A BIOSYSTEMATIC STUDY OF THE GENUS SUTHERLANDIA Br. R.
by
DINEO MOSHE
DISSERTATION
presented in fulfilment of the requirements for the degree of
MAGISTER SCIENTIAE
in
BOTANY
at the
FACULTY OF NATURAL SCIENCES
of the
RAND AFRIKAANS UNIVERSITY
SUPERVISOR: PROF B-E VAN WYK
CO-SUPERVISOR: MRS M VAN DER BANK
DECEMBER 1998 OPSOMMING
'n Biosistematiese studie van die genus Sutherlandia (L.) R. Br., 'n relatief onbekende genus met verwarrende geografiese vorme, word aangebied. Die spesies van Sutherlandia is almal endemies aan Suidelike Afrika. Die spesies is naverwant en probleme rondom hul taksonomie word bespreek. Enkele morfologiese kenmerke wat nutting is om spesies te onderskei, word geIllustreer en in detail bespreek. Morfologiese inligting word gebruik om infrageneriese verwantskappe te ondersoek in 'n fenetiese ontleding van 51 geografies-geIsoleerde bevolkings.
Sutherlandia het tradisionele gebruike, hoofsaaklik as behandeling teen interne kankers en as 'n algemene tonikum. 'n Ondersoek van chemiese verbindings is gedoen en die resultate word geIllustreer en in tabelle aangebied. Die aard van hierdie studie het nie gedetailleerde mediese ondersoeke toegelaat nie, maar die medisinale waarde van Sutherlandia en die waargenome chemiese verbindings word toegelig. Daar word voorgestel dat die anti-kanker aktiwiteit hoofsaaklik toegeskryf kan word aan die hob vlakke van kanavanien, 'n nie-proteIen aminosuur, in die blare van die plant. Kanavanien, 'n analoog van arginien, is bekend vir sy aktiwiteit teen gewasvorming. Die waarde van die plant as 'n bitter tonikum hou waarskynlik verband met die teenwoordigheid van triterpendiede, sommige waarvan waarskynlik ook ander voordelige uitwerkings het.
Ensiem-elektroforese is gedoen om genetiese verwantskappe tussen die talryke streeksvorme van Sutherlandia te ondersoek. 'n Studie van 19 bevolkings het aangetoon dat hulle almal naverwant is en dat 'n meer konserwatiewe behandeling van die taksons nodig is. Die aantal taksons word dus verminder.
'n Volledige taksonomiese hersiening van die genus word aangebied. Die aantal spesies word verminder van ses na twee, naamlik S. frutescens en S. tomentosa. Eersgenoemde word in drie subspesies verdeel, naamlik subsp. frutescens, subsp. microphylla en subsp. speciosa. Sommige streeksvorme word beskryf en gelllustreer, maar hulle word nie as formele taksons erken nie. 'n Sleutel tot die spesies, subspesies en streeksvorme word voorsien, en die nomenklatuur, tipifisering, beskrywing en geografiese verspreiding van elke takson word verskaf. Die multi-dissiplinere benadering van hierdie studie het gelei tot 'n beter begrip van die morfologiese, chemiese en genetiese variasie in hierdie onbekende, maar potensieel waardevolle medisinale en sierplant.
II A BIOSYSTEMATIC STUDY OF THE GENUS SUTHERLANDIA R. Br. (FABACEAE, GALEGEAE)
SUMMARY
A biosystematic study of the genus Sutherlandia (L.) R. Br., a poorly studied genus with confusing geographical variants, is presented. The species of Sutherlandia are all endemic to southern Africa. The species are very closely related and problems regarding their taxonomy are discussed. A few morphological characters that are useful in distinguishing amongst species are illustrated and discussed in detail. Morphological data are used to investigate infrageneric relationships in a phenetic analysis of 51 geographically separated populations.
Sutherlandia has traditional medicinal uses, mainly as an anti-cancer treatment for internal cancers and as a general tonic. A survey of chemical compounds was done and the results are illustrated and presented in tables. The nature of this study did not allow detailed medical investigations, but the medicinal value of Sutherlandia and the compounds detected are highlighted. It is suggested that the anti-cancer activity can mainly be ascribed to the high levels of canavanine, a non-protein amino acid, in the leaves of the plant. Canavanine, an arginine analogue, is known for its antitumourigenic properties. The value of the plant as a bitter tonic is probably related to the presence of several triterpenoids, some of which may well also have other beneficial effects.
Enzyme electrophoresis was done to explore genetic relationships amongst the numerous regional forms of Sutherlandia. A study of 19 populations showed that they are all closely related and that a more conservative treatment of the taxa is called for. As a result the number of taxa is reduced.
A complete taxonomic revision of the genus is presented. The number of species is reduced from six to two, namely S. frutescens and S. tomentosa. The former is divided into three subspecies, namely subsp. frutescens, subsp. microphylla and subsp. speciosa. Some regional forms are described and illustrated, but these are not formally recognised as taxa. A key to the species, subspecies and regional forms is provided, and the. nomenclature, typification, description and geographical distribution for each of the taxa are given.
III The multidisciplinary approach of this study provided a better understanding of the morphological, chemical and genetic variation in this relatively poorly known but potentially valuable ornamental and medicinal plant.
IV TABLE OF CONTENTS
OPSOMMING
SUMMARY III
INTRODUCTION. 1
MATERIALS AND METHODS 2 2.1 Morphological characters. 2 2.2 Chemical compounds 4 2.3 Enzyme electrophoresis 10 2.4 Phenetic analysis 14
MORPHOLOGICAL CHARACTERS 15 3.1 Habit 15 3.2 Leaves 17 3.3 Inflorescence and flowers 20 3.4 Fruits and seeds 26
ENZYME ELECTROPHORESIS 31
CHEMICAL CHARACTERS 39 5.1 Alkaloids 40 5.2 Monoterpenoids 40 5.3 Flavonoids 41 5.4 Triterpenoids 43 5.5 Amino Acids 48 5.6 Pinitol 53
MEDICINAL VALUE OF SUTHERLANDIA 66 6.1 Historical background 66 6.2 Importance of triterpenoids 67 6.3 Importance of amino acids 68 Canavanine 68 Arginine and the nitric oxide pathway 69 y-Aminobutyric acid (GABA) 70 6.4 Pinitol 70
PHENETIC ANALYSIS 72 8. GENERAL CONCLUSIONS 82
9.TAXONOMY 84 9.1 Historical overview 84 9.2 The genus Sutherlandia 85 9.3 Key to the species and subspecies of Sutherlandia 86 9.4 The species and subspecies of Sutherlandia 87
ACKNOWLEDGEMENTS 105
10. REFERENCES 106
LJ CHAPTER 1
INTRODUCTION
The genus Sutherlandia R. Br., commonly known as cancer bush, comprises six species, all of which are endemic to southern Africa (Phillips & Dyer, 1934). The taxonomically significant characters to distinguish the species are the habit, the shape of the pods and the shape and pubescence of the leaflets. Sutherlandia species have both medicinal and horticultural uses. Leaf infusions of S. frutescens (L.) R. Br. are used to treat stomach, intestinal and uterine ailments. It is also used as a cough remedy, as a tonic (Smith, 1895; Van Wyk et al., 1997), to relieve eye ailments and chicken pox (Watt & Breyer-Brandwijk, 1962) and most interestingly, to treat internal cancers (Smith, 1895; Dykman, 1908; Van Wyk et al., 1997). The vernacular names refer to the use as a cancer cure, and there are several anecdotes to support this claim, but so far no published scientific evidence (Gabrielse, 1996). Canavanine, a non-protein amino acid with known antitumourigenic properties, has been extracted from seeds of S. frutescens (Bell, 1958) and further investigations seem worthwhile. The species were grown as ornamentals in England as early as 1683 (Curtis, 1792). To this day, Sutherlandia plants are popular in gardens in many parts of the world (Bailey, 1976; Mabberley, 1987; Griffiths, 1992).
Phillips & Dyer (1934) highlighted the problems associated with the taxonomy of this genus. It was still unclear whether there are six species, two, or perhaps only one. The taxonomy of Sutherlandia is confused because the taxa often grade into each other, but at most localities individual species are quite easy to recognise (Phillips & Dyer, 1934). They can only be distinguished by a combination of characters and even then this is not always conclusive (Schrire & Andrews, 1992). The existence of morphologically integrated populations also led Phillips & Dyer (1934) to identify the need for a genetic study of the genus Sutherlandia.
The aim of this study was to revise the taxonomy of all the confusing geographical variants of this relatively poorly known genus, by studying morphological and genetic variation at the population level. A further aim was to study chemical compounds of the genus, not only to find possible chemotaxonomic characters, but also to explore a possible rationale behind the traditional medicinal uses of Sutherlandia.
1 CHAPTER 2
MATERIALS AND METHODS
2.1 Morphological characters
Herbarium material The large collections of specimens in the National Herbarium, Pretoria (PRE) and the Rand Afrikaans University Herbarium (JRAU) were used. Apart from morphological data, the collections provided locality information for plotting distribution maps of the species and infraspecific taxa.
Fieldwork A field trip was undertaken to the Northern, Western and Eastern Cape and Free State Provinces to study plants in situ and to collect fresh material for analysis. Fresh flower, fruit, stem and leaf material was fixed in FAA (Sass, 1958). Voucher specimens were collected at each population. Photographs were taken to illustrate habitat, habit and gross morphology.
Morphology The herbarium specimens from different localities/populations were categorized into operational taxonomic units (OTU's). At least six specimens from each locality/population were selected where possible to represent each OTU (Table 2.1.1). All the taxa hitherto recognised were included. These are S. frutescens (L.) R. Br., S. frutescens var. incana E. Mey., S. tomentosa Eckl. & Zeyh., S. microphylla Burch. ex DC., S. montana Phill. & Dyer, S. humilis R. Br. and S. speciosa Phill. & Dyer. Specimens were studied using a WILD M3Z dissecting microscope and drawings were done with the aid of a camera lucida attached to a Zeiss compound microscope. CHAPTER 2 MATERIALS AND METHODS
Table 2.1.1 List of populations and specimens of Sutheriandia used to record morphological data for phenetic analysis. OTU's Taxa Herbarium voucher specimen(s) Locality / Population S. frutescens 1 Palmer 1 (JRAU) Olifantshoek S. frutescens 2 Palmer 2 (JRAU) Gamsberg S. frutescens 3 Palmer 6 (JRAU) Vanrhynsdorp S. frutescens 4 Palmer 23 (JRAU) Kleinsleutelfontein S. frutescens 5 Palmer 21 (JRAU) Swartberg Pass S. frutescens 6 Palmer 15 (JRAU) Chapman's Peak S. frutescens 7 Moshe, Van Wyk & De Castro 4 (JRAU) Worcester S. frutescens 8 Moshe, Van wyk & De Castro 5 (JRAU) Springbok S. frutescens 9 Moshe, Van wyk & De Castro 14 (JRAU) Aurora S. frutescens 10 Moshe, Van Wyk & De Castro 15 (JRAU) Saldanha S. frutescens 11 Moshe, Van Wyk & De Castro 18 (JRAU) Camp's Bay S. frutescens 12 Moshe, Van Wyk & De Castro 21 (JRAU) Fauresmith S. frutescens var. incana 1 Palmer 9 (JRAU) Aurora S. frutecens var. incana 2 Palmer 10 (JRAU) Aurora S. frutescens var. incana 3 Palmer 12 (JRAU) Blouberg Strand S. frutescens var. incana 4 Moshe, Van Wyk & De Castro 16 (JRAU) Blouberg Strand S. frutescens var. incana 5 Palmer 14 (JRAU) Hout Bay S. frutescens var. incana 6 Palmer 18 (JRAU) Struisbaai S. frutescens var. incana 7 Plowes 7262 (PRE) Port Nolloth S. humilis 1 Henrici 1809 (PRE) Fauresmith S. humilis 2 Palmer 17 (JRAU) Barrydale S. humilis 3 Palmer 22 (JRAU) Meiringspoort S. humilis 4 Palmer 25 (JRAU) Uniondale S. humilis 5 Retief & Reid 570 (PRE) Graaff - Reinet S. microphylla 1 Van Hoepen 1736 (PRE) Rustenburg S. microphylla 2 Palmer 5 (JRAU) Bitterfontein S. microphylla 3 Palmer 3 (JRAU) Khamiesberg S. microphylla 4 Palmer 7 (JRAU) Vanrhynsdorp S. microphylla 5 Moshe, Van Wyk & De Castro 12 (JRAU) Vanrhynsdorp S. microphylla 6 Palmer 27 (JRAU) Cradock 31 . S. microphylla 7 Palmer 24 (JRAU) Kleinsleutelfontein S. microphylla 8 Palmer 26 (JRAU) Unionpoort S. microphylla 9 Moshe, Van Wyk & De Castro 20 (JRAU) Leeuwberg Pass S. microphylla 10 Van Wyk 3804 (JRAU) Elliot S. microphylla 11 Palmer 16 (JRAU) Touwsrivier S. montana 1 Van Vuuren 1818 (PRE) Wolkberg S. montana 2 Van Wyk 2771 (JRAU) Golden Gate S. montana 3 Moshe, Van Wyk & De Castro 13 (JRAU) Piquetberg S. montana 4 De Castro 138 (JRAU) Ledger's Cave S. montana 5 Jacobsz 605 (PRE) Harrismith S. montana 6 Moshe, Van Wyk 7 De Castro 22 (JRAU) Reitz S. montana 7 Van Wyk 3800 (JRAU) Reitz S. speciosa 1 Germishuizen 4724 (PRE) Aus, Namibia S. speciosa 2 Oliver, Tolken & Venter 388 (PRE) Kodas Peak S. speciosa 3 Palmer 4 (JRAU) Khamiesberg S. speciosa 4 Hardy & Bayliss 1097 (PRE) Kamieskroon 47 S. tomentosa 1 Palmer 13 (JRAU) Blouberg Strand S. tomentosa 2 Moshe, Van Wyk & De Castro 17 (JRAU) Blouberg Strand S. tomentosa 3 Palmer 19 (JRAU) Witsand S. tomentosa 4 Van Wyk 3669 (JRAU) Still Bay S. tomentosa 5 Moshe, Van Wyk & De Castro 1 (JRAU) Koeberg Nature Reserve CHAPTER 2 MATERIALS AND METHODS
2.2 Chemical compounds Alkaloids Several alkaloids of the polyhydroxy type, such as swainsonine, salsolidine, cassine, castanospermine, smimovinine, sesbanine and sphaerophysine are known from members of the tribe Galegeae (Southon, 1994; Bruneton, 1995). As a result, it was expected that Sutherfandia would have similar compounds, and that the medicinal activity could possibly be ascribed to the alkaloids. However, preliminary studies by B-E. van Wyk (pers. comm.) using a modification of the method by
Ghosal et al. (1970) showed the total absence of alkaloids and other water-soluble bases in Sutherfandia leaves and seeds. The method used was the following: Pack a small column with Dowex 50W resin (H+ form) using water. Elute column with 1N KOH. Rinse column with distilled water until pH is 5.5 (same as tap water). Elute column with 4N HCI (2 column volumes). Rinse with distilled water until pH is 5.5. Rinse with methanol (11/2 column volumes). Methanolic extracts of air-dried leaves and seeds (1 g in 25 ml methanol) were slowly eluted through the column. Seeds of Castanospermum australe, a well-known source of castanospermine, were used as a positive control. The column was stripped using Water - 25% Ammonia - Methanol (8:1:1) The solvent was then taken to dryness using a rotatory evaporator. In the Sutheriandia samples no alkaloids could be detected whereas C. australe seeds typically gave high yields of castanospermine.
Monoterpenoids Air-dried leaf and stem material was steam distilled using a Clevenger apparatus and the ether collected was analysed by gas chromatography (GC).
Flavonoids Extraction: Approximately 1 g of leaf or seed material was pulverised into powder using sand with a mortar and pestle. About 10 ml of 96% Me0H was added and left to extract for 30 minutes. Two methods were employed for the survey of flavonoids, Thin Layer Chromatography (TLC) and High Perfomance Liquid Chromatography (HPLC).
4 CHAPTER 2 MATERIALS AND METHODS
Thin Layer Chromatography (TLC): Extracts were filtered through celite and cotton wool and applied directly to a silica gel (Merck F560) plate and developed in system 1: Chloroform - Methanol - Water - Acetic acid (6:3:0.8:0.6). Flavonoids were exposed to long wavelength UV (366 nm) for detection on the TLC plates. The plates were sprayed with 5% ethanolic H2SO4 and then with 1% ethanolic vanillin, after which they were developed in the oven at 110 °C for 5-10 minutes.
High Perfomance Liquid Chromatography (HPLC):
Methanolic extracts of dry leaf material were passed through 018 cartridges to remove substances of high retention time. They were then mixed with water (1:1) and subjected to a Beckman System
Gold HPLC analyser (Phenomex IB -Sil column C18 reverse phase column, 0.5 i.trn particle size, 250 mm x 4.6 mm internal diameter, flow rate 1 ml min"'. 20 111 sample loop). The solvent system comprised a 30 -100 % linear gradient of methanol in 1% acetic acid - water over 32 minutes. Detection was by diode array detector using two channels (A: 260-300 nm; B: 295-365 nm).
Triterpenoids Analytical method: Dried leaf material (1 g) was ground with mortar and pestle, pulverised into powder and boiled in water in a waterbath for 60 minutes at 100 °C. The extracts were cooled and filtered through celite
and cotton wool and passed through C18 cartridges. Silica gel (Merck F560) plates with different eluent systems were used. The most useful system was: Chloroform - Methanol - Water - Acetic acid (6: 3: 0.8: 0.6). The spray reagent for detection was 5% ethanolic H2SO4 and 1% ethanolic vanillin. The plates were sprayed and developed in the oven at 110 °C for 5-10 minutes.
Isolation of the two major compounds: A bulk extract (22 g) was prepared by Soxhiet extraction (36 h) of a sample of leaves and stems (1069 g dry weight) from Sutherlandia microphylla (Voucher specimen: B-E. van Wyk 3667 in JRAU). Analysis of the extract was done by TLC using UV detection. Compounds were detected under UV light at 254 nm as brown spots on plates with fluorescent indicator. Isolation of detected compounds from this extract was done by column chromatography (CC) using silica gel as adsorbent and solvent system 1. Two impure fractions containing two unknown compounds X1 and X2 (8.3 g) (Fig. 5.1.1 in Chapter 5), and unidentified flavones Y1-Y3 (3.85 g) (Fig. 5.1.1 in Chapter 5) respectively, were collected. The X1 and X2 mixture was re-chromatographed using flash chromatography, flash silica as adsorbent and the eluent system: Chloroform — Methanol — Water — Acetic acid (6: 3: 0.8: 0.6). The two isolates X1 (67 mg) and X2 (62 mg) were analysed by mass spectrometry (MS). CHAPTER 2 MATERIALS AND METHODS
Amino Acids Extraction and sample preparation: Approximately 1 g of leaf or seed material was pulverised into powder using sand with a mortar and pestle. About 10 ml of distilled H 2O was added and left to extract for 30 minutes at room temperature.
The free amino acid fraction was filtered through celite and cotton wool and passed through 0 18 cartridges.
Thin Layer Chromatography (TLC): Extracts were applied directly to silica gel (Merck F 560) plates and developed in Phenol — H2O (3:1 v/v). Amino acids were revealed using 0.2 % ninhydrin in acetone as spray reagent. For identification purposes, small quantities of authentic reference compounds were co- chromatographed with leaf extracts in the above solvent system. After 2 hours of development, the plate was air-dried, sprayed with ninhydrin and then developed in the oven at 110 °C for 5-10 minutes.
High Perfomance Liquid Chromatography (HPLC): HPLC analyses of amino acids were done at the Department of Chemical Pathology, Medical University of Southern Africa (Medunsa). I thank Mr. Pat Smith of this department for his expert guidance and for doing the analyses.
Extraction and sample preparation: Leaf or seed material (1 g) of each species was ground into powder using a pestle and mortar. Distilled water (5 ml) was added and left to extract for 60 minutes. The samples were filtered through
celite and cotton wool and finally through pre-filter 018 cartridges. The final filtrate was then evaporated to dryness by blowing with nitrogen gas. Once dried, they were re-constituted in 1000 S.D.B. (Sample Dilution Buffer pH 2.2 - Beckman 338084) and vortexed to ensure homogeneity. A
further 1001.1.1 S.D.B., together with 501.11 Internal Standard, was added. These samples were then
re-vortexed and filtered through Millex-HV 0.45 .Lm filter units. The pH of each sample was checked and if necessary adjusted to pH 2.0-2.2 (using 0.6N NCI).
Analysis: The samples were labelled and placed in chronological order on the carousel of the AA6300 amino acid analyser. They were then run in conjunction with a calibrant containing 43 known amino acids and spiked with canavanine.
C.,
6 CHAPTER 2 MATERIALS AND METHODS
The completed chromatograms were examined and where necessary, dilutions were made and diluted samples re-run in order to obtain the required resolution of all peaks for easier identification and quantification.
All samples and calibrant volumes injected were 20 IA. The AA6300 amino acid analyser was connected to a Hewlett-Packard 3390A Integrator (Hewlett-Packard 1000 N.F Circle Boulevard, Oregan 97330) which integrated the peak areas and calculated the results — run in series with Beckmann System Gold.
Apparatus: Amino acid analysis was performed using the Beckmann - High Performance Amino Acid Analyser, AA6300 (Fig. 2.1.1), incorporating ion-exchange chromatography using a 10 cm Lithium column (Beckmann 338051) [Beckmann Instruments Inc. Brea, California].
The system is capable of separating 40 to 53 different amino acids when using a four buffer system of varying pH values and ionic strengths with Lithium as a counter-ion. A temperature gradient is necessary in order to obtain optimum resolution of all the amino acids used in the standard calibrant. It contains 43 amino acids including the "Internal Standard".
A calibration of the instrument and column was done using a calibrant solution made up from a mixture of the following: Hydrolysate amino acid standard (Beckman 338088) Amino acid supplement AN+ (Beckman 338156) Amino acid supplement B+ (Beckman 338157) Glutamine [m.w. 1446.1] 2.5 ilmol/m1 Asparagine [m.w. 132.1] 2.5 1.tmol/m1 Tryptophan [m.w. 204.2] 2,5 plmol/m1 Canavanine [m.w. 176.2] 2.5 [trnol/m1 [5-2-Aminoethyl-L-cysteine HCI [m.w. 200.6] 2.5 i.tmol/m1 CHAPTER 2 MATERIALS AND METHODS
Figure 2.1.1 An automated amino acid analyser used for studying amino acids in leaf and seed samples of Sutherlandia.
To prepare the standard calibrant solution: 200 gl of above solutions were accurately pipetted into a 5.0 ml volumetric flask and made up to the mark with Citrate buffer pH 2.2.
Preparation of working calibrant: Into a 2.0 ml capped micro-tube (Eppendorf), 1000 gl of above standard calibrant solution, 100 gl S.D.B and 50 gl Internal Standard were accurately pipetted and vortexed. A 1:1 dilution was then made (using S.D.B or Citrate buffer) and used as the working calibrant.
Analytical method:
After the column had been equilibrated and correctly calibrated, the prepared samples were injected in sequence. The samples were interspersed with calibrants in order to verify any noticeable shift in retention times of any amino acid peaks and to facilitate easier recognition and identification of the amino acids eluted from the various samples. Using the above mentioned methodology, the order in which the amino acids elute out is as in Table 5.5.2 in Chapter 5.
Having obtained the respective chromatograms, the various amino acids were identified according to the retention times. The recorded values obtained from the integrator printouts were used to calculate the amounts of each amino acid present as mg/g of plant material.
8 CHAPTER 2 MATERIALS AND METHODS
Pinitol Extraction and sample preparation: Air-dried leaf material (5.6 g) of S. microphylla collected from Cradock was used for the pinitol survey. The material was pulverised into powder using sand with a mortar and pestle. About 20 ml of 80% ethanol was added and left to extract for 30 minutes at room temperature. The ethanolic extract was filtered through celite and cotton wool and passed through 018 cartridges.
Thin Layer Chromatography (TLC): About 51.d of the ethanol extract was applied directly to silica gel (Merck F5 60) plates and developed in Acetonitrile — Carbon disulphide — Water — Formic acid (85:5:10:0.5). For identification purposes, small quantities of authentic pinitol were co-chromatographed with leaf extracts in the above solvent system. After development, the plate was air-dried, sprayed with concentrated chromic acid then passed over an open flame. Pinitol appeared as a dark brown to black spot on the plate.
High Perfomance Liquid Chromatography (HPLC): HPLC analyses of pinitol were done at the Department of Botany (RAU) by Mr Alvaro Viljoen using a refractive index detector (Shimadzu C R6A) and a "Waters sugarpack" column, with acetonitrile and water (93:7) as eluent. The dried ethanolic extract (168.3 mg) of Sutherlandia was separated between chloroform and water. The water fraction was filtered through celite and cotton wool and through C18 cartridges. The extract was then re-dissolved with 0.2 ml water and 201.1l was injected. The chromatograms are given in Figure 5.6.1 in Chapter 5.
Authentic standard preparation: The pinitol standard was provided by Prof. Fanie van Heerden (Chemistry Department, RAU).
Approximately 1.4 mg were dissolved in 175 µl distilled water of which 20 µl was injected.
2.3 Enzyme electrophoresis Plant material - Leaf samples of Suthertandia individuals were collected from 19 natural populations: five populations of S. frutescens (N=79), two of S. frutescens var. incana (N=49), two of S. tomentosa (N=76), six of S. microphylla (N=76), two of S. montana (N=40), one of S. humilis (N=20), one of S. speciosa (N=20) and outgroups Astragalus atropilosulus (Hochst.) Burge subsp. burkeanus (Harv.) Gillett (N=2) and a Lessertia sp. (N=5). The latter is a new species, not yet described. Voucher specimens of all these populations were collected and deposited in the Rand Afrikaans University herbarium (JRAU). CHAPTER 2 MATERIALS AND METHODS
Sampling: The nineteen populations sampled are geographically separated and were chosen because they allowed comparisons at population, variety and species levels. They also represent the full range of morphological and geographical variation within the genus (Fig. 2.3.1). Sample size for the ingroup was subject to availability. Sutherlandia populations usually comprise a relatively small number of individuals growing in an area of less than 500 m 2 . Populations and species investigated, their sources and voucher specimens are denoted in Fig. 2.3.1 and listed in Table 2.3.1. The identification of the populations is based on considerable field experience of B-E. van Wyk.
Figure 2.3.1 Localities sampled for enzyme electrophoresis (populations numbered as in Table 2.3.1).
Procedure :
Young leaves were collected from actively growing shoots, placed in cryotubes and immediately k_. submerged in liquid nitrogen (-196 °C) and transported to the laboratory.
Sample preparation:
The cryotubes were retrieved from the nitrogen tank and two grams of leaf tissue was manually ground with a mortar and pestel in 3 ml of Tris-HCI extraction buffer (pH= 7.5), as described by Soltis et al. (1983). The supernatant was then centrifuged at 4000 g for five minutes. The samples were then applied directly onto paper wicks.
10 CHAPTER 2 MATERIALS AND METHODS
Table 2.3.1 Source, location and number of individuals sampled per population of Sutheriandia studied.
Population Taxa Voucher Specimens Abbreviation Source Grid Ref. N
1 Astragalus atropilosulus De Castro 11 A Randburg 2529 AC 1 subsp. burkeanus
2 Lessertia sp. Moshe, Van Wyk & De Castro 6 L Pofadder 2918 CB 5
3 S. frutescens Moshe, Van Wyk & De Castro 14 SF1 Aurora 3318 DC 20
4 S. frutescens Moshe, Van Wyk & De Castro 22 SF2 Fauresmith 2925 AC 20
5 S. frutescens Moshe, Van Wyk & De Castro 15 SF3 Saldanha 3317BC 5
6 S. frutescens Moshe, Van Wyk & De Castro 5 SF4 Camps Bay 3318 CD 4
7 S. frutescens Moshe, Van Wyk & De Castro 4 SF5 Worcester 3319 CC 19
8 S. frutescens var. incana Van Wyk 3668 SV1 Pearly Beach 3419 CB 20
9 S. frutescens var. incana Moshe, Van Wyk & De Castro 16 SV2 Blouberg Strand 3318 CD 20
10 S. humilis Moshe, Van Wyk & De Castro 19 SH Uniondale 3323 CA 20
11 S. microphylla Moshe, Van Wyk & De Castro 12 SM1 Vanrhynsdorp 3118 BC 20
12 S. microphylla Moshe, Van Wyk & De Castro 13 SM2 Vanrhynsdorp 3118 BC 3
13 S. microphylla Moshe, Van Wyk & De Castro 20 SM3 Leeuwberg Pass 3218 BC 20
14 S. microphylla Van Wyk 3667 SM4 Colesberg 3024 DA 20
15 S. microphylla Van Wyk 3801 SM5 Tweeling 2927 CA 7
16 S. microphylla Van Wyk 3804 SM6 Elliot 3028 DD 7
17 S. montana Moshe, Van Wyk & De Castro 10 SMO1 Piquetberg 3218 DD 20
18 S. montana Moshe, Van Wyk & De Castro 23 SMO2 Reitz 2728 CD 20 \ • *;. 19 S. speciosa Moshe, Van Wyk & De Castro 8 SS Khamiesberg 3017 BB 20
20 S. tomentosa Moshe, Van Wyk & De Castro 1 ST1 Koeberg Nature Reserve 3118 AC 33
21 S. tomentosa Moshe, Van Wyk & De Castro 6 ST2 Still Bay 3421 AD 43
11 CHAPTER 2 MATERIALS AND METHODS
Technique:
Horizontal starch gel electrophoresis was utilised as described by Van der Bank et a/. (1995) (Fig. 2.3.2). Twelve percent starch (Sigma: S-4501) gels were used for this study. With lack of published information regarding the isozymes and allozymes of the genus, as many as possible enzymes that would provide informative results were identified. In a preliminary study using one population of S. tomentosa and two populations of S. microphylla, various buffer system and enzyme combinations to determine the best resolution and satisfactory results were used. From this study, both monomorphic and polymorphic enzymes were identified. Only polymorphic enzymes were then analysed in subsequent studies. Sixteen of the 19 enzyme systems used produced interpretable banding patterns (Table 2.3.2).
Electrophoresis:
Four gels could be run simultaneously (using a sample once only and staining for up to 19 enzymes) to minimise the effect of thawing. Wicks were removed after 15 minutes of electrophoresis and plastic film ("Glad Wrap") was used to cover gels during electrophoresis to prevent dehydration. Electrophoresis was conducted at 4 °C for four hours. After electrophoresis, the gel was removed from the cool slab, cut into a rectangular shape by trimming away excess starch from the sides of the gel. The gel was then cut into four slices by pulling a tightened wire horizontally through the gel slab on a slicing bed. Gel slices were placed on staining trays and containers for staining. Isozymes were detected in situ through the use of specific activity stains. About 19 enzymes were stained for and 32 of the 34 enzyme coding loci produced interpretable banding patterns (Table 2.3.2).
12 CHAPTER 2 MATERIALS AND METHODS
Figure 2.3.2 Enzyme electrophoresis laboratory setup, showing horizontal gel slabs during electrophoresis.
Table 2.3.2 List of enzymes studied in Sutherlandia. Locus abbreviations, enzyme commission numbers (E.C. No.), buffers and their pH use are listed after each enzyme.
Enzyme Locus E.C. No Buffer pH
Adenylate kinase AK 2.7.4.3 RW 8.0 Esterase EST-1, -2 3.1.1.- RW 8.0 'EST-3' Glyceraldehyde-3-phosphate dehydrogenase *GAPDH-1, -2 1.2.1.12 HC 5.7 Glucose-6-phosphate isomerase GPI-1, -2 3.5.1.9 RW 8.0 Isocitrate dehydrogenase IDH 1.1.1.42 HC 6.5 Leucine aminopeptidase *LAP 3.4.11.1 MF 8.6 Malate dehydrogenase MDH-1 1.1.1.37 LiOH 8.1 i s *MDH-2, -3 Menadione reductase *MNR-1, -3 1.6.99.- MF 8.6 MNR-2 Manose-6-phosphate dehydrogenase MP! 5.3.1.8 RW 8.0 Peptidase, substrate: Leucine-alanyl 'PEP-C-1,-2 3.4.-.- MF 8.6 Leucylglycylglycine 'PEP-B-1,-2,-3 Leucyl-tryosine 'PEP-S-1,-2 Peroxidase *PER-1, -3 1.11.1.7 MF 8.6 PER-2 6-Phosphogluconate dehydrogenase PGDH-1, -2 1.1.1.44 HC 6.5 Phosphoglucomutase PGM-1, -2 5.4.2.2 MF 8.6 Superoxide dismutase SOD-1, -3 1.15.1.1 RW 8.0 * = Monomorphic loci MF - a continuous buffer (pH = 8.6) system (Marked & Faulhaber 1965) RW - a discontinuous buffer (electrode pH = 8.0; gel pH = 8.7) system (Ridgway et al., 1970) LIOH - a discontinuous buffer (electrode pH = 8.1; gel pH = 8.3) system (Kephart, 1990 ) HC - a discontinuous buffer (elecrode pH = 6.5; gel pH= 6.5) system (Kephart, 1990)
13 CHAPTER 2 MATERIALS AND METHODS
Interpretation:
Genetic interpretation of enzyme banding patterns was based on the subunit structure of the enzymes (Gottlieb, 1982; Kephart, 1990). Locus nomenclature followed Soltis & Soltis (1989), Hillis & Moritz (1990) and Shacklee et al. (1990). Locus abbreviations, monomorphic loci, enzyme commission numbers and buffer systems are given in Table 2.3.2.
Statistical procedure and analysis:
Statistical analysis of allozyme data was executed using B!OSYS-1 (Swofford & Selander, 1981). Allele frequency data for populations of the same species were combined and DISPAN (Ota, 1993) was used for the construction of dendrograms using neighbour-joining and bootstrap tests (1000 replications). Phylogenetic relationships were also determined using Nei's (1978) genetic distance values and UPGMA dendrograms were constructed (Fig. 4.4 to 4.5 & 4.7 in Chapter 4).
2.4 Phenetic analysis
Material examined included representative samples of 51 geographically isolated provenances. Voucher specimens are located at JRAU herbarium and are given in Table 7.1. Authorities of the names of the species are given in Chapter 2 (page 2). The character and character states used in this study are listed in Table 7.2. For the phenetic analysis, a matrix of the 51 provenances as OTU's and 25 characters were used (Table 7.3). The NTSYS-pc 2.01 program was used to generate phenograms (illustrated in Figs. 7.1 to 7.4 in Chapter 7).
14 CHAPTER 3
MORPHOLOGICAL CHARACTERS
3.1 Habit The species of the tribe Galegeae are multi-branched suffrutescent shrubs and subshrubs (Polhill, 1981). Phillips (1926) and Harvey (1862) described Suthedandia species as canescent shrubs. The members of the genus Sutherlandia display considerable variation as regards habit and this character has indeed been used as diagnostic for S. humilis, a small shrublet of less than 0.2 m high (Phillips & Dyer, 1934). The range of variation in habit is shown in Figure 3.1.1. It varies from prostrate and less than 0.2 m high as in S. humilis (a in Fig. 3.1.1) to large and erect reaching 2.5 m as in S. microphylla (b in Fig. 3.1.1). The plants may also be small and procumbent, about 0.3 m high as in some forms of S. frutescens (c in Fig. 3.1.1) or procumbent and spreading, about 0.5 m high and 1 m wide as in S. speciosa (d in Fig. 3.1.1). While most plants are single-stemmed, branching often occurs close to the ground, giving a shrubby appearance. S. humilis is particulary noted for its multi- branched form, often with a number of main stems (Phillips & Dyer, 1934).
Similarities and differences in habit of Sutherlandia species are summarised in Table 3.1.1 and illustrated in Figures 3.1.1 a to d.
Table 3.1.1 Different types of habit found in Sutherlandia species.
Species Height Habit
S. frutescens (Fig. 3.1.1 c) 0.3-1.5 m erect to procumbent
S. frutescens var. incana 0.5-0.7 m erect to procumbent
S. humilis (Fig. 3.1.1 a) less than 0.2 m procumbent to prostrate
S. microphylla (Fig. 3.1.1 b) 0.8-2.5 m erect
S. montana 0.8-1 m erect
S. speciosa (Fig. 3.1.1 d) 0.5 m high, 1 m wide procumbent
S. tomentosa 0.6 m erect to procumbent
V : t 15 CHAPTER 3 MORPHOLOGICAL CHARACTERS
The habit of Sutherlandia species is taxonomically useful to some extent, but there is considerable variation and the character is only of value if used in combination with others.
(a)
(c) (d)
Figure 3.1.1 Variation in the habit of Sutherlandia species, varying from prostrate (a), erect (b) erect to procumbent (c) to procumbent (d). a, S. humilis (Uniondale); b, S. microphylla (Meiringspoort); c, S. frutescens (Worcester); d, S. speciosa (Khamiesberg).
16 CHAPTER 3 MORPHOLOGICAL CHARACTERS
3.2 Leaves Leaves of the Papilionoideae are often imparipinnate (Bentham, 1865; Dormer, 1945, 1946; Polhill 1981a). The same is true for the tribe Galegeae (Bentham, 1865; Polhill, 1981a). In the genus Sutherlandia, the leaves are imparipinnate, entire, petiolate and multi-foliate (Harvey, 1862; Phillips, 1926). In their revision of the genus Sutherlandia, Phillips & Dyer (1934) found that the morphological characters of the leaves rarely exhibited any outstanding or constant features which could be used for grouping the specimens into distinct species. They used leaflet shape and pubescence to a limited extent. In this study, these two characters were also found to be of some taxonomic value (see below).
Leaflet number Phillips & Dyer (1934) found the number of leaflets to be of no taxonomic value in distinguishing Sutherlandia species. The number of leaflet pairs is fairly constant amongst populations and species and also within populations (Fig. 3.2.1). There is considerable overlap between species, however, so that the character has limited diagnostic value. It is interesting to note that S. humilis has the lowest number of leaflets and that the two pubescent taxa (S. frutescens var. incana and S. tomentosa) have the highest number. The intermediate position of the var. incana between typical S. frutescens and S. tomentosa is thus also seen in the leaflet number.
12
10
irs a p t 8 fle a le f o
ber 6 Num
4
2 SP/ SF S1-1 SM SMO SS ST N = 7 12 5 11 5 4 5
(Taxa: SFV = S. frutescens var. incana, SF = S. frutescens, SH = S humilis, SM = S. microphylla, SMO = S. montane, SS = S. speciosa, ST = S. tomentosa)
Figure 3.2.1 Variation in the number of leaflet pairs of Sutherlandia species. The range and mean value are shown for each taxon. The number of populations sampled (N) is given below each species. For each population, between three and seven specimens were sampled, depending on availability.
17 CHAPTER 3 MORPHOLOGICAL CHARACTERS
Leaflet shape Leaflet shape of Sutherlandia plants varies from linear to oblong to ovate (SF, SMO, SH, SM, SFV and SS in Fig. 3.2.2), or obcordate and deeply emarginate (ST in Fig. 3.2.2). All the species have linear to oblong-ovate leaflets except for S. tomentosa which has obcordate or deeply emarginate leaflets. This is a diagnostic feature for the species. In some populations of S. frutescens var. incana, the leaflets (SFV3 in Fig. 3.2.2) may be similar to those of S. tomentosa but they are only slightly notched and never deeply emarginate as in the latter.
SF SH
SM SFV ST
Figure 3.2.2 Variation in the leaflet shape and pubescence of Sutherlandia species. Leaflets are numbered as follows: SF = S. frutescens, 1 = Kleinsleutelfontein (Palmer 23 (JRAU)], 2 = Olifantshoek [Palmer 1(JRAU)J, 3 = Bakoven to Chapman's Peak [Van Wyk & De Castro 367(JRAU)]; )]; SMO = S. montane, 1 = Golden Gate [Van Wyk 2771 (JRAU)], 2 = Piquetberg, De Hoek [Moshe, Van Wyk & De Castro 13 (JRAU)]; SH = S. humilis, 1 = Uniondale [Moshe, Van Kt)* & De Castro 19 (JRAU)], 2 = Barrydale [Palmer 17 (JRAU)], 3 = Meiringspoort [Palmer 22 (JRAU)]; SM = S. microphylla, 1 = Bitterfontein [Palmer 5 (JRAU)], 2 = Kamiesberg [Palmer 3 (JRAU)], 3 = Vanrhynsdorp [Moshe, Van Wyk & De Castro 12 (JRAU)]; SFV = S. frutescens var. incana, 1 = Hout Bay [Palmer 14 (JRAU)], 2 = Aurora [Palmer 9 (JRAU)], 3 = Blouberg Strand [Moshe, Van Wyk & De Castro 16 (JRAU)];SS = S. speciosa, 1 = Kamiesberg [Moshe, Van Wyk & De Castro 11 (JRAU)], 2 = Kamieskroon (Hardy & Bayliss 1097 (PRE)); ST = S. tomentosa, 1 = Blouberg Strand [Moshe, Van Wyk & De Castro 17 (JRAU), 2 = Witsand [Palmer 19 (JRAU)];
18 CHAPTER 3 MORPHOLOGICAL CHARACTERS
Leaflet pubescence The abaxial surface of leaflets of Sutherlandia species are invariably pubescent but the adaxial surface differs between species. Pubescence is a significant diagnostic character for S. tomentosa (ST in Fig. 3.2.2). The leaflets are densely covered with silvery white tomentose hairs, completely obscuring the surface of the lamina. In some forms of S. frutescens (hitherto known as S. frutescens var. incana), the upper leaflet surfaces may also be sparsely tomentose, but the leaf surface epiderm is always visible between the hairs.
Petiole length The petiole length varies amongst individuals within a population and among species but does not show significant differences between species (Fig. 3.2.3). S. microphylla and S. montana are particularly variable. It is interesting to note that the two coastal taxa (S. tomentosa and S. frutescens var. incana) generally have shorter petioles than the other species. The large variation precludes this character from being used.
20
15
E
10
5
SFV SF SH SM SMO SS ST N = 7 12 5 11 5 4 5
(Taxa: SFV = S. frutescens var. incana, SF = S. frutescens, SH = S humilis, SM = S. microphylla, SMO = S. montana, SS = S. speciosa, ST = S. tomentosa)
Figure 3.2.3 Petiole length of Sutherlandia species. The number of populations sampled (N) is given below each taxon. For each population, between three and seven specimens were sampled, depending on availability.
Petiolule length The petiolule length varies between species but no obvious taxonomic trend could be observed. The petiolule of S. tomentosa (ST), however, is generally shorter than that of the other species (Fig. 3.2.4).
19 CHAPTER 3 MORPHOLOGICAL CHARACTERS
1.5
1
E E
0,
0.5
0 I 1 I 1 I I SFV SF SH SM SMO SS ST N = 7 12 5 11 5 4 5
(Taxa: SFV = S. frutescens var. incana, SF = S. frutescens, SH = S. humilis, SM = S. microphylla, SMO = S. montana, SS = S. speciosa, ST = S. tomentosa)
Figure 3.2.4 Variation in petiolule length in Sutherlandia species. The number of populations sampled (N) is given below each taxon. For each population, between three and seven specimens were sampled, depending on availability.
3.3 Floral characters
Inflorescence Racemes are the most common type of inflorescence among Fabaceae (Tucker, 1987). In the tribe Galegeae, the flowers are borne on short axillary racemes (Hutchinson, 1964; Polhill, 1981a). In Sutherlandia, the inflorescence structure is therefore typical short racemes as in the rest of the tribe (Fig. 3.3.1) and the number of flowers varies from two to seven per raceme (Phillips & Dyer, 1934). No taxonomically useful discontinuities could be found in the inflorescence structure of Sutherlandia.
Flowers The papilionoid flower differs from other subfamilies of the Fabaceae in the adaxial petal being outside in bud forming a flag at anthesis and the lower petal housing the fertile parts (Polhill, 1981). Zygomorphic, bilateral papilionoid flowers represent an ecological advancement relative to the Caesalpinioideae and Mimosoideae (Polhill, 1981).
20 CHAPTER 3 MORPHOLOGICAL CHARACTERS
Flower structure Bee flowers take the form of narrowly or widely tubular flowers (Weberling, 1989; Faegri & Van der Pijl, 1979). This is true for many papilionaceous flowers with a landing platform for the flower visitor and which often have some special mechanism for protecting nectar. Bird flowers are common in tropical and subtropical zones and are distinguishable from bee flowers by the absence of a landing platform. Bird-pollinated flowers are tubular and comprise about 320 species in the Cape flora (Rebelo, 1987).
The bird-pollinated flower has a peculiar structure compared to a bee flower (Arroyo, 1981; Faegri & Van der Pjil, 1979). The standard petal (1 in Fig. 3.3.2) is the largest and most conspicuous and overlaps all the others. The lower two petals are united into a large keel (4 in Fig. 3.3.2), which encloses the stamens and the gynoecium and functions along with two lateral wings in bee flowers as landing platform for insects visitors. The wings are often strongly reduced in bird flowers and this is indeed also true of Sutherlandia (2 a to d in Fig. 3.3.2).
The flowers of Sutherlandia species have a tubular structure (a to g in Fig. 3. 3.3) and are pollinated by sunbirds such as the malachite sunbird, Nectarinia famosa, commonly found in southern Africa (Skead, 1967; Arroyo, 1981). The flowers are bright red or scarlet (a in Fig. 3.3.1) usually with white streaks on the standard petal which possibly function as nectar guides for birds. An albino form is also known from cultivation in the Worcester area (b in Fig. 3.3.1). Kay (1987) reports that the visible guidemarks coincides with or is in many papilionoids slightly different to the UV absorbing guidemarks. Most ornithophilous papilionoids, unlike mellitophillous species, lack UV patterns, for example Erythrina (Kay, 1987). This character was not investigated in Sutherfandia. All the species are very similar in flower structure (a to g in Fig. 3.3.3) and have very large and showy standard and keel petals. Flower parts are discussed individually below.
21 CHAPTER 3 MORPHOLOGICAL CHARACTERS
igure 3.3.1 Sutherlandia speciosa (a), showing the bright red flowers typical of all Sutherlandia pecies. Slide of Palmer & Van Wyk 4 (JRAU). An albino form of Sutherlandia microphylla (b), nown only from cultivation. Slide of Conradie s.n. sub Moshe, Van Wyk & De Castro 3.
Calyx The calyx of Sutherlandia is constantly campanulate and it is either strigose or pubescent. The hairs may be black and scattered or white and very dense (3a and 3b in Fig. 3.3.2). The calyx lobes are generally acute and strigose and sometimes pilose within (Phillips & Dyer, 1934).
Standard petals
171 In flowers of Sutherlandia the standard petal has white streaks which contrasts with the bright red colour. The standard petal is as long or slightly shorter than the keel (1 and 4 in Fig. 3.3.2; Fig. 3.3.3).
22 CHAPTER 3 MORPHOLOGICAL CHARACTERS
4 Figure 3.3.2 The typical structure of a Scitherlandia flower, showing the large standard petal and keel, and the much reduced wing petals. Note the equally lobed calyces and the variability of the shape of the wing petals. 1, S. frutescens (Olifantshoek); 2a, S. tomentosa (Witsand); 2b, S. microphylla (Bitterfontein); 2c & 2d, S. humilis (Uniondale); 3a, S. frutescens (Vanrhynsdorp); 3b, S. frutescens var. incana (Aurora); 4, S. frutescens (Kleinsleutelfontein).
Wing petals
The wing petals of Sutherlandia species are highly reduced as found in many bird-pollinated legumes. They vary in size and shape but both these characters appear to be of no diagnostic value at species level. Phillips & Dyer (1934) found the wings to show great variability with a tendency for coastal species to have wings rounded at the apex and for the inland plants like S. microphylla to have the apex oblique and acute. These findings by Phillips & Dyer (1934) were also observed in this study: the wings of S. tomentosa were rounded at the apex (2a in Fig. 3.3.2) compared to the inland S. microphylla which were obliquely pointed (2b in Fig. 3.3.2). The limited value of this character is shown by S. humilis, also an inland species but with wings that vary considerably from rounded (2c in Fig. 3.3.2) to acute (2d in Fig. 3.3.2) at the apex.
23 CHAPTER 3 MORPHOLOGICAL CHARACTERS
Figure 3.3.3 Variation in the shape and size of flowers of Sutherlandia species. Note that the flower structure is remarkably similar in all the species. a, S. frutescens (Olifantshoek); b, S. frutescens var. incana (Aurora); c, S. humilis (Uniondale); d, S. microphylla (Bitterfontein); e, S. montana (Piquetberg); f, S. speciosa (Khamiesberg); g, S. tomentosa (Witsand).
Keel petals The large, boat shaped keel in Sutherlandia is uniform in shape but varies in size (4 in Fig. 3.3.2). S. speciosa and S. montana have larger keel petals than the other species. This character is logically correlated to the flower length (see below).
24 CHAPTER 3 MORPHOLOGICAL CHARACTERS
Flower length In the genus Sutherlandia, flower length varies from 25 mm to more than 50 mm (Fig. 3.3.4). For most species, the flower length is quite uniform and varies between 25 and 35 mm. S. montana and S. speciosa have significantly larger flowers than the other species and the length is often between 35 and 53 mm. It seems that larger flowers are found at high altitudes but this correlation is not very clear. The large flower (more than 35 mm) is a useful character to separate S. montana and S.
speciosa from S. frutescens which would otherwise be difficult if fruits are not available.
55 1.
50
45
m)
40 h (m t ng le r 35 we Flo 30
25
20
(Taxa: SFV = S. frutescens var. incana, SF = S. frutescens, SH = S humilis, SM = S. microphylla, SMO
= S. montana, SS = S. speciosa, ST = S. tomentosa)
Figure 3.3.4 Variation in flower length in Sutherlandia species. The range and mean value are shown for each taxon. The number of populations sampled (N) is given below each species.
Phillips & Dyer (1934) found floral parts to be of little value in distinguishing between species except for the size and shape of wing petals (2a to d in Fig. 3.3.2) and flower size (a to g in Fig. 3.3.3 and Fig. 3.3.4). This study confirms the limited taxonomic value of these characters.
25
CHAPTER 3 MORPHOLOGICAL CHARACTERS
3.4 Fruit Most legume fruits are plastic (Polhill, 1981a) and this is particularly evident in the tribe Galegeae, where the membraneous inflated fruits vary considerably in size and shape and are often diagnostically different at the generic and species levels. Lessertia species, for example, are all very similar and are distinguished mainly by the pods (Harvey, 1862).
Sutherlandia has inflated, papery, membranous pods (Linnaeus, 1753; Brown, 1812; De Candolle, 1825; Harvey, 1862, Phillips, 1926; Phillips & Dyer, 1934 and Hutchinson, 1964). The shape of the pod is one of the most important diagnostic characters in Sutherlandia, not only at generic level, but also to distinguish some of the species (Phillips & Dyer, 1934). Stipe length was used by Phillips & Dyer (1934) to distinguish S. montana from other species, but they seem to have overlooked S. speciosa, which also have very long stipes. Of particular significance is the orientation of the stipe, which seems to have been overlooked by previous workers, possibly because the character is not always obvious on herbarium specimens. In view of the proven utility of pod characters in the Galegeae in general, and in Sutherlandia in particular, much attention was focused on fruit characters in this study.
Figure 3.4.1 The three basic pod types in Sutherlandia (left- lateral view, right — top view) a, Pod rounded and inflated with stipe directed upwards i.e. towards the upper, seed-bearing suture [S. frutescens (Moshe, Van Wyk & De Castro 14 (JRAU)]; L b, pod rounded and inflated with the stipe directed downwards i.e. towards the lower suture [S. speciosa (Palmer 4d (JRAU)] and c, pod narrow and oblong with the stipe in line with the pod [S. microphylla (Palmer 5c (JRAU)].
26 CHAPTER 1 MORPHOLOGICAL CHARACTERS
Fruit shape The fruit shape vanes from ovoid to oblong (Fig. 3.4.1). The pods of all the species have an ovoid- ellipsoid shape (a and b in Fig. 3.4.1) and one with an almost globose shape (S. humilis). In contrast,
S. microphylla has much narrower, oblong pods (c in Fig. 3.4.1). The oblong pod shape (length to width ratio of more than 2:1) distinguishes S. microphylla from all other taxa (Fig. 3.4.2). Note the relative uniformity of the length to width ratio in other taxa (Fig. 3.4.2). 7
6
3
2
1 0 1 1 [11111111 11 E1111 ST 1 11111 SS F1 1111111 1111 SMO 1111 11111 SM 1 SH [1 SF SFV
N = 7 12 5 11 4 5
(Taxa: SFV = S. frutescens var. incana, SF = S. frutescens, SH = S humilis, SM = S. microphylla, SMO = S. montana, SS = S. speciosa, ST = S. tomentosa)
Figure 3.4.2 Pod length to width ratio in populations of Sutherlandia species. Note that S. microphylla has a ratio of more than 2:1 while all the other species have this ratio at or below 2:1. The number of populations sampled (N) is given below each taxon. For each population, between three and seven specimens were sampled, depending on availability.
Fruit stipe orientation The pods of S. frutescens and S. speciosa are superficially identical, but on closer investigation it will be noted that the seed-bearing suture is carried below or above the fruit as a result of the difference in orientation of the stipe (a and b in Fig. 3.4.1). S. frutescens, S. humilis, S. frutescens var. incana, S. montana and S. tomentosa have the stipe directed upwards (closer to the upper, seed-bearing suture) (a in Fig. 3.4.1). S. speciosa however, has a unique stipe which is directed downwards towards the lower suture (b in Fig. 3.4.1). This unique feature of S. speciosa fruits is easily overlooked because the fruit tend to twist around when it becomes heavy, thus appearing superficially similar to those of S. frutescens and other species. When mature S. speciosa pods are viewed in situ, it will be clear that the upper, seed-bearing suture is at the bottom of the fruit rather than at the top as in other species. S. microphylla has the stipe in line with the pod (c in Fig. 3.4.1).
27 CHAPTER 3 MORPHOLOGICAL CHARACTERS
Fruit stipe length Phillips & Dyer (1934) used the long fruit stipes of S. montana as a diagnostic character. The stipe length shows considerable variation but there are some interesting trends (Fig. 3.4.3). S. montana
and S. speciosa have very long stipes of more than 10 mm long (Fig. 3.4.3). S. frutescens, S. frutescens var. incana, S. humilis and S. tomentosa have short stipes (less than 8 mm long), while S. microphylla is intermediate. It is possible for example to distinguish between S. frutescens and S. speciosa (both species have inflated pods, see a and b in Fig. 3.4.1 because S. speciosa has much longer stipes than S. frutescens (generally more than 8 mm long in the former and less than 8 mm long in the latter species) (Fig. 3.4.3).
15
10 E E
5
5
0 SFV SF SH SM SMO SS ST N= 7 12 5 11 5 4 5
(Taxa: SFV = S. frutescens var. incana, SF = S. frutescens, SH = S humilis , SM = S. microphylla, = S. montana, SS = S. speciosa, ST = S. tomentosa)
Figure 3.4.3 Pod stipe length in Sutherlandia species. The range and mean value are shown for each taxon. The number of populations sampled (N) is given below each taxon. For each population, between three and L.. seven specimens were sampled, depending on availability.
Seeds The Papilionoideae seed is characterized by the microscopic structure of the hilum, tracheid bar and rim aril (Corner, 1951). In Galegeae, the hilum is circular or oval and has a micropyle adjacent to the hilum (Lersten, 1979; Lersten & Gunn, 1981a). The testa patterns of genera in this tribe vary from levigate rugulate, lopholate, multi-reticulate to simple foveolate (Lersten, 1981).
28
CHAPTER 3 MORPHOLOGICAL CHARACTERS
In Sutherlandia, the micropyle is not engulfed by the hilum as in other genera of the tribe Galegeae and has a punctate micropyle (Manning & Van Staden, 1987). The foveolate testa pattern found in Sutherlandia seeds is unique. This undulate pattern may be attributed to alterations in the chemistry of the wall and the changes in the length of the epidermal cells (Manning & Van Staden, 1987).
In this study all seed samples of all the species showed a similar surface sculpturing (foveolate to smooth) pattern and no obvious discontinuities were found. There was also considerable variation in the colour of the seeds, varying from light to dark brown, but there was again no obvious differences that could be related to the taxa.
L_
Figure 3.4.5 Seeds of Sutherlandia species. Top row: a, S. frutescens (Gamsberg); b, S. frutescens (Olifantshoek); c, S. frutescens var. incana (Aurora); Bottom row: d, S. microphylla (Vanrhynsdorp). e, S. humilis (Uniondale); f, S. speciosa (Khamiesberg); g, S. tomentosa (Witsand).
29 CHAPTER 3 MORPHOLOGICAL CHARACTERS
General discussion and conclusions
Although S. microphylla can easily be distinguished from the other species by its large erect habit at most localities, it is sometimes easily confused with S. frutescens and S. montana. The same is true for S. humilis, which tends to overlap in size with some small forms of S. frutescens. The overall variation is such that habit is not particularly useful as a diagnostic character or as a key character.
The shape and pubescence of leaflets can be regarded as reliable characters for distinguishing S. tomentosa from the rest of the species. The densely tomentose adaxial surface and the obcordate shape of the leaflets satisfactorily distinguishes S. tomentosa from the rest of the species. It is interesting to note that pubescent leaves (S. frutescens var. incana and S. tomentosa) seem to have shorter petioles.
Flower length proved to be of diagnostic value (particularly when pods are not available) because it is possible to distinguish S. speciosa from S. frutescens. Wing petals exhibited variation in shape but this variation is not taxonomically significant to distinguish species.
Pod characters are traditionally used to distinguish the genera of the tribe Galegeae and are particularly useful in distinguishing between the species. Pod shape, stipe orientation and stipe length are valuable taxonomic characters at the species level in Sutherlandia. Pod stipe length distinguishes S. speciosa and S. montana from the rest of the species.
30 CHAPTER 4
ENZYME ELECTROPHORESIS
Introduction Sutherlandia comprises a taxonomically difficult species complex with no information regarding interspecific relationships. Enzyme electrophoresis was used to try and understand more about the real genetic differences between species and populations of the genus. The technique was also used in an attempt to learn more about phylogenetic relationships and perhaps also modes of evolution.
Results and Discussion Thirty-two enzyme coding loci provided interpretable results in all populations analysed, of which 56.3% displayed polymorphism (Table 4.1). Fifteen of the 32 loci (43.7%) displayed monomorphic gel banding patterns and products of the SOD-3 and EST-3 loci migrated cathodally. In addition to these loci, the following enzymes were stained for: acid phosphatase (E.C. No. 3.1.3.2) and aspartate aminotransferase (E.C. No. 2.6.1.1). These two enzymes did not show sufficient activity or resolution to score them satisfactorily in Sutherlandia samples. At SOD-2, only a few taxa showed activity, hence the exclusion of this locus from Table 4.1. Table 4.1 also shows allele frequencies and Chi-square (X2) values for polymorphic loci in populations studied. The maximum number of alleles for any given taxon was five and the average heterozygosity (H) values ranged between 0.01 and 0.097 (Table 4.2). Loci where significant (P>0.05) deviations of allele frequencies from expected Hardy-Weinberg proportions occurred are listed together with their observed numbers of heterozygotes (Table 4.2).
Astragalus atropilosulus subsp. burkeanus had unique alleles at GPI-1*c, IDH*c (Fig. 4.1), MPI*c, PGDH-2*e, PGM-1*d, and high allele frequencies at SOD-1*b and -3*a as compared to Sutherlandia taxa, with A and B as alternate alleles at the latter two loci. The Lessertia species had a unique (A) allele at IDH whereas the Sutherlandia taxa had only BB and BC phenotypes at this locus.
31 CHAPTER 4 ENZYME ELECTROPHORESIS
Figure 4.1 An example of a gel showing allozyme variation at the IDH locus. Note the unique allele - c in Astragalus atropilosulus subsp. burkeanus. ST = S. tomentosa; SFV = S. frutescens var. incana; SM = S. micopylla; A = Astragalus atropilosulus subsp. burkeanus; a, b and c and three alleles.
Genetic variation within Sutherlandia populations - Of the 19 ingroup populations studied, five had unique alleles. For example, S. microphylla had unique alleles at GPI-2*a (SM5, SM6), MDH-1*b (SM4, SM6) and MNR-2*a (SM1, SM2). Astragalus atropilosulus subsp. burkeanus shared MNR-1*a with SF1 and MNR-3*a with ST1. The percentage of polymorphic loci (P) was the same for three populations, SM1, SH and SS (20.6), and varied from zero to 17.6 for the other populations; it was 8.8 for SF1, SF2, SFV1, SM2 and SM6; 14.7 for SM3, SMO2 and ST2 and 5.9 for SF5, SM4 and L (Table 4.2).
Individual heterozygosity (h) values ranged from 0.05 to 0.6; the mean number of alleles per locus (A) was between 1.0 and 1.3 (Table 4.2); the observed number of heterozygotes (OBS) ranged from zero to 18 and the coefficient of heterozygosity deficiency (d) values were from - 0.029 to 0.900 (Table 4.2). Genotypic frequency deviations from expected Hardy-Weinberg proportions were detected at EST-1 (SFV2; SM1; SF1), EST-2 (ST2; SM5), GPI-1 (ST2; SS),GPI-2 (ST2; SM1-3; SF2; SH; SS), IDH (SM1; SM5; SM6; SF5; SMO1-2; SH; SS), MNR-1 (SS), MNR-3 (L), MPI (SF2; SMO1-2), PER-2 (ST2), PGDH-1 (SFV2; SM1; SM3; SF1; SF5; SMO1), PGDH-2 (ST2; SM1; SM3; SM4; SM5; SM6; SF4; SF5; SMO1; SH; SS), PGM-1 (SM1; SM3; SM4; SMO1-2; SH; SS) and PGM-2 (SM1) (Table 4.1). Deficiencies of heterozygotes occurred at all these loci except for MNR-1 (SS) and PGM-1 (SM3) (Table 4.1). This could be due to small sample size and could not be corrected because the total population was sampled in most cases. The average heterozygosity (H) values per locus (Table 4.2) ranged from zero (A, SF3, ST) to 0.097 (SS); and variation at population level occurred at GPI-1 between two populations of S. microphylla (SM5 and SM6) (Fig. 4.2) and heterozygotes were present at PGM-1 (Fig. 4.3). 32 • CD GO CD • 0 0 Ca C> CD CZ• V. CD 0 Co• • • • CD. cz R Rc, CD. 0 C. CD 1: V. V: v!C 1'1' CV) V. C! N. v. ar: V. V: v: CP CD CD CD CD CD O CD O O imp es CD CD Ca 0 0 C. CD 0 CD 0 00 0 0 el CO 0 0 0 CD el CD ca R 1'1 . . V. V. P . c- a r a 6 6 , r .
CA CD CD Tr CD el es CD CD O CD Ca CD et en MD CI CD CD CD Cr CO 0) MI CD CD CD Ca m. es 0 cc. Ca es e4 Ca elf et cD CD CD cD ul Tr N N Ca Ua of UD CA m- Ca CD Ca eA Ca Ca CD of Ul r en en m- CO CD CD Oa Ca Ca Ca en en V 0 CD CD 6 m- .m. CO en .666..66.66 c-.. .1- V. 0 0
CD m- CO et OD et CO CD CD Ca CD CD C> Ca Ca Ca N CO CD ul Ul CD CD CR O CD ul N N CD CD Ca CD CD CD m- CO N es C4 es CD en ul CD cD Ca CD CD Ca m- CO es el CO m- r. C4 Ca CD Ca Ca CD CD en el 02 m- CO r- Ca 01 CD CD R R V: V. s 0 N 6 6 6 6 6 c- 6 6 s N. V. s V. 6 CD s 6 6 CD CD CD
0,1 e4 OD CD CD CD CP el es CD 0 C4 ts v.' co co CO o o N CD CD CD CD CD el CD CD CD O CD C4 CO m. Ca O O U, es ul CD CD CD N UD CA Ca CD CI Ca c, co O R O CO CV O CO m- Ca Ca O O es 1- m- O 0 C) I I •4-: N N 6 0 00 06 cz 060 cr•
m- el CD CD CD CD CD Ca CD C> el es Ca CO CA CD CD CD ul u) CD Ca Ca CD Ca CD C> C> CD CD O CD CD Ca CA CD Ca CD Ca Ca CD CD M UD CD el CD Ca ul Ul N N en u) u) CD CD R co o C1) 1-1 0:1 •11". O 0) O o CV C. IA Vt ^ c.4 0. 0. o R T- V. V. V) ca T- 000 o eZ 6 T. o Ci 6 00 T 6 T- 6 6 6 6 c, 00