whereas NADP +, monosodium salt, letrahydrate, Mr 837.5, from U.S.B. (U.S.A.) was used for NMR studies and for analytical purpose. Glucose f-phosphate monosodium salt and gluense-6-phosphate dehydrogenase, lyophilized powder (140 U / m g enzyme protein) were from Boehringer (F.R.G.). Gt0-Sephadex was obtained from Pharmacia (Sweden); the resin was swollen in distilled water and stored in 0.1% NaN 3 at 4°C. Before the elution, the packed resin was exhaustively washed with 50 mM aqueous NH 3. Other chemicals, unless otherwise specified, were of analytical grade. Polarograms were recorded with a three-electrode Princeton Applied Researc,h 170 Multipurpose Instrument in a cell kept al 25,0-4-0.1 °C. Apparatus and procedures for v,larographic analysis were as previously de,scribed [14]. The reduction polarogram of NADP + (c=0.5 raM), vw.orded in aqueous 0.1 M NH4CI-NH 3 buffer (pH 10.6) in presence of Et4NCI, final ionic strength 0.5 M, e~Aibits two well-defined In-reduction waves with E~/2 = -1.00 V and -1.60 V vs. S.C.E., respectively. These values are in goc,d agreement with those obtained by Schmackel ct al. [1] in a K~CO3/KHCO3, plus Et4NCI, buffer. Macroscale dec_ trolyses were carried out by using an Am¢! Mud. 555/A Potentiostat equipped with an Amel MUd. 721 Analogic Integrator and aa Omniscribe Recorder. A three-com. partment water-jacketed cell kept at 25.0 + 0A°C, containing a magnetically stirred mercury pool, apparent area 20 cm2, as working electrode, was used. An afar-salt bridge was inserted in the anodic compartment on a sinterized glass frit of medious porosity. The anodic compartment contained a platinum gauze cylinder immersed in a KCi solution. A calomel saturated KCI electrode (Metrohm Herisau, Switzerland) was used as reference electrode. The buffer solution (0.1 M NH4Cb NH 3 (pH 9.3)) was pre-electrolysed at the same potential of the dectrolysis until a constant current value was reached. Nitrogen or argon (99.9% purity) was bubbled continuously through the cell during both pre-electro|ysis and electrolysis. HPLC analyses were performed by a Perkin Elmer Series 3 Liquid Chromatogral~h , equipped with a LC55B spectrophotometric detector, an LC 55S digital scanner, and a 023 recorder. Integrations of peak area were made by means of a Hewlett-Packard HP 3390A integrating recorder. A RP-18 type 250-4 LiChrocart Merck column, particle size 7 lsm, was used for the analyses, monitoring the eluate at 260 or 340 rim. Composition of the mobile phase (v/v): solvent A, redislilled water
(30%), LiChrosolv ethanol (12%) and aqueous 0.1 M NH4HCO ~ (58%); and solvent B, 0.07 M NH4HCO ~ in redistilled water~ The f~lowing solvent program was used: (1) an isocratic step with 3% of solvent A (5 min); (2) a linearly increasing gradient step from 3 to 6~g of solvent A (20 vain); (3) an equilibration step with 3% of solvent A (8 min). Flow, 1.0 ml/min. After each set of analysis the column was regenerated with redistilled water (200 ml), then with MeOH-H20 (15%, v/v) at a flow of 0.1 ml/min overnight. Before each set of analysis the column was conditioned with redistilled watersolvent B (1:1, v/v, 50 ml, flow: 1.0 ral/min.), then replacing water with solvent A for further 30 rain. Finally, the initial conditions of the solvent program can be used for equilibration time. All the solvents were pre-treated with the RP-18 phase (10 v.m of particle size), filtered on Millipore HAW 0.45 .am fillers and degassed. A precolumn (25 x 4 ram, i.d.) filled with RP-18 phase was applied before the injection valve. Ultraviolet spectra were measured according to the stopped-flow technique with digital scanner, after calibration performed in the same conditions of the analytical program. The presence of NADP + was detected either by comparison with an authentic sample, or by enzymatic test in presence of gtucose-6-phosphate dehydrogeaase. Preparative scale HPLC separations were carried out by using the pumps system of a Waters 500A preparative chromatograph and a stainless column of 400 rnm length x 40 ram, id., fdted with LiChroprep RP-18 (15-25 /~m of particle size, 30 g) suspended in MeOH (150 nil). Composition of the mobile phase (v/v): distilled water (30%), aqueous 0.1 M NH~HCO 3 (67.6~,) and absolute ethanol (2.4%). The column was conditioned with the mobile phase for 30 min before the injection, flow about 6 mi/rnin. The effluent was monitored by the R.I. detector system of the Waters 500A. Ultraviolet spectra and enzymatic analyses were performed on a Perkin Elmer 555 UV-VIS Spectrophommcter in aqueous buffer at neutral or slightly alkaline pH. NMR spectra were recorded with Broker CXP-300 and AM-500 spectrometers. Chemical shifts are in ppm (~) values from external 3-(trimethylsilyl)propane-1sulphoaic acid sodium salt hydrate (DSS), estimated accuracy_+ 0.005 ppm. Coupling constants are in Hz, accuracy +0A. Hz, unless specified in the Tables. Data for compounds IIl-Vlli were obtained from HPLC fraction 1, those for VIII and IX from fraction 2, and those for X! from fraction 3. The spectra were measured (60 mg.m1-1) in D20 at pH 9.2 :i:0.1; the pH values arc uncorrected meter readings (glass electrode) from DzO solution. The solutions were adjusted to the desired pH with NH 3, and kept under nitrogen. The variation in chemical shift with concentration and pH, in the range 80-20 rag-ml" 1 pH 9-10, are within 0.1 ppm. The measurements of chemical shift values and of the rein-
39 rive areas were performed immediately after the preparation of the solution samples, The dimeric species studied are stable for r~are than 24 h in the experimental conditions used. The two-dimensional COSY spectra were acquired by using ~hk45 ° observing pulse, 256 FIDs zero filled to 1024 points, weighted ~ t h a sine-bell function, and transformed in absolute value. The suppression of the HDO signal was achieved with a long low-power pulse (1 s), before acquisition. The spectra were analysed by usir~g the PANIC program, a LAOCOON-type program included in the Aspect-2000 computer library.
Macroscale electrolysis of NADP + and isola/ion of diilrlers
In a typical run, NADP +, free acid (1 g) was di.ssolved in 50 ml of a previously electrolysed 0.1 M NH4CI/NH 3 buffer (pH 9.3) and then the solution was eleetrolysed at -1.20 V vs. SCE. The electrolysis was complete in about 1.5 h, as shown by the constant value of the current, equal to tha¢ measured at the end of the preelectrolysis. The faradaic value at the end of the electrolysis was in good agreement with the theoretical uptake of one electron per mole of NADP + The solution was lyopkilised and the residue (1.2 g) was chromatographed on a G,o-Sephadex column (800 mm length × 32 mm i.d.) using distilled water containing 0.3~ (v/v) of 33% ammonia (about 50 raM) as ducat. A flow rate of 0.8 ml/min was maintained. The fractions eluted in the volume range of 210-280 ml showed the same absorbance ratio (.~I259/A338) a s the electrolysed solution and were free of chloride and NADP +. These fractions were lyophilised yielding a soft yellow crude product (0.8 g), which must be stored under vacuum at - 3 0 ° C , otherwise a slow autooxidatioo to still enzymatically active NADP + takes place. Calculated for C 4 2 H ~ z N l a O a 4 P t ~ ( N H 4 ) 6 . 12H20, M~ = 1807,26 (crude reaction mixture): % C=27.91; H = 5,58; N ~ 15.50. Found: ~ C ~ 28.15; H ffi 4.82; N = 14.75. ¢zs9=34300 M - t . c m - I ; ~3~s=6250 M 1.
is composed almost exclusively of isomer X], as shown by HPLC (92%) and NMR (84%), An attempt to crystallize it was not successful. Fractions 1, 2 and 3, rapidly frozen and lyophilised, yielded residues of 110 rag, 30 mg and 40 rag, respectively, which showed km~ (259 and 33g am) and chemical reactivity (e.g., backo~idation with K~Fe(CN)6 ) to NADP +) characteristic of the tetrahydrobipyridine system [2-4]. Results and Discussion
Analysis of the reaction mixture The electrolysis of NADP + at - 1.20 V vs. SCE gave a crude reaction mixture, which, after chromatography on a Sephadex column, displays the HPLC elation profile reported in Fig, 1. This shows eleven peaks (I-XI), corresponding to three main products Ill (10%), V (25%), X! (37~) ond to other components in a low IU
80
E o
~6o
41J U
~
t~ g. Q
4o
I
c m - I.,
HPLC analysis and preparative scale separation of dimers The dution profile of a sample of the above lyophilised crude product, monitored at 260 nm. is shown in Fig. 1, while the relative areas of the single peaks are reported in Table I. When the monitoring is made at 340 am, the elation profile shows only seven peaks, with three major components, t R 2.8 mix (lII), 7.9 mix (V) and 20.4 rain (XI). A sample of lyophilised crude product (300 rag) dissolved in the mobile phase (2 nil), was subjected to a preparative scale HPLC separation, as described above. After a dead volume of about 90 ml, three fractions 1, 2 and 3 of 70 ml, 30 ml and 150 ml, respectively, were collected and analysed; their compositions are shown in Table I. In particular, fraction 3
20
~/q v,, IVl
s
Virl
~'2
~#X
I'G
~o 2'4 time (rain)
Fig. 1. HPLC ¢ltttioll profile of the crude rcaetiQn mi~tture obtained from electrochemical redaction of NADP + and monkored at 2r~ nra. Releation times (ta) are given neglecting Ihe dead ~:olume of the ~:olutnn and slart Rora the instant of iajeclion (see arrow)~ The peak with t a of 4.6 rain corresponds to NADP" (see Table I and the ~,~ectionMaterials and Methods).
40 TABLE I Cemposizion of the cnde m ~ m r e and of the itPLC froctwr~' from electrochemical r~'ductron ef N,'tDP *
The crude mixture {M), obtained h-c~mSephadex column, lyophilised and dissolved in the mobile phase (see Materials and Methods) was injected in a v~lume of 4 pl (3 m8. ml t). The separated fxaclions 1.2 and 3 were injured in volumes ranging from ! 0 to 30 pl. The ¢mt,mt was m0nitoted at 260 nm. Peak numhcr '
Compound
!
=
It III IV V V[ VII VIII IX X Xl
= 4,4-direct NADP * 4,4-dimer 4,6-dlmer 4,6-dimer 4.6Mimer
4,6-direct nicotinamide 4~4-dimer
rR ~ M-"~
HPLC p~ak area (%) Mc Id
2e
t.8
4.5
7.0
6.0
2.3 2.8 4.6 7.9 9.0 "[1 5 ]5.6 16 g 18.8
1.0 10.0 6.0 25.0 3.0 6.0
2_0 23.0 5.5 38.0 9.0 13.0 0.5 2.0
1.0 3.0 19.0 6.0 7.0 4.0 320 18.0
~.4
4.0 3.0 0,5 37.0
NMR signal area (%) Me 1d
2d
3d
2.0
e
•
e
~ 10.5 10.5 8.7 4.7
t
L0 L0 0.5
. 11.6 2,8 25,4 5.0
37.2
r
22.3 m 6s
r
3~
1• 0.5 2,0 92.0
4.0
~
.
:~.5
. 29.1 3.5 44.2 9.3 11.6
3 ~
I
3.9 1.6 40.9
.
2s ~ ~
6.3 r
r r
9.5 84.2
a See Fig. I.
b The retention times (min) ~re given ne~ecti,g the dead volume of the column_ " Ctttde reaction ndxmm. a Fractions of pruF,arativ¢ HPLC. = Not identified. f Absem or le~ than 0.5~, e Partlaily overlapped b5 od~er z,i3.'~s.
q u a n t i t y . T h e ultraviolet s p e c t r u m o f t h e c r u d e r e a c t i o n m i x t u r e exhibits a b s o r p t i o n m a x i m a at 259 a n d 338 n m , b u t n o t b e y o n d 400 n m , t h u s e x c l u d i n g s t r u c t u r e s inv o l v i n g d i m e r i z a t i o n at p o s i t i o n 2. T h r e e m a i n fractions, 1, 2 a n d 3 w e r e o b t a i n e d b y p r e p a r a t i v e liquidliquid c h r o m a t o g r a p h y ; t h e H P L C a n a l y s i s s h o w s t h a t f r a c t i o n 1 is e n r i c h e d o f c o m p o n e n t s IIl, V, V1 a n d V I I ,
fraction 2 in VIII and IX, ,~hile fraction 3 ¢on'aius almost exclusively compound XI (Table I). The ~H-NMR spectrum of the crude mixture shows the main products III (11%), V (25%), XI (40%) and severat other components in low abundance, in agreement with the chromatoBraphyc results (Table I). Amon 8 the latter one is always NADP +, often appearing to-
H
NH 2 (~0 H
H (111, V, XI)
~H 2 ,~ON-f
2
-e
~CONH
2
x2
+e
NH2
S.C.E.
¢
H
~0
R--kf A ~
H
H B/)2-CONH 2
(Vl, VII, VIII, IX)
Scheme I. Formation of 4,4- and 4,6-1i~ed NADP dimers from one.electron electrocheafi ~zl reduction of NADP +. R = adenosine (2.phmphate)diphosphale-ribo~e.
41
A2
A8 N2
N2
N 6 A~"
N4
N6
rd4h
V+XI
NS
V
~
a aba
,
-
-
~
~
III
b
I ....
t .... g.o
' ....
I 8.0
'
I 7.0
I .... §.O
'
'I 5.0
"
' '
~ 4.0
'
I
~.o
'
8 ppm
'I
2.0
Fig. 2. I H - N M R s p ~ r m n (300 MHz, D z O / N H 3 (pH 9.3))of the crude roction mlxtme from electrochemicalreduction of N A D P +. The sign~Is -marked wilh (a) and {b) a ~ N A D P + and nicotinamide,r~p~tively.
gether with nicotin,mfide. The signals of both these compounds are clearly visible at low field in the proton spectrum reported in Fig. 2. The assignment was made by means of el',¢mical shifts and coupling constant values, compared with authentic samples. The coupling constants are the following: J(N2,N4)= 1.7, 2~5 Hz; J(N2,NS)= 0.0, 0.g I-Iz; J(N4,NS)= 8.1, 8.3 Hz; J(N4,N6) ~ 1.3, 1.7 Hz; and J(N5,N6)- 6.3, 5.0 Hz, for NADP + and nicotinamide, respectively. NADP + is not starting material unreacted, but comes from autoxidation of dimers, because it increases with time and appears in all the three fractions obtained by preparative HPI.C. Nicotinamide oriSJnates from NADP ÷, as proved by the 1H specmun of NADP measured in the same conditions. A few protons of the main components give spectral patterns which can be easily identified, namely N2 * and N4 of dihydropyridine tings at about 7 ~ and 3 f$, respectively, A2 of adenines at about
* Numbered As and Ns indicale the hydrogen atoms of the adenine a t u c l e o t i d e and those of nicofinamide nucleotid¢ unit, resp~lively. Unspecified numbers indicate atom,~ of bolh unit~.
8 8. The other protons partially overlap with ribose nuclei, N6 with AI' and N5 with NI', or give broad peaks of low intensity, such as A8 of adenines, due to exchange with the deuterium of the solvent. Nevertheless, almost all protons of the bases and part of those of ribose moieties were identified, by means of syst ,. ~tie spin decoupling experiments. The analysis of the proton spectrum of HPLC fraction 1 confirmed the identification of Ill and V, while the four compounds, VI-IX, detected in small quantities in the crude mixture, were identified by examining fraction 1 (VI and VII), and fraction 2 (VIII and IX). For the most abundant one, XI, the NMR study was performed on fraction 3, where it amounted to 80-905. tH.NMR spectra, structure and absolute configuration o~ 4,4-dimers IlL V and X I Fraction 3 was analysed first, because the abundance of isomer XI allowed to obtain the complete set of chemical ~hifts (Table If) and eoupfing constants (Table III) for all protons. The conneetivities of nuclei were found by using two-dimensional COSY (Fig. 3) and
42 TABLE In
TABLE l[ #lt fhemwal shi~r values for 4,4-d~mers II1. V and XI
H,H couphng consra,t values far 4.4-dtmers IlL V and XI
Measured in ppm (61 relative In external DSS; DzO solutions adJ:l-ir,t wit!i NH) to pH 9.3. E~limated accuracy +0.005 ppm unless specified, Similar values in parentheses may be interchanged,
J values are measmed in Hz and me given without sign. Eslimaled accuraey+0.'l Hz. For the ribose protons of Ill ~nd V, aecnraey4-O.2 Hz.
III ~
V
XI
A b
Bb
(8.129) (&359) 6.951 3.191 4.669 6_108 6.131 (4 rg00) ¢
{g.113} {g.383} 6.93(I 2.909
(4.561) ¢ u
(4~54) • ,J
(4.5~la) c d
d a
d d
NI'
(4.720) ¢
(4.7221 ~
N2'
(4.071) c
N3"
(4.134)"
(3.966)': (4.148)~ 3.98-4.01 d
d a (4.659) ~ (4.030) ' (4.119) ~
A2 A8 N2 N4 N5
N6 At" A2' A3" A4' A5' A5"
N4"-SL5*"
8.101 $.M.~ 6.964 2.27? 4.409 6.122 6.059 (4.917)
c
4.530
6.026 6102 (4.8391 ~
J
Ill
V
XI
A a
8.118 8.380 %081 2.916 4.400 6.008 6.100 4,$76
N4,N4 t~ N4,N5 N5.N6 N2.N4 N2,N5
2.2 5.1 E..2
0.5" 0.5 ¢
N2,N6
1.4
N4.N6
0-5 ~
NI',N2' N2".N3'
7.1 ~ 5.5 •
4~335 4.245 4.160 * 4.729 4.080
N3',N4' At',A2." A4'.A5'
4.[41
A4".A5"
4.529
• The assignment of [11 with respect to V followed hy comparing spectra with different proportions el the two isomers. Values for the two non.equivalent halve.g. ¢ Acc,tra~y:t0.0t ppm~ 'J Not analysed, Partially nvcdal0pcdby N2' signal; value obtained by the calculated spectrum.
B a
3.2 5,5 8/) 0.9 0.5
5.6 %9 0.8 0,3 ,l
2.3 4.8 7_9 0.5 = 0.5 =
2.0
1,5
1,4
0.5 (7,2)
0.5
0.5 =
(5.1)
('L4} 5.4
7.1 5.4
2.0 • 15.|)
(1.9) (4.9)
(2.3} (5.0)
2.5 4.9
AZ',A3'
4.8
5.0
4.g
4.8
5.1
t
4.8 4.8 i
4.8
A3',A4'
I
3.0
# f
t f
i I
4.3
AS".AS"
11.5
a Values for two non-equivelent halves; A and B values belong to the signals o[ eolnrrmsA and B, respectively, of Table 11. Coupling constant belween protons at the ring-nngjunstion. Not measured, But delO~ledlby decoapling experlmenls. The reported value g~v¢ the best fit with the calculated spectrnm. Line width 0.5 tlz. Tentative attribution. r Not analysed.
spin dccoupling experiments, some of which are reported (Fig. 4), whereas the coupling constants were measured from one*dimensional spectra. Fig. 3a shows the cross.peaks corresponding to the coupled protons in the dihydronicotinamide and ribos¢ fragments of XI. The sequence NdvNS-N6 was readily identified, and N2 was assigned by the long-range coupling with N6. This inleraction, although small, is clearly visible in the right inset of Fig. 3. The hi#l-field shift (2.916 8) and the second-order pattern of N4 indicates a 4,4-lctrahydrobipyridin¢ structure. This signal belongs to two protons, which arc chemical shift equivalent, but magnetically non-equivalenL i.e., they are the AA" part of an A A ' X X ' Y Y ' Z Z ' spin-s~stem [ l l ] , Figs. 4a and b show the expcrimental spectrum a n d the calculated patterns for N4 and N5 protons. Owing to the second order effect, we obtained the AA' coupling conszant, i.e., J(N4,N4), which proves that dimerization occurs with junction at C-4. Also the other protons at corresponding positions in the two halves are chemical shift equivalent.
The anomeric protons, A I ' a n d NI" of the ribose moieties were assigned in the following way: starting from A 2 ' pattern at 4.876 6 (Fig, 4a), which shows a three-bond interaction of 6,8 Ii~ with phosphorus, we decoupled A I ' at 6.100 8 (Fig. 4c). In the inverse experiment, not reported here, A I ' donblel collapsed into a singlet. N I ' protons, expected as a dou[/xt, were identified at 4.729 8, after pre-saturation of the water peak, and decoupling of N 2 ' protons at 4.080 & The four-line signal of N 2 ' in turn becomes a doublet of 5.5 Hz by irradialion of N I ' (Fig. 4d and scheme ll). The connectivitics with the other protons, which complete the sequence of nuclei in the ribose units, were detected through the two-dimensional spectrum, while the analysis of H-4' and H-5" patterns (Fig. 4a and e) aimed at obtaining the coupling constant values, was performed on the one-dimensional spectrum at 300 and 500 MHz frequency. In the case of the A ribose rings the protons are well resolved, because they are sufficiently spread out, whereas N4', N5' and N S " are so
Fig. 3. la) Two-dimcnslonal Ill-COSY and (b) one-dimensional ?':MR spectra (300 MI-Iz,D20/NH~ (pH 9.3}1of dimer X! from fraction 3, t~gion from 2.8 to 6A 8. The sing,lets of A2 and A8 at 8.118 and 8.3gl}~ are not rel~orted.Left inset: expansion from 4.0 to 4.5 8; right inset: cross-pc~s from N2 and Nb protons,
43
(b)
510
31o
...~--~.d.~:
6'
O
P
./
.
4~4
4-2
,
q
o
4.0
0
j-
(a) ,
t
i
g
""
r
UP :r~o
~.o
ppm
44
Il~II m2
ml"
At"
~ ,[d}
let"
m2"
t~¢
m~J
?~ml
1110
IIom
•
ITI
• 7ae
@~Ul
4215
I
I~
4141
I~I
l'gll
I @p
Fig. 4. III-NMR sp~lrum of 4,4~dimer XI from fraction 3 (~0~ MHz. D~O/NH~ (pH 9.3}). (a) expanded pattet-~ of the most important sig~]s: (b) caicuimmt spectrum of N4 and N5 protons (AA' and XX' ;:~at'lof ~ ' XX"~'~r' ~ ' spin system}; (c) A2' d~couplcd from A t ' upon irradiation at 6.100 ~l; (d) N2" d¢coupled from N I ' upon inadiation at 4.729 ~: (e) calculated spectrdm of A4', A5' and AS", with J(P, A 4 ' ) = 2.0, .I(P,AS") = 5.0 and J(P.AS') = 5,4 Hz.
close t3gmher even at 5Q0 MHz, that the coupling constants involving these nuclei could not be obtained in D20/'NH t solution. The degeneracy was removed by using sedium tetraborate buffer, and the results are reported in the accompanying paper, together with the values of the hydrogen-phosphorus coupling constants. The NMR analysis of compounds Ill and V was performed on the reaction mixture and on HPLC fraction 1, wHch is enriched of these components. The adenine A2 and A8 singlets of the three main isomers
H
CONH2
are close together, but detectable even in the spectrum
~f the crude mixture (Fig. 2). Two protons in the spectrum of fraction 1 are clearly visible and correspond to N2 and N4 of |II and V (Fig. 5); the other protons of the nicotinamide rings were assigned by decoupling experiments, whereas a few ribos¢ resonances were detected but not analyzed. The doubling of N2 and N4 signals of V indicates that the two halves are diastereotopic, whereas for Ill, the correspo.ding nuclei of the two halves are chemical shift equivalent.
N
N H
H
s~
H
I H
t
H
H
s
ONH4 0
'/
o Io "~/
"12 H 2 0
II,o
-,,p ~.s'
;L i" 4' o" ,
a'
0,.~
i+ O
tl OH
O--P--ONH
4
I Scheme11.Molecular~.lm¢(urcof 4,4-[inkcclNADP dimcrs,(NADP)2.
45
N4B(V)
N4A(V)
N 201|)
N4(lll)
N 2 (V)
(0 i
J ~ ~ ( d )
~,
(b)
I••
I
I
6.9~4
8,95|
6.930
3,191
(a)
(a) ppm
2,909
I 2.2?7
Fig. 5~ 1H-NMR spectrum of 4,4-directs III and V from fracdon t (309 MHz; DzO/NH s (pH 9,3)), expanded patterns of (a) N4 and (b) N2 protons; (c} calculaled spectrum of N4 for Ill; (d) inadialioo of N4A proton of V at 3,191 8; (e) irradiation o[ N4B proton of V at 2.909 & (f) irradiation of N2 protons at 6.93-6.95 8; (g,h,i) kradialion of N4 protons at 2,909, 3.191 and 2.2"/'/~, resp~tivdy.
The double resonance experiments reported in Figs. 5d and e show that the N4 protons of V at 3.191 and 2.909 8 are connected by a vieinal coupling; those in Figs. 5f-i prove the long-range interactions between N2, N4 and N6. Note the absence of coupliug between N4 and N6 (Fig. fit'), as already observed for Xl. The experiments in Fig. 5g-i were used to assign N2 signals. The NMR data, reported in Table II and Ill, confirm that the three compounds are dimers and prove that the dim¢fization occurs at position 4. The symmetry and the high-field shift of N4 protons are indication of this junction, but conclusive e,ddenc~ for the 4,4-tetrahydrobipyridine structure came from the analysis of all proton signals o1' ihe pyridine fragmenL In particular J(Nf,N6) = 8.0 Hz appears to be diagnostic for a 1,4-dihydronicotinamide ring. We carl the~fore conclude that the two halves of isomers Ill and XI are identical and that the configuration of the chiral (2-4 centres is 4R, 4R and 4S, 4S, or vice-versa. As a consequence the stereochemistry of isomer V must be 4R, 4S. Since the configuration of the diphosphate ribose moieties is re-
tained, neither enantiomers nor mcso isomers, but only three diastereoisomers can be obtained.
~H spectra and structure determfnation of 4,6.dimees VI1X
VIII and IX isomers were identified in the HPLC fraction 2 and also in the crude reaction mixture through detection or the N4 and N6 proton signals for the 1,6-dihydropyridine ring (A) and 1,4,dihydropyridine ring (B), respectively. The chernical shift values (Table IV) of these protons are diagnostic for the ring.-ring junction, as well as tbe vicinal couplings betwce,t protons on the w bond (Table V). The N4 hydtogens of dngs A lie at about 3 ~, whereas the N6 of rings B are deshielded to about 4 8 by the viclnal nitrogen atom. The N4A resonances of both VIII and IX isomers (Fig. 6a) appear at 3.330 and 3.280 ~ as double doublets due to ~/(4A,SA) and ~/(4A,6B), with each line further split or broadened by allylic couplings with N2 and N6 of the same ring (see the irradiation of N2 in Fig. 6d).
46
N6B
NSB
(,[) V I I I .1- I X 4111
IX IX
VIII
'~~1 (~-) c) .098
4.97!
J
3.890
i 3.887
ppm
3.330
3.280
Hg. 6.1H-NMRspcctrnmof 4,6-dimcrsViii aud IX fromfraction2 (300 Ml-tz,I~O/NH 3 ~pH 9.2)),expandedpatterns of (a) N4A,(b) N6B and (¢) NSB prot~az~;(d) N4A decoupledfromN2A upon irradiationat 7.05 3; (c) N6B dccouplcdfromN4A upon irradiationat 3,3 6; (~ NSB al~d(g) I,,14Ade,coupled fromN69 upon irradiationat 3.88&
The identification of N6B signals close to the strong absorption or c - 5 ' protons was done by irradiation of N4A and viceversa (see Fig. 6b, c and g). The analysis of the signals at 3.87-3.89 8 (Fig. 6b) was not easy, but the coupling constants reported on Table V are accurate, because the N5B protons are visible at 5.028 8 and 4.971 8 (Fig. 6c); they decoupled upon irradiation at 3.87 ~ (Fig. 6f). These data prove the 4,6 ring-ring junction and consequently the 4,6-tetrahydrobipyridine structure for VII! and IX. It dearly appears that the vicinai couplings between the protons on the ~r bond arc different for the 1,4- and ],6-dihydropyridine systems: J(SA,6A) measures 8.0 Hz for 1A-dihydro structures, whereas for 1,6-dibydro structures J(4B, SB) is 10 Hz, in line with th© modal compounds obtained by reduction of 1-benzyl-3-carbamoylpiddinium chloride [12]. We might be tempted to interpret these values with a stronger delocalization effect in 1,4- than in 1,6-dihydropyddine system; but the lower value of the olefinic
interaction for ring A can also be explained with the inductive substitugnt effect of the nitrogen atom on the ~icinal proton coupling EllE. The four-bond interne tions, J(2,6) for ring A and J(2A) for ring B, cannot be used to irderprct delocalization effects, because their values (1.5 Hz) are well explained with a u-mechanism through a W path of coupling Ell]. The assignment of VIII vs. IX was performed by comparing the intensities of N2 signals (Fig. 7b), taking into account ~ a t in fraction 2 isomer Vill is more abundant than IX on the basis of HPLC analysis. In the frequency range of N2 protons (7-7.5 8), there are a total number of nine peaks. Excluding those at 7.088 (dimer XI), at 6.951 and 6.930 8 (dimer V), two peaks of the remaining six decoupled upon irradiation of N4A at 3.3 8 (Fig. 7c), and two others upon irradiation of N6B at 3.88 6 (Fig. 7d). This allowed to assign the protons of ring A vs. those of ring B and to identify the signals of dimers VIII and IX.
47
~(
~ (f)
{e)
II4"V (a (')
)
?.165
7.086
.~VIII)N
7.014
.2.(v,,.)
N2B (IX)
(d)
t ~1 " l ' " "rl 7,111 7.101 7.088 7,023 6,D51 6.930 ppm 7.305 7.234 7.171 Fig, 7, 1H,NMR spectra(N2 protons)of (a) HPLC fraction ] and (b) fra¢lion2; (c) N2A signalsof VIII and IX decaapl~df~oraN4A upon irradiationat 3.3 ~; (d) N2B signalsof VIII and 1XdocoupL',dfromN6B upon irradiationat 3.88 ~ (e) r42Asignalsor vI and VIi deconpledfrom N4A uponirradiational 3.0 & (f) N2Bsignalof VII decoupledfromN6B upon irradiationat 3,7298.
The low intensity N2 signals at 7.234 and 7 . t l l were assigned to the two remaining diastereoisomers VI and VII, which are slightly more abundant in fraction 1 (Table I). The NMR spectrum of fraction 1 was thus used for the structure determination of these dimers. Fig, 7a shows the N2 peaks of VI and VII, well separated from those of the major components III and V; decoupling experiments, some of which are reported in the insets (e) and 03, confirm the 4,6-tetrahydrobipyridine structure also for isomers VI and VII. The relative proportions of the components given in Table I were calculated through the integral of N2 signals. For all the 4,6-directs, ~he N6 protons of ring A and the N4 ones of tin 8 B lie in the 6 ppm region and are both connected to N2 through long-range couplings. However the N6A and N4B patterns are difficult to analyse, due to overlap with the anomeric AI' protons. As a consequence, in the case of the minor isomer V[
and Vll these signals could not be detected; whereas for VIll and IX the chemical shift and coupling constant values (Tables IV and V) were obtained through decOupiing experiments~ and from the analysis of the related NdA. NSB and N2 spectral patterns.
Conclusions The one-electron electrochemical reduction of NADP ~ !~tds to the formation of dimers with the ring-ring junction preferentially at the C-4 positions, the reaction thus appearing to be regiospecific. The relative amounts of the three 4,4-diastereoisomers indicate that the dimerization process is also stereosetec~ve, Actually the two faces of the nicotinamide ring are diaslereotopic and the approach, for instance, of the Re
48 TABLE IV
JH e,~emical shi/r vMues for 4,6-dimerr V1- IX Measured in ppm (8) relative In external DsS: Dan solutions adjusted with NH 3 to pH 9.2. Eslireated accuracy4-0.005 pprn unless specified. Similar values in parentheses may be inlerchanged. VI
VII
VIIi
IX
RingA a N2 N4 N5 N6
7_014 2.966 b b
7_086 3.003 b b
7_07.3 3.280
7_101 3.330
6.095
Ring B" N2 N4 N5 N6
7_229 6.177 b 3.89 ~
7,165 b 5.018 3.729
7+111 (6.136) d 5.028 {3.8~) a
A2
8.119 f 8+139 r
8.0~0 t
8.091 t
18-113)
(8.122)
A8
(8.43) s
{8.,IB) ~
8.396 8,420
8.390 8+410
4.656 + 5.935 7.305 (6.103) a 4.971 (3.867) d (8.103)
A gad B corr~poad to the 1.~diydro- and 1,6-dihydtonicotinamid¢ rings, respectively. b Not delected. Only one signal was detected, but could not be assigned to VII[ or XL a Accuracy+O.Ol ppm. c P~rlialty ovedupged by the N5"-5" signals. t Tentative identification. s Broad signal. TABLE V
tt, H cogpling coezstaol values/or 4,6~dJmors V I - I X J values are measured in Hz and are given without sign; estima:ed accuracy:L0.1 Hz, unless sla~ified. A and B correspond ~o the hydrogen alums of *he lA-dihydro and 1,6-dihydrenicotinamide rings. respecliveiy. J'
VI
vII
VIII
IX
2A,4A 2A.SA 2A,6A 4A,SA 4A,6A
0.8 A 1.5 5.5 0.5 a 1.8 0.9 a 9.9 a ~ 3.8
0.8 a 1.5 5.7 0,5 a 2.0 0.7 1.0 9.9
1.0 , 1_5 5.5 0.5
1.0 a 1.5 5.0 0,5
8.0 b
8.0
1.3 0.7 0.7 9.9
1.5 0.6 0.7 9.9
0.3
0.5
5.6 2.3
5.6 2.8
0.5 S.4
5A.6A 2B.4B
2B,SB 2B,rB 4B,SB 4B.rB 5B,rB 4A.6B a Not delectcd.
Accuracy ±0.2 Hz.
3.8
face t o w a r d R e , o r S i t o w a r d S i , l e a d s for the 4,4-dim e r s to t h e 4 R , 4 R o r 4 S , 4 S c o n f i g u r a t i o n , respectively, T h e s t e r e o c h e m i e a l choices m a d e b y t h e d e h y d l ~ g e n e s e s d e p e n d e n t on n i c o t i n a m i d e c o f a e t o r s is well k n o w n , b u t also in this n o n . e n z y m a t i c r e a c l i o n a dia s t e r c o f a c e d i f f e r e n t i a t i o u occurs, a n d leads p r e f e r entially to R e - R e a n d S i - S i d i a s t e r e o i s o m e r s . T h e low a b u n d a n c e o f t h e 4 , 6 - d i m e r s i n d i c a t e s that, n o t w i t h s t a n d i n g t h e steric h i n d r a n c e a t t h e r i n g - r i n g j u n c l i o n site, t h e re.action m a y h o w e v e r occur, w h e r e a s t h e a b s e n c e o f 2,4- a n d 6 , 6 - d i r e c t s s h o w s t h a t these j u n c t i o n s are n o t a l l o w e d . F r o m t h e a p p r o x i m a t e l y similar q u a n t i t i e s o f these i s o m e r s , f o u n d in t h e c r u d e r e a c t i o n m i x t u r e , it w o u l d a p p e a r t h a t t h e 4 , 6 - d i m e r i z a tion p r o c e s s is less stereoselective. T h e i d e n t i f i c a t i o n o f t h e tetrahydrobipyridine species
was carried out by using the H,H coupling constant values of the dihydronicotinamide rings. Some of these couplings, namely the vicinal ones between the protons on the ~r bond appeared to be diagnostic for the ringring junction. The oilier coupfing constants can be used for t h e c o n f o r m a t i o n a l a n a l y s i s o f these systems. References
1 Sclunacke.], C.O., Santhanam, K,SN, and I~lving, PJ. (1975) J. Am. Chem. Soc. 97, 5083-5092, and references quoted therein. 2 Carelli. V.. Liberatore, F., Casini, A., Monddli, IL, Amone, A., Carelli, I.. Rotilio, G. and Mavelli, L (1980) Bior8. Chem. 9, 342-351, and references quoted therein+ 3 FinazTJ-Agr~, A-. Avigliano, L., Carelll, V.. Libe~awre, F. and Casini, A. (1981) Bioc.him. Bioph'is, Acta 661,120-123. 4 Chan, S,S., Nordluad, T.M., Frauenfelder, H., Hen'|son, J.E. and Gunsalus, LC. (1975) J. Biol. Chem. 250, 716-719. 5 CarellL I., Ro~ati, R. and Casini. A. (1981) F.Icctrechim. Acta 26, 1695-1697. 6 Avigliano, L , Carelli, V=I Cas~, An, Fina22i-Ag~, A, and Lib¢ratore, F. (19[~3) Biochire. 5iuphys. Aria 723, 372-3"75, A,Jigliano. L . Carelli, V., Casini. A.. Finazzi-Agrr, A., Liberator¢. F. and Rossi, A. (1986) Bicchem. J. 237, 919-922. 8 Avigliano, L , Car¢lii. V.. Casini, A., Finazz.i-Agrr, A. and Liberatore, F. (1985) Biochem. J, 226, 391-395. 9 Cunninghare. A.J. and Underwood, A.L. (1967) Biochemistry 6, 266-271. tO Kovar, J. and Klukaaova. H, (1984) Biochim. Bioph~s. Aeta 788, 98-109. 11 GOnter, H. (198U) NMR 8pectrosan~y, John Wiley and Sons, Chichester. 12 Carelli, t., Cardinal|, M.E., Casini, A. and Amone, A. (1976) J. Ors. Chem. 4.], 3967-3969, t3 Burg|, H,-B. and Dunitz. J.D. (1987) J. Am. Chem. So¢. 109, 2924-2926. t4 Carelii, L, Cardinal|, M E and Michel¢lli Moreoci. F. (1980) .I. Electtoanel. Chem. 107, 391-404.
37
BBAPRO33799
~H-NMR. study and structure determination of 4,4- and 4,6-dimers from electrochemical reduction of NADP ÷ Enzio Ragg 1, Leonardo Scaglioni t, g o s a n n a Mondelli 1 Vincenzo Carelli :, ltalo Carelli 3, Antonio Casini z, Alessandro Finazzi-Agrb 4 Felice Liberatore : and Silvano Tortorella 2 J Diparlimento d3 Scien~ Mnlecolari Agroahmeatari, Oaiversit~ di Milang Milano (Italy). ' Dipartimento di Studi di Chimiea Tecnolagia delle Sostanze glotogtcamente Attwe, UnJ~ersllhdi Roma "La Sapiep~za'. Roma (Italy). ~ Dipanimrnt~ .~.iChinffca, LC.M., Universitd de L'Aquda, L'Aquila (Italy) and ~ Dipartimento di Med~rina Sfc*i,~itntate e Scienze Blochirniche, Uni~ersitd di Roam Tot Vergata. Roma (Italy)
(Received9 August19~)
Keywords: NMR,1ft.-:Nucleolidestructure; Electrochemicalreduction;Nicolinamideadeninenucleotide The products arising from ono-electrun electrochemical reduction of the eoenzyme nicolinamide adenine dinucleollde phosphate (NADP +) have been studied by ItPLC chromatography and 1H-NMR spectroscopy. I-1PLC and NMR analyses have shown seven dimefie species, the most abundant of which (40%) has been isolated and has resulted to be an NADP 4,4~linked dimer. The other two diaster~isomeric 4,4-directs present for the 25% and 10%, respectively, have been detected in the crude reaction mixture, but have not been isolated. The 4,4-tetrahydrobipyridine s~ructme and the stereochemistry at the ring-ring junction [or these three isomers have been determined on the basis of their NMR parameters. Preparative HPLC chromatography also led to two fractions enriched in another four dimers, present in the crude mixlure, which turned out I~ have a 4,6-tetrahydrobipyridino slructure. All the chemleal shifts and the It,H coupling ~onstants of the 4,4- and 4,6-tetrahydrobipyridine systems have been ohlained for the seven COmlmunds. For the most abundant among the 4,4-directs Ihe NMR analysis also gave the coupling constant values of the ribose-diphosphate chain.
ln~eduetlen The electrochemical reduction of the coenzyme nientinamide adenine dinudeotide (NAD ÷) has been the subject of several investigations [1,2], The uptake of two electrons by the pyridinium cation present in the coenzyme gives rise to the formation of dihydropyridines, while the uptake of one electro.n,prc~duces a free radical species which after~vards dimerizes. The Facile back oxidation of (NAD) 2 obtained by chemical [3,4], electrochemical [5] and photochemical [61 methods proves the occurrence of an in vitro redox cycle NAD+/(NAD)2 analogous to the couple NAD+/ NADH. Furthermore, the oxidation of (NAD)z to NAD ÷ by molecular oxygen in the presence either of some proteins [7,8], or of the supematant of a standard
Correspondence: R. Mondelli,Dipartimcntodi Scienz¢Mol~r.olari Agroalimentar/,ViaCeloria2, 20133, I~lano, Italy.
liver mylochnndrial preparation [2], and the ability of (NAD) z to bind some NAD-dependent dchydrogenases [3]. might be of biological interest. However. dimerie species of NAD(P) have not so far been identified in viva. Other authors have also studied the electrochemical reduction of the other coenzyme, the nicotinamide adenine dinucleotide phosphate, (NADP+), and the binding of NADP dimers to NADP-dependent dehydrogenas~s [1,9,10], but none of these works deals with the isolation and structural characterization of the reduction products. In the present paper we report on the results of an NMR investigation, performed on the crude reaction mixture obtained from one-electron dectrochemical reduetion of NADP + and on fractions obtained by preparative HPI.C, which led to the structure determination of the 4A-linked and the 4,6-1inked dimers. The NMR parameters have also been used for the molecular eonformationa] study reported in the accompanying paper.
38 Materials and Methods ~-NADP +, free acid, 96% pure, Mr 797.4 (trihydrate) was purchased from Sigma (U.S.A.). Its molar absorption coefficient was found to be 3,2~0 - 18000 4400 M 1. cm (water). This was the starting product used for the preparation of the directs, whereas NADP +, monosodium salt, letrahydrate, Mr 837.5, from U.S.B. (U.S.A.) was used for NMR studies and for analytical purpose. Glucose f-phosphate monosodium salt and gluense-6-phosphate dehydrogenase, lyophilized powder (140 U / m g enzyme protein) were from Boehringer (F.R.G.). Gt0-Sephadex was obtained from Pharmacia (Sweden); the resin was swollen in distilled water and stored in 0.1% NaN 3 at 4°C. Before the elution, the packed resin was exhaustively washed with 50 mM aqueous NH 3. Other chemicals, unless otherwise specified, were of analytical grade. Polarograms were recorded with a three-electrode Princeton Applied Researc,h 170 Multipurpose Instrument in a cell kept al 25,0-4-0.1 °C. Apparatus and procedures for v,larographic analysis were as previously de,scribed [14]. The reduction polarogram of NADP + (c=0.5 raM), vw.orded in aqueous 0.1 M NH4CI-NH 3 buffer (pH 10.6) in presence of Et4NCI, final ionic strength 0.5 M, e~Aibits two well-defined In-reduction waves with E~/2 = -1.00 V and -1.60 V vs. S.C.E., respectively. These values are in goc,d agreement with those obtained by Schmackel ct al. [1] in a K~CO3/KHCO3, plus Et4NCI, buffer. Macroscale dec_ trolyses were carried out by using an Am¢! Mud. 555/A Potentiostat equipped with an Amel MUd. 721 Analogic Integrator and aa Omniscribe Recorder. A three-com. partment water-jacketed cell kept at 25.0 + 0A°C, containing a magnetically stirred mercury pool, apparent area 20 cm2, as working electrode, was used. An afar-salt bridge was inserted in the anodic compartment on a sinterized glass frit of medious porosity. The anodic compartment contained a platinum gauze cylinder immersed in a KCi solution. A calomel saturated KCI electrode (Metrohm Herisau, Switzerland) was used as reference electrode. The buffer solution (0.1 M NH4Cb NH 3 (pH 9.3)) was pre-electrolysed at the same potential of the dectrolysis until a constant current value was reached. Nitrogen or argon (99.9% purity) was bubbled continuously through the cell during both pre-electro|ysis and electrolysis. HPLC analyses were performed by a Perkin Elmer Series 3 Liquid Chromatogral~h , equipped with a LC55B spectrophotometric detector, an LC 55S digital scanner, and a 023 recorder. Integrations of peak area were made by means of a Hewlett-Packard HP 3390A integrating recorder. A RP-18 type 250-4 LiChrocart Merck column, particle size 7 lsm, was used for the analyses, monitoring the eluate at 260 or 340 rim. Composition of the mobile phase (v/v): solvent A, redislilled water
(30%), LiChrosolv ethanol (12%) and aqueous 0.1 M NH4HCO ~ (58%); and solvent B, 0.07 M NH4HCO ~ in redistilled water~ The f~lowing solvent program was used: (1) an isocratic step with 3% of solvent A (5 min); (2) a linearly increasing gradient step from 3 to 6~g of solvent A (20 vain); (3) an equilibration step with 3% of solvent A (8 min). Flow, 1.0 ml/min. After each set of analysis the column was regenerated with redistilled water (200 ml), then with MeOH-H20 (15%, v/v) at a flow of 0.1 ml/min overnight. Before each set of analysis the column was conditioned with redistilled watersolvent B (1:1, v/v, 50 ml, flow: 1.0 ral/min.), then replacing water with solvent A for further 30 rain. Finally, the initial conditions of the solvent program can be used for equilibration time. All the solvents were pre-treated with the RP-18 phase (10 v.m of particle size), filtered on Millipore HAW 0.45 .am fillers and degassed. A precolumn (25 x 4 ram, i.d.) filled with RP-18 phase was applied before the injection valve. Ultraviolet spectra were measured according to the stopped-flow technique with digital scanner, after calibration performed in the same conditions of the analytical program. The presence of NADP + was detected either by comparison with an authentic sample, or by enzymatic test in presence of gtucose-6-phosphate dehydrogeaase. Preparative scale HPLC separations were carried out by using the pumps system of a Waters 500A preparative chromatograph and a stainless column of 400 rnm length x 40 ram, id., fdted with LiChroprep RP-18 (15-25 /~m of particle size, 30 g) suspended in MeOH (150 nil). Composition of the mobile phase (v/v): distilled water (30%), aqueous 0.1 M NH~HCO 3 (67.6~,) and absolute ethanol (2.4%). The column was conditioned with the mobile phase for 30 min before the injection, flow about 6 mi/rnin. The effluent was monitored by the R.I. detector system of the Waters 500A. Ultraviolet spectra and enzymatic analyses were performed on a Perkin Elmer 555 UV-VIS Spectrophommcter in aqueous buffer at neutral or slightly alkaline pH. NMR spectra were recorded with Broker CXP-300 and AM-500 spectrometers. Chemical shifts are in ppm (~) values from external 3-(trimethylsilyl)propane-1sulphoaic acid sodium salt hydrate (DSS), estimated accuracy_+ 0.005 ppm. Coupling constants are in Hz, accuracy +0A. Hz, unless specified in the Tables. Data for compounds IIl-Vlli were obtained from HPLC fraction 1, those for VIII and IX from fraction 2, and those for X! from fraction 3. The spectra were measured (60 mg.m1-1) in D20 at pH 9.2 :i:0.1; the pH values arc uncorrected meter readings (glass electrode) from DzO solution. The solutions were adjusted to the desired pH with NH 3, and kept under nitrogen. The variation in chemical shift with concentration and pH, in the range 80-20 rag-ml" 1 pH 9-10, are within 0.1 ppm. The measurements of chemical shift values and of the rein-
39 rive areas were performed immediately after the preparation of the solution samples, The dimeric species studied are stable for r~are than 24 h in the experimental conditions used. The two-dimensional COSY spectra were acquired by using ~hk45 ° observing pulse, 256 FIDs zero filled to 1024 points, weighted ~ t h a sine-bell function, and transformed in absolute value. The suppression of the HDO signal was achieved with a long low-power pulse (1 s), before acquisition. The spectra were analysed by usir~g the PANIC program, a LAOCOON-type program included in the Aspect-2000 computer library.
Macroscale electrolysis of NADP + and isola/ion of diilrlers
In a typical run, NADP +, free acid (1 g) was di.ssolved in 50 ml of a previously electrolysed 0.1 M NH4CI/NH 3 buffer (pH 9.3) and then the solution was eleetrolysed at -1.20 V vs. SCE. The electrolysis was complete in about 1.5 h, as shown by the constant value of the current, equal to tha¢ measured at the end of the preelectrolysis. The faradaic value at the end of the electrolysis was in good agreement with the theoretical uptake of one electron per mole of NADP + The solution was lyopkilised and the residue (1.2 g) was chromatographed on a G,o-Sephadex column (800 mm length × 32 mm i.d.) using distilled water containing 0.3~ (v/v) of 33% ammonia (about 50 raM) as ducat. A flow rate of 0.8 ml/min was maintained. The fractions eluted in the volume range of 210-280 ml showed the same absorbance ratio (.~I259/A338) a s the electrolysed solution and were free of chloride and NADP +. These fractions were lyophilised yielding a soft yellow crude product (0.8 g), which must be stored under vacuum at - 3 0 ° C , otherwise a slow autooxidatioo to still enzymatically active NADP + takes place. Calculated for C 4 2 H ~ z N l a O a 4 P t ~ ( N H 4 ) 6 . 12H20, M~ = 1807,26 (crude reaction mixture): % C=27.91; H = 5,58; N ~ 15.50. Found: ~ C ~ 28.15; H ffi 4.82; N = 14.75. ¢zs9=34300 M - t . c m - I ; ~3~s=6250 M 1.
is composed almost exclusively of isomer X], as shown by HPLC (92%) and NMR (84%), An attempt to crystallize it was not successful. Fractions 1, 2 and 3, rapidly frozen and lyophilised, yielded residues of 110 rag, 30 mg and 40 rag, respectively, which showed km~ (259 and 33g am) and chemical reactivity (e.g., backo~idation with K~Fe(CN)6 ) to NADP +) characteristic of the tetrahydrobipyridine system [2-4]. Results and Discussion
Analysis of the reaction mixture The electrolysis of NADP + at - 1.20 V vs. SCE gave a crude reaction mixture, which, after chromatography on a Sephadex column, displays the HPLC elation profile reported in Fig, 1. This shows eleven peaks (I-XI), corresponding to three main products Ill (10%), V (25%), X! (37~) ond to other components in a low IU
80
E o
~6o
41J U
~
t~ g. Q
4o
I
c m - I.,
HPLC analysis and preparative scale separation of dimers The dution profile of a sample of the above lyophilised crude product, monitored at 260 nm. is shown in Fig. 1, while the relative areas of the single peaks are reported in Table I. When the monitoring is made at 340 am, the elation profile shows only seven peaks, with three major components, t R 2.8 mix (lII), 7.9 mix (V) and 20.4 rain (XI). A sample of lyophilised crude product (300 rag) dissolved in the mobile phase (2 nil), was subjected to a preparative scale HPLC separation, as described above. After a dead volume of about 90 ml, three fractions 1, 2 and 3 of 70 ml, 30 ml and 150 ml, respectively, were collected and analysed; their compositions are shown in Table I. In particular, fraction 3
20
~/q v,, IVl
s
Virl
~'2
~#X
I'G
~o 2'4 time (rain)
Fig. 1. HPLC ¢ltttioll profile of the crude rcaetiQn mi~tture obtained from electrochemical redaction of NADP + and monkored at 2r~ nra. Releation times (ta) are given neglecting Ihe dead ~:olume of the ~:olutnn and slart Rora the instant of iajeclion (see arrow)~ The peak with t a of 4.6 rain corresponds to NADP" (see Table I and the ~,~ectionMaterials and Methods).
40 TABLE I Cemposizion of the cnde m ~ m r e and of the itPLC froctwr~' from electrochemical r~'ductron ef N,'tDP *
The crude mixture {M), obtained h-c~mSephadex column, lyophilised and dissolved in the mobile phase (see Materials and Methods) was injected in a v~lume of 4 pl (3 m8. ml t). The separated fxaclions 1.2 and 3 were injured in volumes ranging from ! 0 to 30 pl. The ¢mt,mt was m0nitoted at 260 nm. Peak numhcr '
Compound
!
=
It III IV V V[ VII VIII IX X Xl
= 4,4-direct NADP * 4,4-dimer 4,6-dlmer 4,6-dimer 4.6Mimer
4,6-direct nicotinamide 4~4-dimer
rR ~ M-"~
HPLC p~ak area (%) Mc Id
2e
t.8
4.5
7.0
6.0
2.3 2.8 4.6 7.9 9.0 "[1 5 ]5.6 16 g 18.8
1.0 10.0 6.0 25.0 3.0 6.0
2_0 23.0 5.5 38.0 9.0 13.0 0.5 2.0
1.0 3.0 19.0 6.0 7.0 4.0 320 18.0
~.4
4.0 3.0 0,5 37.0
NMR signal area (%) Me 1d
2d
3d
2.0
e
•
e
~ 10.5 10.5 8.7 4.7
t
L0 L0 0.5
. 11.6 2,8 25,4 5.0
37.2
r
22.3 m 6s
r
3~
1• 0.5 2,0 92.0
4.0
~
.
:~.5
. 29.1 3.5 44.2 9.3 11.6
3 ~
I
3.9 1.6 40.9
.
2s ~ ~
6.3 r
r r
9.5 84.2
a See Fig. I.
b The retention times (min) ~re given ne~ecti,g the dead volume of the column_ " Ctttde reaction ndxmm. a Fractions of pruF,arativ¢ HPLC. = Not identified. f Absem or le~ than 0.5~, e Partlaily overlapped b5 od~er z,i3.'~s.
q u a n t i t y . T h e ultraviolet s p e c t r u m o f t h e c r u d e r e a c t i o n m i x t u r e exhibits a b s o r p t i o n m a x i m a at 259 a n d 338 n m , b u t n o t b e y o n d 400 n m , t h u s e x c l u d i n g s t r u c t u r e s inv o l v i n g d i m e r i z a t i o n at p o s i t i o n 2. T h r e e m a i n fractions, 1, 2 a n d 3 w e r e o b t a i n e d b y p r e p a r a t i v e liquidliquid c h r o m a t o g r a p h y ; t h e H P L C a n a l y s i s s h o w s t h a t f r a c t i o n 1 is e n r i c h e d o f c o m p o n e n t s IIl, V, V1 a n d V I I ,
fraction 2 in VIII and IX, ,~hile fraction 3 ¢on'aius almost exclusively compound XI (Table I). The ~H-NMR spectrum of the crude mixture shows the main products III (11%), V (25%), XI (40%) and severat other components in low abundance, in agreement with the chromatoBraphyc results (Table I). Amon 8 the latter one is always NADP +, often appearing to-
H
NH 2 (~0 H
H (111, V, XI)
~H 2 ,~ON-f
2
-e
~CONH
2
x2
+e
NH2
S.C.E.
¢
H
~0
R--kf A ~
H
H B/)2-CONH 2
(Vl, VII, VIII, IX)
Scheme I. Formation of 4,4- and 4,6-1i~ed NADP dimers from one.electron electrocheafi ~zl reduction of NADP +. R = adenosine (2.phmphate)diphosphale-ribo~e.
41
A2
A8 N2
N2
N 6 A~"
N4
N6
rd4h
V+XI
NS
V
~
a aba
,
-
-
~
~
III
b
I ....
t .... g.o
' ....
I 8.0
'
I 7.0
I .... §.O
'
'I 5.0
"
' '
~ 4.0
'
I
~.o
'
8 ppm
'I
2.0
Fig. 2. I H - N M R s p ~ r m n (300 MHz, D z O / N H 3 (pH 9.3))of the crude roction mlxtme from electrochemicalreduction of N A D P +. The sign~Is -marked wilh (a) and {b) a ~ N A D P + and nicotinamide,r~p~tively.
gether with nicotin,mfide. The signals of both these compounds are clearly visible at low field in the proton spectrum reported in Fig. 2. The assignment was made by means of el',¢mical shifts and coupling constant values, compared with authentic samples. The coupling constants are the following: J(N2,N4)= 1.7, 2~5 Hz; J(N2,NS)= 0.0, 0.g I-Iz; J(N4,NS)= 8.1, 8.3 Hz; J(N4,N6) ~ 1.3, 1.7 Hz; and J(N5,N6)- 6.3, 5.0 Hz, for NADP + and nicotinamide, respectively. NADP + is not starting material unreacted, but comes from autoxidation of dimers, because it increases with time and appears in all the three fractions obtained by preparative HPI.C. Nicotinamide oriSJnates from NADP ÷, as proved by the 1H specmun of NADP measured in the same conditions. A few protons of the main components give spectral patterns which can be easily identified, namely N2 * and N4 of dihydropyridine tings at about 7 ~ and 3 f$, respectively, A2 of adenines at about
* Numbered As and Ns indicale the hydrogen atoms of the adenine a t u c l e o t i d e and those of nicofinamide nucleotid¢ unit, resp~lively. Unspecified numbers indicate atom,~ of bolh unit~.
8 8. The other protons partially overlap with ribose nuclei, N6 with AI' and N5 with NI', or give broad peaks of low intensity, such as A8 of adenines, due to exchange with the deuterium of the solvent. Nevertheless, almost all protons of the bases and part of those of ribose moieties were identified, by means of syst ,. ~tie spin decoupling experiments. The analysis of the proton spectrum of HPLC fraction 1 confirmed the identification of Ill and V, while the four compounds, VI-IX, detected in small quantities in the crude mixture, were identified by examining fraction 1 (VI and VII), and fraction 2 (VIII and IX). For the most abundant one, XI, the NMR study was performed on fraction 3, where it amounted to 80-905. tH.NMR spectra, structure and absolute configuration o~ 4,4-dimers IlL V and X I Fraction 3 was analysed first, because the abundance of isomer XI allowed to obtain the complete set of chemical ~hifts (Table If) and eoupfing constants (Table III) for all protons. The conneetivities of nuclei were found by using two-dimensional COSY (Fig. 3) and
42 TABLE In
TABLE l[ #lt fhemwal shi~r values for 4,4-d~mers II1. V and XI
H,H couphng consra,t values far 4.4-dtmers IlL V and XI
Measured in ppm (61 relative In external DSS; DzO solutions adJ:l-ir,t wit!i NH) to pH 9.3. E~limated accuracy +0.005 ppm unless specified, Similar values in parentheses may be interchanged,
J values are measmed in Hz and me given without sign. Eslimaled accuraey+0.'l Hz. For the ribose protons of Ill ~nd V, aecnraey4-O.2 Hz.
III ~
V
XI
A b
Bb
(8.129) (&359) 6.951 3.191 4.669 6_108 6.131 (4 rg00) ¢
{g.113} {g.383} 6.93(I 2.909
(4.561) ¢ u
(4~54) • ,J
(4.5~la) c d
d a
d d
NI'
(4.720) ¢
(4.7221 ~
N2'
(4.071) c
N3"
(4.134)"
(3.966)': (4.148)~ 3.98-4.01 d
d a (4.659) ~ (4.030) ' (4.119) ~
A2 A8 N2 N4 N5
N6 At" A2' A3" A4' A5' A5"
N4"-SL5*"
8.101 $.M.~ 6.964 2.27? 4.409 6.122 6.059 (4.917)
c
4.530
6.026 6102 (4.8391 ~
J
Ill
V
XI
A a
8.118 8.380 %081 2.916 4.400 6.008 6.100 4,$76
N4,N4 t~ N4,N5 N5.N6 N2.N4 N2,N5
2.2 5.1 E..2
0.5" 0.5 ¢
N2,N6
1.4
N4.N6
0-5 ~
NI',N2' N2".N3'
7.1 ~ 5.5 •
4~335 4.245 4.160 * 4.729 4.080
N3',N4' At',A2." A4'.A5'
4.[41
A4".A5"
4.529
• The assignment of [11 with respect to V followed hy comparing spectra with different proportions el the two isomers. Values for the two non.equivalent halve.g. ¢ Acc,tra~y:t0.0t ppm~ 'J Not analysed, Partially nvcdal0pcdby N2' signal; value obtained by the calculated spectrum.
B a
3.2 5,5 8/) 0.9 0.5
5.6 %9 0.8 0,3 ,l
2.3 4.8 7_9 0.5 = 0.5 =
2.0
1,5
1,4
0.5 (7,2)
0.5
0.5 =
(5.1)
('L4} 5.4
7.1 5.4
2.0 • 15.|)
(1.9) (4.9)
(2.3} (5.0)
2.5 4.9
AZ',A3'
4.8
5.0
4.g
4.8
5.1
t
4.8 4.8 i
4.8
A3',A4'
I
3.0
# f
t f
i I
4.3
AS".AS"
11.5
a Values for two non-equivelent halves; A and B values belong to the signals o[ eolnrrmsA and B, respectively, of Table 11. Coupling constant belween protons at the ring-nngjunstion. Not measured, But delO~ledlby decoapling experlmenls. The reported value g~v¢ the best fit with the calculated spectrnm. Line width 0.5 tlz. Tentative attribution. r Not analysed.
spin dccoupling experiments, some of which are reported (Fig. 4), whereas the coupling constants were measured from one*dimensional spectra. Fig. 3a shows the cross.peaks corresponding to the coupled protons in the dihydronicotinamide and ribos¢ fragments of XI. The sequence NdvNS-N6 was readily identified, and N2 was assigned by the long-range coupling with N6. This inleraction, although small, is clearly visible in the right inset of Fig. 3. The hi#l-field shift (2.916 8) and the second-order pattern of N4 indicates a 4,4-lctrahydrobipyridin¢ structure. This signal belongs to two protons, which arc chemical shift equivalent, but magnetically non-equivalenL i.e., they are the AA" part of an A A ' X X ' Y Y ' Z Z ' spin-s~stem [ l l ] , Figs. 4a and b show the expcrimental spectrum a n d the calculated patterns for N4 and N5 protons. Owing to the second order effect, we obtained the AA' coupling conszant, i.e., J(N4,N4), which proves that dimerization occurs with junction at C-4. Also the other protons at corresponding positions in the two halves are chemical shift equivalent.
The anomeric protons, A I ' a n d NI" of the ribose moieties were assigned in the following way: starting from A 2 ' pattern at 4.876 6 (Fig, 4a), which shows a three-bond interaction of 6,8 Ii~ with phosphorus, we decoupled A I ' at 6.100 8 (Fig. 4c). In the inverse experiment, not reported here, A I ' donblel collapsed into a singlet. N I ' protons, expected as a dou[/xt, were identified at 4.729 8, after pre-saturation of the water peak, and decoupling of N 2 ' protons at 4.080 & The four-line signal of N 2 ' in turn becomes a doublet of 5.5 Hz by irradialion of N I ' (Fig. 4d and scheme ll). The connectivitics with the other protons, which complete the sequence of nuclei in the ribose units, were detected through the two-dimensional spectrum, while the analysis of H-4' and H-5" patterns (Fig. 4a and e) aimed at obtaining the coupling constant values, was performed on the one-dimensional spectrum at 300 and 500 MHz frequency. In the case of the A ribose rings the protons are well resolved, because they are sufficiently spread out, whereas N4', N5' and N S " are so
Fig. 3. la) Two-dimcnslonal Ill-COSY and (b) one-dimensional ?':MR spectra (300 MI-Iz,D20/NH~ (pH 9.3}1of dimer X! from fraction 3, t~gion from 2.8 to 6A 8. The sing,lets of A2 and A8 at 8.118 and 8.3gl}~ are not rel~orted.Left inset: expansion from 4.0 to 4.5 8; right inset: cross-pc~s from N2 and Nb protons,
43
(b)
510
31o
...~--~.d.~:
6'
O
P
./
.
4~4
4-2
,
q
o
4.0
0
j-
(a) ,
t
i
g
""
r
UP :r~o
~.o
ppm
44
Il~II m2
ml"
At"
~ ,[d}
let"
m2"
t~¢
m~J
?~ml
1110
IIom
•
ITI
• 7ae
@~Ul
4215
I
I~
4141
I~I
l'gll
I @p
Fig. 4. III-NMR sp~lrum of 4,4~dimer XI from fraction 3 (~0~ MHz. D~O/NH~ (pH 9.3}). (a) expanded pattet-~ of the most important sig~]s: (b) caicuimmt spectrum of N4 and N5 protons (AA' and XX' ;:~at'lof ~ ' XX"~'~r' ~ ' spin system}; (c) A2' d~couplcd from A t ' upon irradiation at 6.100 ~l; (d) N2" d¢coupled from N I ' upon inadiation at 4.729 ~: (e) calculated spectrdm of A4', A5' and AS", with J(P, A 4 ' ) = 2.0, .I(P,AS") = 5.0 and J(P.AS') = 5,4 Hz.
close t3gmher even at 5Q0 MHz, that the coupling constants involving these nuclei could not be obtained in D20/'NH t solution. The degeneracy was removed by using sedium tetraborate buffer, and the results are reported in the accompanying paper, together with the values of the hydrogen-phosphorus coupling constants. The NMR analysis of compounds Ill and V was performed on the reaction mixture and on HPLC fraction 1, wHch is enriched of these components. The adenine A2 and A8 singlets of the three main isomers
H
CONH2
are close together, but detectable even in the spectrum
~f the crude mixture (Fig. 2). Two protons in the spectrum of fraction 1 are clearly visible and correspond to N2 and N4 of |II and V (Fig. 5); the other protons of the nicotinamide rings were assigned by decoupling experiments, whereas a few ribos¢ resonances were detected but not analyzed. The doubling of N2 and N4 signals of V indicates that the two halves are diastereotopic, whereas for Ill, the correspo.ding nuclei of the two halves are chemical shift equivalent.
N
N H
H
s~
H
I H
t
H
H
s
ONH4 0
'/
o Io "~/
"12 H 2 0
II,o
-,,p ~.s'
;L i" 4' o" ,
a'
0,.~
i+ O
tl OH
O--P--ONH
4
I Scheme11.Molecular~.lm¢(urcof 4,4-[inkcclNADP dimcrs,(NADP)2.
45
N4B(V)
N4A(V)
N 201|)
N4(lll)
N 2 (V)
(0 i
J ~ ~ ( d )
~,
(b)
I••
I
I
6.9~4
8,95|
6.930
3,191
(a)
(a) ppm
2,909
I 2.2?7
Fig. 5~ 1H-NMR spectrum of 4,4-directs III and V from fracdon t (309 MHz; DzO/NH s (pH 9,3)), expanded patterns of (a) N4 and (b) N2 protons; (c} calculaled spectrum of N4 for Ill; (d) inadialioo of N4A proton of V at 3,191 8; (e) irradiation o[ N4B proton of V at 2.909 & (f) irradiation of N2 protons at 6.93-6.95 8; (g,h,i) kradialion of N4 protons at 2,909, 3.191 and 2.2"/'/~, resp~tivdy.
The double resonance experiments reported in Figs. 5d and e show that the N4 protons of V at 3.191 and 2.909 8 are connected by a vieinal coupling; those in Figs. 5f-i prove the long-range interactions between N2, N4 and N6. Note the absence of coupliug between N4 and N6 (Fig. fit'), as already observed for Xl. The experiments in Fig. 5g-i were used to assign N2 signals. The NMR data, reported in Table II and Ill, confirm that the three compounds are dimers and prove that the dim¢fization occurs at position 4. The symmetry and the high-field shift of N4 protons are indication of this junction, but conclusive e,ddenc~ for the 4,4-tetrahydrobipyridine structure came from the analysis of all proton signals o1' ihe pyridine fragmenL In particular J(Nf,N6) = 8.0 Hz appears to be diagnostic for a 1,4-dihydronicotinamide ring. We carl the~fore conclude that the two halves of isomers Ill and XI are identical and that the configuration of the chiral (2-4 centres is 4R, 4R and 4S, 4S, or vice-versa. As a consequence the stereochemistry of isomer V must be 4R, 4S. Since the configuration of the diphosphate ribose moieties is re-
tained, neither enantiomers nor mcso isomers, but only three diastereoisomers can be obtained.
~H spectra and structure determfnation of 4,6.dimees VI1X
VIII and IX isomers were identified in the HPLC fraction 2 and also in the crude reaction mixture through detection or the N4 and N6 proton signals for the 1,6-dihydropyridine ring (A) and 1,4,dihydropyridine ring (B), respectively. The chernical shift values (Table IV) of these protons are diagnostic for the ring.-ring junction, as well as tbe vicinal couplings betwce,t protons on the w bond (Table V). The N4 hydtogens of dngs A lie at about 3 ~, whereas the N6 of rings B are deshielded to about 4 8 by the viclnal nitrogen atom. The N4A resonances of both VIII and IX isomers (Fig. 6a) appear at 3.330 and 3.280 ~ as double doublets due to ~/(4A,SA) and ~/(4A,6B), with each line further split or broadened by allylic couplings with N2 and N6 of the same ring (see the irradiation of N2 in Fig. 6d).
46
N6B
NSB
(,[) V I I I .1- I X 4111
IX IX
VIII
'~~1 (~-) c) .098
4.97!
J
3.890
i 3.887
ppm
3.330
3.280
Hg. 6.1H-NMRspcctrnmof 4,6-dimcrsViii aud IX fromfraction2 (300 Ml-tz,I~O/NH 3 ~pH 9.2)),expandedpatterns of (a) N4A,(b) N6B and (¢) NSB prot~az~;(d) N4A decoupledfromN2A upon irradiationat 7.05 3; (c) N6B dccouplcdfromN4A upon irradiationat 3,3 6; (~ NSB al~d(g) I,,14Ade,coupled fromN69 upon irradiationat 3.88&
The identification of N6B signals close to the strong absorption or c - 5 ' protons was done by irradiation of N4A and viceversa (see Fig. 6b, c and g). The analysis of the signals at 3.87-3.89 8 (Fig. 6b) was not easy, but the coupling constants reported on Table V are accurate, because the N5B protons are visible at 5.028 8 and 4.971 8 (Fig. 6c); they decoupled upon irradiation at 3.87 ~ (Fig. 6f). These data prove the 4,6 ring-ring junction and consequently the 4,6-tetrahydrobipyridine structure for VII! and IX. It dearly appears that the vicinai couplings between the protons on the ~r bond arc different for the 1,4- and ],6-dihydropyridine systems: J(SA,6A) measures 8.0 Hz for 1A-dihydro structures, whereas for 1,6-dibydro structures J(4B, SB) is 10 Hz, in line with th© modal compounds obtained by reduction of 1-benzyl-3-carbamoylpiddinium chloride [12]. We might be tempted to interpret these values with a stronger delocalization effect in 1,4- than in 1,6-dihydropyddine system; but the lower value of the olefinic
interaction for ring A can also be explained with the inductive substitugnt effect of the nitrogen atom on the ~icinal proton coupling EllE. The four-bond interne tions, J(2,6) for ring A and J(2A) for ring B, cannot be used to irderprct delocalization effects, because their values (1.5 Hz) are well explained with a u-mechanism through a W path of coupling Ell]. The assignment of VIII vs. IX was performed by comparing the intensities of N2 signals (Fig. 7b), taking into account ~ a t in fraction 2 isomer Vill is more abundant than IX on the basis of HPLC analysis. In the frequency range of N2 protons (7-7.5 8), there are a total number of nine peaks. Excluding those at 7.088 (dimer XI), at 6.951 and 6.930 8 (dimer V), two peaks of the remaining six decoupled upon irradiation of N4A at 3.3 8 (Fig. 7c), and two others upon irradiation of N6B at 3.88 6 (Fig. 7d). This allowed to assign the protons of ring A vs. those of ring B and to identify the signals of dimers VIII and IX.
47
~(
~ (f)
{e)
II4"V (a (')
)
?.165
7.086
.~VIII)N
7.014
.2.(v,,.)
N2B (IX)
(d)
t ~1 " l ' " "rl 7,111 7.101 7.088 7,023 6,D51 6.930 ppm 7.305 7.234 7.171 Fig, 7, 1H,NMR spectra(N2 protons)of (a) HPLC fraction ] and (b) fra¢lion2; (c) N2A signalsof VIII and IX decaapl~df~oraN4A upon irradiationat 3.3 ~; (d) N2B signalsof VIII and 1XdocoupL',dfromN6B upon irradiationat 3.88 ~ (e) r42Asignalsor vI and VIi deconpledfrom N4A uponirradiational 3.0 & (f) N2Bsignalof VII decoupledfromN6B upon irradiationat 3,7298.
The low intensity N2 signals at 7.234 and 7 . t l l were assigned to the two remaining diastereoisomers VI and VII, which are slightly more abundant in fraction 1 (Table I). The NMR spectrum of fraction 1 was thus used for the structure determination of these dimers. Fig, 7a shows the N2 peaks of VI and VII, well separated from those of the major components III and V; decoupling experiments, some of which are reported in the insets (e) and 03, confirm the 4,6-tetrahydrobipyridine structure also for isomers VI and VII. The relative proportions of the components given in Table I were calculated through the integral of N2 signals. For all the 4,6-directs, ~he N6 protons of ring A and the N4 ones of tin 8 B lie in the 6 ppm region and are both connected to N2 through long-range couplings. However the N6A and N4B patterns are difficult to analyse, due to overlap with the anomeric AI' protons. As a consequence, in the case of the minor isomer V[
and Vll these signals could not be detected; whereas for VIll and IX the chemical shift and coupling constant values (Tables IV and V) were obtained through decOupiing experiments~ and from the analysis of the related NdA. NSB and N2 spectral patterns.
Conclusions The one-electron electrochemical reduction of NADP ~ !~tds to the formation of dimers with the ring-ring junction preferentially at the C-4 positions, the reaction thus appearing to be regiospecific. The relative amounts of the three 4,4-diastereoisomers indicate that the dimerization process is also stereosetec~ve, Actually the two faces of the nicotinamide ring are diaslereotopic and the approach, for instance, of the Re
48 TABLE IV
JH e,~emical shi/r vMues for 4,6-dimerr V1- IX Measured in ppm (8) relative In external DsS: Dan solutions adjusted with NH 3 to pH 9.2. Eslireated accuracy4-0.005 pprn unless specified. Similar values in parentheses may be inlerchanged. VI
VII
VIIi
IX
RingA a N2 N4 N5 N6
7_014 2.966 b b
7_086 3.003 b b
7_07.3 3.280
7_101 3.330
6.095
Ring B" N2 N4 N5 N6
7_229 6.177 b 3.89 ~
7,165 b 5.018 3.729
7+111 (6.136) d 5.028 {3.8~) a
A2
8.119 f 8+139 r
8.0~0 t
8.091 t
18-113)
(8.122)
A8
(8.43) s
{8.,IB) ~
8.396 8,420
8.390 8+410
4.656 + 5.935 7.305 (6.103) a 4.971 (3.867) d (8.103)
A gad B corr~poad to the 1.~diydro- and 1,6-dihydtonicotinamid¢ rings, respectively. b Not delected. Only one signal was detected, but could not be assigned to VII[ or XL a Accuracy+O.Ol ppm. c P~rlialty ovedupged by the N5"-5" signals. t Tentative identification. s Broad signal. TABLE V
tt, H cogpling coezstaol values/or 4,6~dJmors V I - I X J values are measured in Hz and are given without sign; estima:ed accuracy:L0.1 Hz, unless sla~ified. A and B correspond ~o the hydrogen alums of *he lA-dihydro and 1,6-dihydrenicotinamide rings. respecliveiy. J'
VI
vII
VIII
IX
2A,4A 2A.SA 2A,6A 4A,SA 4A,6A
0.8 A 1.5 5.5 0.5 a 1.8 0.9 a 9.9 a ~ 3.8
0.8 a 1.5 5.7 0,5 a 2.0 0.7 1.0 9.9
1.0 , 1_5 5.5 0.5
1.0 a 1.5 5.0 0,5
8.0 b
8.0
1.3 0.7 0.7 9.9
1.5 0.6 0.7 9.9
0.3
0.5
5.6 2.3
5.6 2.8
0.5 S.4
5A.6A 2B.4B
2B,SB 2B,rB 4B,SB 4B.rB 5B,rB 4A.6B a Not delectcd.
Accuracy ±0.2 Hz.
3.8
face t o w a r d R e , o r S i t o w a r d S i , l e a d s for the 4,4-dim e r s to t h e 4 R , 4 R o r 4 S , 4 S c o n f i g u r a t i o n , respectively, T h e s t e r e o c h e m i e a l choices m a d e b y t h e d e h y d l ~ g e n e s e s d e p e n d e n t on n i c o t i n a m i d e c o f a e t o r s is well k n o w n , b u t also in this n o n . e n z y m a t i c r e a c l i o n a dia s t e r c o f a c e d i f f e r e n t i a t i o u occurs, a n d leads p r e f e r entially to R e - R e a n d S i - S i d i a s t e r e o i s o m e r s . T h e low a b u n d a n c e o f t h e 4 , 6 - d i m e r s i n d i c a t e s that, n o t w i t h s t a n d i n g t h e steric h i n d r a n c e a t t h e r i n g - r i n g j u n c l i o n site, t h e re.action m a y h o w e v e r occur, w h e r e a s t h e a b s e n c e o f 2,4- a n d 6 , 6 - d i r e c t s s h o w s t h a t these j u n c t i o n s are n o t a l l o w e d . F r o m t h e a p p r o x i m a t e l y similar q u a n t i t i e s o f these i s o m e r s , f o u n d in t h e c r u d e r e a c t i o n m i x t u r e , it w o u l d a p p e a r t h a t t h e 4 , 6 - d i m e r i z a tion p r o c e s s is less stereoselective. T h e i d e n t i f i c a t i o n o f t h e tetrahydrobipyridine species
was carried out by using the H,H coupling constant values of the dihydronicotinamide rings. Some of these couplings, namely the vicinal ones between the protons on the ~r bond appeared to be diagnostic for the ringring junction. The oilier coupfing constants can be used for t h e c o n f o r m a t i o n a l a n a l y s i s o f these systems. References
1 Sclunacke.], C.O., Santhanam, K,SN, and I~lving, PJ. (1975) J. Am. Chem. Soc. 97, 5083-5092, and references quoted therein. 2 Carelli. V.. Liberatore, F., Casini, A., Monddli, IL, Amone, A., Carelli, I.. Rotilio, G. and Mavelli, L (1980) Bior8. Chem. 9, 342-351, and references quoted therein+ 3 FinazTJ-Agr~, A-. Avigliano, L., Carelll, V.. Libe~awre, F. and Casini, A. (1981) Bioc.him. Bioph'is, Acta 661,120-123. 4 Chan, S,S., Nordluad, T.M., Frauenfelder, H., Hen'|son, J.E. and Gunsalus, LC. (1975) J. Biol. Chem. 250, 716-719. 5 CarellL I., Ro~ati, R. and Casini. A. (1981) F.Icctrechim. Acta 26, 1695-1697. 6 Avigliano, L , Carelli, V=I Cas~, An, Fina22i-Ag~, A, and Lib¢ratore, F. (19[~3) Biochire. 5iuphys. Aria 723, 372-3"75, A,Jigliano. L . Carelli, V., Casini. A.. Finazzi-Agrr, A., Liberator¢. F. and Rossi, A. (1986) Bicchem. J. 237, 919-922. 8 Avigliano, L , Car¢lii. V.. Casini, A., Finazz.i-Agrr, A. and Liberatore, F. (1985) Biochem. J, 226, 391-395. 9 Cunninghare. A.J. and Underwood, A.L. (1967) Biochemistry 6, 266-271. tO Kovar, J. and Klukaaova. H, (1984) Biochim. Bioph~s. Aeta 788, 98-109. 11 GOnter, H. (198U) NMR 8pectrosan~y, John Wiley and Sons, Chichester. 12 Carelli, t., Cardinal|, M.E., Casini, A. and Amone, A. (1976) J. Ors. Chem. 4.], 3967-3969, t3 Burg|, H,-B. and Dunitz. J.D. (1987) J. Am. Chem. So¢. 109, 2924-2926. t4 Carelii, L, Cardinal|, M E and Michel¢lli Moreoci. F. (1980) .I. Electtoanel. Chem. 107, 391-404.
