nmr of carotenoids—new experimental techniques

Pure&AppI. Chem., Vol. 57, No. 5, pp. 801—821, 1985.Printed in Great Britain.© 1985 IUPAC

NMR of carotenoids—new experimental techniques

Gerhard Englert

Central Research Units, F. Hoffmann—La Roche & Co. , Ltd. , CH—4002 Basle

(Switzerland)

Abstract - In continuation of my previous work further 1H—NMR data ofnumerous end groups of retinoids and carotenoids are presented. In ad—dition, the 'C—NflR shifts and assignments of more than 60 end groupsand derivatives thereof are compiled. Several new experimental NMRtechniques which are extremely useful for the elucidation of molecularstructures have been applied to carotenoids of known structure in orderto demonstrate their utility and great power. Included are Double IndorDifference (DID) for the assignment of signals in complicated 1H—NMRspectra, the pulse sequence DEPT for the investigation of the multipli—city of carbon signals, and some recent two—dimensional (2D) NMR tech-niques. Examples are given for the heteronuclear J—resolved 2D, thehomonuclear (COSY) and heteronuclear chemical shift correlated 2D, andthe 2D-INADEQUATE experiments. The last provides information on thecarbon framework of the molecule. The advantages and disadvantages ofthese experimental techniques are discussed with special emphasis onthe spectra of carotenoids.

INTRODUCTION

During the last three years again a number of new or improved NMR techniques has been madeavailable to the practicing spectroscopist which will no doubt further enhance the import-ance and efficiency of NMR in solving complicated structural problems. It is the main ob-jective of this paper to present some selected examples of the application of such newtechniques in order to demonstrate and discuss their relative advantages and drawbacks withspecial emphasis on the spectra of carotenoids. In this context the following methods willbe discussed:— A simple modification of the long—known INDOR experiment called Double Indor Difference

(DID) which significantly simplifies the assignment of proton signals in strongly over-lapping spectra typical for carotenoids.

— A pulse sequence called DEPT, which is frequently applied now in 13C—NMR spectroscopyto determine the multiplicity of the signals, i.e. the number of attached protons.

— Several two—dimensional (2D) NMR methods which were found extremely useful in the NMR ofcarotenoids. Although these methods are in part highly sophisticated from an experimen-tal as well as a theoretical point of view, there is no doubt that they will beincreasingly applied in this field in the near future. Due to lack of space and time Ishall be unable, unfortunately, to discuss the physical background of these new methodsand will merely reference the pertinent literature.

NMR DATA OF END GROUPS

Spectroscopic methods such as IN, MS and NMR are well—known, effective tools for the eluci-dation of the structure of carotenoids. An important starting point in this task, namelythe determination or identification of the structure of the specific end groups, can bevery efficiently achieved by 1H— or 13C—NMR. I previously presented 1H—NMR data on 48end groups with 84 derivatives including some cis isomers (Ref. 1). In Table 1 are shownfurther 1H—NMR shifts for 26 additional end groups and 52 derivatives; thus data are nowcompiled for more than 70 different end groups and 136 derivatives.Since 13C—NMR has also become increasingly applied in the field of retinoids and carote—noids a similar compilation of '3C chemical shifts is given in Table 2, containing 92 de-rivatives of 62 end groups. The data of Tables 1 and 2 were extracted from numerous spectraof retinoids and carotenoids which were all run and assigned in our laboratory. However, Ishould like to mention here that part of the data resulted from a very fruitful cooperation

801

802 G. ENGLERT

1Table 1. Characteristic H-NMR data of some typical end groups of natural and synthetic ca—

rotenoids (solvent CDC13).*

(ic)1 04

16

1.96 1.51

(5e)

1.09br.578

22

—i.37 1.56

4.02 a 1.9952.35ct

(5f)1.09 1.14

2

21

1.55

5.09 2.0752.4lcz

(14a)2.01

1.20 1.15

1.41 H 2.075399 2.43c

(14b)2.00

HO 1.971.41 ''ax 2 09

4.01 2.4ócx

(14c)2.01

AcO A 1.912.04

''ax 2.1355.04 2.49a

(14d)2.00

AcO 1.962.05 ax 2.165

5.06 2.52c

(43b)

1.426.91

2176.49

49a

2.02

2.051.9

49b2.01

::2.08 1.98

50

0.95 2.20 1.92

:::21?.;;5.60 1.73

51

1.97

-2.67 1.78

52

-1 44 1.04 1.09 1.97

HQo1Hax (JJ6.12 6.16

1.755 -2.15 1.711 .85cx

53

1 312.04

6.29

6.98 2.08

54a 54b

552.04

:::

0

56a ss a)

1.08 CH2OH

-1.30 Hax 2.0854.03 2.39ct

57b b)

1 17 0 99 1.93

2.021.19

1.7852.4lcx

57c

1 17 1 01 1.93

%i931.213.92 1.635

2.4Ocx

58a b)

1 01 1 161.93

HO—1.65 =1.19

3.88 1.8952.2Ocx

58b

0 99 1 21 1.93

2.02 =1.17

4.92 1.8752.30 cx

59a1 21 1 14

1.74

HOA AO#H—1.37 H 1.45 5.19

4.02 j 0 Hz2.32cx 7,8

59b1 21 1 19 1.74

AcOA £0 H2.04 H '1.49 5.19

5.10 1.6950 H

2.3Oa 7,8

60a1.79

HOA AO Hjj

=1.48 5.114.02 1.515 J 1 8 H

2.28a 7,8

60b1 23 1 19 1.79

AcO AO H2.03 1.52 5.11

5.11 1.585 J 1.8 Hz2.24a 7,8

Table 1. Continued

NMR of carotenoids—new experimental techniques 803

61a b)1.34 1.20 1 81

148c

1.8O8?16.19HO'tAO H1.68 5.07

4.24j

2.118 7,81.9 Hz

61b1.28 1.20 1.81

6.50

1.885 ' 18

AcO9A>cT''2.07 Het\1•62 5.075.19 l.87cz

2.198 J7,81.9 Hz

62a b)1.33 1 17 1.72

1.51>i%!._::.1.768HO 0

HeqN11625.17

4.24 1.99

2.128 J78 — 0.7 Hz

62b

1.26 1.18 1.72

1.\ 5.291 6.49

49cx1.885 1 I /6.19AcOT'AO H

2.07 1.58 5.18

5.21 1.95cc1 Hz

2.248 7,8

63

2.06

1.32

2.37 /'%.650O1.89

64

1.190.85 1.35 1.99

2:091

HO H \oHo7.24

1.49 1.552.88

65

1.86

1.18 6 42 6.571 21

2.43' 6.212.41'—. .''L."

5.92 1.97

66a

2.03 1.18 1.93 --1.98

1 24 /1 6.38 6.20 6.39

HO 571 I 659

OH—2.05

66b

—1.72 1.23 1.91 —1.98

>JJ1.21HO

6 41 6.21 6.395.19''OAc

2.11

67a

—1.44 1 24 1 93 —1 98HOi24>1"L''L:

2.31 6.14

6.21 6.36

67b

—1.46 1.26 1.94

HO.J 5 81

1.26 6.03—2.38 6.76 I

6.74

1.98

68 c)

1.73 1.82 1.96

1-177 I 64614.90 /1 \ 5.94 6.244.96 /OAc \

5.16 2.07 —2.06

69a

1.43 1.98 1.98

o 6.46

—6.62

69b

2.08 1.53 1.94 1.98

AcO>J,JJA45

—6.58

702.09

1.18 1.92 —1.98

HOJ (5 50 I 6.60 1

'J\\ \

1.67 6.176.37

71

1 89 1.99 —2.00

I 669 I 667 I1 .83 '__(_—_._._\

II 6.29 6.36

1.691.75

72a

2.231) 2.06

2.17 I 666 I

J,,,J,6. 23 6.19

HO 6 53 2.24-4.66

72b

2.24 2.07

2.15...

H3CO

6.24 6.20

3.82 6.60

72c

2.24 2.07

-L.•206 6.53

Ac ._.JL6.26 6.212.32 6.73 2.25

73a

2.08

I 661 I

HO" L2 221

2.191)

73b2.08

'%%H3CO''6.18 6.19

2.221)3.672.191)

73c2.08

%%•AcO

6.18 6.20

2.362.21

2.06

*Numbering continued from Table 1 of Ref. 1. Numbers in parantheses are supplements to the

previous Table;1)corresponding assignments may be interchanged; br. broad signal, tempera-ture dependent; n.a. not assigned. Note a): see Ref. 2. Note b): see Ref. 3. Note C): seeRef. 4.

Table 2. Continued.


16

156 12 639.5 P.A23.3I 124 6

125.9 134.7

186.7 130.8

17a

161.0I 12.6

27.7 '124 3 124.8

37:61 P141.21134.4

>u—.

34 3134.8

198.9 130.0

17b

161.535.8

27.7/i 25 71

34:31 II i33.3t ,

/ 20.4

3 4

123.2199.5° 130.0 133.1

17c

161 7 124.435.9 27.6

198.8 130.9

18a162.1

36.8 30.7 26.2'12.6

. 31 124 4

H0!114O 134.3200.4

18b160.4

30.5 26 3/ 12.6

/\\/1232 124 4

H3CCO'('140 134.520.8 0 o\ 128.4194.0

19a

161.412.6

39 2 28.1 23 0 124.7

I Ii 142.4t135 4

HO")ff37 .134.6

182.40128.3

19b

158.0 12.6

21 124 6141.6

20.4 143

H3CCO'fl13 6 134.5

168.4 0 /0 130.6179.7

2067.22)

28.1 / 11 8

HO 134.5

1)

41 •71 )/21 .4 197.8

66.32)

2155.2

24.4 29.5' 13.134.1 ¼/$\/128.81 124.944 7

SHOO1'T65 9135

i24.61 22.8137.9

2255.1

27.1 29 3 / 13.134.8 ./$\/129.8I 124.8

66.71 ,136.71 130.841.1''%.'.'

35.2HO12471 22.7

138.0

23

156.0 12.579 ./22 ol

/1 138.8132 331.5137CCH3135:01;8.8 o 30.2

24

\j J119 9 124 9134.8 171 P312.4

52 240.6'/ 28.3 135.4

25108.3 23.0

159.3 21.041 .2 2?? _L,,393

50 .0.0 .207.4126:7

9.9118.0

26169.6

26.0 / 12 420.7

21 124 8

135.1201.30 / 9.6

132.3

27a107.9 17.9

143.8

39.8 27.\3/156.1 38.0126 S8.lt/ll84

'fl'14.3185.30 138.3

27b103.0

143.5 23.3

27?'

\—L's39.8

156.Oq 1193.5 1

U 14.3 126.1126 I

185.30 119.4

28168.2

140 7 12.5J '119 0 124.4¼%187.9

203..2 135.0.

I 146.51137.0

29

124 0141.734.0 29.1

134.791.3 121.0

32 i1 22 8.123.5

30a 111 0147.7 \ 17.6

31.0 26.)\,J91140.544.3

88 1117 6

-1041(414469.3 Q 133.5199.5

30b105.9

147.0 23.0

31.2 26.3"\,L139.34469 31 11934

HO14.4126.8

199.30 134.0 118.3

30c 110.8

146.2 17.630.8 26.\ I 123.7

36.8 \\41 5117.320 9 707.8 1140.5

H3CCO14.5

170.2Q0 I 135.1193.0

30d106.0

145.8 22.930.9 26')\

+1139 236.9

20.8 70 91 1111541.6

H3CCO(146 126.7170.0 0192.60 135.3

31a108.0

147.4\\j,,,39.7 28.3125.8 38.4145

.2 1H01\14Q

A. iJ. 119.8

181.5 0 135.8

806 GENGLERT

Table 2. Continued.

31b103.6

144 223.2

29

39.149:5 I i 93.6'tL6:5138 0

123

H3CO 14.555.0 180.2 0 138.0

119. 5'

32

25.3 24.9 19 9 12.81 24 641.2

141 5'%

216. 48.3\ 201.133.955.3 148.1

33a

25:9 21.4 12.8

44.0 725

70 4 40.51.0 ••fl

HO1 34.1

59.0 202.91468

33b

24.8 25.7 20.8 12 8

43 .7 \ '120.8l 124.6

140.647.6

H3CCO 42 202.2\ 133.7170.6

21.3 73.658.5 147.0

34a25.725.0 19.5 12.8

45.8

51. 4b4L::.70.2. \41.1HO4 46.7\ 206.1\134.1

58.8 147.2

34b26.224.7 21.1 12.8

43.0I120.8

124.647.

21.3 7341.9\ ° \ 134.0H3CC9 58.6 \1472

o 170.9 201.7

35a

18.1 23.8 21.3 12.8

HO30.881.8 41.0

32.O334.158.1 146.7

35b

18.9 24.3 21.2 12 8

21.1 46.7H3CCO

170.3 83 2.3133.928.4 58.3 147.0

36 117.5 103.3 37a

/14.0

146.1I 125.7 21 7 I 12 1

134.1._I• /120.3 I

1128.5?45 521.1 682r'2.3 132.6 I I 124.3

H3CC2921)23.0 29 0135.5

0I 72.2

45.

37b

147.4 12.1

. .36.3124.1

37.2

23.9 .

25.5132.5?

38a142.722 2 / 12 4

126.01990 8.31 138.0

154.6 1 A28 21 126:9

o.ç7[io.o 141.3

38.7

38b

143 8/ 12 4

52./,I1 5139.6199

00 128.81 25.4155.5

39168.3

14.6 1 12 4148.41 /120 7 126.9

2110' I126.61137.7jl7/.o 140 8

46.3

4025.1 70.329 6 / 12 9

35.4\/1252( 125.2479630

41.8A 67.0 \134.520.3 137.2

41a25.1 70.412 9

472

64.O\t132.2HO •/2o.1 \134.368.8 137.5

41b26.8 71.124.8 / 12 8

35.2$/1J//,,24.7

H3CCO)21.3

170.7 66./K21.4 \134.1138.435.6 64.5

42 153.3

111118.831.3 I /

30.61)//13.4

138.734 2

47:,/>HO 88.467.9/ 4 87.2/ /47.5 28.2

43 154.1

1)! 120.029 129:01)! /12.633.7

47.s2\.k::..6771 I /\127.2HO' 'A\O 87.8

2)31.41)\86.9468.

44

1526 120.5

30.7 26 /12.6

21.3 68.61 I

H39COV'.?O\ 137.4

170.3 o 46.11)127.188.3

87.3

45151 8 119.5

29.7 28.61;//12.61124.22) 6

21.0 68 9 126.4

H39CdvQol37.6?170.3 o 46.61'T7.4 88.8I 1)26.4

46119 7 103.1

31.5' 14 230.0\

34.7$- # -46.7

ZZ.34.067 4

HO'4281"/99 200.825.9

47117 9 103.4

05.4) %28.0

21.2 7OO'\133.0H3CCO"170.4 40.91 19.7 200.7

0 29.3

48a124.3 98.6 1

30.5

28J36.6 /[\135.1

64.91 1189.1 119.1

HO41.51 22.5

137.3

Table 2. Continued.


55

18 8 17 2 12.9

69.3/ 1126.4/36.8 136.01

137.9 136.6

56

203.7 119.9

/ 12.9019.3

24.5"\ .

29.71I147.8r:2084 039.4 134.0

46.5

57

12.7 131.5

1270 11284H

193.1156.11141.1 1

133.5 135.3 12.9

58

198.0 12.7 12.8125 ol 124 0 I

147.61140.4 1 135.7133.2 142.2

59

9.6 131.9 141.2128 0

194.1 u t148.4

137.3 137.413.0

60

9.7 12.8122.4

H ..

94..8t1137.801 /136.6 146.0

134.9

61

168.3 12.8 12.8

60 3 123.1HCHOC..'L.'L

14 4 1138.9 1.01261 144.0

62

27.4 12.8HO 29.8 125 072.7

1;3.0 I 28.8 138.5

27 0138.01131

25.8132.0

1Corresponding assignments may be reversed.

48b 48c 48d 49

17 0 12.71265 124820 41

127.5 20.9 135.5?

126.5

18.6 13.0 12.9

136

26.2 131.2131.21

134.9 137.5137.1

808 GENGLERT

with the groups of Dr. Britton (Liverpool), Prof. Eugster (Zurich), Prof. Jensen (Trond—heim), Prof. Kleinig (Freiburg i.Br.) and Dr. Pfander (Berne), who kindly provided quite anumber of interesting samples.The assignments given in Table 1 were supported by the comparison of a great number ofreference spectra, by decoupling and DID experiments (see example below) and by numeroushomonuclear Overhauser difference experiments, a few examples of which were discussedpreviously (Ref. 1). Similarly, the assignments in Table 2 were based on the application ofshift reagents, measurement of CW—offset and 1H—coupled spectra (see Refs. 5, 6), and inseveral cases on the application of 13C—1H chemical shift correlated 2D experiments(see example below). In both Tables the chemical shifts of the first protons or carbons ofthe side—chains were included because they give additional information on the specific endgroup. Since these data were partly extracted from the spectra of retinoids and relatedcompounds with only 4 or 5 conjugated double bonds the shifts of protons H(lO) and H(ll) aswell as of carbons C(9) to C(ll) had to be corrected in some of these cases by simple addi—tivity rules in order to take into account the influence of strongly anisotropic terminalgroups such as carbonyls or carboxyls.The assignments were given as complete as possible and were carefully checked forconsistency within the entire set of spectra available in our laboratory.I hope that this extended chemical shift compilation will be of practical use in the iden—tification of retinoids and carotenoids.

NEW EXPERIMENTAL TECHNIQUES

Double Indor Difference — Recently, the utility of the interproton Overhauser differencetechnique for the assignment of proton signals of carotenoids and the solution of stereo—chemical problems was strongly emphasized (Ref. 1). In this technique a sufficiently isola—ted signal (singlet or multiplet) is strongly irradiated for a few seconds by a secondradiofrequency. After switching off the irradiation, the observation pulse is applied andthe free induction decay (FID) is accumulated as usual. If a second FID is subtracted withthe pre—irradiation set outside the absorption range the difference spectrum contains onlythose signals which have altered intensities in the two FID's. Signals with increased (po—sitive) intensities are caused by the Overhauser effect and reflect a close spatial proxi—mity of the two protons.The long—known Fourier INDOR difference experiment (Ref. 7) uses a similar experimentalset—up, however, only a short ( 0.5 sec) and very weak (j H2 1 Hz) pre—irradia—tion is applied to one selected component of a multiplet. This causes a change of the popu—lation of these two energy levels which is reflected by a change in intensity of all othertransitions which start or end on one of the two previous levels. In the INDOR differencespectrum only these so—called connected transitions within a coupled spin system appear aspositive or negative signals reflecting an increase or decrease of the excess population ofthe corresponding transitions. The observation of these difference signals allows one tolocate approximately the signals of coupled protons. The main advantage of INDOR is thefact that in contrast to decoupling or Overhauser experiments the perturbation isselectively applied only in a very narrow frequency range of the order of 1 Hz. Thisstrongly extends the applicability of the method in crowded spectra.Kessler and co—workers recently proposed a simple modification of the INDOR difference ex-periment called Double Indor ifference (DID; Ref. 8). A first modification is that thedifference FID is obtained from two on—resonance experiments, namely by irradiating twosuitably chosen components of a given multiplet, e.g. the first and second component of adoublet, or the first and fourth component of a doublet of doublets. The second modifica-

tion is that instead of positive and negative signals, only positive ones are produced byapplying a magnitude calculation to the final difference spectrum. The main advantage ofDID is the fact that in simple, coupled spin systems as frequently observed in carotenoidsthe true multiplet structure of the coupled protons appears in the difference spectrum. Aswill be shown, this method greatly simplifies the assignment of protons in crowded spectraof carotenoids.As an example shown in Fig. 1, DID is used to perform or confirm some assignments in theolefinic part of the 400 MHz spectrum of (5'R)—cryptocapsone (Ref. 6). As usual the assign-ments of the protons near the end groups are easily and unambiguously performed by compari-son with the data of the two symmetrical compounds 13—carotene and (5R,5'R)—capsorubone (seeTables 1 here and in Ref. 1). These comprise H(7), H(8), H(lO), H(l4), H(l4') and H(8).Additional assignments can be obtained by a few selected DID experiments. The first experi-ment, namely pre—irradiation of H(8') near 7.4 ppm as indicated by the two arrows, resultsin the appearance only of the doublet for H(7') near 6.5 ppm. Irradiation of the first andfourth component of the doublet of doublets near 6.7 ppm gives rise to a further doublet ofdoublets and the broadened doublet of H(l4), thus the two former multiplets must beassigned to H(l5) and H(l5') as it is confirmed by the fourth and sixth experiment. In thethird experiment the signals of H(ll), H(l2) and H(lO) are linked. From these and the fol—


19%1O

151 141

11

4'ci20LL 18

18'

2.0 1.5

10 '11'12

48 7'1L_ _________

-----A

—--------------

14 1.99 ppmH20

14' 10.-

H1:+H20,

7:0 6.5 6.lppm

Fig. 1. Top: High and low—field sections of the 400 MHz 1H—NMR of(5'R)—cryptocapsone in CDC13. Below: Sequence of DID experiments withpre—irradiation of 0.5 sec duration at positions marked by arrows;*impurity.

810 G.ENGLERT

lowing experiments It Is been that by a series of DID experiments all but one sub—group ofthree coupled protons can be easily identified. These are the three very strongly coupledprotons H(lO'), H(ll) and H(12') all absorbing near 6.6 ppm as revealed by the integral.It is worthwile mentioning here that DID can also be successfully applied to detect con—nectivities between the protons of the in—chain methyl groups and their olefinic couplingpartners although the corresponding 4—bond coupling is only ca. 1.3 Hz. This is demonstra—ted in the last two experiments of Fig. 1. A difference spectrum obtained by pre—irradia—tion of the methyl signal at 1.992 ppm (H(20)) just before and after the peak maximum(difference ca. 1 Hz) gives rise to the appearance of the signal of H(l4). Similarly, pre—irradiation of the 6—proton singlet near 1.98 ppm reveals connectivities of the correspond—ing methyl protons H(l9) and H(20') with H(lO) and H(l4') leaving the assignment of thesignal of the remaining in—chain methyl protons H(l9') at slightly higher field.In conclusion it is seen that DID is a very selective method for locating the signals of

coupled spin systems in strongly overlapping spectra as those of carotenoids. Assuming nor—mal sample quantities of a few mg a fully automated sequence of experiments can be per—formed in less than one hour (see example of Fig. 1). If longer accumulation times areacceptable, experiments in the pg—range are also possible. We have found that a fewselected DID experiments which can be performed even on older instruments can replace, incertain cases, a much more time—consuming and complicated 2D experiment.

g! — The assignment of 13C—NMR signals to the corresponding carbons in complicatedmolecules relies heavily on the correct determination of their multiplicities reflectingthe number of directly attached protons. This information was previously derived fromproton coupled spectra or from CW—offset decoupled spectra. The former method gives addi—tional information on the proton—carbon couplings J, the latter on relative chemicalshifts of the corresponding protons if the magnitude of the couplings are roughly pre—dictable. However, it is frequently difficult, even at high magnetic fields, to identifythe different multiplets in crowded spectra and hence reliably to derive the individualmultiplicities. In addition, a considerable loss of intensity is implied due to the split—ting of the carbon signals into several components; thus much longer acquisition times areneeded.In the last few years a number of multi—pulse sequences has been proposed for the rapid de—termination of the multiplicities of 13C signals (Refs. 9, 10). For almost two years wehave routinely applied the pulse sequence called DEPT (Distortionless Enhancement byPolarization Transfer; Ref. 10) which has several advantages compared to previous methods(Ref. 11). In this technique which uses two carbon—13 and three proton pulses, polarizationis transferred from protons to carbons, i.e. the excess population between the C levelsand hence the intensity of these transitions is enhanced considerably, thus reducing theacquisition time. The separating delay between the pulses is set equal to l/(2J). Thisone—bond 13C—1H coupling may vary in carotenoids between ca. 130 and 170 Hz. A furtheradvantage of DEPT is that the results do not critically depend on the value used for Jand therefore a typical average value of ca. 140 Hz is suitable. If the usual broad—bandproton decoupling is switched on during the accumulation of the FID, the intensity of thedecoupled 13C singlets depends on the magnitude of the last proton pulse in the pulsesequence, called theta pulse P0, and on the number of protons attached to the carbons.This dependence is caused by the different rotation frequencies and phases of the trans-verse magnetizations of the different types of carbon—l3 due to the couplings to the

attached protons (Ref. 10).A typical example is shown in Fig. 2 wherein trace A represents the normal l3C NMR spec-trum of astaxanthin diacetate, the assignments of which are known (Ref. 1). Also presentedis a set of typical DEPT spectra obtained for 90° (trace B), 45° (trace C) and 135° thetapulses (trace D). The first observation is that in all these DEPT spectra the quaternarycarbon signals (designated by s in trace A) including the solvent signals are eliminated.In addition, in trace B only GI signals are retained, in trace C all prtonated carbons areseen with positive intensities, in trace D CH and CH3 are positive, CH2 negative. Thus,the clear—cut differentiation of all the signals according to their multiplicities ispossible even in very crowded spectra. The normal l3C—NMR was obtained from 128 scans,the DEPT spectra from 256 scans with a relaxation delay of 4 sec. It is clearly seen thatonly a small fraction of this number of scans is actually needed for a signal—to—noiseratio comparable to the normal spectrum.Useful for daily routine applications is the fact that, in general,a single experiment,namely the one in trace D (P0 = 135°), is sufficient to give all needed information, ifthe CH and CH3 signals can be differentiated by consideration of their chemical shifts atlow and high field, respectively.However, if all three DEPT spectra are available one can calculate linear combinations of

the three spectra yielding so—called edited DEPT spectra presenting separated sub—spectrafor all three types of carbon as shown in Fig. 3. Thus, trace C of Fig. 3 with the CH2sub—spectrum was obtained by substracting trace D from C of Fig. 2 with a suitable scalingfactor. Similarly, the CH3 sub—spectrum was calculated from all three spectra of Fig. 2.


0

0

0 ppm

Fig. 2. A: 100 MHz 13C—NMR of astaxanthin diacetate (60 mg in 0.4 mlCDC13, 128 scans, 45° flip angle); S quaternary carbons. B,C,D: DEPTspectra obtained for proton pulses P0 = 90°, 45° and 135°; 256scans, 4 sec relaxation delay, J 130 Hz.

0 ppm

Fig. 3. 100 MHz 13C—NMR and edited DEPT spectra of astaxanthin diacetate.

C P0=451LIi

CH, CH2, CH3I-

D P0=135°

CH2j

150 100 50

A

L J ii

13

1fliJt10

I H10 50 0 ppm

CCH2

DCm3

10 13 50

812 G. ENGLERT

It should be mentioned that the DEPT sequence can be used also for the measurement ofproton—coupled 13C—spectra if the decoupler remains switched off during the accumulationof the FID. Several other pulse sequences have been published including those for thegeneration of l3C sub—spectra of quaternary carbons only (Ref. 12). This might be usefulin cases where quaternary and other carbons have exactly identical chemical shifts.

Two—dimensional NMR (2D) — The basic idea of 2D NMR was first proposed in 1971 by J. Jeener(Ref. 13). In more recent years this promising technique was mainly developed and extendedby the research groups of R. Ernst and K. Wüthrich in Zurich and R. Freeman in Oxford (foran introduction see Ref. 14).In a normal lD Fourier experiment a short rf—pulse is applied to the spin system. Theresulting x,y—magnetization rotating around the magnetic field axis z is detected as a

function of time, the acquisition time t2, and stored in the computer memory. By aFourier transform with respect to t2 the intensities and frequencies of the normal 1Dspectrum are calculated by the computer within a few seconds. The greater the number N ofdata points used in the experiment. the better is the digital resolution of the spectrummeasured in Hz per data point.In a 2D experiment two or more pulses are applied for the preparation of the spin systemduring a first time domain, called evolution time t1, before the free induction decay FIDis acquired during t2. In addition, the timing of the pulses and hence the evolutionperiod t1 is systematically incremented, thus N subsequent FID's are now acquired insteadof only one. In practice N is chosen mostly between ca. 50 and 1000 depending on the typeof experiment and the desired resolution. It is clearly evident that if the number of ex—

perirnents and the digital resolution is not kept reasonably small the storage requirementN * N of the data matrix as well as the acquisition and processing times may be excessive.The processing consists of two consecutive steps. First, N subsequent Fourier transforms ofall the rows of the data matrix are performed with respect to t2 yielding a set of Nspectra. Since t was systematically incremented the amplitudes and phases of a specificsignal in subsequent rows can also vary as a function of t1, i.e. the signal along aspecific column may be modulated e.g. by spin couplings, either directly (J—resolved 2D) orindirectly due to transfer of polarization between coupled spins (çrrelated 2Dpectroscop, COSY). The frequencies f1 involved in the amplitude variation of the dif—ferent columns are calculated by subsequent Fourier transforms with respect to ti. Ingeneral, a magnitude calculation of the data is performed to obtain only positive intensi—ties. The result, namely the intensities as a function of the two frequencies f1 andf2, is called a 2D spectrum. It can be presented as an impressive stacked plot as shownin Fig. 4. Here, a section of a proton—proton chemical shift correlated 2D spectrum isshown. The two axes cover the range between S = 6.1 and 6.8 ppm and thus contain the che-mical shift range of all olefinic protons except one (H(lO')). The normal spectrum is con-centrated along the diagonal running from top—right to bottom—left. In this homonuclearCOSY spectrum, which is symmetrical with respect to the diagonal, cross—peaks at coordi-nates 6A in f1 and 6B in f2 (as well as 6B In f1 and A in 2)reflect the fact that the two protons at chemical shifts 6A and 6B are spin—coupledto each other. Stronger couplings generally give more intense cross—peaks. This spectrumwill be discussed later in more detail.A stacked plot is not normally suited for extracting the information contained in a 2Dspectrum, but it gives a good picture of the different intensities and the achieved signal—to—noise ratio. The information is much better revealed by so—called contour plots wherethe intensities are shown as closed contour levels of equal intensity, starting from asuitably chosen minimum level above the noise and incrementing normally by a factor of 2from level to level.

J—resolved 2D — In this relatively simple 2D technique only the chemical shift informationis retained in the horizontal f2—axis, the coupling information solely along the perpen-dicular f1—axis. It appears as if the spin multiplets were simply rotated by 900 into asecond frequency axis. The technique can be applied to homonuclear (mostly proton—proton)and heteronuclear (mostly carbon—proton) systems. Two examples of the former were alreadydiscussed previously (Ref. 1). It is felt that the homonuclear variant will not find manyuseful applications in the field of carotenoids, because severe difficulties are en-

countered in strongly coupled spin systems.More useful in specific cases is the heteronuclear J—resolved 2D experiment, an example ofwhich is shown in Fig. 5. Here, the contour plot of the 13C—1H J—resolved 2D of ethyl8'—apo--13—caroten—8'—oate is presented on bottom. Since the width of the horizontal

f2—axis is more than 18 kHz according to the wide range of 13C chemical shifts, thesingle contour lines representing the different intensities levels are, unfortunately, notresolved in this picture. The perpendicular f1—axis containing the coupling informationcovers a range of 290 Hz. Although 128 experiments from 128 different t1 increments wereperformed the digital resolution in f1 (after zero—filling to 256 points) was only 1.1 Hzper data point. This is more than sufficient clearly to distinguish the different singlets,doublets, triplets and quartets in the contour plot but is insufficient for the detection

of very small long—range couplings.

NMR ofcarotenoids—new experimental techniques 813

Fig. 4. Stacked plot (section between 6.1 and 6.8 ppm) of the sym—metrized proton—proton chemical shift correlated 2D (400 MHz, COSY—90,N—type) of ethyl 8—apo—13—caroten—8'—oate (22 mg in 0.4 ml CDC13).Experimental data: Spectral width W 601.7 x 601.7 Hz, 2K x 470, zero—filled to 1K, 470 t1 increments, 0.6 and 0.6 Hz per data point, po-larization transfer delay 0.1 sec. relaxation delay 2.3 sec, 16 scansand 1 dummy scan per increment, sine—bell windows, magnitude spectrum.9.5 h acquisition time, 1024 K file size. A corresponding contour plotis shown in Fig. 7.

6.1

The actual measurement of the resolved couplings and the visual presentation of the dif-ferent multiplets can be achieved in the following way. First, a horizontal projection ofthe 2D spectrum is calculated (see top of Fig. 5). It looks very similar to the usualbroad—band decoupled 13C spectrum but its digital resolution is. in general. necessarilysmaller (here 9 instead of ca. 1 Hz per point). Since the maxima of the horizontal projec-tion represent the chemical shift of the different carbons, the corresponding multipletscan be plotted as so—called J—cross—sections as normal intensity plots along the f1—axis.All J—cross—sections (except COO) which were resolved are shown in Fig. 6 together with theassignments, the intensity scaling factor, and the chemical shifts. In carotenoids the1J—values of olefinic carbons are in a rather narrow range between ca. 149 and 155 Hzand, therefore, they are not very useful for the purpose of assignment. However, the numberof long—range couplings or the width of the unresolved components of a multiplet can berather characteristic. Thus C(7). c( 11) and the corresponding c( 11 ) usually appear assharp doublets with high intensities. They can already be distinguished from other carbonsby inspection of the horizontal projection as seen in Fig. 5.A further point should be noted here. If a singlet and a doublet appear at (almost) thesame chemical shift they can be completely separated in a J—resolved spectrum as is seenfrom two examples in Fig. 6. The intensities and the magnitude of the measured J—coup—ling allows a safe differentiation from a single triplet signal.It is evident from this example that any meaningful application of this special 2D methodwill be essentially limited to those cases where the number and magnitude of the carbon—proton couplings are of importance for the solution of a structural problem and the normal

1H—coupled 13C—spectrum is too crowded for a clear—cut differentiation.

20

18

6.7

H

I6.5 6.3

6.7 ppm

H

6.1 ppm

G. ENGLERT

f.i

50 0

Fig. 6. J—cross—sections of the J—resolved 2D spectrum of Fig. 5. Theassignments are given together with the chemical shifts and the inten-sity scaling factors. The COO—signal is not shown.

150 100 50

814

0 ppm

f

—100 $

I liii III IIII t ii

0 IIIt I III I

I 1111 111111I i ii

I•

100 I

Hz

613c150 100 ppm

Fig. 5. Ethyl 8'—apo—13—caroten—8'—oate (100 mg in 0.5 ml CDC13):Contour plot and horizontal projection of the 13C-1H J-resolved 2Dspectrum. Experimental conditions: V 18518 x 290 Hz, 4K x 128 datapoints, zero—filled to 256 points, 128 t1—increments, digitalresolution 9 and 1.1 Hz, 3.5 sec. relaxation delay, sine—bell windows,magnitude spectrum, file size 512 K, ca. 14 h acquisition time.

C ,C 12.78 ppm 14.35 19.31 ppm

C2Q,C,CH3

A A

C7

150.4H5 129.33

C10130.76

___ ___J/2 •_•__•_-. .

100 0 -100Hz

C18 21.71 C16 28.96 C4 33.13C14 131.92 C3 135.26 C14, 135.70

x2 _. —,--- .,

C1 34.28 C2 39.73 OCH2 60.34C9 136.35 C12 136.98 C8 137.68

...T

:9JJL_


Homonuclear chemical shift correlated 2D spectroscopy (COSY) — In this experiment a spincoupling between two protons is revealed by the existence of cross—peaks in the 2D plot(Ref. 14). Since it is actually the homonuclear spin coupling which causes the cross—peaksthe method is also called homoscalar correlated 2D. It is often possible to derive the full

sequence of coupled protons even in a complicated compound, thus providing importantinformation on its molecular structure. As will be shown the coupling pathway or connec—tivity of protons can even in such cases be detected where the couplings are very small.The previous 400 MHz COSY spectrum of Fig. 4 from ethyl 8'—apo—13—caroten—8'—oate is npresented in Fig. 7 as a contour plot which is much more appropriate for the deduction ofthe connectivities. Since the olefinic part is highly crowded a very good digital resolu—tion of only 0.6 Hz per point was used. The normal 1D spectrum and the assignments aregiven in the upper 1D spectrum. The 2D contour plot is symmetrical with respect to the dia—gonal on which the normal spectrum is seen. Those peaks which are very close to the dia—gonal are caused by transfer of magnetization between the transitions of the same proton.Thus, the doublets appear as 2x2, all doublets of doublets as 4x4 signals. More distantfrom the diagonal are found the cross—peaks indicating that the two corresponding protonsare coupled to each other. Thus, the two off—diagonal peaks at coordinates 6A'6Band (6B'6A) prove the presence of a J—coupling between proton A at (6A'6A)and B at (6B'6B)• One of the advantages of a COSY in the case of highly crowdedsignals near the diagonal is the fact that the multiplet structures can be derived byinspection of the cross—peaks which are often sufficiently separated.If a good signal—to—noise ratio has been achieved in a COSY spectrum by application of alonger accumulation time it is often possible to detect very weak cross—peaks due to long—range couplings. As will be seen later in this COSY spectrum a long—range coupling of theorder of 0.1 Hz or less extending over 7 intervening bonds can be detected although suchsmall couplings can not normally be seen or resolved in the lD spectrum. In Fig. 7 alllong—range connectivities were designated in the upper—left triangle, the 3—bond connecti—vities in the lower—right triangle of the symmetrical 2D spectrum.The procedure for revealing the sequences of coupled protons can start e.g. from the isola—ted signal of H(lO') at lowest field. Horizontal and vertical lines lead to the strongcross—peaks with H(ll'), the diagonal signal of which is now found at the intersections ofthe horizontal and vertical lines with the diagonal. A further much weaker cross—peak dueto 4J10. ,12' identifies the signal of H(l2'). It can also be located from the muchstronger, direct cross—peaks between H(ll') and H(l2') via 3J11.12.. Further connecti—vities are visible between H(l5), H(l5') and H(l4), between H(ll), H(12) and H(lO), betweenH(15) and H(14') etc. Some further long—range connectivities are marked in the upper—lefttriangle of Fig. 7. Interestingly, a cross—peak H(15)/H(ll') is clearly visible,corresponding to a 7—bond coupling between these two protons (see also Fig. 8). From the 2Dspectrum of Fig. 7 the whole sequence of protons between H(lO) and H(lO') is deduced. Theconnectivity between H(7) and H(8) can also be found in an expanded plot of the right uppercorner (not shown) although their chemical shifts are very close to each other.A very convenient way of visualizing the different connectivities is to plot selected rows(or columns) of the 2D data. Three selected rows chosen at 6.59 ppm through H(l2'), at 6.74ppm through H(l5') and 7.30 ppm through H(lO') (see arrows in Fig. 7) are presented asexamples in Fig. 8. The form and intensities of the diagonal peak and its cross—peaks areclearly seen as well as the above—mentioned long—range connectivity.In a completely analogous way the connectivities between the protons of in—chain methylgroups and olefinic protons are detectable if the experiment is run with a spectral widthcovering the whole spectrum. However, in this case the digital resolution has to be reducedin order to avoid an excessive size of the data matrix. Despite several drawbacks of this

technique, namely long accumulation and processing times and large data storage require-ments, it is strongly recommended for the solution of complex structural problems ifsimpler methods such as DID cannot provide all necessary information.

Heteronuclear chemical shift correlated 2D — This technique (Ref. 15) which is mostlyapplied to carbon—l3 and protons is even more useful in practice since it provides connec—tivities between 13C and 1H chemical shifts. The two portions of a heteronuclear 2Dfrom (5'R)—cryptocapsone are presented in Fig. 9 and 10, showing separately the olefinic—olefinic and the aliphatic—aliphatic sections of the same experiment. The proton spectrumruns along the vertical f1—axis, the carbon spectrum along the horizontal f2—axis. Forconvenience, both lD spectra of the two nuclei and the horizontal projection of the 2Dspectrum are also presented. Since the delay for polarization transfer between the twonuclei was chosen to enhance connectivities by one—bond couplings of 140 Hz, cross—peaksare seen only between carbons and their directly attached protons. Correspondingly, in the

horizontal projections only signals of protonated carbons are detectable. Thus, startingfrom a known proton signal in f1 one obtains in a simple and mostly unambiguous way theassignment of the corresponding carbon in f2. In a strongly crowded proton spectrum itfrequently suffices to know the relative order of the chemical shifts to assign the carbonsignals. Conversely, if the carbon assignment is known the chemical shift of the coupledproton is deduced.

816 G. ENGLERT

12' 12

11 101cJ111

10' 15 ' 14'rLl

rrL1r11ñ 14I HJJ L

7,4 65 ppm

Fig. 7. Top: Olefinic range of the resolution enhanced 400 MHz

1H—NMR spectrum of ethyl 8'—apo—13-caroten—8'—oate. Bottom: Contourplot of the symmetrized COSY—90 spectrum. Experimental conditions see

Fig. 4; * folded signal. Arrows indicate selected rows shown in Fig. 8.


20

II

12' 12iif'i

71

10

I I I I J I I

70 6.5 6.0

10'14 l2 ,11'

6.59 ppm________ 11'

3J I

1515,15

19L 3Jiol ,11I

7.30_..!JJL10 ,12'

7.0 6.5 6.0 ppm

Fig. 8. Top: Olefinic section of the 400 MHz 1H—NMR spectrum of ethyl8'—apo—B—caroten—8'—oate. Bottom: Selective horizontal rows of the COSYspectrum of Fig. 7 plotted at 6.59, 6.74 and 7.30 ppm (see arrows in

Fig. 7) indicating all cross—peaks of H12', H15 and H10',respectively.

In the case of ambiguities important aid is given by additional plots of horizontal rows atselected proton chemical shifts, revealing the position of the connected carbon signal, orconversely, the plot of selected columns yielding the multiplet structure of the signal ofthe proton coupled to a specific carbon.A very useful supplement to the one-bond coupling experiment is obtained by a subsequentexperiment tuned to connectivities mediated by two— and three—bond couplings which arefrequently between 5 and 10 Hz. This gives further relevant information since connectivi—ties between more distant nuclei, including the quaternary carbons, are involved. As anexample, the aliphatic—aliphatic portion of the 13C—1H correlated 2D of the same com-pound is shown in Fig. 11. The delay for polarization transfer was chosen to enhance cross—peaks caused by long—range couplings of ca. 8 Hz. However, some of the cross—peaks causedby the much larger one—bond couplings can be still detectable.For purposes of interpretation it is important to note that only two— or three—bond carbon—

proton couplings are sufficiently large to cause long—range cross—peaks. Applying this ruleand the results from Fig. 10, all the cross—peaks of Fig. 11 can be readily assigned. It isevident that not only the assignments of proton and carbon signals can be deduced or con-firmed, but also important information on the molecular structure is revealed by the con—nectivities caused by the two— and three—bond couplings. Thus, as an example, in thissection of the 2D plot, protons H(l6) and H(l7) of the 13—ring have cross—peaks with C(2),C(l) and C(l7), C(l6); protons H(l8) only with C(4) and C(l8). In the 5—membered ring, onthe other hand, protons H(l6') and H(l7') have cross—peaks with C(5'), c(2') and C(l'),protons H(l8') with C(S), C(4') and C(l').Further relevant information of this type can be deduced from the other sections of theheterocorrelated 2D, namely the aliphatic—olefinic and olefinic—olefinic sections which arenot shown here. This includes e.g. cross—peaks between H(l9) and C(8), C(9) and C(lO) aswell as between H(20) and C(l2), C(l3) and C(l4).

12

Fig. 9. A: 100 MHz 13C—NMR spectrum (olefinic section) of (5'R)—cryp—tocapsone (45 mg in 0.5 ml CDC13); *impurity. B: Horizontal projec-tion of the 13C—1H correlated 2D spectrum. C: 400 MHz 1H—NMRspectrum measured with the decoupler coil. Bottom: 13C—1H correla-ted 2D spectrum (olefinic—olefinic section); Experimental conditions:W 21740 x 3334 Hz, 4K x 256, zero—filling to 512, 10.6 and 6.5 Hz perdata point, 2.5 sec relaxation delay, polarization transfer delay 3.6msec (J 140 Hz), 48 scans, 1 dummy scan, sine—bell windows,magnitude spectrum, 1024 K file size, 9 h acquisition.

In conclusion it is found that the heteronuclear chemical shift correlated 2D is extremelyuseful for the assignment of signals and the deduction of structural information on thenuclear framework and hence for the elucidation of molecular structures. In practice, wenormally acquire both the one—bond and the long—range type of spectra in two subsequent ex-periments normally during an overnight run from about 20 to 50 mg of sample.

INADEQUATE—2D — This technique, also called Carbon—Carbon Connectivity (CCC), is a 2Dvariant of the former iD—INADEQUATE experiment which was used to measure 13C—13Ccouplings constants in natural abundance (Ref. 16).In normal 13C—NMR the spectrum from molecules containing 13C at ca. 1 % naturalabundance is acquired. The probability of having two adjacent 13C nuclei in a molecule isabout 100 times less. Since two neighbouring carbons are spin—coupled the resulting twodoublets have roughly 200 times smaller intensities. Despite this extreme loss of intensity

818 G. ENGLERT

19 20

t16

18

I

150

B

120 ppm

10 io ppm

C

14

10' 11' 12

6

6.5

'7

.11

ii

7.5

150 140 130 120 ppm


Fig. 10. Aliphatic—aliphatic section, continued from Fig. 9.

it is possible in favourable cases to acquire an INADEQUATE spectrum with a reasonable

signal—to—noise if very long acquisition times can be accepted and sufficiently largesample quantities of several 100 mg are available and soluble in ca. 2 ml of solvent. Inthe 1D version, all the pairs of doublets are superimposed along the only frequency axisf2. Furthermore, the much more intense singlet signals of molecules with only one

carbon—l3 are partly superimposed because they are mostly incompletely suppressed by appli-cation of complicated phase cycles.In the 2D—INADEQUATE the different AB or AX spectra are separated along the secondfrequency axis f1 according to their so—called double quantum frequency. This is the sumof their frequency distances from the transmitter frequency which is normally placed in themiddle of the f2—axis.

An example, using a quite useful experimental variant called refocused 2D—INADEQUATE (Ref.17), is shown in Fig. 12. It was obtained from 640 mg of astaxanthin diacetate with anacquisition time of 64 hours. This long acquisition time is mainly caused by the fact thata relaxation delay of (at least) 2.5 sec had to be interleaved after each scan because the13c spin lattice relaxation time T1 of the unprotonated carbons of this compound canexceed 3 sec (Ref. 1). An average value for J of 55 Hz was adopted for this measurementalthough the actual coupling constants vary between roughly 35 and 65 Hz. The pairs ofdoublets are symmetrically displaced with respect to the diagonal. This simplifies theidentification of the different connectivities and, in addition, artifacts having no sym-metrical counter—parts can be removed by symmetrization. Unfortunately, the doublets arevery poorly resolved in this picture since the total range of f2 covers 20.8 kHz. On thediagonal some stronger residual artifacts are seen which are caused by imperfectly

suppressed singlets of slowly relaxing carbon—l3 nuclei.The connectivities of the carbon—carbon skeleton are revealed by the procedure shown inFig. 12. Thus, starting e.g. from the quaternary carbon C(l), its neighbours C(l7), C(l6),C(2) and C(6) are clearly revealed by the existence of the corresponding pairs of doublets.Similarly, C(2) is connected to C(3), C(3) to C(4), C(4) to C(S), C(S) to C(6), and soforth. The CCC is interrupted at C(9)/C(lO), which have almost identical chemical shifts,

S

A I II jOppm

16 19

BILJJIILLJ

19 20

C

0

1

2

3

16

19

ppm

820 G. ENGLERT

1920

C

16 17

Ai [ sj

B

50

Fig. 11. (5'R)—cryptocapsone, aliphatic sections of the 100 MHzNMR (A), the horizontal projection (B) of the 13C—1H correlated 2D(bottom—right). C: aliphatic section of the 400 MHz 1H—NMR. Polariza-tion transfer delay chosen to enhance long—range couplings 1J

8 Hz. Other experimental conditions as in Figs. 9 and 10.

and at the intervening oxygen to the substituent at C(3). The position of the acetyl groupis, however, evidenced by consideration of the chemical shift. This example clearly demon-strates the great power of this 2D method to reveal in a fully unambiguous way not only allthe assignments but also the full sequence of carbons in the molecular structure includingring size, position of double bonds and of methyl groups. It should be mentioned that expe-riments with reduced sample quantities or reduced accumulation time of the order of 24 hcan also be quite useful. However, in this case only connectivities between protonated car-bons with short relaxation times are likely to be detected as e.g. the olefinic carbonswith T1s of 0.3 sec or less.

CONCLUSION

From these selected examples it is clearly evident that NMR spectroscopy has again madesome steps forward increasing its power and utility for the solution of structural problemsin the field of carotenoids. Although some of the recent 2D techniques have becomeincreasingly complex and sophisticated, there is no doubt that they will be more and moreapplied even on a routine basis. All the experiments described here were performed in afully automated manner under computer control on a Bruker WM 400 FT NMR spectrometer usingstandard software.

Acknowledgement —I am indebted to several colleagues who provided the compoundsused in this study. I should like to specifically mention Drs. K. Bernhard, H.Mayer, A. Rüttimann, E. Widmer, R. Zell and Mrs. K. Schiedt. Thanks are also dueto Mr. W. Grunauer for his skilful technical assistance.

REFERENCES

1. G. Englert, in Carotenoid Chemistry & Biochemistry, G. Britton and T.W. Goodwin,Editors, Pergamon Press, 107—134 (1982).

2. E. Märki—Fischer, P.—A. Bütikofer, R. Buchecker, and C.H. Eugster, Helv. Chim. Acta, 1175—1182 (1983).

ill 11

16

20 ppm

19 20'

50 40 30 20 ppm


16 17 19 2011 15

HCCO 180

Fig. 12. Top: 100 MHz 13C—NMR spectrum of astaxanthin diacetate.Bottom: Contour plot of the symmetrized, refocused 2D—INADEQUATE. Thecarbon connectivities are indicated starting from C(l). Experimentalconditions: 640 mg in 1.8 ml CDC13, W 20.8 x 10.4 kHz, 4K x 256,zero—filling to 2K, 10.2 and 5.1 Hz digital resolution, relaxationdelay 2.5 sec, echo—delay 4.5 msec (CC = 55 Hz), '3C pulses 90°(14 psec), 180° and 120°, Lorentz—to—Gauss filtering, LB = —4,GB = 0.1, LB1 = —4, GB1 = 0.1, 320 scans per 4—increment, 4096 Kdata file, acquisition time 64 h.

References (contd.)

3. E. Märki—Fischer, R. Buchecker, C.H. Eugster, G. Englert, K. Noack, and M. Vecchi,Helv. Chim. Acta 65, 2198—2212 (1982).

4. W. Eschenmoser, P. Uebelhart, and C.H. Eugster, Helv. Chim. Acta 66, 82—91 (1983).5. G. Englert, Helv. Chim. Acta 58, 2367—2390 (1975); K. Bernhard, G. Englert, H. Mayer,

R.K. Muller, A. Rüttimann, M. Vecchi, E. Widmer, and R. Zell, Helv. Chim. Acta 64,2469—2484 (1981).

6. A. Rüttimann, G. Englert, H. Mayer, G.P. Moss, and B.C.L. Weedon, Helv. Chim. Acta 66,1939—1960 (1983).

7. K.G.R. Pachler and P.L. Wessels, J. Chem. Soc., Chem. Commun., 1038—1039 (1974).8. H. Kessler, G. Krack, and G. Zimmermann, J. Maqn. Res. 44, 208—211 (1981).9. G.A. Morris and R. Freeman, J. Am. Chem. Soc. 101, 760—762 (1979).10. D.M. Doddrell, D.T. Pegg, and M.R. Bendall, J. Maqn. Res. 48, 323—327 (1982).11. 0.W. SØrensen and R.R. Ernst, J. Magn. Res. 51, 477—489 (1983).12. M.R. Bendall, D.T. Pegg, D.M. Doddrell, S.R. Johns, and R.I. Willing, J. Chem. Soc.,

Chem. Commun., 1138—1140 (1982).13. J. Jeener, Ampere International Summer School, Basko Polje, Yugoslavia (1971).14. A. Bax, Two—Dimensional Nuclear Magnetic Resonance in Liquids, Delft University Press,

D. Reidel Publishing Company, Boston, London (1982); R. Benn and H. Gunther, Angew.Chem. 95, 381—411 (1983).

15. A.A. Maudsley and R.R. Ernst, Chem. Phys. Lett. 50, 368—372 (1977).16. A. Bax, R. Freeman, and S.P. Kempsell, J. Am. Chem. Soc. 102, 4849—4851 (1980).17. D.L. Turner, J. Maqn. Res. 49, 175—178 (1982).

13C 100MHz

190 150 100 5b Ôppm

I10416Hz

nmr of carotenoids—new experimental techniques

Documents