Biochemical
Syst emat ics
Pergamon
and
Ecology.
Vol. 24 ,
No.
718,
pp.
725-733.1996
Copyright
Q 1996Elsetier
Science Ltd
Printed
in Great Britain.All rightsreserved
0305-1978/ 96
~15.00+0.00
PII: SO305-1979(96)00079-S
Phylogenetic Relationships among Chilean Sophora Species
RAIL C. PEfiA* and BRUCE K. CASSELST
lEscueia
de Quimica y Farmacia, The Herbarium, Casilla 6084, Santiago 22, Chile;
tDepartment of Chemistry, Facultad de Ciencias, Casilla 653, Santiago, Chile
Key Word
Index--Sophora;
Papilionaceae;
Quinolizidine
alkaloids; cladistics.
Abstract-Phylogenetic
affinities among Chilean Sophora species are not clear. We suggest a new
hypothesis for the origin of the section Edwardsa on the basis of parsimony analysis, which allows a South
American origin to be established for the species of this section. The seed alkaloid composition did not
provide useful information for the filiation of Edwardsia species, and the shortest tree was obtained using
morphological characters only. Two branches are clearly distinguishable by the pubescence of the leaflets
and the flag/wings length ratio: one of them includes S. chrysophylla, S. fetraptera, S. toromiro, S. howinsula
and S. denudata; the other one includes S. macnabiana, S. microphylia, S. masafuerana, S. prostrata and S.
fernandeziana. In contrast, S. macrocarpa, an ancient element of the South American flora, is closely related
to species belonging to the section Sophora represented in the region by S. fomentosa, S. knearifoba and S.
rhynchocarpa. Sections Calia and Sfyphnolobium are clearly related to each other, both morphologically
and chemically. Copyright Q 1996 Elsevier Science Ltd
Introduction
The genus Sophora comprises about 45-50 species, 10 of which are included in
the section Edwardsia. A segmentation of Sophora was suggested originally by
Yakovlev (1967), who revived the genera Calia and Styphnolobium primarily on
the basis of morphological
arguments. More than a decade later Tsoong and Ma
(1981) merged Calia into Styphnolobium.
Sousa and Rudd (1993), in their
revision of Styphnolobium,
have acknowledged
that this taxon is close to Calia
although
its chromosome
number 2n=28
contrasts with 2n= 18 shared by
Sophora and Calia (cf. Palomino et al., 1993).
Bailey (1974) found support for this segregation based on the polysaccharide
composition of the seeds, which appear to be a reliable material for chemotaxonomic studies, as their composition is presumably quite stable over time and does
not depend on the vegetative state of the plant. Bailey recognized Calia, Sophora
(most of them representing the section Edwardsia) and Styphnolobium as separate
genera. The former two genera contain no galactomannans, which again distinguish them from Styphnolobium (S. affinis Torrey et A. Gray). Calia (S. secundiflora
Lag. ex DC) seeds contain starch or amyloid and polymers of galactose and
arabinose, and Sophora tomentosa seeds (section Sophora) contain galactosearabinose polymers. Recently, Professor A. S. Cerezo (personal communication)
confirmed that the Chilean Edwardsias S. macnabiana and S. macrocarpa also
contain arabinogalactans as the major seed polysaccharides.
Murray (1986) studied the electrophoretic
patterns of seed proteins from S.
microphylla
and the Chilean S. macrocarpa, finding them undistinguishable.
Nevertheless, this author had noted earlier that the protein pattern of Pisum was
very similar to that of S. macrocarpa (Murray and Porter, 1980), thus showing, in
our opinion, that seed protein electrophoretic
patterns are of little use to distinguish anything much below the family level. Therefore, although Murray (1986)
(Received 23 January 1996; accepted 20 May 1996)
725
Ft. C. PEfiA
726
AND
B. K. CASSELS
used these data to derive both Chilean species, S. macnabiana and S. macrocarpa,
from taxa growing in New Zealand, it seems more reasonable to refrain from
drawing any conclusion other than that seed protein electrophoresis shows strong
similarities between rather distantly related Papilionaceous genera. In line with
these conclusions,
Markham and Godley (1972) had found no support for the
segregation of S. macnabiana from Chile and S. microphylla from New Zealand
considering the similarity of their leaf flavonoid composition.
On the other hand, pioneering chemotaxonomic
studies carried out by Briggs
and Ricketts (1937) and Briggs and Mangan (1948) recognized S. microphylla
var. folvida, S. chathamica and material from Anawhata, New Zealand as separate
species on the basis of their seed alkaloid composition. Urztia and Cassels (1970),
using the same criterion and methodology, found the Chilean S. tetraptera sensu
Reiche (=S. macnabiana non S. microphylla) to differ from the previously studied
material from New Zealand. These conclusions, however, although they are the
direct precedent of the present chemical study, can no longer be considered
reliable due to the coarseness of the analysis (by vacuum distillation of the alkaloid
mixture) and the fact that they were not based on individual samples.
The present study has attempted to integrate morphological characters and seed
alkaloid composition, analyzed by GLC at the individual level, into a scheme providing a more secure foundation for our hypothesis regarding the South American
origin and affinities of the Chilean Sophora species (PeAa et a/., 1993). zyxwvutsrqponmlkjihg
M a t e ria ls a nd M e t hods
See Table 1. The following taxa were studied, all of which have specific status according to Allan (1961),
Yakovlev (1967), Green (1970) and lsely (1981): S. c/xysop/ty//i (Salisb.) Seem, (including S. ukokak
(Rock) Deg. et Scherff.), S. denudata Borv, S. femandeziana (Phil.) Skottsb., S. howinsula (Oliv.) Green, S.
hacnakana (Grah.) Skottsb., S. macrocaipa J.E. Sm., S. maiafkana (Phil.) Skonsb., S’micr~phylla Ait.
(including S. chathamica), S. prostrata Buchan., S. telraptera J. Mill., S. toromiro (Phil.) Skonsb. (all
included in sect. Eokardsia), S. tomentosa L., S. linearifo/ia Griseb., and S. rhynchocarpa Griseb. (sect.
Sophora), S. secundifora Lag. ex DC. (sect. Calia) and S. japonica (sect. Styphnolobium). Voucher specimens were deposited in the Herbarium of the Escuela de Quimica y Farmacia, Universidad de Chile
(SQF). Ripe seeds of all species from the localities indicated in Table 2 were used in extraction and
alkaloid determination procedures.
Extraction and anahsis. Air-dried seeds (one to five, depending on their size) from each individual plant
were coarsely ground and extracted with methanol until a fresh extract gave no Dragendorff reaction. The
concentrated residue was dissolved in 1% sulfuric acid, saturated with NaCI, and extracted with
TABLE
1. DATA
MATRIX
FOR
SOPHORA SPECIES
STUDIED:
9 INDICATES
UNDETERMINED
11
14
OR
INAPPLICABLE
CHARACTERS
1
2
3
4
5
MACRO
0000010100
CHRYS
110001111101111100221
DENUD
101001111100011101211
MACNA
111110111111111102200
FERN
101011010100001110221
MASAF
101111010101101199999
TORO
101011111111111112201
HOW
100011110111111999999
TETRA
110001210111111102200
MICRO
111111111111111112201
PROST
001111011100110100220
LINEA
010000000000000100020
RHYNC
011000000000109199999
TOMEN
000000000101009112101
SECUN
100000000030000100020
JAPON
100020000000100000000
Character
states 1-21
6
7
8
am given at the end of the paper.
9
10
12
13
00001100220
15
16
17
18
19
20
21
PHYLOGENETIC
RELATIONSHIPS
SOPHORA
AMONG
727
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONM
0110011110010+1001011
ooololloololl zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
I lo o
lo
100000
I I I I I10000
I lo
+::++++
+++++
++++++++++++
++++++++t+tt+++t++tttt+++++++t+
++++tt+++++tttttt+++++++ttt+t+++~++
I10000000
I
+ + +
tt
+
+
tl++ltt++++t+++ttt++++++tt+t~ttl+++
t
:
++:+++~+tttttt++
++x++++++++++ttt+++t+++:::::i++:tt
+
+ + +
:z
728
R . C . PEI~IA A N D B. K. C A S S E L S
+
I I Io
ooooo++
oool
I+÷÷+
Io++o
I I÷++
+
ooo
Io
+ooo+I
+
I
Ioi
+
+
I I I I÷o÷÷ooo
I
Ioo+ooo+oo+
+
+ + + +
+
+
+ + + + + + + + +
+ + + +
o o + + + + + + ÷ + + o o + + + +
++
÷+
+ + + + + + o o o + +
J
÷
J + + + +
+
+
++
++
+÷
+ ÷ + + ÷ ÷ + + ÷ + +
+÷
+ + ÷ + ÷ + + ÷ + + + + + + + + +
8
&
o
~A
uJ
z
0
~
-~
~
~zz
eq +
~+
!,
++
+
PHYLOGENETIC
RELATIONSHIPS
AMONG
729
SOPHORA
chloroform. The aqueous phase was made basic with sodium bicarbonate and partitioned into chloroform.
The chloroform extract of the alkaline aqueous phase was dried with sodium sulfate and concentrated to
dryness.
GLC was performed on a 20 m SE-30 capillary column at a constant temperature of 120°C for 3 min.
then with a constant rate of increase (G”C/min) up to 210°C. followed by a plateau lasting 20 min. The
N/P selective detector temperature was 250°C. Caffein was injected with each sample as internal standard.
Authentic samples of the following alkaloids were available: baptifoline, cytisine, N-methylcytisine,
matrine and 5hydroxymatrine (sophoranol). Anagyrine and rhombifoline were identified tentatively by
comparing relative retention times with published data.
The characters were polarized by the out-group technique. Character states are given at the end of the
paper.
Results
The cladogram
The data matrix shown in Table 1 was analyzed using Swofford’s
(Swofford,
1985) PAUP 2.4 program package. Styphnolobium was chosen as the ancestor
and the Lundberg option was applied. This technique minimizes the homoplasies
due to the out-group. The shortest tree is 68 steps long, consistency index 0.412,
for 21 characters and 16 taxa (Fig. 1).
A heuristic analysis requiring only 52 steps, 16 fewer than in the original tree,
was obtained using 16 characters, neglecting 5,7,9 and 13. This strategy supports
a New Zealand origin for the section Edwardsia.
Another heuristic development using only the morphological characters (l-l 6)
further reduced the length of the tree to 46 steps, 22 less than the original for 21
characters (Figs 1 and 2), consistency index 0.435 with a topology showing two
clades in the Edwardsia group (2,9),(7,8)3)
and (4,10)6)11)5).
The percentages of N-methylcytisine,
cytisine, matrine, sophoranol, anagyrine,
rhombifoline and baptifoline in the diferent specimens were subjected to principal
component analysis. Baptifoline was not recognized as a principal component in
this treatment. The distribution of specimens into the planes Fl x F2 (accounting
for 52.11% of the total variation) revealed three groups corresponding
to Calia,
Edwardsia and Sophora (Fig. 3). No additional groups could be distinguished
considering factor 3 (F3 = 10.72% of the variation).
MACROCARPA
CHRYSOPHYLLA
MACNABIANA
MICROPHYLLA
TOROMIRO
HOWINSULA
TETRAPTERA
DENUDATA
PROSTRATA
MASAFUERANA
FERNANDEZIANA
LINEARIFOLIA
RHYNCHOCARPA
SECUNDIFLORA
TOMENTOSA
165
13
JAPONICA
FIG.
1. A WORKING
MORPHIES
CLAOOGRAM
AS UNDERLINED
FOR SOPHORA
CHARACTERS,
SECTION
PARALLELISMS
EDWARDS/A,
AS 11. Chemical
SHOWING
REVERSION
characters:
17-21.
AS X, SYNAPO-
R. C. PEAA AND 6. K. CASSELS
730
zyxwvutsr
CHRYSOPHYLLA
TETRAPTERA
TOROMIRO
HOWINSULA
DENUDATA
MACNABIANA
MICROPHYLLA
6 8 15
MASAFUERANA
PROSTRATA
FERNANDEZIANA
MACROCARPA
TOMENTOSA
LINEARIFOLIA
RHYNCHOCARPA
SECUNDIFLORA
JAPONICA
FIG. 2. A REDUCED CLADOGRAM FOR SOPHORA SHOWING RELATIONSHIPS GROUNDED ON MORPHOLOGY
BETWEEN SECTIONS EDWARDS/A, SOPHORA (S. TOMENTOSA. S. LlA’EARlFOUA.S. RHYNCHOCARPA) AND STYPHNCLOBIUM SENSU TSOONG (SOPHORA SECUNDIFLORA AND S. JAPONICA).
zyxwvutsrqponmlkjihgfedcbaZYXWVU
\
el
g,,
i
bl
c2
\
b6
I2 f
f6g;
:6
c5
b8
g4
c9
e e8
\
a6
d3dl
c4
g7
t9
e6
dS
:7
c8
EDWARDSIA
82
Fi
FIG. 3. TWO PLANE Fl x F2 PRINCIPAL COMPONENT ANALYSIS FOR QUINOLIZIDINE ALKALOID CHARACTERS
CORRESPONDING TO SEEDS OF SECTIONS CALIA, EDWARDS/A ANDSOPHORA.
PHYLOGENETIC RElATIONSHIPS AMONG SOPHORA
731
Disc ussio n
A cladogram including seed alkaloid composition shows the greatest number of
homoplasies. Section Edwardsa is represented as a group with the following
characteristics: 6, 8, and 15 (extended flag, exerted stamens, and petals without
auricules). Character 15 is reversed in S. prostrata. Two branches are defined by
characters 5 and 7 (pubescence of the leaflets and flag/wings ratio): the tetraptera
clade, with leaflets hairy below, and the microphylla clade with the petals of equal
length (with the only exception of S. microphylla itself whose petals are unequal).
Characters 4 and 13 (short leaflets and fruits less pubescent) join S. prostrata to
this branch. Characters 11, 12 (color and number of seeds), and 13 join the tetraptera group, excluding S. denudata, to the microphylla group, through S. toromiro.
Character 3 (length of leaflets) is reversed in the tetraptera branch, when S.
denudata, with longer leaflets, is excluded. Characters 1, 3, 9, 10, and 14 (trees,
with little leaflets, winged and constricted fruits, and small seeds) delimit
Edwardsa, excluding S. macrocarpa.
In a biogeographical discussion, one could recognize as the South American
group of Sophora sect. Edwardsia, the continental S. macrocarpa and the species
from Juan Fernandez Islands, S. fernandeziana and S. masafuerana. Another branch
associates S. macnabiana and S. toromiro to S. microphylla. Finally, S. chrysophylla
and S. denudata, with some ancestral characters (orange or reddish seeds, low
relative cytisine content) are left out. Characters 11 and 18 (yellow-ocher seeds
and a greater content of cytisine), are common to the the New Zealand species,
including S. howinsula, S. toromiro and S. macnabiana, which appear to be more
advanced, with S. prostrata being a sister group for the latter clade.
Although principal component analysis does not show any difference between
the Edwardsias from the South American continent, S. macrocarpa and S.
macnabiana,
morphologically S. macrocarpa represents a transition between
sections Sophora and Edwardsia, with strong affinities to Argentinian species
belonging to the former section. Considering the flora of the Juan Fernandez
Islands, S. fernandeziana seems closely related to S. macrocarpa (2, 4, 6-8, IO19), but parallel evolution appears to be the simplest explanation for this.
Similarly, S. masafuefana is closer to the microphylla group. The palynological
characteristics of the first pair, S. femandeziana and S. macrocarpa (exine heterobrocate), contrast with the homobrocate exine of S. macnabiana, S. masafuerana,
and most of the insular Edwardsa species (Pena et al., 1993). Therefore, derivation
of the insular species, particularly S. femandeziana, from South American species
cannot be dismissed. Similar suggestions have been previously raised for the flora
of Juan Fernandez (see Hoffmann and Marticorena, 1987). The other species
from these islands, S. masafuerana, is here considered as derived from New
Zealand stocks related to the microphylla group. Chemical characteristics could
clarify this point.
In 1979, Godley had speculated ” . . . that S. microphylla arose in New Zealand,
derived some of its present variation from S. prostrata, and that the large-leafleted
southern species, S. chrysophylla of Hawaii, S. macrocarpa of central Chile, and S.
tetraptera-S. howinsula of New Zealand and Lord Howe Island, each distinct, are
older”. This view is supported by our own initial analysis including for the first time
morphological and chemical characters (Fig. 1). Nevertheless, when the seed
alkaloid composition is neglected, a more parsimonious cladogram is obtained
supporting the alternative hypothesis, i.e. that only S. macrocarpa is ancient and
close to S. tomentosa, S. linearifolia, and S. rhynchocarpa belonging to section
Sophora. S. macnabiana and all the insular species, including those from New
Zealand, thus appear to be derived from an ancestor common to them and S.
macrocarpa.
zyxwvutsrq
R. C. PEIilAAND 6. K. CASSELS
732
S. microphylla and S. macnabiana, in agreement with Godley’s observations, are
very closely related. S. prostrata, a taxon from New Zealand, retains some unspecialized characteristics, particularly auricules on the wing petals and an almost
unwinged pod. It is thus possible that it represents a relictual species in the South
Island. We refer to our 52 steps long heuristic cladistic treatment, in which this
species appears at the bottom of the Edkardsia branch of Sophora. Cockayne
(1912, fide Godley, 1979) regarded it as a juvenile form of S. microphyla. Quinolizidine alkaloid contents also separate S. prostrata from S. microphylla. Godley
(1979) speculated that S. microphylla had parents in an old hybrid swarm of S.
tetraptera and S. prostrate. Our data do not support this filiation: S. prostrata is
closer to S. tetraptera than to S. microphylla. The seed alkaloid composition of these
last two species is also closer to that of insular species, like S. fernandeziana or S.
toromiro, with higher contents of tricyclic cytisine derivatives.
AcknowledgemenbeWe
thank Mr Pedro Le6n and Dr Alan Walkowiak for excellent technical assistance. A sample of baptifoline was generously provided by Dr M. Silva and matrine, cytisine, methylcytisine and 5-hydroxymatrine (sophoranol) were supplied by Professor Rosa Negmte. Professor M. Castillo
kindly allowed access to the GLC equipment.
Character
states
Coding numbers, characters and states.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Life form: shrub: 0, tree: 1.
Number of leaflets: less than 15: 0,15-25: 1.
Length of leaflets: over 1 cm: 0, lass than 1 cm: 1.
Length/width of leaflets ratio: more than 2: 0, less than 2: 1.
Pubescence of the leaflets: face and bade 0, back only: 1.
Directfon of flag: erect: 0, extended: 1.
Flag/wings ratio: 1 (same length): 0, longer flag: 1, much longer. 2.
Exertion of stamens: no: 0, yes: 1.
Strangulation of the lomentum: no: 0, yes: 1.
Length of seed: 1 cm or more: 0, less than 1 cm: 1.
Seed color: brown: 0, ochec 1, orange: 2, reddish: 3.
Number of seeds per fruit: 26: 0,6 or more: 1.
Pubescence of the lomentum marked: 0, slight: 1.
Presence of wings on the fruit no: 0, yes: 1.
Presence of auricules on the petals: yes: 0, no: 1.
Presence of stipulae: yes: 0, no: 1.
Methylcytisine percentage: below 30%: 0, over 30%: 1.
Cytisine percentage: below 30%: 0, over 30%: 1.
Sparteina percentage: over 1%: 0, below 1%: 1.
Matrine percentage: below 30%: 0, over 30%: 1.
Methylcytisine/Cytisine ratio: less than 0.5: 0, more than 0.5: 1.
References
Allan, H. H. (1961) Leguminosae. In F/ore of New Zealand (Owen, K. E., ed.), Vol. 1, pp. 480627. Wellington, NZ.
Bailey, R. W. (1974) Galactomannans and other soluble polysaccharides in Sophora seeds. NZJ. Sot 12,
131-136.
Briggs, L. H. and Ricketts, J. (1937) Sophora alkaloids. Part I. The alkaloids of the seeds of Sophora
microphylla An. J. Chem. Sot., 1796-l 798.
Briggs, L. H. and Mangan. J. (1948) Sophora alkaloids. Part V. The alkaloids of the seeds of a possible
new species from Anawhata, New Zealand. J. Chem. Sot., 1889-l 891.
Godley, E. (1979) Leonard Cockayne and evolution. NZ J. Sot 17,197-215.
Gcdley, E. (1985) Path to maturity. NZJ. BOC 23.687-706.
Greene, P. S. (1970) Notes to the floras of Norfolk and Lord Howe Islands. J. Am. Arbor. 51.204-220.
Hoffmann, A. J. and Marticorena, C. (1987) La Vegetation de las lslas Oce&nicas Chilenas (The Vegetation of the Chilean Oceanic Islands). In l&s Oce&icas Chilenax Conocbnknto Cien&o y Necesidad
de Inves&aciones (Castilla, J. C., ed.), Univenidad Catblica de Chile, Santiago.
PHYLOGENETIC RElATIONSHIPS AMONG SOPHORA
733
Isely, D. (1981) Leguminosae of the United States. Ill Subfamily Papilionoideae: Tribes Sophoreae,
Podalyrieae, Loteae. Mem. NY Bat Gard 25, l-284.
Markham, K. R. and Godley, E. J. (1972) Chemotaxonomic studies in Sophora. An evaluation of Sophora
microphylla Ait. NZ J. Bat. 10,627-640.
Murray, D. R. (1986) Seed DispersaL Academic Press, Sydney.
Murray, D. R. and Porter, I. J. (1980) A comparative electrophoretic study of seed albumins from Sophora
microphylla and Pisum sadvum cultivar Greenfast (Leguminosae). Pt. Syst. Evol. 134, 207-214.
Palomino, H. G., Martinez, M. P., Bernal, M. C. and Sousa, S. M. (1993) Diferencias cromosbmicas entre
algunas especies de 10sgeneros Sophora L. y Styphnolobium Schott. Ann. Missouri Bot. Gard. 80, 284
290.
Peiia, R. C., Ituniaga, L., Mujica, A. M. and Montenegro, G. (1993) Analisis micromorfolrjgico de polen de
Sophora (Papilionaceae). Hipotesis filogenetica sobre el origen de la seccibn Edwardsa.
Gayana
(Botdnica)
50. 57-65.
Sousa, S. M. and Rudd, V. E. (1993) Revision del genera Styphnolobium (Leguminosae: Papilionoideae:
Sophoreae). Ann. Missouri Bot. Gard 80, 270-283.
Swofford, D. L. (1985) PAUP: Phylogenetic anahsis using parsimony user’s manual. Illinois Natural History
Survey, Champaign, Illinois.
Tsoong, P.-Ch. and Ma, Ch.-Y. (1981) A study of the genus Sophora Linn. Acta Phytotaxonomica Sinica
19,1-22 and 143-l 66.
U&a, A. and Cassels, B. K. (1970) Alkaloids of Sophora tetraptera, sensu Reiche. Phytochemistry 9,23662367.
Yakovlev, G. H. (1967) Systematical and geographical studies of genus Sophora and allied genera. Proc.
Leningrad Chem.-Phamwceut
Inst. 21,42-62.
Raul Peña