Prenylflavonoid variation in Humulus lupulus: Distribution and taxonomic significance of xanthogalenol and 4'-O-methylxanthohumol

Article abstract of DOI:10.1016/S0031-9422(00)00005-4

The resins produced by either lupulin or leaf glands of over 120 plants of Humulus lupulus and one plant of H. japonicus (Cannabinaceae) were analyzed for the presence of prenylated flavonoids. The H. lupulus taxa investigated were H. lupulus var. lupulus

Full text of DOI:10.1016/S0031-9422(00)00005-4

Phytochemistry 53 (2000) 759±775
www.elsevier.com/locate/phytochem
Prenyl¯avonoid variation in Humulus lupulus: distribution and
taxonomic signi®cance of xanthogalenol and 4'-O-
methylxanthohumol
Jan F. Stevensa, 1, Alan W. Taylora, Gail B. Nickersonb, Monika Ivancicc,
John Henningb, Alfred Haunoldb, Max L. Deinzera,*
aDepartment of Chemistry, Oregon State University, 153 Gilbert Hall, Corvallis, OR 97331, USA
bDepartment of Crop Science, Crop Science Building 109, Corvallis, OR 97331, USA
cDepartment of Biochemistry and Biophysics, ALS Building 2011, Corvallis OR 97331, USA
Received 3 September 1999; received in revised form 29 November 1999
Abstract
The resins produced by either lupulin or leaf glands of over 120 plants of Humulus lupulus and one plant of H. japonicus
(Cannabinaceae) were analyzed for the presence of prenylated ¯avonoids. The H. lupulus taxa investigated were H. lupulus var.
lupulus from Europe, H. lupulus var. cordifolius from Japan, and H. lupulus from North America. Fifty-two of the plants
examined were cultivars of European, American, and Japanese origin. Twenty-two ¯avonoids were detected in the glandular
exudates of H. lupulus by HPLC-MS±MS. Xanthohumol (3'-prenyl-6'-O-methylchalconaringenin) was the principal
prenyl¯avonoid in all H. lupulus plants and was accompanied by 11 structurally similar chalcones. Ten ¯avonoids were identi®ed
as the ¯avanone isomers of these chalcones. Three other prenylchalcones were isolated from H. lupulus cv. `Galena', one of
which was identi®ed as 3'-prenyl-4'-O-methylchalconaringenin (named `xanthogalenol'). The distribution of three 4'-O-
methylchalcones, i.e. xanthogalenol, 4'-O-methylxanthohumol, and 4',6'-di-O-methylchalconaringenin, was found to be limited
to wild American plants from the Missouri±Mississippi river basin, H. lupulus var. cordifolius, and most of their descendents.
These 4'-O-methylchalcones were absent from cultivars of European origin, and from wild hops from Europe and southwestern
USA. The ¯avonoid dichotomy (presence versus absence of 4'-O-methylchalcones) indicates that there are at least two
evolutionary lineages within H. lupulus (European and Japanese±American), which is in agreement with morphological,
molecular, and phytogeographical evidence. Leaf glands of H. japonicus from eastern Asia did not produce the H. lupulus
prenyl¯avonoids. 7 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Humulus lupulus; H. japonicus; Cannabinaceae; Hops; Prenylated ¯avonoids; Xanthohumol; Xanthogalenol; Mass spectrometry; Che-
motaxonomy
1. Introduction
female in¯orescences (commonly referred to as `hop
cones' or `hops'), which are used in the brewing indus-
try to add bitterness and aroma to beer. The female
¯ower clusters are partly covered with lupulin glands
while male ¯owers have only a few glands in the crease
of their anthers and on their sepals. The resin secreted
by these glands contains bitter acids, essential oils and
prenylated ¯avonoids. In a limited prenyl¯avonoid sur-
vey of hops, Stevens and co-workers (Stevens, Ivancic,
Hsu & Deinzer, 1997) found no signi®cant di€erence
between nine female hop cultivars of European origin.
The hop plant (Humulus lupulus L., Cannabinaceae)
is cultivated in the temperate zones of the world for its
* Corresponding author. Tel.: +541-737-1773; fax: +541-737-
0497.
E-mail address: Max.Deinzer@orst.edu (M.L. Deinzer).
1 Present address: Department of Organic and Inorganic Chem-
istry, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam,
The Netherlands.
0031-9422/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved.
PII: S0031-9422(00)00005-4

760
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
Xanthohumol (1, see Fig. 1 and Table 1 for structures)
was the principal ¯avonoid in all samples examined,
while eight other prenylated ¯avonoids were found
only in small quantities. Following the isolation and
identi®cation of the novel prenylated chalcone, xantho-
galenol (2), and the known chalcones, 4'-O-methyl-
xanthohumol (3) and ¯avokawin (4), from the
American cultivar Galena, we conducted a more com-
prehensive ¯avonoid survey with three wild taxa and
52 cultivars of European, American and Japanese ori-
gin. The distribution of 1±4, as well as 18 other ¯avo-
noids (5±22) is compared with the hybridization
patterns and morphological and molecular (rDNA)
variation in H. lupulus.
Although some authors have argued against the
splitting of H. lupulus into varieties, subspecies, or
even separate species (Burgess, 1964; Neve, 1991; and
refs. cited), we decided to follow largely the infraspeci-
Fig. 1. Prenyl¯avonoids from Humulus lupulus. For compound names, see Table 1.

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
761
®c classi®cation according to Small (1978, 1980),
which is primarily based on geographic distribution
and leaf morphology. Wild hops of Europe are rela-
tively uniform and therefore grouped in one taxon, H.
lupulus var. lupulus. Small (1978) recognized three var-
ieties endemic to North America: H. lupulus var. lupu-
loides E. Small (central and eastern N. America), H.
lupulus var. pubescens E. Small (midwestern USA), and
H. lupulus var. neomexicanus Nelson and Cockerell
(western N. America). Due to sympatry and morpho-
logical intergradation of these groups in most parts of
central and midwestern USA, and for reasons dis-
cussed below, they are here referred to as wild Ameri-
can hops. The indigenous wild hops of Japan and
perhaps eastern mainland Asia are known as H. lupu-
lus var. cordifolius (Miquel) Maximowicz.
with authentic ¯avonoids obtained in the present study
and during previous work (Stevens et al., 1997; Ste-
vens, Taylor & Deinzer, 1999a; Stevens, Taylor, Claw-
son & Deinzer, 1999b). Of the ®ve newly isolated
chalcones from H. lupulus cv. `Galena', compounds 2,
11, and 12 are new natural products. Mass fragmenta-
tion of prenylated chalcones and ¯avanones by col-
lision-induced dissociation (CID; also referred to as
collision-activated decomposition, CAD) has been dis-
cussed in previous work (Stevens et al., 1997). As
pointed out in that report, chalcones cannot be distin-
guished from their isomeric ¯avanones by mass spec-
trometry due to thermal isomerization in the ion
source. Distinction between both ¯avonoid types was
made by UV spectroscopy (Markham, 1982): chal-
cones have UV maxima near 370 nm while those of
the ¯avanones are near 290 nm (Table 1).
2. Results
2.1. Identi®cation of ¯avonoids
Twenty-two ¯avonoids (Table 1) were detected in
methanolic extracts of hops by HPLC-MS comparison
The acetone extract of `Galena' hops was fractio-
nated on Sephadex LH-20 with methanol as the eluent.
Table 1
Chromatographic and spectral data of prenyl¯avonoids from Humulus lupulus
No. Trivial name
Rt
UV lmax
APCI-MS, m/z
(min)a (nm) in MeOH ꢀlog e) (rel. int.)b
[MH]+ Daughter ions of [MH]+
Chalcones
Xanthohumol
Xanthogalenol
1
2
3
4
5
6
7
8
9
15.1
15.9
18.6
14.0
13.2
16.9
17.0
19.2
12.7
368 (4.56)
365 (4.49)
367 (4.55)
367 (4.51)
366
355
355
369
301
341
409
409
423
371
179 (100)
179 (100)
4'-O-Methylxanthohumol
4',6'-Di-O-methylchalconaringenin
Desmethylxanthohumol
3'-Geranylchalconaringenin
3',5'-Diprenylchalconaringenin
5'-Prenylxanthohumol
Xanthohumol B
193 (100), 163 (13)
181 (100), 166 (11), 147 (10)
183 (58), 165 (100)
367, 310 sh
372
371
285 (7), 183 (22), 165 (100)
297 (14), 233 (49), 195 (47), 177 (100), 121 (15)
247 (19), 191 (100)
370 (4.35), 275sh
(3.92), 243 (4.09)
371, 298sh, 285
369 (4.42)
369
251 (45), 233 (100), 191 (24), 179 (93), 147 (30), 109 (10)
10
11
12
Xanthohumol C
Xanthohumol D
Xanthohumol E
18.0
12.7
17.9
353
371
407
233 (100), 191 (19), 109 (10)
251 (6), 233 (89), 191 (26), 179 (100), 109 (13)
231 (100), 189 (22), 135 (7)
Flavanones
Isoxanthohumol
13
14
15
16
17
18
19
20
21
22
9.6
17.3
16.3
287, 325sh
291, 339sh
291, 339sh
287, 322sh
283 (4.27)
291, 334sh
291, 334sh
295, 347sh
294, 336sh
294, 338sh
355
355
355
369
301
341
341
409
409
409
179 (100)
7-O-Methyl-6-Prenylnaringenin
7-O-Methyl-8-Prenylnaringenin
179 (100), 149 (12), 121 (6)
179 (100), 149 (10), 121 (5)
193 (100), 163 (21)
5,7-Di-O-methyl-8-Prenylnaringenin 13.2
5,7-Di-O-methylnaringenin
6-Prenylnaringenin
8-Prenylnaringenin
8.2
14.1
12.4
17.4
17.9
16.0
181 (100), 166 (8), 147 (12), 119 (8)
183 (18), 165 (100)
183 (21), 165 (100)
6,8-Diprenylnaringenin
6-Geranylnaringenin
8-Geranylnaringenin
297 (19), 233 (44), 177 (100), 121 (18)
285 (20), 183 (17), 165 (100)
285 (10), 183 (19), 165 (100)
a HPLC separations were achieved on a 5 mm LiChrosphere (4 Â 250 mm) column using a linear solvent gradient starting from 40% MeCN
1
.
(B) in 1% aq. HCOOH (A) to 100% B over 15 min, then 100% B for 7 min, at 0.8 ml min
b Atmospheric pressure chemical ionization-mass spectrometry. Daughter ions were obtained in the tandem MS mode with the collision energy
set at 30 V (Table 1), 15 V (Section 2), or 11 V (Fig. 2).

762
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
Table 2
1H-NMR spectroscopic data for compounds 2, 3, 9, 11, and 12 ‰d ppm, mult. (J in Hz)]
H
2
3
9
11
12
a
b
2,6
3,5
5'
10
7.98 d (15.5)
7.69 d (15.5)
7.54 d (8.4)
6.84 d (8.4)
6.11 s
7.75 d (15.5)
7.69 d (15.5)
7.59 d (8.3)
6.84 d (8.3)
6.27 s
7.77 d (15.5)
7.70 d (15.5)
7.58 d (8.7)
6.84 d (8.7)
6.01 s
7.75 d (15.5)
7.66 d (15.5)
7.57 d (8.6)
7.90 d (15.5)
7.68 d (15.5)
7.54 d (8.6)
6.86 d (8.6)
6.84 d (8.6)
6.06 s
3.13 d (7.0)
3.16 d (7.0)
2.75 dd (6.7, 13.4)
2.64 dd (6.9, 13.4)
4.22 t (6.7)
3.21 d (7.0)
20
40
5.09 t (7.0)
1.69 s
5.09 t (7.0)
1.70 s
5.12 t (7.0)
1.71 s
2.71 dd (5.2, 16.5)
2.38 dd (7.0, 16.5)
3.66 dd (5.3, 6.9)
1.29 s, 1.22 s
4.64 s, 4.60 s
50
60-Mes
41
1.60 s
1.60 s
1.72 s
1.62 s
6.66 d (9.9)
5.60 d (9.9)
1.49 s
51
61-Mes
4-OH
2'-OH
4'-OH
6'-OH
4'-OMe
6'-OMe
10.07 s
14.17 s
10.09 s
14.20 s
10.09 br s
14.78 s
10.06 s
14.68 s
10.55 s
9.95 s
14.38 s
10.11 s
10.93 s
3.80 s
3.91 s
3.98 s
3.88 s
3.87 s
In the fractions richest in xanthohumol (1), compound
2 was detected, which eluted slightly later than xantho-
humol (1) on a reversed-phase C18 column, and which
showed an MS±MS fragmentation pattern identical to
that of xanthohumol (1). Both compounds yielded
fragment ions with m/z 299 [MH±C4H8]+ and the
retro Diels±Alder fragments, m/z 235 ‰1,3A]+ and m/z
179 ‰1,3A C4H8]+, indicating that 2, like 1, contains a
phloroglucinol A-ring with one C-prenyl and one O-
Table 3
13C-NMR spectroscopic data for compounds 2, 3, 9, 11, and 12 ꢀd
ppm)
C
2
3
9
11
12
C1O
a
b
192.4
123.8
142.6
126.0
130.5
116.0
159.9
104.8
162.6
106.9
163.0
90.7
192.3
123.7
143.0
125.9
130.6
116.0
160.0
105.5
162.8
108.4
163.1
87.7
191.8
123.5
143.0
126.0
130.6
116.0
159.7
104.9
164.7
100.4
160.0
91.7
191.7
123.8
142.5
126.0
130.5
116.0
159.9
104.5
165.2
105.2
163.3
91.1
191.9a
123.6
142.7
126.0
130.3
116.1
160.1
105.1a
163.4
108.4
157.4a
102.5
153.7
21.3
1
methyl substituent. The H- and 13C-NMR spectra of
1
2,6
3,5
4
1 (Stevens et al., 1997) and 2 (Tables 2 and 3) were
also very similar. Because 2 showed a low-®eld hydro-
gen-bonded hydroxyl (OH-2' at dH 14.17) in the 1H
spectrum, two structures seemed possible: (i) 3'-prenyl-
4'-O-methylchalconaringenin and (ii) 5'-prenyl-6'-O-
methylchalconaringenin. Possibility (ii) was excluded
1'
2'
3'
4'
5'
6'
10
20
30
40
50
60
1
by H±13C HMBC spectroscopy as follows. The identi-
160.3
20.9
161.1
21.0
160.5
160.6
28.8
73.7
ties of C-1', C-2' and C-3' followed from their cross
peaks with the OH-2' resonance at dH 14.17. Since the
H-10 protons of the prenyl substituent showed inter-
actions with C-30, C-20, C-2', C-3' and C-4', the pre-
nyl group was located at C-3'. The methoxy protons
interacted with only one carbon atom, which was
identi®ed as C-4' by its other cross peaks with H-10
and H-5'.
123.0
129.9
17.6
122.7
130.2
17.6
122.9
130.2
17.8
148.0
109.8
17.3
25.1
67.3
78.8
25.5
25.4
25.5
41
51
61
60-Me
117.0
125.0
77.2
25.3
21.0
For the other A-ring hydroxyl, interactions were
observed with C-1', C-5' and C-6': this hydroxyl was
therefore located at C-6'. Chalcone 2 was thus ident-
i®ed as 3'-prenyl-4'-O-methylchalconaringenin. The tri-
vial name, `xanthogalenol', is proposed for this
compound to express its structural resemblance with
61-Me
4'-OMe
6'-OMe
27.3
55.5
55.9
56.2
56.0
55.7
a These 13C shift values were obtained from the HMQC and
HMBC spectra because the corresponding signals did not exceed the
noise level in the 13C 1D spectrum.

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
763
xanthohumol and its ®rst isolation from H. lupulus cv.
`Galena'.
less favorable. Distinction between 6- and 8-substituted
¯avanoids by APCI tandem mass spectrometry has
previously been observed for 6/8-prenylnaringenins
(Stevens et al., 1997), 6/8-geranylnaringenins (Stevens
et al., 1999b), vitexin/isovitexin (Rong, Stevens, Dein-
zer, de Cooman & de Keukeleire, 1998), and for 6/8-
methoxy apigenins (Stevens, Wollenweber, Ivancic,
Hsu, Sundberg & Deinzer, 1999c).
Xanthogalenol (2) was converted into two ¯avanone
isomers, 14 and 15, upon heating in aqueous solution.
This may be taken as additional proof for the position
of the O-methyl at C-4', leaving two free hydroxyls
ortho to C-1' available for cyclization.
Xanthohumol (1) forms only one ¯avanone, iso-
xanthohumol (13), because the second ortho hydroxyl
is methylated and not available for ring closure. The
position of the C-prenyl function in the two isoxantho-
galenols was determined by the intensity of the [MH±
C4H8]+ ion with m/z 299 in the daughter ion spectra
of 14 and 15. Cleavage of the prenyl group is much
more prominent in the 6-prenyl¯avanone 14, presum-
ably as a result of para quinonoid stabilization (Fig. 2).
In the ortho quinone-methide fragment ion of the 8-
prenyl¯avanone 15, quinonoid stabilization is less
extended, and prenyl cleavage is therefore intuitively
Compound 3 was also characterized as a chalcone
1
on the basis of its UV maximum at 367 nm and its H
spectrum with two mutually coupled bridge protons (J
= 15.5 Hz, H-a and H-b). Collision-induced dis-
sociation of its molecular ion, m/z 369 [C22H25O5]+,
yielded the following fragment ions in its MS±MS
spectrum: m/z 313 [MH C4H8]+, m/z 249 ‰1,3A]+ and
m/z 193 ‰1,3A C4H8]+. This fragmentation pattern is
consistent with chalconaringenin bearing one C-prenyl
and two O-methyls on the A-ring. 1H-NMR spectro-
scopic analysis revealed the presence of a low-®eld res-
Fig. 2. Distinction between isoxanthogalenols (14 and 15) by APCI-CID-MS (collision energy, 11V).

764
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
posed structure has been con®rmed by NMR ꢀ H, 13C,
1
onance which was attributed to OH-2' (Table 2). As a
consequence, the two methoxy substituents had to be
at C-4' and C-6'. In the HMBC spectrum of 3, the
identity of C-1', C-2' and C-3' readily followed from
their cross peaks with the OH-2' signal. The aromatic
A-ring proton ꢀdH 6.27) interacted with all A-ring car-
bons but C-2', and was therefore identi®ed as H-5',
leaving the prenyl substituent to be located at C-3'.
This was con®rmed by interactions between H-10 of
the prenyl group and carbons 2', 3' and 4'. The struc-
ture of 3 was now determined to be 4'-O-methyl-
1H±1H COSY, H±13C HMBC and HMQC, Tables 2
1
and 3). It was later isolated again from hops by
Tabata and co-workers (Tabata, Ito, Tomoda &
Omura, 1997), who named it `xanthohumol B'. We
decided to adopt Tabata's name for compound 9 to
avoid confusion, and now propose the name `xantho-
humol C' for compound 10 which was earlier named
dehydrocycloxanthohumol (Stevens et al., 1997).
The second isomer ꢀMr 370, C21H22O6), xanthohu-
mol D (11), is a new natural product. The CID mass
spectrum resembled that of xanthohumol B (9). Both
chalcones had the following peaks in common: m/z
353 [MH H2O]+, m/z 299 [MH C4H10O]+, m/z 251
xanthohumol. Compound
3 has previously been
isolated from hops collected in northern Japan by Sun
and co-workers (Sun, Watanabe & Saito, 1989). These
authors did not report 13C shift data, and therefore
compound 3 is included in Table 3. Chalcone 3 pro-
duced one isomeric ¯avanone (16) when heated in a
1:1 mixture of methanol and water.
‰
1,3A]+, m/z 233 ‰1,3A H2O]+, and m/z 179 ‰1,3A
C4H10O]+. From these diagnostic fragment ions it
was deduced that 11, like 9, is a derivative of xantho-
humol with one oxygen in the prenyl group. Since
chalcone 11 showed three phenolic hydroxyls (OH-2',
Compound 4 was identi®ed as 2',4-dihydroxy-4',6'-
1
dimethoxychalcone
(=4',6'-di-O-methylchalconarin-
OH-4' and OH-4) in its H-NMR spectrum, the possi-
1
genin) on the basis of UV, MS and H-NMR spectro-
scopic data. MS±MS fragmentation of the [MH]+ ion
(m/z 301) yielded two diagnostic RDA fragment ions,
m/z 181 ‰1,3A]+ and m/z 147 ‰1,4B]+, indicating that
two of chalconaringenin's A-ring hydroxyls were meth-
ylated. A free OH-2', apparent from a low-®eld reson-
ance at dH 13.65, left the methoxy substituents to be
bility was excluded that the prenyl substituent was
®xed in a pyrano or furano ring with OH-4' delivering
the hetero atom. The side-chain resonances were
attributed to two ole®nic protons (1CH2, dH 4.64 s
and 4.60 s ), two methylene protons ꢀdH 2.75 dd and
2.64 dd, J = 13, 7 Hz), one ±CH$% proton ꢀdH 4.22,
apparent t, coupled with the methylene protons, J = 7
Hz), and three methyl protons ꢀdH 1.72 s). These sig-
nals are consistent with a 2-hydroxy-3-methyl-but-3-
enyl substituent (Seo et al., 1997). Structure 11 was
1
located at carbons 4' and 6' ꢀ H-NMR spectral data
listed under Section 4.7). Isomerization of chalcone 4
yielded 5,7-di-O-methylnaringenin (17), which was
identical to an authentic sample of 17 by LC-MS.
Chalconaringenin-4',6'-dimethyl ether (4) has been
reported from the fern Notholaena dealbata (Wollenwe-
ber & Roitman, 1991). Sun et al. (1989) also isolated a
dimethyl ether of chalconaringenin from H. lupulus cv.
`Shinshuwase'. They placed the O-methyls at positions
2' and 6' on the basis of EI MS data, that is, they sta-
ted that absence of retro Diels±Alder fragmentation
indicated the absence of a free hydroxyl group at C-2'
and C-6' in the chalcone. However, the authors did
observe a base peak at m/z 181 (100%) which one
could identify as the RDA fragment, ‰1,3A]+, accord-
ing to Claeys' proposed nomenclature for RDA frag-
ments of ¯avonoids (Ma, Li, van den Heuvel &
Claeys, 1997). This implies that Sun's chalconarin-
genin-dimethyl ether and compound 4 could be identi-
cal. In the present survey, only one chalcone with Mr
300 was detected in Shinshuwase hops, which was
identi®ed as 2',4-dihydroxy-4',6'-dimethoxychalcone,
by retention time and MS comparison with compound
4.
1
fully supported by 2D NMR spectroscopy ꢀ H±1H
13
1
COSY, C, H±13C HMQC and HMBC, Tables 2 and
3).
Compound 12 ꢀMr 406, C25H26O5) showed a UV
maximum at 369 nm, and was therefore also character-
ized as a chalcone. Collisional activation of the [MH]+
ion yielded daughter ions with m/z 351 [MH C4H8]+,
m/z 287 ‰1,3A]+, m/z 231 ‰1,3A C4H8]+, and m/z 189
[231-C3H6]+. This fragmentation pattern pointed to a
chalconaringenin derivative with two C-prenyl substi-
tuents on the A-ring, one of which is cyclized with an
adjacent phenolic hydroxyl to form a dimethyl chro-
meno ring (Fig. 3). These assumptions gained support
from 1H-NMR spectroscopic analysis (Table 2), but
1
the H spectrum o€ered no clue as to which of the two
possible hydroxyls (4' or 6') provided the hetero atom.
It showed three phenolic hydroxyls, two of which were
identi®ed as OH-2' ꢀdH 14.38) and OH-4 ꢀdH 9.95) by
HMBC spectroscopy. For OH-2', interactions were
observed with carbons 2' ꢀdC 163.4), 1' and 3' ꢀdC
105.1 and 108.4) in the HMBC spectrum. Since H-10
of the prenyl substituent also interacted with the car-
bon resonating at dC 108.4, this carbon signal was
assigned to C-3' and the other ꢀdC 105.1) to C-1'. The
third hydroxyl resonance ꢀdH 10.11) showed cross
peaks with dC 102.5 (C-5') and dC 108.4 (C-3'), but
The resin fraction of Galena hops yielded two chal-
cones with Mr 370. The ®rst isomer, compound 9, was
earlier obtained from `Saazer hops' in small amounts
and identi®ed as the hydrate of compound 10 by CID
mass fragmentation (Stevens et al., 1997). The pro-

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
765
not with C-1', and was thus identi®ed as OH-4'. From
these assignments it became clear that the oxygen
atom attached to C-6' was part of the dimethyl chro-
meno ring. Structure 12, as depicted in Fig. 1, was
consistent with other interactions observed in the
HMBC spectrum. The methine proton H-41 of the
dimethyl chromeno moiety interacted with C-61 ꢀdC
77.2) and C-6' ꢀdC 153.7), while the other methine sig-
nal (H-51) showed cross peaks with C-61 and C-5'.
Assignments of all proton and carbon signals are listed
in Tables 2 and 3. Although not directly a derivative
of xanthohumol, compound 12 was named `xanthohu-
mol E'.
analysis, see below) in a wild hop plant from Mon-
tana, despite the fact that the leaves (early Summer
1999) and hop cones (late Summer 1998) were col-
lected in di€erent seasons and years. Based on this ob-
servation, we decided we could include leaf gland
samples from other hop plants in this study. These
plants had not yet developed lupulin glands at the
time the analyses were performed, and therefore leaf
gland sampling o€ered an alternative way to investi-
gate their prenyl¯avonoid chemistry. Similarity
between both types of secretory structures has also
been observed with humulones and lupulones (Nicker-
son, unpubl. results).
The most conspicuous ®nding was the dichotomous
distribution (presence versus absence) of the 4'-O-
methylchalcones 2±4 (wild hops and cultivars are listed
accordingly in Tables 4 and 5). A sample of the data
set is given in Table 6 to illustrate the di€erences
between the ¯avonoid pro®les observed for plants of
di€erent geographic origins. Fig. 4 shows a typical
LC-MS±MS chromatogram.
2.2. Flavonoid survey
The LC-MS±MS analyses of the plant materials
yielded a data matrix listing the amount of 19 ¯avo-
noids as per cent of xanthohumol (1) for each of 121
samples covering three wild taxa and 52 cultivars. The
leaf gland ¯avonoids and lupulin gland ¯avonoids
showed similar pro®les (by principal components
Principal components analysis (PCA) of the whole
Fig. 3. Proposed APCI-CID (15V) mass fragmentation of xanthohumol E.

766
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
data set revealed three groups of hop plants: (i)
cultivars and wild hops of European origin and
southwestern US wild hops, (ii) hybrids between
European cultivars and central US or Japanese wild
hops, and (iii) central US and Japanese wild hops
(Fig. 5). In PCA, a set of independent linear com-
binations of the original data vectors (e.g., the ¯a-
vonoid composition) are calculated in such a way
Table 4
Hop varieties containing xanthogalenol (2), 4'-O-methylxanthohumol (3), and 4',6'-di-O-methylchalconaringenin (4)
Sample
code
Variety
USDA acc.
no.a
Parentage (female  male), remarksb
Collection site and yearc
H. lupulus
var. cordifolius
Cr
Originally from Japan
Leiden Univ., 1999 (leaf
glands)
Co2
Co3
Co4
Io
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
wild American var.
60026M
60028M
60031M
21565M
21605
Originally from Colorado
Originally from Colorado
Originally from Colorado
Originally from Iowa
OSU,e1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1998
Mi1
Ms
M1
M2
M3
M4
M5
M6
M7
M8
N2
Originally from Minnesota
Originally from Missouri
Originally from Montana
Originally from Montana
Originally from Montana
Originally from Montana
Originally from Montana
Originally from Montana
Originally from Montana
Originally from Montana
Originally from Nebraska
Originally from North Dakota
21558M
21585M
21586
OSU, 1999 (leaf glands)
OSU, 1998
OSU, 1998
21586
OSU, 1999 (leaf glands)
OSU, 1998
OSU, 1992
21595M
21601M
21602
OSU, 1998
OSU, 1999 (leaf glands)
OSU, 1998
21580
21580
21550M
21574M
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1998
N7
No
Wn
9032-028Md Originally from North Dakota
21117M
Originally from Wisconsin
OSU, 1999 (leaf glands)
Cultivars
Aquila
Banner
Aq
Ba
21222
21287
19001
21116
64100
66054
21226
65102
65102
65102
21182
21182
21182
21182
21699
21039
21709
21286
21225
21115
60042
21403
Brewer's Gold (USDA 19001) seedling  OP
Brewer's Gold (USDA 19001) seedling  OP
Wild Manitoba BB1 Â OP
WSUf, 1998
WSU, 1998
Bg
Bv
Brewer's Gold
Brewer's Gold, v.f.
Bullion
1997
OSU, 1998
virus free clone of Brewer's Gold (19001)
Wild Manitoba BB1 Â OP
Bu
Ca
1998
OSU, 1998
Calicross
Chinook
Cluster
California Cluster  Fuggle seedling
Petham Golding  (Brewer's Gold  Wild Utah hop)
Probably derived from an American Wild hop
Probably derived from an American Wild hop
Probably derived from an American Wild hop
Brewer's Gold (USDA 19001) Â OP
Brewer's Gold (USDA 19001) Â OP
Brewer's Gold (USDA 19001) Â OP
Brewer's Gold (USDA 19001) Â OP
Virus free clone of Galena (21182)
Mutant of Shinshuwase
Ch
Cl1
Cl2
Cl3
Ga1
Ga2
Ga3
Ga4
Gv
Gs
OSU, 1998
Idaho, 1997
Cluster
Cluster
Washington, 1997
Washington, 1998
Idaho, 1997
Galena
Galena
Washington, 1998
OSU, 1998
Galena
Galena
Washington, 1997
OSU, 1998
Galena, v.f.
Golden Star
Kirin C-601
Kirin II
OSU, 1999 (leaf glands)
OSU, 1998
OSU, 1998
Kc
Ki
1/2 Toyomidori, 1/8 Kirin II, 1/8 OP (Japan)
Clonal selection from Shinshuwase hops
3/4 Brewer's Gold
O
Pt
Olympic
OSU, 1999 (leaf glands)
WSU, 1998
Pocket Talisman
Shinshuwase
Sticklebract
Selected from Talisman (Late Cluster  OP)
Saazer  wild Japanese male (OP)
S1, S2
Sk
OSU, 1999 (leaf glands)
OSU, 1998
Tetraploid First Choice (USDA 66055, Californa
Cluster  OP)  OP (New Zealand)
Late Cluster  OP
Ta
Talisman
65101
OSU, 1998
a USDA accession nos. ending with `M' are male plants.
b OP: open pollination.
c Collection sites of commercial samples are not given. Lupulin glands of the ¯owers were used for analysis (unless otherwise mentioned).
d USDA nursery number.
e Oregon State University.
f Washington State University.

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
767
Table 5
Hop varieties lacking xanthogalenol and other 4'-O-methylchalcones
Sample code Variety
USDA acc. Parentage (female  male), remarksb
Collection site and yearc
no.a
H. lupulus
var. lupulus
E1
E2
Seeds from Humboldt Univ., Berlin
Originally from Reeuwijk, The Netherlands
Originally from Boskoop, The Netherlands
Plant raised from seeds of E1
Leiden University, 1997
Leiden University, 1998
Leiden University, 1997
Leiden University, 1997
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
var. lupulus
var. lupulus
var. lupulus
E3
E4
Az
Wild American var.
Wild American var.
Wild American var.
Wild American var.
Wild American var.
60013M
60023M
60016
Originally from Arizona
Co1
Nm1
Nm2
Nm3
Originally from San Juan Riv.Valley, Colorado
Originally from New Mexico
Originally from New Mexico
60020
60021
Originally from New Mexico
Cultivars
Apolon
Atlas
Ap
At
Cc
Cb
Eg
Er
21051
21052
56013
21040
21678
21183
48209
Brewer's Gold  Yugoslavian wild male
Brewer's Gold  Yugoslavian wild male
[Fuggle  (Serebrianka-Fuggle Seedling)]  OP
Tetraploid Fuggle (USDA 21003) Â Fuggle seedling
Japanese cv  Wye OB79 (USDA 64103M)
Brewer's Gold (USDA 19001) Â OP
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
Washington, 1998
OSU, 1998
Cascade
Columbia
Eastern Gold
Eroica
WSU, 1998
OSU, 1998
OSU, 1998
Fh
Fuggle H
Selected from Fuggle (USDA 19209). Pedigree of Fuggle
unknown, originally from Kent, UK
Clonal selection of Fuggle material
Fu
Fuggle N
Hallertauer
21016
56001
21014
21672
21516
21278
OSU, 1998
Idaho, 1997
Ha
Hm
Ht
Probably derived from an old German land race
Selected from an old German land race
Hallertau Gold  a German male
Hallertauer mittelfruh
Hallertauer Tradition
Hersbrucker
OSU, 1998
1997
OSU, 1998
He
Ke
Probably derived from an old German land race
OP seedlings of New Mexican wild female
OP seedlings of New Mexican wild female
Japanese cultivar
Keyworth's Early
OSU, 1999 (leaf glands)
OSU, 1999 (leaf glands)
WSU, 1998
Ks
Keyworth's Midseason 21279
Km
Lc
Kitamidori
21677
19005M
21670
21670
64107
Late Cluster seedling
Magnum
Magnum
Late Cluster  OP
OSU, 1992
1997
Ma1
Ma2
Nb
Galena  a German male
Galena  a German male
Washington, 1998
Germany, 1997
Northern Brewer
Canterbury Golding (old English hop) Â OB21
(seedling of Brewer's Gold  Wild American)
USDA 65009 Â 63015M; 5/8 Brewer's Gold
USDA 65009 Â 63015M; 5/8 Brewer's Gold
USDA 65009 Â 63015M; 5/8 Brewer's Gold
USDA 65009 Â 63015M; 5/8 Brewer's Gold
USDA 65009 Â 63015M; 5/8 Brewer's Gold
Northern Brewer (64107) Â a German male
Northern Brewer (64107) Â a German male
Northern Brewer (64107) Â a German male
Pride of Kent (Brewer's Gold  OP)  OP (seedling
raised and comm. grown in Australia)
Selected from an old Czechoslovakian land race
Virus free clone of Saazer (USDA 21077)
Virus free clone of Saazer (USDA 21077)
Virus free clone of Saazer (USDA 21077)
Virus free clone of Saazer (USDA 21077)
Virus free clone of Saazer (USDA 21077)
Swiss Tettnanger  (Tetraploid Hallertauer m.f.
 USDA 21381M) = 61021  21618M
Swiss Tettnanger  (Tetraploid Hallertauer m.f.
 USDA 21381M) = 61021  21618M
California cluster  OP
Nu1
Nu2
Nu3
Nu4
Nu5
Pe1
Nugget
Nugget
Nugget
Nugget
Nugget
Perle
21193
21193
21193
21193
21193
21227
21227
21227
66052
OSU, 1998 (stored 158C)
OSU, 1998 (stored +208C)
Oregon, 1997
Washington, 1997
Idaho, 1997
Germany, 1997
Pe2
Pe3
Perle
Perle
Pride of Ringwood
Oregon, 1997
Washington, 1997
WSU, 1998
Pw
Sa
Saazer
21077
21521
21521
21521
21522
21532
21664
1997
Sb1
Sb2
Sb3
Sc
Saazer-36
Saazer-36
Saazer-36
Saazer-38
Saazer-72
Santiam
OSU, 1998 (stored 158C)
OSU, 1998 (stored +208C)
Idaho, 1997
Idaho, 1997
Idaho, 1997
Washington, 1997
Sd
Sn1
Sn2
Santiam
21664
Oregon, 1997
Sm
Ss
Smooth Cone
Spalter
66056
21186
21674
21674
21689
OSU, 1999 (leaf glands)
Selected in the Spalt area of Germany from an old land race Germany, 1997
Sp1
Sp4
St
Spalter Select
Spalter Select
Sterling
Cross between two German breeding lines
Cross between two German breeding lines
1/2 Saazer-38, 1/4 Cascade, 1/8 Brewer's Gold (19001),
1/8 Zattler seedling
1997
OSU, 1998
Oregon, 1997
Te1
Tettnanger
21015
Selected in the Tettnang area of Germany from
an old land race
1997
(continued on next page)

768
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
Table 6 (continued )
Sample code Variety
USDA acc. Parentage (female  male), remarksb
Collection site and yearc
no.a
Te2
Ty
Tettnanger
21015
21676
Selected in the Tettnang area of Germany from
an old land race
Oregon, 1997
WSU, 1998
Toyomidori
Northern Brewer (64107) Â Wye OB79 (USDA 64103M);
Japanese cultivar
Us
USDA 21120
Willamette
Willamette
Willamette
Willamette
21120
21041
21041
21041
21041
Has Fuggle and Late Cluster in ancestral line
Tetraploid Fuggle (21003) Â Fuggle seedling
Tetraploid Fuggle (21003) Â Fuggle seedling
Tetraploid Fuggle (21003) Â Fuggle seedling
Tetraploid Fuggle (21003) Â Fuggle seedling
WSU, 1998
Idaho, 1997
Wi1,3
Wi2
Wi4
Wi5
Oregon, 1997
Washington, 1997
Washington, 1998
a,b,cSee footnotes in Table 4.
that the ®rst linear combination, or principal com-
ponent, accounts for as much of the variation in the
original data as possible. The second PC accounts for as
much of the remaining variability as possible, then a
third PC is calculated, and so on. In e€ect, a new set of
axes is created on which the data can be plotted that
might better reveal patterns. In this case, the separation
of the three groups along the axis of the ®rst principal
component, accounting for 68% of the variance in the
original data, is largely due to the amounts of the 4'-O-
methylchalcones 2±4. The dispersion of the observations
along the second PC axis, which accounts for 17% of
the variance, is associated with the relative amounts of
desmethylxanthohumol (5) and, to a lesser extent, the
prenylnaringenins 18 and 19. For instance, the sample
of Cascade hops (Cc, in top region of the European clus-
ter) contained a signi®cantly higher percentage of 5
(44%) than the samples of Willamette hops, which were
placed near the bottom of that cluster (range 1.7±6.1%,
average 3.8%).
As a group, the ¯avanones had little impact on
the separation of the three clusters due to their low
abundance. However, the progressive isomerization
of chalcones into ¯avanones during storage did in-
¯uence separation within the clusters. For example,
a 1997 sample of a given hop cultivar will be
located lower on the second PC axis than its corre-
sponding 1998 sample. The levels of the prenylnar-
ingenins 18 and 19, the geranylnaringenins 21 and
22, and 6,8-diprenylnaringenin (20) were elevated in
the 1997 samples. The increased levels were also
seen in 1998 samples stored at room temperature
versus those stored in the freezer immediately after
harvest (Table 6). These changes lend support to
Table 6
Prenyl¯avonoid concentration as percentage of xanthohumol
Compound
number
Japanese
wild var.
cordifolius
Japanese cultivar
Kirin II
American wild
Montana 8
American
cultivar
Cluster
European
wild
Reeuwijk
European
cultivar
European
cultivar
Saaz-36, Stored 158C
Saaz-36, Stored 208C
1
100
100
100
100
100
100
100
2
36.94
75.36
6.36
7.71
10.11
2.16
0.04
0.28
0.19
0.00
0.56
1.35
0.40
0.34
0.31
0.08
0.24
0.27
0.08
4.06
8.28
1.16
4.95
0.69
0.12
0.05
0.00
1.74
0.03
1.74
0.76
0.10
0.16
0.63
0.18
0.18
0.12
0.00
16.77
27.25
5.13
16.76
2.63
2.46
0.45
0.18
0.10
0.13
1.41
2.31
0.34
0.73
1.60
0.46
0.59
0.29
0.09
3.64
8.27
0.69
6.49
1.44
0.74
0.11
0.00
0.21
0.09
1.65
0.45
0.10
0.11
0.71
0.21
0.10
0.18
0.05
0.00
0.00
0.00
7.99
0.21
0.53
0.07
0.92
0.48
0.06
1.96
0.00
0.00
0.00
0.69
0.26
0.28
0.03
0.00
0.00
0.00
0.00
29.98
2.10
3.85
0.06
0.06
0.22
0.20
0.81
0.00
0.00
0.00
1.45
0.49
0.37
0.13
0.13
0.00
0.00
0.00
17.61
0.84
1.79
0.17
0.24
0.63
0.19
3.48
0.00
0.00
0.00
2.87
1.06
1.12
0.32
0.09
3
4
5
6
7
8
9
10
12
13
14
16
17
18
19
20
21
22

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
769
Fig. 4. LC-MS±MS chromatogram of the lupulin resin of a wild H. lupulus plant from Montana. For key to peak identities, see Fig. 1 and Table 1.

770
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
Fig. 5. Principal components analysis of the ¯avonoid data. The plants are plotted on the ®rst two PC axes; see Section 3. For sample abbrevi-
ations, see Tables 4 and 5.
our earlier hypothesis that the ¯avanones of hops
are formed chemically from their isomeric chalcones
during drying and storage, and not enzymatically in
the lupulin glands (Stevens et al., 1997). The results
obtained with the leaf glands provide additional evi-
dence for absence of chalcone isomerase in these se-
cretory structures. Prenyl¯avanones were virtually
absent in leaf glands from young leaves which were
collected in Oregon State University's hop yard on
the same day the analyses were performed.
The leaf glands of H. japonicus Sieb. & Zucc.
did not contain the prenylated ¯avonoids found in
the glandular exudates of H. lupulus. Only one plant
of H. japonicus was examined, and clearly more
material (more plants, lupulin) needs to be analyzed
before the exudate chemistry of this species can be
established. The data collected for this plant (with the
absence of all 22 ¯avonoids) were not included in the
PCA.
3. Discussion
The dichotomous distribution of the 4'-O-methyl-
chalcones 2±4 correlates well with leaf morphology
and hybridization patterns in H. lupulus. It is generally
accepted that the early European cultivars originated
by local domestication of wild hops (H. lupulus var.
lupulus ). The American and Japanese cultivars are
considered to be hybrids between introduced European
cultivars and local wild (male) plants (Small, 1980;
Neve, 1991). (Note: some American-developed culti-
vars, for example, Santiam, Sterling, and Willamette,
contain only European germplasm and are here con-
sidered ``European''). The origin of the three groups of
cultivars is in good agreement with their ¯avonoid pat-
terns. Both wild hops and cultivars of European origin
lack xanthogalenol (2), and both groups were other-
wise also very similar with regard to ¯avonoid chem-
istry. The hybrid origin of the American cultivars is

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
771
re¯ected by the placement of these varieties in the
PCA between the European cultivars and the Ameri-
can wild plants from the Missouri and Mississippi
river basins (Fig. 5). The latter group and one plant of
H. lupulus var. cordifolius contained the highest
amounts of xanthogalenol, 4'-O-methylxanthohumol,
and 4',6'-di-O-methylchalconaringenin, while inter-
mediate levels of these 4'-O-methylchalcones were
found in the hybrid cultivars. The plants from Ari-
zona, New Mexico and southern Colorado lacked the
4'-O-methylchalcones, and were consequently grouped
together with the European hops in the PCA scatter
plot. The cultivars, Keyworth's Midseason and Key-
worth's Early, are the `granddaughter' and `great-
granddaughter' of a wild female plant from New Mex-
ico. Each generation was raised from open pollinated
¯owers at Wye College, England. The absence of 4'-O-
methylchalcones in these cultivars re¯ects the absence
of biosynthetic capability in the ancestral lines.
direct descendent from a lineage that never possessed
the character.
In a study of genetic variation using restriction frag-
ment length polymorphisms of the ribosomal RNA
genes, Pillay and Kenny (1996) also recognized two
major groups within H. lupulus, (1) wild and cultivated
European hops and wild hops from western China
(Xinjiang province), and (2) wild American hops. In
contrast to our ¯avonoid variation data, Pillay & Ken-
ny's analysis of rDNA variation did not reveal di€er-
ences between wild hops from the Missouri±
Mississippi river basins and those from Arizona and
New Mexico (Pillay & Kenny, 1996). Wild H. lupulus
plants from eastern China and Japan were not
included in their studies.
4'-O-methylchalcones are quantitatively well rep-
resented in the leaf gland exudate of H. lupulus var.
cordifolius of Japanese origin. Many of the Japanese
cultivars presumably received gene material from male
plants of this taxon by wind pollination. The inter-
mediate position of cv. `Golden Star' and the Kirin
cultivars from Japan is apparently the result of inheri-
tance of the 4'-O-methylchalcones from H. lupulus var.
cordifolius through the classic Japanese cultivar, `Shin-
shuwase'. The lupulin resin of the latter cultivar is
known as a source of 4'-O-methylxanthohumol and
xanthohumol (Sun et al., 1989), and the leaf gland
exudates of both the Shinshuwase hop plants examined
contained the three 4'-O-methylchalcones and xantho-
humol. Although Shinshuwase has been recorded as a
hybrid between the European cultivars, `Saazer' and
`White Vine', these ancestors would not account for
the presence of the 4'-O-methylchalcones. In fact, mor-
phological evidence indicates that the female Saazer
plant was open pollinated by a wild Japanese plant, a
hypothesis supported by the ¯avonoid composition.
Ono (Ono, 1959; cited by Neve, 1991) noted that Shin-
shuwase resembles the American hop more than the
European one from a morphological point of view.
Golden Star has been reported as a mutant form of
Shinshuwase, while `Kirin II' is a clonal selection from
Shinshuwase, and as such is virtually identical to Shin-
shuwase (Neve, 1991). The other Japanese cultivar
containing the three 4'-O-methylchalcones, Kirin C-
601, received 1/8 gene material from Kirin II and 1/8
from an unknown male plant, both of which could
have passed on the capability to accumulate 4'-O-
methylchalcones.
Small's classi®cation of wild North American hops
was based on morphological characters and geographic
distribution, and he recognized the complications of
sympatry and intergradation (Small, 1978). Our study
did not include morphological examination, nor did
we know if our selection of material coincided with
Small's. Except for one sample originally collected in
Wisconsin, all wild American hops in this study were
from west of the Mississippi River. Excepting those
from Arizona and New Mexico, they could, by geogra-
phy, be any of Small's varieties. Many specimens were
originally collected west of the 100th meridian, a
region in which Small considered H. lupulus var. neo-
mexicanus to be largely allopatric. In contrast, the pre-
sence or absence of the 4'-O-methyl chalcones divides
the wild hops in this study into one group from the
Missouri±Mississippi drainage, and a second one south
and west of that basin, including Arizona, New Mex-
ico, and southwestern Colorado. Thus, we refer to
these hops as wild American hops with two geo-
graphic/chemical groups.
The phylogenetic relationship between H. lupulus
from the southwestern USA, most likely of var. neo-
mexicanus, and the other taxa of H. lupulus is not
clear. It could have been derived from H. lupulus of
central or midwestern USA, or perhaps directly from
an east Asian taxon (either capable or incapable of ac-
cumulating 4'-O-methylchalcones), if one considers the
Rocky Mountains as a physical barrier between the
two American taxa. The Rocky Mountains continued
to be uplifted since early Tertiary times, and it became
an increasingly e€ective east±west barrier for plants
growing at lower elevations (Morain, 1984). It is there-
fore not clear whether absence of 4'-O-methylchal-
cones in H. lupulus of the American southwest is the
result of loss of the trait, or if this taxon is a more
The classic American cultivars are the Clusters.
Although the origin of the Clusters is not entirely
clear, hybridization between introduced European cul-
tivars and American male plants most likely occurred
during the settlement of the eastern seaboard in the
17th century. It is therefore assumed that the Ameri-
can Clusters inherited the 4'-O-methylchalcones from
local wild males. The `Early Cluster' (Oregon) and

772
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
`California Cluster' arose on the west coast as clonal
selections of the `Late Cluster' that the settlers brought
with them in the mid to late 1800s (Neve, 1991).
(Collinson, 1989), indicating that speciation was most
vigorous during the mid to late Tertiary. In this
period, the temperate zone of the northern hemisphere
was extended towards the pole (Takhtajan, 1969), and,
although the sea level was rising, the Bering Strait con-
sisted of a dry land bridge (Pielou, 1979) or a chain of
islands and archipelagos (Morain, 1984), for extended
periods of time. These conditions allowed some species
to use the Bering Strait as a land route for migration
between Eurasia and North America (Pielou, 1979;
Morain, 1984). Moreover, Raven and Axelrod (1974)
argued that the Bering land bridge has been more im-
portant as a land route for migrating species than the
North Atlantic passage via Greenland and Iceland,
since the middle of the Eocene. Examples of genera
that presumably owe their present disjunct distribution
(eastern Asia and N. America) to the Bering land
bridge include Carya (hickory), Acer (maple), Epigaea
(trailing arbutus), and Stewartia (shrubs of the tea
family) (Pielou, 1979).
Many of the xanthogalenol-containing cultivars
listed in Table 4 are derived from H. lupulus cv.
`Brewers Gold', for example, the American cultivars
`Aquila' and `Galena'. The varieties `Brewer's Gold'
and `Bullion' are selections from seedlings of a hybrid
between a wild female plant from Manitoba and an
unknown European male plant. This hybrid resulted
from an open-pollination event at Wye College, Eng-
land, in 1918 (Neve, 1991). There is little doubt that
the 4'-O-methylchalcones were passed on to `Brewer's
Gold', `Bullion' and their descendents, by this Mani-
toba female plant, which is from the same geographic
region as the wild hops from Montana, North Dakota
and Minnesota examined in this study. However, some
cultivars contain wild American gene material, notably
Apolon, Atlas, Hallertauer Magnum, Nugget, and
Smooth Cone, yet they lack the capability to produce
4'-O-methylchalcones. A possible explanation is loss of
the trait by genetic drift due to selection for other
characters. This is presumably also true for the culti-
vars, Northern Brewer, Pride of Ringwood, and Toyo-
midori, all of which contain less than 1/16th wild
American gene material.
4. Experimental
4.1. Plant material
The distribution of xanthogalenol and 4'-O-methyl-
xanthohumol leads us to suggest that 4'-O-methylation
is a primitive character in H. lupulus. The most parsi-
monious explanation for the observed dichotomy is
that 4'-O-methylation of chalcones has been retained
(or arisen during the early stages of infraspeci®c di€er-
entiation) in the Japanese and American lineages, and
was lost in the European lineage (var. lupulus ). The
present-day disjunct distribution of H. lupulus is con-
sidered to be the result of its eastward migration to
Japan and America and westward to Europe. Geo-
graphic separation of the species has earlier been as-
sociated with the development of two genetically
di€erent populations, and may explain why wild hops
of American origin are more similar to wild Japanese
hops than to wild European plants, with regard to
pubescence and shape of the leaves (Neve, 1991; Small,
1980).
These assumptions are in line with phytogeographi-
cal and palaeobotanical data. Three species of Humu-
lus are known as endemics of China (H. lupulus var.
cordifolius, H. japonicus [syn. H. scandens (Lour.)
Merr] and H. yunnanensis Hu), which led Neve (1991)
to suggest that this part of Asia is the center of origin
of the genus. Most of the fossil fruits of Humulus (H.
irtyshensis, H. rotundatus, H. scandens ) and the pre-
sumably closely related Humularia (now believed to be
extinct), were found in China and Russia. These
records all date back to the Oligocene, Miocene, and
Pliocene periods (35 million to 2 million years ago)
Hop cones and male lupulin materials were obtained
from various sources; their origin, collection site and
year of harvest are listed in Tables 4 and 5. The hop
cones were dried at ca. 508C to ®nal moisture contents
of ca. 12% and stored at 158C, until analysis (unless
otherwise stated in Tables 4 and 5).
Leaves were collected at the experimental hop yards
of Oregon State University (OSU) and the National
Clonal Germplasm Repository, Corvallis, in early
Summer 1999. Leaf samples were processed fresh
shortly after collection of the leaves. Humulus japonicus
(originally from China) was grown in the greenhouses
of the NCGR at Corvallis.
4.2. Reference ¯avonoids
Flavonoids 1, 5, 8, 9, 10, and 13 were obtained
during earlier work or isolated as described by Stevens
et al. (1997). Flavanones 18 and 19 were prepared by
prenylation of naringenin (Stevens et al., 1999a); the
geranyl¯avanones 21 and 22 were synthesized from
naringenin and linalool (Stevens et al., 1999b). The
remaining ¯avonoids were obtained as described in
Section 4.7.
4.3. Sample preparation for LC-MS
Female plant material (one hop cone or ground
hops, typically 0.1 g) was immersed in 100 ml of
MeOH±HCOOH (99:1 by vol.). The extraction mix-

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
773
ture was gently swirled several times over 2 h and then
allowed to settle for 2 h. A portion of the cleared
extract was carefully removed by pipetting and directly
analyzed by LC-MS. Male lupulin samples were pre-
pared as follows: ca. 0.1 mg resin was immersed in
MeOH±HCOOH (99:1, 1 ml) at room temperature for
2 h with occasional agitation. A portion of the super-
natant was analyzed by LC-MS.
of ion pairs scanned in a single HPLC run was limited
to six in order to retain sucient sensitivity for detec-
tion of each daughter ion. The remaining ion pairs
were scanned in a second HPLC±MRM run for each
sample. Scanning of the ion pair, m/z 355 4 179, was
included in both runs in order to be able to use
xanthohumol (1), xanthogalenol (2), and isoxanthohu-
mol (13) as internal calibrants.
Leaf gland extracts of female or male plants were
prepared as follows: leaf glands (ca. 30) were removed
from the bottom side of young leaves with a solid nee-
dle at 7Â (stereomicroscope), transferred to a 0.25 ml
sample vial, and extracted with MeOH±HCOOH
(99:1, 0.1 ml) at room temp for 2 h with occasional
agitation.
4.5. Data analysis
The chromatographic peak areas of the ¯avonoids
analyzed by HPLC±MRM were determined using PE
Sciex's MacQuan 1.5 software. Compounds 9 and 11
coeluted and were treated as a single amount. Com-
pound 15 was not detected in the survey of hop var-
ieties. If the peak areas of xanthohumol (1),
xanthogalenol (2), and isoxanthohumol (13) from the
two chromatographic runs for each hop sample were
comparable, the data were pooled and used to calcu-
late the amount of 19 ¯avonoids relative to xanthohu-
mol (1) for that sample.
The SAS System 6.12 statistics software was used to
calculate a covariance matrix from these relative
amounts and then to perform a principal component
analysis. Xanthohumol (1), serving as an internal stan-
dard, had a relative amount of 100%, and was not
used in the PCA calculation. The scatter plot (Fig. 5)
was generated from the ®rst two PCs.
4.4. LC-tandem MS analysis
Hop extracts were analyzed for prenyl¯avonoids by
LC-MS as described previously (Stevens et al., 1999a).
In short, hop extracts were separated on a 5 mm C18
column (250 Â 4 mm) using a linear solvent gradient
from 40 to 100% B (MeCN) in A (1% aq. HCOOH)
over 15 min, followed by 100% B for 7 min. The ¯ow
rate was 0.8 ml min 1. The HPLC instrument was con-
nected to a PE Sciex API III Plus triple-quadrupole
mass spectrometer, equipped with an atmospheric
pressure chemical ionization (APCI) source. The APCI
source was operated in the positive ion mode. The
HPLC e‚uent was introduced into the mass spec-
trometer via a heated nebulizer interface at 5008C.
Ionization of the analyte vapor mixture was initiated
by a corona discharge needle at ca. 8 kV and a dis-
charge current of ca. 6 mA. For MS±MS, argon±nitro-
gen (9:1) was used as target gas in the collision cell
(ca. 1.8 Â 1014 atoms cm 2). The collision energy was
set at 15 V to obtain diagnostic daughter ions for
structure determination (Section 2), and increased to
30 V to enhance the yield of the ‰1,3A C4H8]+ type
ions for semi-quantitation by multiple reaction moni-
toring (MRM, see below).
4.6. NMR spectroscopy
Nuclear magnetic resonance spectra were recorded
on a multinuclear Bruker DRX 600 instrument at 600
MHz for H and at 150.9 MHz for 13C: Samples were
1
dissolved in DMSO-d6 and analyzed at room tempera-
ture. DMSO resonances ꢀdH 2.50 and dC 39.51) were
1
used as internal chemical shift references. H±13C
HMQC and HMBC experiments were performed using
standard pulse sequences.
The 22 ¯avonoids (Table 1) were selectively detected
by MRM in the tandem MS mode. In MRM, speci®c
daughter ions were detected that were produced by
collision-induced dissociation of selected [MH]+ ions.
That is, for each [MH]+ ion, the most intense ion in
the full-scan daughter ion spectrum was chosen for
detection (cf. Table 1): m/z 355 4 179 for compounds
1, 2, 13, 14 and 15; m/z 369 4 193 for 3 and 16; m/z
301 4 181 for 4 and 17; m/z 341 4 165 for 5, 18 and
19; m/z 409 4 165 for 6, 21 and 22; m/z 409 4 177 for
7 and 20; m/z 423 4 191 for 8; m/z 371 4 233 for 9 +
11 (indistinguishable from each other); m/z 353 4 233
for 10; and m/z 407 4 231 for chalcone 12. Although
the MRM software allows rapid sequential scanning of
up to 10 ion pairs in a single HPLC run, the number
4.7. Extraction and puri®cation of ¯avonoids
Whole hop cones (cv. `Galena', 301 g) were
immersed in Me2CO for a few minutes, and then the
extraction solvent was decanted and ®ltered. After a
second extraction with Me2CO, the ®ltrates were com-
bined and concentrated on a rotary evaporator at
458C. The oily residue was fractionated on a Sephadex
LH-20 column with MeOH as eluent. Fractions of ca.
10 ml were collected and monitored with TLC (silica
gel plates run in toluene±dioxane±HOAc, 18:5:1).
Chalcone-containing fractions were combined and
fractionated again by column chromatography on
Sephadex LH-20. Fractions of similar composition,
monitored by HPLC-UV (370 nm), were combined to

774
J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
give three main fractions: (i) containing compounds 1,
3, 4, 9, 10, 11, and 12, (ii) containing 1, 2, and 6, and
fraction (iii) with mainly 1 and 5. These compound
mixtures were separated by preparative HPLC on a 10
mm Econosil RP-18 column (22 Â 250 mm) using a lin-
ear gradient starting from 40 to 100% MeCN in 1%
aq. HCOOH over 40 min at a ¯ow rate of 11.2 ml
min 1. The UV trace was recorded at 370 nm. Peak
fractions were collected manually and dried by rotary
evaporation and lyophilization. Mixed peak fractions
were re-chromatographed on the same column using
di€erent solvent gradients or isocratic elution (MeCN±
1% aq. HCOOH).
3.22±3.15 (5H, H-3ax, H-10 and H-11), 2.72 (1H, dd, J
= 17.0, 3.1 Hz, H-3eq), 1.70, 1.61, 1.59, and 1.52
(each 3H, 4s, H-40, H-50, H-41 and H-51); UV and
MS data are listed in Table 1.
3',5'-Diprenylchalconaringenin (7) was prepared by
treatment of ¯avanone 20 (4 mg) with 5% NaOH in
MeOH (4 ml) under re¯ux conditions for 30 min. The
reaction mixture was poured into 25 ml of 2N HCl.
The aqueous layer was extracted with Et2O (25 ml)
and discarded. The Et2O layer was washed with water
(3 Â 25 ml) and evaporated to near-dryness. A small
amount of 3',5'-diprenylchalconaringenin (ca. 0.1 mg)
was obtained from the residue by preparative HPLC
(UV 370 nm). UV and MS data are listed in Table 1.
4',6'-di-O-methylchalconaringenin (4). Yellow pow-
der after lyophilization. UV and MS data, Table 1; H
1
-NMR spectral data dH: 13.65 (OH-2'), 10.08 (OH-4),
7.62 (2H, s, H-a and H-b), 7.58 (2H, d, J = 8.6 Hz,
H-2/6), 6.83 (2H, d, J = 8.6 Hz, H-3/5), 6.15 (1H, d, J
= 2.3 Hz, H-5'), 6.11 (1H, d, J = 2.3 Hz, H-3'), 3.90
and 3.82 (each 3H, s, 2 Â OMe).
Acknowledgements
The authors thank the following persons for provid-
ing some of the plant materials: Gerard W. Ch. Lem-
mens (Morris Hanbury Inc., Yakima, WA), Dr.
Stephen T. Kenny (Washington State University),
Wim Snoeijer and Prof. Rob Verpoorte (Leiden Uni-
versity, The Netherlands). The Sciex API III Plus mass
spectrometer was purchased in part through grants
from the National Science Foundation (BIR-9214371)
and from the Anheuser-Busch Companies. This work
was supported by the Anheuser-Busch Companies,
Miller Brewing Company, the Hop Research Council,
and the United States Department of Agriculture.
4.8. Conversion of chalcones into ¯avanones
Chalcones 2, 3, and 4 (ca. 0.2 mg) were dissolved
separately in a mixture of MeOH±H2O (1:1, 3 ml) and
kept overnight at 608C. This treatment resulted in
complete conversion of the chalcones, monitored by
HPLC-MS. The ¯avanone products (14/15 from 2; 16
from 3; and 17 from 4) were isolated from the reaction
mixtures by preparative HPLC and analyzed by UV
and MS (Table 1). Compound 17 was identical with a
sample of 4'-hydroxy-5,7-dimethoxy¯avanone (15,7-
di-O-methylnaringenin), purchased from Indo®ne
(Somerville, NJ), by UV and HPLC-MS.
References
Burgess, A. H. (1964). Hops: Botany, cultivation and utilization. New
York: Interscience.
4.9. Preparation of 6,8-diprenylnaringenin and 3',5'-
diprenylchalconaringenin
Collinson, M. E. (1989). The fossil history of the Moraceae,
Urticaceae (including Cecropiaceae), and Cannabaceae. In P. R.
Crane, & S. Blackmore, Evolution, systematics, and fossil history
of the hamamelidae, vol. 2: `Higher' Hamamelidae (pp. 319±339).
Oxford: Clarendon Press.
Prenylated naringenins were obtained by treatment
of naringenin with 2-hydroxy-2-methyl-but-3-ene in
dioxane/BF3 etherate as described by Jain, Gupta and
Sarpal (1978). A portion of the reaction products was
separated on a 10 m Econosil RP-18 column (10 Â 250
mm) using a linear gradient starting from 40 to 100%
MeCN in 1% aq. HCOOH over 60 min at a ¯ow rate
of 5.0 ml min 1. The UV trace was recorded at 290
nm. Peak fractions were collected manually and dried
by rotary evaporation and lyophilization. 8-Prenylnar-
ingenin (13 mg), 6-prenylnaringenin (5 mg), and 6,8-
diprenylnaringenin (4 mg) were obtained as white pow-
ders.
Jain, A. C., Gupta, R. C., & Sarpal, P. D. (1978). Synthesis of lupi-
nifolin, di-O-methylxanthohumol and isoxanthohumol and re-
lated compounds. Tetrahedron, 34, 3563±3567.
Ma, Y. L., Li, Q. M., van den Heuvel, H., & Claeys, M. (1997).
Characterization of ¯avone and ¯avonol aglycones by collision-
induced dissociation tandem mass spectrometry. Rapid
Communications in Mass Spectrometry, 11, 1357±1364.
Markham, K. R. (1982). Techniques of ¯avonoid identi®cation.
London: Academic Press.
Morain, S. A. (1984). Systematic and regional biogeography. New
York: Van Nostrand Reinhold.
Neve, R. A. (1991). Hops. London: Chapman and Hall.
Ono, T. (1959). The hop in Japan. Bulletin of Brewing Science
Tokyo, 5, 1±12.
1
6,8-Diprenylnaringenin (20). H-NMR spectral data;
Pielou, E. C. (1979). Biogeography. New York: Wiley.
Pillay, M., & Kenny, S. T. (1996). Structure and inheritance of ribo-
somal DNA variants in cultivated and wild hop Humulus lupulus
L. Theoretical and Applied Genetics, 93, 333±340.
dH 12.39 (OH-2), 9.61 and 9.54 (OH-4 and OH-4'),
7.30 (2H, d, J = 8.5 Hz, H-2'/6'), 6.78 (2H, d, J =
8.5 Hz, H-3'/5'), 5.39 (1H, dd, J = 12.6, 2.9 Hz, H-2),
5.08 and 5.05 (2H, 2t, J = 7.2 Hz, H-20 and H-21),
Raven, P. H., & Axelrod, D. I. (1974). Angiosperm biogeography

J.F. Stevens et al. / Phytochemistry 53 (2000) 759±775
775
and past continental movements. Annals of the Missouri Botanical
Garden, 61, 539±673.
noids in hops and beer by liquid chromatography-tandem mass
spectrometry. Journal of Chromatography, A 832, 97±107.
Stevens, J. F., Taylor, A. W., Clawson, J. E., & Deinzer, M. L.
(1999b). Fate of xanthohumol and related prenyl¯avonoids from
hops to beer. Journal of Agricultural and Food Chemistry, 47,
2421±2428.
Rong, H., Stevens, J. F., Deinzer, M. L., de Cooman, L., & de
Keukeleire, D. (1998). Identi®cation of iso¯avones in the roots of
Pueraria lobata. Planta Medica, 64, 620±627.
Seo, E. K., Silva, G. L., Chai, H. B., Chagwedere, T. E.,
Farnsworth, N. R., Cordell, G. A., Pezzuto, J. M., & Kinghorn,
A. D. (1997). Cytotoxic prenylated ¯avanones from Monotes eng-
leri. Phytochemistry, 45, 509±515.
Stevens, J. F., Wollenweber, E., Ivancic, M., Hsu, V. L., Sundberg,
S.,
Chrysothamnus. Phytochemistry, 51, 771±780.
Sun, S.-S., Watanabe, S., Saito, T. (1989). Chalcones from
Humulus lupulus. Phytochemistry, 28, 1776±1777.
Tabata, N., Ito, M., Tomoda, H., Omura, S. (1997).
& Deinzer, M. L. (1999c). Leaf surface ¯avonoids of
Small, E. (1978).
A
numerical and nomenclatural analysis of
&
morpho-geographic taxa of Humulus. Systematic Botany, 3, 37±
76.
&
Small, E. (1980). The relationships of hop cultivars and wild
variants of Humulus lupulus. Canadian Journal of Botany, 58,
676±686.
Xanthohumols, diacylglycerol acyltransferase inhibitors, from
Humulus lupulus. Phytochemistry, 46, 683±687.
Takhtajan, A. (1969). Flowering plants: origin and dispersal.
Washington: Smithsonian Institution Press (translated edition by
C. Je€rey).
Stevens, J. F., Ivancic, M., Hsu, V. L., & Deinzer, M. L. (1997).
Prenyl¯avonoids from Humulus lupulus. Phytochemistry, 44,
1575±1585.
Wollenweber, E., & Roitman, J. N. (1991). New frond exudate ¯avo-
noids from Cheilanthoid Ferns. Zeitschrift fur Naturforschung,
46c, 325±330.
Stevens, J. F., Taylor, A. W.,
Quantitative analysis of xanthohumol and related prenyl¯avo-
&
Deinzer, M. L. (1999a).
È