A BIOSYSTEMATIC STUDY OF THE GENUS SUTHERLANDIA Br. R.

(FABACEAE, GALEGEAE)

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 111111111 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

CD Ca O CD Ca Ca el es Ca CD CD Ca CD Ca N CO Of m- CD CD C) CD Ul et CD CD CD CD Ca CD es C4 co m- CD CD C4 C) eD Ca O CD CD 0 of 0 CO r O C! LL Ca Ca Ca Ca O R O O O D R O V. I V. o N. N I. o V. V. V. o s V. CD Ci V. g V. V. e V. V. V. CD CD CD 6 V. I

0 o o 0 0 0 0 0 0 0 0 0 0 0 o o o 0 0 0 o 0 o o o O O o cD Ca Ca u) en Ca Ca LL en 0. R R R O. a. R R (/) l. . s V. N. . I. . o T. 1. g V. s V. V. 1. V. I. T. 6 6 V. V. o T.

cD CD O Ca Ca CD CD CD CD CD 0 CD C) Ca CD CD CD CD CD Ca CD CD Ca C> o CI Ca CD Ca Ca CD CD CD C) CD Ca O 0Ca Ca U- CD Ca CD Ca Ca O CR Ca a. R v. V. I V. o V. V. (1) N: T. V: g V. V. o s V. T. r. N. V. V.

CD CD O F) es Ca O CD CD Ca 1- Cn 0 CD CD Ca et N et CD Ca CD CD O el UD Ca Ca CR CR CR CI UR O CD O Ca en r CO Ca CD Ca N o co o O o U- CD CD O of GO CR CR O Ca O C4 e, 0000. NeD V I 6 6 V. V. o V. V. T. 6 CD a T. V. V. a r 6 6 6 c-• I

en ul Ca CD et CD C) CD Cp Ca Ca CD Ca en eD en ul CD es el CD elf el CO CD CP CD es N 0 CD es N Ca CD CD 0 CD CD CD es N N es C3 UD el CD es UD en CD CD CD 01 CD 0 Ca C! CD CD C! CAD el e4 Ca Cc) O CD Co em et en Ca cD es m- Ca Ca Ca ci 6 Ci 6 Co m- V. 6 6 6 o V. s 6 6 6 CD V. CD 6 I T. 6 6 6 v. s V. 1 V.

CD C2 CD CD el es C> Ca Ca CO N CD cD el et of CD CD CD Ca UD 0 CD Ca Ca CD et en CD CD CD cD el CD CD CD Tr Tr CD CD CD CO Ca co Ca CR Ca m- es r- C! 2 © o ..- a. 0. a a. VI N r e•. 64 . r I 1- 6 6 6 I 1. 6 O 6 6 6 a r I I

V. CD el Is 0 0 0 0 CD CD 0 ra is. CD 1`... CD 0 0 0 O CD 1... C., CD ca c, CD en et 0 CD Tr en CD CD C> CD CD CD 0 et en CO en et CD CD CD m- CO CV 0 ,- 0 0 0 Ca 0 02 m- Ca CD m- co CD a. a. a. a r . v 6 6 r 1 r• . . 6 6 V. a o N. s V. . V. T. I o T. i r 6 6 C5 6 ci V I V a . N:

C> Ca Ca CD CD CD O cD UD et CD CD O CD et CO et 03 CO C) CD CD C> C) CD CD CD O Ca CD es e4 CD O CD O co m- Co u, en Ca CD Ca Ca CD Ca Ca Ca O Ca CD 0) 0 CD CD C) CD (JD of 03 Ca Ca Ca Ca Ca T. a A.: • V. T. V. V: . o 6 6 6 6 6 o . V:

CD CD CD en Ul 02 N O Ca Ca CD es el es of N CO CD CD CD es el C4 m- es CD CD CD O es N es N CD CD CD CD m- 03 CD el UD F) Ca CD CD U, rl C4 es CD CD Ca CD Ca 01 CR es C4 Ca Ca p, el 04 Ca Co CD Cn CD Ca ul cr) e4 es el ul m- CD CD CD V.: 0 0 I s C5 C5 V. T. o O 6 6 6 1 6 6 6 6 I N. 6 CD 6 CD O

CD of es 0 CD CD CD CD CD cD Ca CD CD Ca 0 Ca CD 0 el es 0 C. 0 r) CD 0 0 0 0 0 0 o 0 0 0 0 0 0 0 0 el CO 0 0 0 CD T.- o o o 03 Csi Ca Ca en Li, o cD ca Ca 0 CD co e- CD C! R ....: a . N.. 6 O ..- . ..4 , , .4 . . 6 6 . c- .- 4 I V CD r I I T. ', I 0 ,. 0 I V 0 6 6 V . V.

CD m- of CD Ca es el CD CD Ca c> O CD 0 0 CD CO e4 ul ul CD CO U, eD ul cD CD CD of ;JD CD O CD F) Lc) cal 0 CD Ca CD CD CD CD CO m. N P.. CO es el Of CD CD eD Ca C4 es Ca Ca U, el e, 04 0 en ul CD O o C! 1.11 et v. CO el et m- co m- CD ■ I. 6 CD V. . V. ot 6 6 6 r- 6 6 r- I r r- I r- 6 CD 6 6 6 CD 6 6 6 m- ' .1

0 CD CD CD et CD CD Ca CO el CD Ca Tr CD CD Ca CD CD 0 CR O CD CD CD Ca Ca CD m- CO CD C) CD of u, CD CD N es CD ul an CD CD u) ul CD CD CD CD CO C4 Cp el UD Ca Ca Ca Cn CD cD CD CO CD C2 es N cR C4 CD m- O O Ca 0. 6 .4 6 6 r- V. V. I CD 6 V V. CD CD v. CD 6 T. 6 CD CD V. V. o V.

C) CA of es e4 CO CO N CD CD Ca CD et U2 CD Ca CD CD es el C) CD CD CD OD CV OD r m. CO et U7 CD O CD CD UD el O CD Ca CD CD el O Ca O CD Cr) CD en Tr 02 m- Cn CD Ca Ca eR 01 ca e) o o 6 6 6 6 c5 6 6 . . . . . .0 a r: 6 .0 v. i V.

02 tO of 0 ONCON- Of N UD CD CD 0) v. C) CO et N CO CD sr UD CO m- CD CD C> CO V. MI 41 el u, et en el Cn CD CD CD et MD CD CO V. el co CD es N et en CD CD CD CO CD CA e, U) el m- CO Ca cD Ca Ca CD anCa Ca 010 OMR Cn 6 CO Ca 6 o 6 6 6 o 6 6 6 6 . CR 6 V. I. g V: 00 V: CD CD 00 V. . 0 0 1 0 V. s V:

CD CD Ca CD CD CD Ca O C) Ca O cD CD CD Ca cD CD CD 0 CD CD CD CD Ca CD CD CD CD C> Ca CD CD CD CD CD CD Ca Ca Ca CD CD CD Ca C> Ca O CD o R . r . I V. o N. V. V' V V I. V. V. a o N. V. V. V. I V.

.1 03 4 e .4 03 cc e 0 .1 0 0 0 .4 03 0 03 .1 a) .1 03 () 03 .4 03 (-) .4 OD () .4 00 () .4 p2 ct E2 5 122 2 E2 U a e (1 E2 .rt e ct e 4 co

4-7 ... ,... (:. c;i ',;' ,- ..) ,.... l'e lE i.1 1 Cc CC cc 0 0 Z O el 0 2 0-. 0 0 0 x 0 h §: a ?-i ,... .L.. Lu 0 0 0 _ ._

CreiT N O • 'cr • 0 62. OS

e- O CV ;! R6 6.Ze 0 o . CD E.

■ Cl) .■ N. Nel c, CO 0 10 C.. 0 0 el • 6 C3 R9 0 r° N 6.1 .00 -.J 0 re 6 ■ ■ w N •• N . C1,..r Om 0Wine,'0 too >- 2 N • R6 R00 1- 0 ,, e, , d_ .066 Cl) 0 (9 r,..- w.,.... ).- 6 CF17 we, owcoel 0 0 oR .-000 N 2 " • r.: 0 0 ,e , 6° .6666 Ce W

I- ■•■■ W V '''. .1. 0.4e.,M - N c, t...00Nis. = O N ‘-.-0.--.-- 2 .-... co R6 0 0 . a c 0 --. . 6 6 6 6 6 W > Ce W 03'...... •OWN1, 00 a 03 CI a C1(4 .0VCOC11 01 11 CDI■ CDCD, 03 0 ,.2. ; 62. .666666 0 W 0 ..... ,t- qt woow.--om 0 N mco.- NcowooG,Q.c, cc 2 ..- • C) o • oc).- ..- W 0 ... 2. wri o° . 6 6 6 6 6 6 6 ort Z co •- covocomcv.o, el T wwmpswmcm .er c, r... ma, 00000000 2 in . et' ri Cf) r: a .-- 6 e .66666666 1-W E2 - Cn es N to el C) 1.- cn e- () c,- aO N c, en 03 03 I. el N Cn CO e- co eg .- . co CD CD C3 v. v- 1.- CD CD T- CO R do N. y: 2. 0 0-- , 6 6 6 6 6 6 6 6 6 6 W = en ---' et CD 01 en N CA N CD CD et 1- e- a co 0 CO c M 41001 .0 N. LOCON. to 0 (16 6 R6 00.-0.-0000R ZZ 0 '...... N 0-- .6666666660 (70 DP mvc000r,coacor,r cl:z We''' r. 0 go csie con.-wcoNe r, 0 = "6 6 R6 oR000.-- ooRRo 3,7 o ..-- N c,- .60666666006 W JO

00 000NNNV00 'CZ Mg. C.,0 ,- EW N tsc, VV040tOt erMT- M 0 (C! c,o0000,-c,c,c,00 a3 > N 1, Rd w o ,- , Qv .666666666666 0- 2 w ).- w .-. a ..t. C) CD CID e- UD mr ‘- I.- ^ es v- ul el Is• -Jo wo et e- CO CV ul 0 en CM eA eV UD el eA 0 oppe-, U. ; N • CO 0 • 000000 IL 0 v;2 co 62. .6006666666666 LL • OW .... (11-1 0-"'" 0.4)4100MCOrN07- VOLC) M to O T- omomco.- wvmr-commw et < u- .-R m o• ..- 0 .-. 0 0000.0 I-1-- 0 .2. W 6 6 .66606666666606 Zw ww Ow ■■•• - vot...-..--wmcownwel.-cow re j in" nr ,- e OINCOetMWWNesN0e-MWM W R0.o 000000.- .- ,- NeooeR 1LZ u. 0 .-.2 mcsi de .666666666666660 <00 0‹ a coorlmr,omN.ommr-we40.- D8 el a c0mm coN0mw04woc4 (J-, u. 0. RR RoR0000000.-R0000 _,c)> 0 ,--.e) -.-, .-.-.- 0 .0606666666606666 ced ■ w. M ---- et CO C4 CD N en et es en Cn C3 e- 0 es el et CO a. to N to N to e- C3 en m- co es in es o F) c, CM Cn el CO CO el 0 w U. I- • to R, 0.- 0000R0 0000 tli 0 ..:2. 0 oa .66666606666666666 -i W -i -l < _.J> N. cocoowelmv.-ecow w .o ui .L.-- ,..,a ...... -N.,...-w. 0 U. r w .• 00.- 0.- 000R0.- .- .- 00.- 00 u. 0 .- 01 CO cia .666666660666666606 2 c. I-- 6 0 ..... CO a wiz wi,oac.w.rewomvor-oco,r... c4 c, t110.1041043 .00Wc.C1W1■ 04W cp R cn CD • 1■ 0404110404 ..... (..‘-.4.- 4,-. MO -i Tze. W .6 ° .6666666666666666666 Z Z W Z .•■■ w w a.... inesele-ese-OetOeluleINWOr1001e-N 0 e- el Ps CA CA el CD CO UD et es CO es es Co e- I- to 01 eV t) CD • CD CD• U)Qe)F) ' ,e d 6 2 .0 0 660006600606000660 '..e'n 0 . < 4m LuL".. Iii .-e4 -11.1"5 M 0. „,,,,,,,,,,,. N-NF)V‘0000 ,..c.4 .... u-wwu-u.>> 22222222 1-1-- ail 0 - ^ -,- _ ENZYME ELECTROPHORESIS

Figure 4.2 Example of a gel showing variation at population level at the GPI-1 locus between two populations of S. microphylla (SM5 and SM6), a and b are the two alleles.

Figure 4.3 Example of a gel showing variation at population level at the PGM-1 locus. Note the heterozygote bands. SFV = S. frutescens var. incana (Pearly Beach); ST1 = S. tomentosa (Koeberg Nature Reserve); SM4 = S. microphylla (Colesberg). tolliir I CM 4 ENZYME ELECTROPHORESIS

The values of population variability obtained in the present study (0-20.6%, 1-1.3, 0.01-0.097 respectively) are quite low as compared to the reported values for vascular plants in Hamrick (1979) (P=22.0-75.3%, A=1.35-2.56, H=0.079-0.354). The lower allozyme variation between taxa might be attributed to various factors. For example, the breeding system is taxonomically important for three reasons: the extent of interbreeding largely defines the pattern of variation and hence the delimitation of taxa; a knowledge of the breeding system frequently helps to understand complexity, although often it does not solve the problems associated with it; and a study of the breeding system is often vital in unveiling evolutionary pathways (Stace, 1980). Sutherlandia taxa are pollinated by the malachite sunbird (Nectarinia famosa) commonly found in southern Africa (Skead, 1967; Arroyo, 1981). The bird can travel long distances and is able to transfer genetic material from different geographically isolated populations of taxa, causing the recombination of alleles. Flowers may also become cleistogamous, thus enabling self-pollination (B-E. van Wyk, pers.obs.). This will result in a small population becoming larger up to a point where pollinators will be attracted.

Cleistogamy would result in a morphologically uniform population which might later, during favourable seasons, be pollinated by a bird with pollen from a geographically distant population. The resulting seeds will have a combination of genetic material and this could result in high levels of integration and recombination. The high degree of integration among the taxa studied (Fig. 4.4) makes it difficult to support, on the basis of enzyme patterns, the hypothesis that the genus comprises six species. Distance from root 0.0 0.04 0.09 0.13 0.18 0.22 0.27

1 90 lfrutescens' frutescens 3 13 microphylla 3 humilis 4 frutescens 2 frutescens var. 31 incana 2 speciosa frutescens var. incana 1 13 13 tomentosa 2 montana 2 12 montana 1 43 frutescens 4 70 25 frutescens 1 frutescens 5 59 microphylla 1 microphylla 2 13 microphylla 4 microphylla 5 59 t microphylla 6 LESSERTIA ASTRAGALUS

Figure 4.4 Dendrogram based on Nei's (1978) unbiased genetic distance data showing relationships between nineteen populations from seven taxa of Sutherlandia and two outgroups (A. atropilosulus subsp. burkeanus and a Lessertia sp.).

34

lollAir I CR ENZYME ELECTROPHORESIS

The average D value between populations was 0.077 and a UPGMA dendrogram derived from D values shows the relationships and integration of nineteen populations from six taxa of Sutherlandia and outgroups (Fig. 4.4). The grouping of populations ST1 and SF3 was mainly due to lack of heterozygotes in both populations and this grouping had the highest bootstrap value (90). Most of the groupings of populations were not well supported as indicated by very low bootstrap values (Fig. 4.4). It is interesting to note that all pairs of populations with relatively high bootstrap values (SM5 and SM6, SM1 and SM2, SFV1 and ST2, and ST1 and SF3) respectively originated from the same geographical areas. This suggests that the geographical component of the allozyme pattern is somewhat stronger that the taxonomic component. A similar result was obtained by Hornero and Perez (1997) in Colutea.

Genetic differentiation - Using limited numbers of populations when investigating genetic variation may lead to incorrect conclusions about the taxa studied. In a preliminary survey, only seven Sutherlandia populations (SF4, SF5, ST1, ST2, SM4, SM5, SM6) were studied. Figure 4. 5 illustrates the relationship among the four taxa using A. atropilosulus subsp. burkeanus (A) as the outgroup. Except for S. frutescens all the taxa could be distinguished by unique alleles: S. frutescens var. incana (PGDH-1*c), S. microphylla (MNR-2*c, PGDH-2*a and PGM-1*b), S. tomentosa (GPI-2*a and PER-2*a) and A. atropilosulus subsp. burkeanus (IDH*c, MDH-1*b, MNR-1*a, MPI*c, PGDH-2*e, PGM-1*d, PGM-2*b and SOD-1*b). From this cluster three groupings were evident and the grouping of taxa agreed with the classification based on morphology. The cophenetic index value was 99,9%. More populations were then investigated, including all taxa, and it was evident that the populations graded into each other because it was not possible to differentiate between taxa as a result of the lack of unique alleles (Fig. 4.4).

10 0.04 0.09 0.13 0.18 0.22 I 0.27 I I I I I I I

ST

SFV

SF

SM

A

Figure 4.5 Dendrogram based on Nei's (1978) unbiased genetic distance data showing relationships between four taxa of Sutherlandia [ST = S. tomentosa; SFV = S. frutescens var. incana; SF = S. frutescens; SM = S. microphylla] and A = A. atropilosulus subsp. burkeanus as outgroup.

35

CHAPTER 4 ENZYME ELECTROPHORESIS

Nei's (1973) coefficient of gene differentiation relative to the total population (Gs1) is equivalent to Wright's (1978) fixation index (F ST). These values averaged 0.643 and 0.638, respectively, for the Sutherlandia populations and species studied (excluding outgroups), and reflect the substantial amount of differentiation also observed between the SF (F ST = 0.767) and SM (Fsr = 0.597) populations. The FST values indicate large genetic differentiation and reduction in heterozygosity between subpopulations studied due to random genetic drift. Values of FST are close to zero if all subpopulations are in Hardy-Weinberg equilibruim with the same allele frequencies (Nei, 1986). The FIT values of 0.876, 0.920 and 0.877 for the above-mentioned populations respectively, which quantifies inbreeding due to population subdivision, is indicative of low levels of gene flow between the populations studied, but it may reflect adaptation to different ecological conditions.

Average D values (Table 4.2) between taxa (0,032) is lower than between populations (0.077), and genetic distances of the same magnitude between three taxa of Sutherlandia (S. frutescens, S. microphylla and S. montana) were obtained. Taxa differences were identified between ST1, SFV1, SM4 and A at GPI-1*a, *b and *c (Fig. 4.6). Despite the taxa differences observed (Fig. 4.6) there are no distinct allozymes (or allele frequencies) or geographic patterns of allele frequencies between taxa at different populations to distinguish them.

L r.:,_., ti _ - Ancra._ er"Is ft->p•' .17.:z. 1 . 0- 4,1/444-4, :1,...::,..fZricSt.2..,..: : V. , . .,„.-j. ;k1.2,,I;1,;,1:."-;,::::7

Figure 4.6 An example of a gel showing species differences at the GPI- 1 locus. ST1 = S. tomentosa (Koeberg Nature Reserve); SFV1 = S. frutescens var. incana (Pearly Beach); SM4 = S. microphylla (Colesberg) and A = A. atropilosulus subsp. burkeanus as outgroup.

36 %-orliAr I cn ENZYME ELECTROPHORESIS

The phylogenetic tree (Fig. 4.7) illustrates the genetic relationships between Sutherlandia taxa; S. tomentosa and S. frutescens var. incana formed a consistent group (92% bootstrap value). This grouping is sensible because S. frutescens var. incana is morphologically an intermediate between S. frutescens and S. tomentosa. These two taxa co-occur along the coastal regions and both have densely pubescent leaves. Sutherlandia speciosa (SS) is endemic to Namaqualand (Northern Cape and Namibia) and is grouped with S. humilis (SH) (Fig. 4.7) from the eastern Karoo regions. Morphologically, S. speciosa is a spreading shrub of 0.4 m with large flowers and the stipe curving downwards in the mature pods. It is similar to S. frutescens (SF), except for the stated pod character and the larger flowers and habit. In contrast, S. humilis is a prostrate shrublet of less than 0.2 m tall and is also similar to S. frutescens, again except for the habit.

0.0 0.04 0.09 0.13 0.18 0.22 0.27 I I I I I I Distance from root S. montana. 36 S frutescens 24 S. humilis 33 45 1 S. speciosa

91 S. microphylla

S. frutescens 92 var. incana

S. tomentosa

LESSERT1A

ASTRAGALUS Figure 4.7 Dendrogram based on Nei's (1978) unbiased gentic distance showing relationships of seven taxa. The bootstrap numbers are presented at the nodes.

It seems likely that S. humilis (SH) and S. speciosa (SS) are mere subspecies or varieties of S. frutescens (SF), adapted to different geographical and environmental conditions (Fig. 4.5). The same may be true of S. speciosa (SS) and S. montana (SMO), the only two taxa found at high altitudes, both with exceptionally large flowers. Sutherlandia microphylla (SM) is grouped separately from the other taxa on the basis of the oblong pods (length to width ratio of more than two) and narrow leaflets.

Liston (1992) studied Astragalus taxa and concluded that morphological and allozyme divergence patterns are concordant and that infraspecific taxa are characterised by slight morphological and little or no genetic differentiation. The taxa however, exhibited more pronounced morphological and allozyme differentiation as compared to conspecific populations. Although these observations

37 lorlAr I CrS ENZYME ELECTROPHORESIS were in agreement with the taxonomic treatment of Barneby (1964) and supported the validity of his taxa concepts in A. sect. Leptocarpi subsect. Califomici, the results of the UPGMA phenogram suggested that taxonomic categories above species level do not reflect genetic relationships. Liston's (1992) findings are similar to the results shown in Fig. 4.5, but not to those in Figs. 4.4 and 4.7. It seems that there could be lack of agreement between morphology and isozyme patterns in other genera of the tribe Galegeae. Hornero and Perez (1997) were able to illustrate geographical relationships among 23 Colutea populations using cluster analysis, but a strong correlation was lacking between genetic and geographic distance within each area. Populations from the same area were clustered together and two major gene pools appeared among the Iberian Colutea populations studied, but these did not agree with the classification proposed on the basis of morphological criteria.

The paper by Hornero and Perez (1997) seems particularly relevant for comparison since Colutea is the closest relative of Sutherlandia for which published information could be found. It is interesting to note the same lack of agreement between the current taxonomy of Sutherlandia and the enzyme patterns.

Conclusions The results indicate that the Sutherlandia taxa have very low genetic differentiation as compared to populations. This observation is congruent with the taxonomic treatments of Meyer (1836) and Harvey (1862) who regarded the South African Sutherlandia genus as one variable taxon, S. frutescens. On the other hand, Phillips and Dyer (1934) were inclined to the view that they were dealing with a large taxon complex which was separating into definite taxa in different geographical areas. In each of these areas they found that the fruits had differentiated sufficiently to justify their grouping at specific rank.

This survey of populations from all the taxa illustrated an integration of populations and taxa which cannot be distinguished from each other by allozyme data. The results of this study, therefore, suggest that a more conservative taxonomic classification system for the genus Sutherlandia is called for. It is clear that the specific rank is inappropriate in many cases, and that the rank of subspecies or variety will give a better reflection of the lack of genetic differentiation and the lack of morphological discontinuities between some of the forms hitherto accepted as distinct species.

38 CHAPTER 5

CHEMICAL CHARACTERS

Introduction

Secondary metabolites of plants have provided systematists with additional and useful characters to distinguish between families, subfamilies, tribes, genera and species. In the Fabaceae, the most useful compounds include alkaloids, flavonoids and to some extent amino acids (Southon, 1994).

The genus Sutherlandia is commonly known as cancer bush and early colonists used it to cure internal cancers and other ailments (Van Wyk et a/., 1997). However, nothing is reported or recorded regarding the chemical compounds of the genus except for the presence of canavanine, which was detected in seeds of Sutherlandia frutescens (Harborne, 1973).

Some unpublished information (in the form of theses) are available (Snyders, 1965; Viijoen, 1969; BrOmmerhoff, 1969 and Gabrielse, 1996) where attempts have been made to isolate and identify chemical compounds from Sutherlandia. Apart from pinitol, no chemical compound has been successfully characterised. Pinitol was extracted from S. microphylla and a yield of 2.8 mg/g was obtained (Snyders, 1965). In this study pinitol was also investigated and high yields were obtained.

In view of the traditional medicinal importance of Sutherlandia, it is surprising that practically no information on the chemistry is known and this work was done to explore the presence of various classes of compounds. Due to the exploratory nature of this investigation, most of the work was focussed on those classes of compounds which proved to be significant because of their high levels in the plant and their known medicinal importance.

A broad survey of various classes of compounds resulted in the discovery of significant quantities of triterpenoids and amino acids and the results are summarised in Table 5.1.1. The detailed results for each of the classes of compounds in Table 5.1.1 are discussed below.

39 ,..- 1/11- IL.Il N CHEMICAL CHARACTERS

Table 5.1.1 Compounds investigated in a chemotaxonomic survey of Sutherlandia.

Class of compounds

Plant Alkaloids Monoterpenoids Flavonoids Triterpenoids Amino Acids S. frutescens + + + +++ S. frutescens var. incana - _ + + + + ++ S. humilis _ _ + + + + ++ S. microphylla _ + + + +++ S. montana _ - + + + +++ S. speciosa _ _ + ++ + ++ S. tomentosa _ + + + +++

Notes: + Clearly detectable ++ Present in substantial quantities +++ Present in high yields - Not detectable

5.1 Alkaloids

More than 350 different alkaloids have been identified from about 60 genera of the legume subfamily Papilionoideae (Mears & Mabry, 1971; Kinghorn & Smoleski, 1981; Southon, 1994; Bruneton, 1995). The possible presence of polyhydroxy alkaloids (castanospermine-type alkaloids) was studied, because compounds such as swainsonine, cassine, castanospermine, smirnovinine, sesbanine and spherophysine have been found in other members of the tribe Galegeae (Southon, 1994; Bruneton, 1995). Since these alkaloids are known for their biological activity, it was speculated that they may be linked to the reported antitumourigenic properties of

Sutherlandia.

Polyhydroxy alkaloids are not easily detected and isolated because the numerous hydroxyl groups make them behave like sugars and the normal method of extracting them as salts does not work. The methods of Ghosal et al. (1970) was therefore used. This method involves cation exchange resin (see materials and methods).

No alkaloids were detected in Sutherlandia leaf and seed samples investigated. Amino acids (e.g. canavanine) seems to replace alkaloids (Table 5.1.1), which are so commonly found in the seeds of many other genera of legumes.

5.2 Monoterpenoids Above-ground parts were steam distilled to see if there were any volatile oils or compounds in

Sutherlandia. No monoterpenoids could be detected by Gas Chromatography (GC) (Table 5.1.1).

40 Unlit- 1CM CHEMICAL CHARACTERS

5.3 Flavonoids Flavonoids are present in all vascular plants and are structurally derived from the parent flavone (Harborne, 1973). The flavonoids are widespread in the Papilionoideae and are found in all tribes (Harborne, 1971a; Gomes, Gottlieb and Salatino, 1981; Southon, 1994; Bruneton, 1995). Some classes of flavonoids are more widely distributed than others; flavones and flavonols, for example, are almost universal. Flavonoids are present in plants as mixtures and it is very rare to find only a single flavonoid component in a plant tissue.

In the tribe Galegeae, flavonoids are widespread in many genera and it seemed worthwhile to investigate the flavonoids of Sutherlandia. The most common flavonoids found in the tribe Galegeae are kaempferol, quercitin, isoquercitin, isorhamnetin, maackiain, pisatin, variabilin and rutin (Southon, 1994).

In Sutherlandia species (Table 5.3.1), low concentrations of six unidentified flavones were detected using both Thin Layer Chromatography (TLC) (Fig. 5.3.1) and High Performance Liquid chromatography (HPLC) (Fig. 5.3.2). All six compounds showed the same characteristic UV absorption spectrum of flavones, with absorption maxima at UV 260 nm and UV 365 nm. The flavones did not show any taxonomic trend to distinguish between species thus seems to be of taxonomic value at generic level only. Since these compounds were not only invariable within the genus, but also present in low concentrations, it was decided not to isolate and identify them.

Table 5.3.1 List of Sutherlandia specimens investigated for flavonoids.

Sample No. Species Voucher Locality 1 S. frutescens Palmer 1 Olifantshoek 2 S. frutescens Palmer 2 Gamsberg 3 S. frutescens Palmer 6 Vanrhynsdorp 4 S. frutescens var. incana Palmer 9 Aurora 5 S. frutescens var. incana Palmer 18 Struisbaai 6 S. microphylla Palmer 5 Bitterfontein 7 S. microphylla Palmer 7 Vanrhynsdorp 8 S. microphylla Palmer 27 Cradock 9 S. humilis Palmer 17 Barrydale 10 S. humilis Palmer 25e Uniondale 11 S. humilis Palmer 25f Uniondale 12 S. montana Palmer 8 Piquetberg 13 S. speciosa Palmer 4 Khamiesberg 14 S. tomentosa Palmer 19b Witsand 15 S. tomentosa Palmer 19d Witsand

41 -• •••• • -- CHEMICAL CHARACTERS

3 4 5 67 7 8- 9 TO 11 12 13 14 15

Figure 5.3.1 A TLC plate of methanolic leaf extracts from Sutherlandia species. Note the uniformity of the flavones, visible as yellow bands (Y1-Y3) near the origin of the plate. The purple spots are triterpenoids (see 5.4). Tracks 1 to 15 correspond to sample numbers 1 to 15 in table 5.3.1. CE N SORBA AB

k-/L) TIME

Figure 5.3.2 An HPLC profile of six unidentified flavones of Sutherlandia species.

42 CHAPTER 5 CHEMICAL CHARACTERS

5.4 Triterpenoids Triterpenoids are the most ubiquitous non-steroidal secondary metabolites in terrestrial and C- marine flora and fauna (Mahato et al., 1992). Although medicinal uses of this class of compounds have been rather limited, considerable recent work in this regard strongly indicates their great potential as drugs (Bruneton, 1995). Moreover, despite the remarkable diversity of skeleton types that is already known to exist among triterpenes, new variants continue to emerge (Mahato et al., 1992). Some medicinally important triterpenoids have been isolated from legumes (Harborne, 1971b). These include oleanic acid, putranoside, swartziasaponin, soyasapogenol and medicagenic acid (Southon, 1994; Bruneton, 1995).

In the Galegeae, some terpenoids have been isolated and identified. Ursolic acid was isolated from Alhagi and dustanin and oleanic acid from Sesbania (Southon, 1994). Due to the known medicinal properties of these triterpenoids, it seemed worthwhile to investigate their possible presence in Sutherlandia. Viljoen (1969) and BrOmmerhoff (1969) have also isolated two triterpenes from Sutherlandia microphylla. Gabrielse (1996) isolated a triterpenoid ("bitter compound") from S. microphylla. This compound was studied by HPLC-MS and had a Rt of 11.9 minutes in his system.

Several plants and populations were included in a survey (Table 5.4.1). In this sample, all the species and infraspecific taxa were included, as well as different plants from the same locality. The sampling was designed to enable an evaluation of the chemotaxonomic potential of triterpenoids.

Results BrOmmerhof (1969), in his unpublished Ph. D. thesis, isolated and proposed structures for two tetracyclic triterpenes. On hydrolysis of the two glycosides, he obtained two aglycones with

molecular formulae C30H4804 and C30H50O5. The molecular masses for the two aglycones were 472 and 490 respectively.

In this study, the presence of triterpenoids in Sutherlandia was investigated using TLC (Fig. 5.3.1). The concentration of these compounds was relatively high (Table 5.1.1). Two triterpenoids were isolated (X1 and X2 in Figure 5.3.1). Compound X1 was studied by mass spectrometry by Prof. F. Van Heerden at the Chemistry Department and had a molecular mass of 562. The initial results indicated that this compound is different to those isolated by Br0rnmerhof (1969) and Gabrielse (1996).

Considerable variation between species and populations (Fig. 5.3.1) was found in triterpenoids. Some species appeared to be different, notably S. humilis (track 9 in Fig. 5.3.1) and S. montana (track 12 in Fig. 5.3.1). However, it was interesting to note that tracks 1 to 2 (S. frutescens),

43 CHAPTER 5 CHEMICAL CHARACTERS tracks 6 to 8 (S. microphylla) and track 13 (S. speciosa) in Figure 5.3.1 had similar band patterns despite the fact that these populations are of different species. Similarly, tracks 4 to 5 (S. frutescens var. incana) and 14 to 15 (S. tomentosa) had similar patterns and these two taxa are known to co-occur at most localities, both being coastal in their distributions (Fig. 5.3.1).

Population variation The lack of taxonomic patterns became obvious when a more detailed survey was done at population level (Fig. 5.4.1). About three chemical groups were identified based on the pattern of spots on the TLC plate (Fig. 5.4.1); those with two major compounds at Rt ca. 0.7 (group 1, the most common pattern); those with one major compound at RI ca. 0.45 and one minor compound at R1 ca. 0.3 (see tracks 7, 10,12, 23) and those lacking the major compounds of groups 1 and 2 (tracks 13, 14, 15). These differences appeared to be random, not linked to any particular taxon.

Interesting interpopulation variation was observed, for example, Sutherlandia frutescens populations 1 to 15 in Fig. 5.4.1. Note that populations 6, 7, 9, 10 and 12 had different triterpenoid patterns from other populations. The same applied to S. frutescens var. incana, track 20; S. humilis populations, tracks 21 to 22 in Figure 5.4.1, have the same pattern as the populations of S. frutescens, tracks 1 to 3.

Triterpenoids exhibited random distribution within the genus Sutherlandia with the observed variation only at the population level rather than at the species level. No triterpenoid pattern could be ascribed to any particular taxon and thus these compounds were of no taxonomic significance.

44 (MAY I LK 0 CHEMICAL CHARACTERS

C - -

Table 5.4.1 List of Sutherlandia specimens investigated for population variation in leaf chemistry. Sample No. Species Voucher Locality 1 S. frutescens Palmer 1 Olifantshoek 2 S. frutescens Palmer 2 Gamsberg 3 S. frutescens Palmer 6 Vanrhynsdorp 4 S. frutescens Palmer 15 Chapman's Peak 5 S. frutescens Palmer 23 Kleinsleutelfontein 6 S. frutescens Moshe, Van Wyk & De Castro 15 Saldanha 7 S. frutescens Moshe, Van Wyk & De Castro 21 Jagersfontein 8 S. frutescens Moshe, Van Wyk & De Castro 5 Camps Bay 9 S. frutescens Moshe, Van Wyk & De Castro 4 Worcester 10 S. frutescens Palmer 21 Swartberg Pass 11 S. frutescens Moshe, Van Wyk & De Castro 14 Aurora 12 S. frutescens Moshe, Van Wyk & De Castro 22 Fauresmith 13 S. frutescens Moshe, Van Wyk & De Castro 5 Camps Bay 14 S. frutescens Moshe, Van Wyk & De Castro 5 Camps Bay 15 S. frutescens Moshe, Van Wyk & De Castro 5 Camps Bay 16 S. frutescens var. incana Palmer 12 Blouberg Strand 17 S. frutescens var. incana Palmer 9 Aurora 18 S. frutescens var. incana Palmer 18 Struisbaai 19 S. frutescens var. incana Palmer 14 Hout Bay 20 S. frutescens var. incana Van Wyk & De Castro 3668 Pearly Beach 21 S. humilis Palmer 17 Barrydale 22 S. humilis Palmer 25 Uniondale 23 S. humilis Palmer 22 Meiringspoort 24 S. microphylla Palmer 5c Bitterfontein 25 S. microphylla Palmer 7 Vanrhynsdorp 26 S. microphylla Palmer 27 Cradock 27 S. microphylla Van Wyk 3804 Elliot 28 S. microphylla Palmer 24 Kleinsleutelfontein 29 S. microphylla Moshe, Van Wyk & De Castro 20 Leeuwberg Pass 30 S. montana Palmer 8 Piquetberg 31 S. montana De Castro 138 Ledger's cave •1i 32 S. montana Moshe, Van Wyk & De Castro 23 Reitz 33 S. speciosa Palmer 4 Khamiesberg 34 S. tomentosa Palmer 19 Witsand 35 S. tomentosa Moshe, Van Wyk & De Castro 1 Koeberg Nature Reserve 36 S. tomentosa Moshe, Van Wyk & De Castro 17 Blouberg Strand 37 S. tomentosa Moshe, Van Wyk & De Castro 6 Still Bay

45 CHEMICAL CHARACTERS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Figure 5.4.1 Variation in the major triterpenoids of Sutherlandia populations.

46 CHAV I tK CHEMICAL CHARACTERS

Within-population variation Intrapopulation variation in triterpenoids was also investigated. The samples extracted are given in Table 5.4.2. The sampling was designed to reveal plant to plant variation within a given population. It was interesting to see that there was absolutely no evidence of plant to plant variation (Figure 5.4.2). For example, note the distinct discontinuities between tracks 1 and 2, 3 to 5, 6 to 8 and 9 to 15.

These results show that the major discontinuities are at the population level; individuals within a population are practically identical while no consistent species differences could be found.

Table 5.4.2 List of populations and species of Sutherlandia used for investigating plant-to-plant variation in triterpenoids.

Sample Species Voucher specimen Locality

1 S. frutescens (plant 1) Palmer 1 b Olifantshoek

2 S. frutescens (plant 2) Palmer 1 c Olifantshoek 5 (plant 1) 3 S. frutescens (plant 1) Moshe, Van Wyk & De Castro Camps Bay . 5 (plant 2) 4 S. frutescens (plant 2) Moshe, Van Wyk & De Castro Camps Bay (plant 3) 5 S. frutescens (plant 3) Moshe, Van Wyk & De Castro 5 Camps Bay (plant 1) 6 S. frutescens var. incana (plant 1) Moshe, Van Wyk & De Castro 16 Blouberg Strand

7 S. frutescens var. incana (plant 2) Moshe, Van Wyk & De Castro 16 (plant 2) Blouberg Strand 16 (plant 3) 8 S. frutescens var. incana (plant 3) Moshe, Van Wyk & De Castro Blouberg Strand (plant 1) 9 S. tomentosa (plant 1) Moshe, Van Wyk & De Castro 17 Blouberg Strand & 17 (plant 2) 10 S. tomentosa (plant 2) Moshe, Van Wyk De Castro Blouberg Strand (plant 3) 11 S. tomentosa (plant 3) Moshe, Van Wyk & De Castro 17 Blouberg Strand 1 (plant 1) 12 S. tomentosa (plant 1) Moshe, Van Wyk & De Castro Koeberg Nature Reserve (plant 2) 13 S. tomentosa (plant 2) Moshe, Van Wyk & De Castro 1 Koeberg Nature Reserve (plant 3) 14 S. tomentosa (plant 3) Moshe, Van Wyk & De Castro 1 Koeberg NatureReserve Resee 1 (plant 4) 15 S. tomentosa (plant 4) Moshe, Van Wyk & De Castro Koeberg NatureReserve Resee

47 - • • - • • • _ . • CHEMICAL CHARACTERS

ti

h!,

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 5.4.2 Intrapopulation variation in the major triterpenoids of Sutherlandia. Track numbers correspond to the sample numbers in Table 5.4.2.

5.5 Amino acids Introduction Early workers in chemotaxonomy largely disregarded amino acids. The development of methods to study amino acids began in 1820 (Braconnot, 1820) and since then numerous different amino acids have been isolated and identified (Fowden, 1964). In addition to the 20 common amino acids, some rare compounds such as citrulline, B-pyrazol-1-ylalanine and its y-glutamyl peptide, and various asparagines have been extracted and isolated (Dunhill & Fowden, 1965). Some protein and non-protein amino acids such as canavanine, homoserine and o-acetylornithine were isolated from various sources, including Astragalus seeds (Dunhill and Fowden, 1967). The introduction of chromatography has opened new doors for amino acid analysis. Thin layer chromatography and ninhydrin have made the detection of amino acids quick and easy. The method has enable various taxonomists to find useful chemotaxonomic information. Canavanine has been used, for example, to distinguish between certain tribes of Papilionoideae (Bell, 1966) although the variation is mostly random within family, tribe, genera and species.

48 CI-1AF I LK 5 CHEMICAL CHARACTERS

In the Fabaceae, canavanine is one of the most important and widespread non-protein amino acid in seeds (Bell, 1958; 1960;1980;1981). Bell et al. (1978) have reported various species of the family to have high levels of non-protein amino acids, especially canavanine. The distribution of uncommon amino acids (Fowden et al., 1967) and canavanine have been reviewed by Bell et. al. (1978). In the tribe Galegeae, most of the genera are reported to have canavanine as the main non-protein amino acid in seeds (Bell, 1981)

A detailed comparative study was undertaken to explore the presence and potential taxonomic value of amino acids in the various taxa of Sutherlandia. In view of the medicinal importance of Sutherlandia, such a survey was expected to also provide a better understanding of the therapeutic value of the traditional medicine.

Thin Layer Chromatography (TLC) Standard methods were used for the extraction of amino acids as discussed in Chapter 2. About eight common amino acids were detected using Thin Layer Chromatography (TLC) (Fig. 5.5.1). They are proline, alanine, valine, leucine, threonine, aspartic acid, phenylalanine and arginine. Proline was distinguishable from the rest of the amino acids as it turns yellow on spraying with ninhydrin, unlike the others that turn violet or violet-grey (Fig. 5.5.1 and Table 5.5.1).

High Performance Liquid Chromatography (HPLC) Due to poor resolution of the TLC, accurate analyses were done for identifying amino acids in leaves and seeds of Sutherlandia samples using an HPLC amino acid analyser connected to "System Gold" computer software. Some typical chromatograms are presented in Fig. 5.5.2 a-g, 5.5.3, 5.5.4, 5.5.5 and 5.5.6. The HPLC results revealed complex mixtures in Sutherlandia. No less than 33 different amino acids were detected.

49 CHAPTER 5 CHEMICAL CHARACTERS

Results Table 5.5.1 List of specimens of Sutherlandia used for the amino acid survey. Samples 8 to 18 are pure reference samples of amino acids.

Sample No. Species & ReferenceSamples Herbarium voucher Locality 1 S. frutescens Palmer 1 Olifantshoek 2 S. frutescens var. incana Van Wyk & De Castro 3668 Pearly Beach 3 S. humilis Palmer 17 Uniondale 4 S. montana Van Wyk 3800 Reitz 5 S. montana Palmer 8 Piquetberg 6 S. speciosa Palmer 4 Khamiesberg 7 S. tomentosa Van Wyk & De Castro 3669 Still Bay 8 Canavanine 9 Arginine 10 Proline 11 Valine 12 Leucine 13 Phenylalanine 14 Histidine 15 Threonine 16 Alanine 17 Serine 18 Aspartic acid

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Figure 5.5.1 A TLC plate of the leaf amino acids detected in Sutherlandia species, together with several reference samples of amino acids. Track numbers correspond to sample numbers in Table 5.5.1. 50 CHAP I tti 5 CHEMICAL CHARACTERS

Chemical structures of all the 33 amino acids detected in Sutherlandia samples are given in Table 5.5.2. The main compounds were asparagine, aspartic acid, glutamine, threonine, proline, y-amino- butyric acid (GABA), canavanine and arginine (Table 5.5.4). The chromatographic profiles of these compounds identified in the Sutherlandia samples are given in Figures 5.5.2 a-g, 5.5.3, 5.5.4, 5.5.5 and 5.5.6). Of particular interest was the discovery of large concentrations of canavanine in the leaves of Sutherlandia and this is reported here for the first time. Since canavanine is a seed metabolite, only trace amounts are usually found in leaves, for example Sesbania - Shqueir & Brown, 1989 ; Medicago (own results showed levels of 0.5 mg/g in leaves of Medicago sativa).

Levels Considerable quantities (in mg/g dry wt ) of 33 amino acids (Table 5.5.3) were detected in leaves of Sutherlandia using an amino acid analyser. The analyses were done at Medunsa by Mr Pat Smith (Department of Chemical Pathology). The yields of these compounds were high although some were relatively low (Table 5.5.4). There was considerable variation in the yields, for example, aspartic acid varied from 1.5 mg/g (SF2) to 46.3 mg/g (SF3) (Table 5.5.4). A seed metabolite, canavanine, was found in leaves of Sutherlandia species in variable yields. S. speciosa (SS) had low levels (2.0 mg/g) compared to S. frutescens var. incana (SFV1) and S. tomentosa (ST6) with high levels (56.0 and 58.7 mg/g respectively) (Table 5.5.4; a and b and g in Fig. 5.5.4).

Species-to-species variation

Variation in the amino acids of samples from the different species are clearly visible in the chromatograms presented in Figure 5.5.2 (results summarised in Table 5.5.4). For example, S. frutescens (SF4) had low levels of asparagine (22.4 mg/g) compared to S. tomentosa (ST4 and ST8) with very high levels (166.0 and 109.1 mg/g respectively). Interestingly, there seems to be high levels of canavanine in coastal species, S. frutescens var. incana (SFV1) and S. tomentosa (ST6) (Table 5.5.4). Although there is considerable variation in the yields amongst species, amino acids appear to be of no taxonomic value to distinguish between taxa as far as can be judged from the relatively small sample size.

Population variation

Considerable variation between populations was found in amino acids (Figures 5.5.3 to 5.5.5). Aspartic acid, asparagine, proline, valine, phenylalanine, canavanine and arginine (Table 5.5.4) are usually the main compounds. Population differences between S. frutescens populations (SF1 to SF4) were observed. For example, S. frutescens (SF1) had higher levels of most amino acids compared to the other populations. This population had higher levels of asparagine (60.8 mg/g) compared to (23.2 and 31.5 mg/g), glutamine (25.0 mg/g), proline (36.8 mg/g), canavanine (17.9 mg/g) and arginine (11.6 mg/g). Although populations SF2 to SF4 had lower levels of most amino

S 1 CHAY I tK CHEMICAL CHARACTERS acids they had 46.3 mg/g of aspartic acid, 10.4 mg/g of valine and 13.3 mg/g of phenylalanine. S. frutescens var. incana populations were comparable in most amino acids, for example, they had almost equal levels of threonine and asparagine. SFV1 had higher level of glutamine and a- aminoadipic aid (35.9 and 37.4 mg/g respectively) and was comparable to SFV3 in canavanine levels (56.0 and 49.9 mg/g). Populations of S. tomentosa ST1 and ST2 to ST4 had considerable variation with ST1 showing lower levels of amino acids except in arginine (17.0 mg/g) compared to 8.3 and 14.5 mg/g. The S. tomentosa Blouberg Strand populations (ST2 to ST4) had higher levels at aspartic acid (42.9 mg/g), threonine (45.4-87.8 mg/g), asparagine (89.4-166.0 mg/g), proline (43.3-65.9 mg/g) and canavanine (26.0-32.5 mg/g). The Koeberg Nature Reserve populations of S. tomentosa only had high levels of aspartic acid (30.9-40.3 mg/g), asparagine (109.0 mg/g), proline (54.4 mg/g) and canavanine (58.7 mg/g). The variation between populations is random and no obvious taxonomic pattern could be found.

Within - population variation

Plant to plant variation seemed to be higher than interpopulation variation (Figures 5.5.3 to 5.5.5). Within S. frutescens populations, SF3 had higher levels of most amino acids than SF2 and SF4 (these are two other plants from the same population). Interesting variation was observed in S. frutescens var. incana populations, where SFV3 had higher levels of canavanine (49.9 mg/g) than the plant growing right next to it (SFV2), which had much lower levels (4.7 mg/g). The same was true for S. tomentosa populations from Koeberg Nature Reserve (ST5 to ST8). For example, ST2 had low levels of aspartic acid (2.9 mg/ g) and ST4 growing next to it had high levels (10.7 mg/g). ST4 had 65.9 mg/g proline compared to 10.0 mg/g detected in ST3. This remarkably high intrapopulation variation showed that each plant has its own quantitative pattern of amino acids. This is in contrast to other characters (eg. triterpenoids), where the largest part of the variation can be ascribed to population differences.

Leaves vs Seeds The leaf samples of some Sutheriandia species had high levels of canavanine. The levels found in leaves were as high as in the seeds (SFV and ST in Table 5.5.4). However, only one seed sample of S. frutescens was investigated for amino acids. This seed sample was compared to the leaf sample of S. frutescens to determine differences between seeds and leaves (Fig. 5.5.6). S. frutescens seeds had higher levels (60.3 mg/g) of canavanine than leaves (18.0 mg/g) (Table 5.5.4). The leaves of Medicago sativa showed much lower levels of canavanine (0.5 mg/g) (d in Figure 5.5.6) compared to 17.7 mg/g in seeds (c in Figure 5.5.6). Most legume seeds have very high levels of canavanine compared to leaves, but S. frutescens var. incana and S. tomentosa showed levels of 56.6 and 58.7 mg/g respectively in leaves! Remarkably, it seems that the

52 CHAPTER 5 CHEMICAL CHARACTERS

concentration of canavanine in leaves of Sutherlandia can reach levels comparable to that normally found in seeds.

Canavanine concentrations i-

) Quantitative differences were evident between Sutherlandia species. S. frutescens, S. frutescens var. incana, S. montana and S. tomentosa had reasonably high levels of canavanine in their leaves, but these differences may relate to population and plant differences rather than real species differences. Intrapopulation variation in canavanine levels was evident in most of the investigated populations. For example, one plant of S. tomentosa (ST5) had low levels of canavanine (2.3 mg/g) while the plant growing right next to it (ST6) had very high levels (58.7 mg/g).

5.6. Pinitol

Previous workers (Snyders, 1965; Viljoen, 1969 and Brummerhoff, 1969) isolated a cyclitol called pinitol from the leaves of S. microphylla. The objective of this survey was to confirm the presence of pinitol and to determine the approximate concentration of pinitol in Sutherlandia.

Thin Layer Chromatography (TLC)

About 5 1_11 of the Sutherlandia ethanolic extract was applied to F560 Merck silica gel TLC plates and developed in acetonitrile — carbon disulphide — water — formic acid (85:5:10:0.5). The plates were sprayed with chromic acid and burned over an open flame. Brummerhoff (1969) reported pinitol at Rf ca. 0.20; in this study pinitol was at Rf ca. 0.18.

The presence and concentration of pinitol could not be clearly determined using TLC, thus HPLC was used.

High Performance Liquid Chromatography (HPLC) Accurate analyses were done for determining the retention time of pinitol and subsequently its presence in Sutherlandia. The HPLC results revealed pinitol at a retention time of 5.823 (a in Figure 5.6.1). A peak was obtained at the same retention time from the Sutherlandia sample (b in Figure 5.6.1) and the yield in this sample was calculated approximately 14mg/g of dry leaves. The typical chromatograms are presented in Fig. 5.6.1.

53 l.11/4r I Cr% CHEMICAL CHARACTERS

Table 5.5.2 Chemical structures of amino acids detected in Sutherlandia leaf samples.

NAME STRUCTURE RETENTION TIME (min)

0 Phosphoserine H2NCH 2OPO,H2COOH 1.67

0 Taurine CH,NH,CH,SO,H 2.64

0 Aspartic acid HO,CCH,CH(NH2)CO21-1 10.11

0 Threonine CH,CH(OH)CH(N11 2)CO2H 13.41 .._ 0 Serine CH 2(OH)CH(NH 2)C0,11 14.72

0 Asparagine H2NOCCH,CH(NHOCO,H 17.13

0 Glutamic acid HO2CCH 2CH,(NH,)CO2H 18.26

Glutamine H2NCOCH,CH,CH(NH,)CO 2H 19.62

Q Sarcosine CH,NHCH,C0,14 21.95

(1) 24.26 cc-Aminoadipic acid HO,CCH2CHICH2CH(NH2)CO2H

H2C CH2 (1) Proline I I 27.51 H2C\N H — COOH H

0 Glycine H2NCH,CO,H 29.62

L: (E) Alanine HINCH(CH,)CO,H 31.35

HACNHCH2CH:CH2CH(NNCO2H 0 Citrulline 33.26 OII 412) a-Amino-n-Butyric acid H2NCH3CH2CO2H 34.89

0 Valine (C1-13),CHCH(NH2)C0,1-1 36.95

6 Cystine H2NSCH2SCH2CH(NH2)CO2HCHCO2H 38.01

(1) Methionine CH3SCH2CH2CH(NH2)CO2H 39.28

ED Cystathionine H2NCH2CH2SCH,CH(NH2)CO2HCHCO2H 41 .39

54 CHAPTER 5 CHEMICAL CHARACTERS

Table 5.5.2 (continued) Chemical structures of amino acids detected in Sutheriandia leaf samples

RETENTION NAME STRUCTURE TIME (min)

0 Isoleucine CH ICH,CH(CH,)CH(NH 2)CO,H 42.65

4) Leucine (CH 3),CHCH,CH(NH 7)CO,H 43.98

e Tyrosine 46.75 HO CH,CH(NH 7)CO21-1

COON 42) Phenylalanine 51.33 H2N H

CH2 0

6,,, fl-alanine CHANNCH,CO2H 62.99

Camino-iso-butyric acid CHANNCH(CH,)CO,H 68.82

0 y -Aminobutyric acid (GABA) CHANNCH,CH,CO,H 80.32

6 Ethanolamine HOCH,CH2N1-17 85.20

COOH H2N H 0 Tryptophan 86.42 CH2

O ° N I H

Ornithine 98.82 CHANI-12)CH2CH2CH(NH,)CO2H

Lysine J CHANF17)CH,CH,CH,CH(NNCO,H 101.88

Histidine 0 H C— —CH,CH(NF12)CO2H I 105.68 HN\C' N H

55 cliAl" I tt( CHEMICAL CHARACTERS

Table 5.5.2 (continued) Chemical structures of amino acids detected in Sutheriandia leaf samples.

NAME STRUCTURE RETENTION TIME (min)

6:415 Canavanine H2NyNHOCH,CH,CH(NH,)CO,H 110.37 NH

0 Arginine HN—CN(NH,)HCH,CH,CH,CH(NH,)CO,H 123.06

56 CHAPTER 5 CHEMICAL CHARACTERS

Table 5.5.3 Leaf samples of Sutherlandia extracted for amino acids.

Species Abbr. Voucher specimen Locality S. frutescens SF1 Palmer 6 Vanrhynsdorp S. frutescens SF2 Moshe, Van Wyk & De Castro 5 (plant 1) Camps Bay S. frutescens SF3 Moshe, Van Wyk & De Castro 5 (plant 2) Camps Bay S. frutescens SF4 Moshe, Van Wyk & De Castro 5 (plant 3) Camps Bay S. frutescens var. incana SFV1 Van Wyk & De Castro 3668 Pearly Beach S. frutescens var. incana SFV2 Moshe, Van Wyk & De Castro 16 (plant 1) Blouberg Strand S. frutescens var. incana SFV3 Moshe, Van Wyk & De Castro 16 (plant 2) Blouberg Strand S. frutescens var. incana SFV4 Moshe, Van Wyk & De Castro 16 (plant 3) Blouberg Strand S. humilis SH Palmer 17 Uniondale S. microphylla SM Palmer 5 Bitterfontein S. montana SMO Palmer 8 Piquetberg S. speciosa SS Palmer 4 Khamiesberg S. tomentosa ST1 Van Wyk & De Castro 3669 Still Bay S. tomentosa ST2 Moshe, Van Wyk & De Castro 17 (plant 1) Blouberg Strand S. tomentosa ST3 Moshe, Van Wyk & De Castro 17 (plant 2) Blouberg Strand S. tomentosa ST4 Moshe, Van Wyk & De Castro 17 (plant 3) Blouberg Strand S. tomentosa ST5 Moshe, Van Wyk & De Castro 1 (plant 1) Koeberg Nature Reserve S. tomentosa ST6 Moshe, Van Wyk & De Castro 1 (plant 2) Koeberg Nature Reserve S. tomentosa ST7 Moshe, Van Wyk & De Castro 1 (plant 3) Koeberg Nature Reserve S. tomentosa ST8 Moshe, Van Wyk & De Castro 1 (plant 4) Koeberg Nature. Reserve

57 I aDle T lelOS or amino aCICI5 kmgrg ury wergnt) UUteC;LeU Irl led! bdi I Iple kat Ma ulic I M - - 2 o a) 2 0 ca a)

i-N 14 p2 47 col-) 1-0)0.NC0 wo u)o4 I-VI V) 1.- Nt.a) 40 el 000 V) 00Y-. 0 i-NNts'.000-. 2 ch41.- 000 2 co.— u 4 >MM „ 4, ;!.86. > (4 0 ) V .-- i 0 , (nc.i u. 00• i.,0 . u.ce)• 0 C N i7c4 N 4 u_ ,00020202 44 i.NI:Mvt.N.Y-Mc) r ) , J., ViC5 c 0 N ST4 1 ST2 N y- N 0 ,- 00g,? .-6M cid ,66T 0 ino a po ,Occscsictioi 00

e of am O Nam oo 1- r- 66 ' ,1-0W.0.00Nrs..NN00 000 00 ,6076466,6A,Nro '-- 00080;30o ,6c4pc.illu5c\i 6 0 r... s1 1. 70 2.50 ine hoser lo Phosp ,0000,0 ,_ ,, 0 6 e4tr) • t N 0 1.20 a 666 D „loop_ fl 0 O (m ,7.6 aoo -11 MNN 0000 *4'01- NON.0 , 0mc4 o T.; ow oo 0 00 1— Taunne r y-NNCIDMW 000000 oo Nr N.-T-r.- CD 0000 n0 oy-c.c,6 mcl,,ca Y-VT-0 ( m o 4/1.MI-C4.--NOM 0 466FIcom 0000 6,66 vcoNm ,cOeicsic.6666 a; ... • .0 v a . . """n w a . co CC"--L'Y)S, 6 vt0v 00, 6 4

4M2N 2?; co 2 c a V ■ No 6 ° 4 c oo m a fl .WMT- r -.1.-- • - -Nc° . ._ I- . .0 . 4 2 6

ac) 6 1--- 05 ,V4ES00 0 N. Vi Y-CC) oo 6 000 0;c4 o y- C6 020000 a) 0W00000 Q5 . - 0 i 0 p; , a 0 x- 2: .43. c• cj .- U5 6 ^ m 6 - a N 1 --. . fh _. . a 'Di w cr• a 6 oi MMP-53SteiSMN. r 00 05 r 6 0 n, 64o6" E , c5 , 6 0 ....o oo a) o r -.WM-;WWYtV). 1- igs - . - 9.90 3.80 11 7.70 1 4. 30 5.80

8.30 6 1 0 c6 T- 6. 1-- moom0000m 1-

4.80 L 4 ,

Glutamic acid p NOMp o 4000 c4 o M . in co cn° -N' ' 1 25. 1 0 1 3.00 8.90 10.20

1 7.60 90 35. 0466 2.90 csi 25.00 csiciAy:606" 000 0 Glutamine ric.i, c ) 0 0 1.80 1. 1 0

0 0 esi 0 0 C 1-.N co r---nN0 0,-,00

0 62,csi 0 0) Sarcosine U) CIDOIrs o ■ , a i

. 0 ; 0 2.00 2. 90 0. 60 37.40 2. 1 0 a c4 0 0 id e dipic ac CINN:66t 0 -aminoa 0. 30 0 a CD Y- 0 3 666 cvNm0,0m70,07 ,coo ) .

28.40 ' 18. 50 6. 50 43.30 65. 90 6. 80 2. 1 0 Mi-viNsop y- 0000 Y- 36. 80 1. 0000°

Proline ,c46 000

60 2.60 . 1 0. 80 2.60 3.20 3. 90 2.40 5. ui ( Ni 1. 30 CNI et N

6 Glycine ANM.( 1.90 4.90 1 0. 1 0 14. 70

30

6. 1 0.70 0.40 0 4.60 6 o Alanine M0M040N4NM o 0 (N 00 46 My-

0.80 0 0.40 ! 22.30 1. 30 0. 1 0 0.70 c5 oo 6 er)

1 0 , 6 T- 0 ul(c) t-- 0. N0p1"- 00 O Citrulline , a 0 6 0 0 0.20 0.60

0 20 6

90 , I 0 8. 0 " 00 0 id c ino-n-butyric ac 0 r... -am 6 a m 1) 1 ) doo oom :

22. 50 5. 00 7.40 22.60 28.00 p . 0. 1 0 7.60 p 0 TV 6 N 0 Valine D 0 N 0 M

0. 1 0 0.20 0

0

0. 80 o 6 ci

0.80 Ul 0 Nt 6 c5 0 N o st 0

0. 20 Y- c° c) CD 0 Cystine o 1. 30 0. 50 1. 60 0. 90 1.70 30

6. a : 3. 30 a .30 a OD 0 Methionine 07 0 O

tr oi o; o

0 0 1-:66° NONV)01 00000 uivieioici NOWLOW 00000 NI ui COMV)0 00000 00 0 ,c) 'Yr gi rsi gl 2 A e 0

tathionin 00 Cys 6600 MNMP-- ' mmor.... aaao

6. 30 5.60 0.80 5.20 2.60 6.30

, 6. 50 Vi 2. 90 r.: 6 V)

ine o Isoleuc 6 c) 0 , , .c.!70(mco 2. 50 9.70 1.408. 60 2.70 1 2.60 3. 50 ei 6 U5- 0,0 ine 2.20 4 Leuc a a -, 5.20 0 1.60 ; 4. 00 2.507. 40 8.70 lf"D„CONNI-N..

I -- 90 8.40 2. 6 30 a) 3 6

ine N Thyros 1 CD 56

6.206. 90 7.20 j 25. 70 27. 50

23. 90 1 0 0.20 0. 6 ci , el 8. 30 I , Phenylalanine c 4'

1.70 oh 1.00 0.07 0.50 1. 30 1. 50 5.70 66,66,:co 0 p --

50 I 0 p-alanine 0. a CD v-

'A R 0. 80 0. 090 0. 1 0 0. 80 1.20

0 NY- 000000 .4000oo

4.60 1-,c4,,c 011- y-

id 00)1 00000000 butyric ac _01 oi p-aminoiso 6,r.:6c461 o .-a 0000002a oo VI 0 ..- Wo o 1-0 O a m 4.40 1 7.40 1 7.50

05. 30 acid 0 y- ino-n-butyric 8 6 y-am 0 c0 incly e5 000000 .7 . ri ,- 0.90 ; 1. 30 0. 09 2.40 1.30 1 8.80 , 2 R 8F1 4.00 0.60 Ni e 0i csi ■ Ethanolamin pi 0 a co m • 3. 50• 4. 1 09. 70 9.40 0 ! 0 4. 5 0.40 6 6 Y-

X4.90 ; N '

tophan 66 0 Y-LI1 Tryp dei NV) oo CV 08 c5 1-y-t... 000 7 6,y.: : 0 00. 1 0; 0. 1 0 00. 50 ! y-: 7 ,

0.20; NN-:Cf) 0 0000 0 OONM opoo 0 hrs1-. .--,-.m(i Ornithine v- ..... ( N N . i 80 i 2.20I 0.60 4.90 5.70 2. ci : 2.20 0_ 1- MOW 1.20 cNi N 0 Lysine 6 , 6 c.) ,

1.000. 34— 4.208. 90 1 3.20 2.70 4.20 ; N N csi

30 cv N 05 2. mm His tidine oo - 46 2 6 0 M m 7cNI D 6 PI 56.004. 70 4.992. 04 0 2. 30 1 9.1 0 26.00 0 0 O ei •72' 0 csi 1 7. 9 LO CANAVAN INE 0 o 0 6 Nt m 4 2

14.50 IAR GIN IN E 1 1 1.601 9.00] 3.90 1.420. 44 1.70 8.30 _ 1

10 = not detected. tr = trace ( .; 0.05) r. •

CHEMICAL CHARACTERS

Sutherlandia frutescens - Vanrhynsdorp

16

1 3 5 17 5 .c 13 32 25 33 2 f 23 27 29 1 28 30 31 J 9 10 415 t31:AkikA2 24

Sutherlandia frutescens var. incana - Pearly Beach 6 16

3 32

1 13 17 12 23 10

Sutherlandia humilis - Uniondale 3

16 6 8 13 5 II 17 111 15 1 22 23 10 24

Sutherlandia microphylla - Bitterfontein

3 d

Figure 5.5.2 Typical chromatographic profiles (HPLC) of amino acids extracted from leaves of Sutherlandia species. The most abundant and important amino acids in the extracts, peaks 1 through 33 were identified. Chemical structures are provided in Table 5.5.2 and samples agree with those listed in Table 5.5.3.

58 CHAPTER 5 CHEMICAL CHARACTERS

Sutherlandia montana - Reitz

16

27 29 31 23 25A vI 2 A i 33 22 24 26 I

Sutherlandia speciosa - Khamiesberg

16

Sutherlandia tomentosa - Still Bay

16 3 5 1 4 1

Figure 5.5.2 (continued) Typical chromatographic profiles (HPLC) of amino acids extracted from leaves of Sutherlandia species. Amino acids are numbered as in Table 5.5.2 and samples agree with those listed in Table 5.5.3.

t.,

59

VII if,. I CHEMICAL CHARACTERS

Vanrhynsdorp 6 1

3

Camps Bay (plant 1)

Camps Bay (plant 2)

16

1 13 1 8 12 10

Figure 5.5.3 Typical chromatographic profiles (HPLC) of Sutherlandia frutescens populations, showing differences between populations and individual plants from the same population. Amino acids are numbered as in Table 5.5.2 and samples agree with those listed in Table 5.5.3.

60

Ldlur- I cm u CHEMICAL CHARACTERS

Pearly Beach 16

11 13 17 25 272 29

Blouberg Strand (plant 1)

b

Blouberg Strand (plant 2)

6

1„. 2

Blouberg Strand (plant 3)

16

5 11

Figure 5.5.4 Typical chromatographic profiles (HPLC) of Sutherlandia frutescens var. incana, showing differences between populations and individual plants from the same population. Amino acids are numbered as in Table 5.5.2 and samples agree with those listed in Table 5.5.3.

61 CHAPTER 5 CHEMICAL CHARACTERS

Koeberg Nature Reserve (plant 1)

Koeberg Nature Reserve (plant 2)

Koeberg Nature Reserve (plant 3)

Koeberg Nature Reserve (plant 4)

16

17 23 25 31 32 33 202122 24

Figure 5.5.5 Typical chromatographic profiles (HPLC) of Sutherlandia tomentosa populations, showing differences between populations and individual plants from the same population. Amino acids are numbered as in Table 5.5.2 and samples agree with those listed in Table 5.5.3.

62 CHAPTER 6 CHEMICAL CHARACTERS

Blouberg Strand (plant 1)

6

25 ) 7 21 23 27 2 31 2C1 32 t1131.41, 22 24 2514,

Blouberg Strand (plant 2) 16

(77 3 5 17 13 / 21 23 0 1 1 22 A 24

Figure 5.5.5 (continued) Typical chromatographic profiles (HPLC) of Sutherlandia tomento sa populations, showing differences between populations and individual plants from the same population. Amino acids are numbered as in Table 5.5.2 and samples agree with those listed in Table 5.5.3. CHAPTER 5 CHEMICAL CHARACTERS

Sutherlandia frutescens - leaf sample (SF1)

6

11 2 3 5 17 13 f U I 1 25 32 12 1° 23 27 29 1 A11415 111.2gn 3 91 24 30

Sutherlandia frutescens - seed sample (SF1S)

20 27 4 11 13 16 21 2

Medicago saliva - seed sample

U 30 31 33

Medicago saliva - leaf sample

Figure 5.5.6 Typical chromatographic profiles (HPLC) of Sutherlandia frutescens above and E. Medicago sativa (below) showing differences in the relative quantities of amino acids between leaf and seed. Note the relative high level of canavanine in Sutherlandia leaves. Amino acids are numbered as in Table 5.5.2 and samples agree with those listed in Table 5.5.3. CHAPTER 5 CHEMICAL CHARACTERS

START t 4.29 0. 823

10 jOP

START

b •••■••■=st.-.., 5.823 8.332 10 - 10.203

Figure 5.6.1 Typical chromatographic profiles (HPLC) of pinitol (8 mg/ml, top) and Sutherlandia microphylla (bottom). Note the presence of a relative high level of pinitol (ca. 14 mg/g dry wt of leaf material) in the Sutherlandia sample at a retention time of 5.823 minutes.

Conclusions

Amino acids do not seem to be of any chemotaxonomic value. This is clear from Table 5.5.3 (and also from Figures 5.5.2 to 5.5.6) where plant to plant variation (within a single population) often exceeds the interpopulation variation. Each plant therefore seems to have a unique amino acid pattern, regardless of population or species. '

The discovery of large concentrations of amino acids in the leaves of Sutherlandia was one of the most important new insights gained during this study. In view of the medicinal value of Sutherlandia, these compounds are important because they provide a rationale behind the reported beneficial effects in the use of Sutherlandia as a general tonic and in cancer treatment (see Chapter 6).

The presence of pinitol is also of some interest since this compound is known to have antidiabetic properties. A relatively easy techinque to detect and quantify pinitol in Sutherlandia samples has been developed and it would be worthwhile to explore the presence and levels of the compound in future studies. CHAPTER 6

MEDICINAL VALUE

6.1 Historical background

The genus Sutherlandia, commonly known as the cancer bush, was used medicinally by the Khoi people and the early colonists as a cure for cancer. Leaf infusions are also used internally for stomach and intestinal ailments, uterine troubles, as a cough remedy, and as a tonic (Smith, 1895; Dykman, 1908; Watt & Breyer-Brandwijk, 1962; Archer, 1990; Van Wyk et al., 1997). Examples of the traditional medicine is shown in Figure 6.1.1.

Sutherlandia infusions is an old Cape remedy for cancer and other ailments (Dykman, 1908). In Namaqualand, Leliefontein inhabitants are highly dependent on their traditional medicine. Leaf and flower infusions of Sutherlandia are sometimes mixed with the stomach contents of the porcupine (Hystrix erica australis) and the mixture is used for stomach complaints (Archer, 1990). Sutherlandia frutescens is used as a basic ingredient in most medicines. It has an extremely bitter taste and this is seen as indicative of the potency of the medicine (Archer, 1990). Gabrielse (1996) reports of a patient diagnosed with pancreas cancer, who later after drinking Sutherlandia tea two times a day (45 days), has been well for four years. Gabrielse (1996) conducted studies on leaf extracts of Sutherlandia using HPLC and isolated a peak at a retention time of 11.9 minutes. His findings suggested that this peak contained the "bitter components" isolated by BrOmmerhoff (1969). Gabrielse (1996) could not prove any anti-cancer properties.

Despite the alleged anticancer activity of Sutherlandia, nothing is known or recorded regarding the chemical composition of the plant except for the non-protein amino acid canavanine which was extracted from seeds of Sutherlandia frutescens (Bell, 1958) and the presence of pinitol which has been isolated from S. microphylla growing in the Free State Province (Viljoen, 1969; Brummerhoff, 1969). Investigations of the main chemical compounds of the genus were done and detailed information is given in Chapter 5. The leaves of the plant were found to contain a large number of common and some uncommon amino acids. Pinitol was also present in relatively high yields and these might be related to the alleged medicinal activity of the plant.

66 MEDICINAL VALUE

Gans ies

Ons Ounes net nit get rai. .ir

LI JIC Clestrent en i ge k.aal

lidt Zrek 1 Eetle0e1 a00gell 20 eat te I KOok-,i , Drink Vas 3-; mas, on, Crittesc,:lis

-ALAND PLAASMUSEUM l',U1.11!111 rARM MUSEUM WIMICES I ER

Figure 6.1.1 Traditional medicines produced from aerial parts of Sutherlandia species.

6. 2 Importance of triterpenoids Triterpenoids are known to be biologically active against various diseases and anticancer activity has also been reported. Some examples of medicinally important triterpenoids include oleanic acid, putranoside, swartziasaponin, asiaticoside, soyasapogenol and medicagenic acid (Southon, 1994; Merck Index, 1989). These compounds usually act as spindle poisons, thereby preventing the proliferation of cancer cells (Bruneton, 1995).

Antitumour and anticancer activity of triterpenoids The relation between chemical structure and anticancer activity of some pentacyclic and tetracyclic triterpenoids was studied by Ling et al., 1982 (in Mahato et al., 1992). The anticancer activity of triterpenoids was tested against various human cancer cell lines (ME-180, u-87MG, SK-HEP-1, CALU-1, CAMA-1 and HEC-1-A). Some triterpenoids showed significant cytotoxic effects, for example ursolic acid in the lymphocytic leukaemia cells P-388, L-1210, as well as the human lung carcinoma cell A-549 (Ling et al., 1982 in Mahato et al., 1992).

In Sutherlandia, two major triterpenoids were isolated but the purificatiom and identification proved to be problematic. There is a possibility that these compounds are responsible for some of the medicinal activity of the plant, but a detailed investigation of triterpenoids was considered impractical and beyond the scope of this study.

67 CHAPJ tK MEDICINAL VALUE

6.3 Importance of amino acids Many amino acids are known to posses growth-inhibitory properties, particularly in microorganisms (Meister, 1965; Rosenthal, 1977; Bruneton, 1995; Rosenthal & Harper, 1996).

Canavanine L- canavanine, 2-amino-4 (guanidinooxy) butyric acid, a structural analogue of arginine, is one of the predominant non-protein amino acids synthesized by many plants of the family Fabaceae (Bell, 1958). Canavanine is a common seed metabolite of most legumes but is also stored in vegetative organs (Park & Kwon, 1990 in Hwang et al., 1996) and in vacuoles of leaves of Canavalia lineata (Yu & Kwon, 1992 in Hwang et a/., 1996).

Canavanine has antimetabolic effects and inhibitory activity against many organisms. Numerous scientific papers have reported the activities of canavanine against cancer, colds and flu viruses (Horowitz & Srb, 1948; Miller & Harrison, 1950; Schwartz et al., 1996; Ranki & Kaariainer, 1969; Neurath, 1970). In addition to storing nitrogen, it also acts as an insecticide (Rehr et al., 1973; Rosenthal, 1977; Rosenthal & Rhodes, 1984; Rosenthal, 1991; Rosenthal, 1992; Rosenthal et a/., 1995).

The antitumor properties of canavanine have been illustrated in mice bearing L1210 leukemic cells where canavanine-treated mice live longer than untreated mice (Green et a1.,1980; 1983). This observation was supported by Thomas et. al. (1986) who found that canavanine inhibited growth of the rat colon carcinoma. Canavanine also inhibited growth of pancreatic cancer cells (Rosenthal et al., 1995).

Due to the unknown biochemical basis for the anticancer effects of canavanine, various hypotheses have been formulated. The first hypothesis involves the replacement of arginine (a structural analogue of canavanine) by canavanine in most metabolic reactions (Rosenthal, 1977; Neurath et al., 1979; Rosenthal & Dahlman, 1991). This substitution results in confusion within the canavanine-sensitive species as they find it difficult to differentiate between the two compounds and as a result structural and functional protein aberrations occur (Attias et al., 1969; Prouty et al., 1975).

Another hypothesis involves the interference of canavanine in RNA synthesis (Schachtele & Rogers, 1965) and the disruption in the reactions of DNA replication and transcription. For example, canavanine treatment of Semilik Forest virus prevented RNA polymerase synthesis (Ranki & Kaariainen, 1969). In adddition, canavanine disrupted the reactions of DNA metabolism, as demonstrated by reduced DNA synthesis in herpes simplex exposed to canavanine (Bell, 1974).

68 CHAPTER 6 MEDICINAL VALUE

Inhibition *of tumor growth can result from canavanine incorporation into tumor proteins, causing the production of aberrant macromolecules exhibiting impaired function (Kruse et al., 1959; Rosenthal et al., 1989). Anomalous canavanyl proteins are degraded more rapidly than their normal counterparts (Knowles et al., 1975). The preferential degradation of proteins in canavanine-treated cells has been demonstrated in prokaryotic and eukaryotic cells, including E. coli (Pine, 1967; Goldberg, 1972; Prouty et al., 1975;), Hela (Hendil, 1975) and hepatoma cells (Knowles et al., 1975). It is possible that in the purging of the aberrant canavanyl-proteins, canavanine-free proteins are degraded inadvertently (Thomas et al., 1986).

Arginine attenuates the growth of several experimental tumors (Cho-Chung et al., 1980; Nakanishi, 1969; Takeda et al., 1975; Weisburger et al., 1969). Both canavanine and arginine may affect growth adversely by similar means; for example Takeda et a/. (1975) have established that polyamine synthesis is curtailed in rats maintained on an arginine-enriched diet. Polyamines are important in rapidly proliferating cells, so that curtailed polyamine synthesis may attenuate tumor growth (Thomas et a/., 1986). Since canavanine competes with arginine in virtually all metabolic reactions in which arginine is a substrate (Rosenthal, 1977), it is possible that canavanine also impedes polyamine synthesis (Thomas et al., 1986).

High levels of canavanine detected in leaves of Sutherlandia provides a highly plausible rationale for the traditional use of Sutherlandia in cancer treatment. It appears to be the first (only?) example of the therapeutic use of canavanine in a traditional phytomedicine.

Arginine and the Nitric Oxide pathway Nitric oxide (NO) is formed from L-arginine (Palmer et al., 1988) by the nitric oxide synthase (NOS) (Stuehr et al, 1991, Das & Khan, 1995) and NO is synthesized by many cells and tissues (vascular endothelium, platelets and macrophages) (Furchgott et al., 1980; Palmer et al., 1988; Ignarro et al., 1987; Xie et al., 1992). Other cells capable of producing nitric oxide include neutrophils, neurons and some tumor cells (Palmer et al., 1987; Hibbs et al., 1988; Garthwaite et al., 1988; Radomski et al., 1991; Cendan et al., 1996).

Nitric oxide is a messenger molecule that has many functions in living organisms, including smooth muscle relaxation (Busse & Mulsch, 1990; Konturek & Konturek, 1995), platelet aggregation (Radomski et al., 1991; Wu et al., 1996), neurotransmission (Bredt et al., 1991), immune cell activation, tumor cell killing (Moncada et al., 1991, Moncada & Higgs, 1993; Wink, 1998), protection against damage to cardiac myocytes (Gorbunov et al., 1998), protectection of endothelial cells and maintainance of vascular wall integrity (Polte et al., 1997). In addition NO h as been found to induce intestinal fluid secretion in human jejunum and prevent intestinal ischaemia (Stark & Szurszewski,

69 GMAYItK 5 MEDICINAL VALUE

1992; Mourad et al., 1996). NO also exhibit potent activities in the central and peripheric nervous system, and also in the cardiovascular system (Furchgott et al., 1980; Ignarro et al., 1987; Moncada & Higgs, 1993;). It serves as an intra- and intercellular mediator (Moncada et al., 1991; Nathan, 1992; Stamler et al., 1992). NO is reported to have 'anti-atherosclerotic effects' (Ross, 1993; Boger et al., 1996). Nitric oxide also plays a major role in dysfunction of septic shock (Kelly et a/., 1995) which results in hypotension, vascular leak and circulatory failure leading to death (Root & Jacobs, 1991).

L-Arginine acts as a common substrate for a number of metabolic reactions in the central nervous system. These include polyamine and nitric oxide production (Das & Khan, 1996). In addition to regulating cell growth and differentiation (Joshi, 1997), polyamines are essential in cell membrane functions (Ramchand et al., 1994; Morgan, 1995). Arginine is also reported to be used in diet supplementation to inhibit atherogenesis (Cooke & Tsao, 1997), shortening tumor regression, retarding tumor growth thereby reducing tumor size (Ma et al., 1996) and it also helps in inhibiting platelet aggregation (Adams et al., 1997; Wascher et al., 1997; Wolf et al., 1997). Although Adams et al., 1997 demonstrated that oral arginine supplemention inhibited platelet aggregation, Khan et al., 1997 concluded that increased oral L-arginine supplementation was unlikely to increase the vasodilator ability in primary Raynaud's phenomenon (RP) patients. The use of Sutherlandia as a general tonic (Archer, 1990; Van Wyk et a/., 1997); may well be partly due to the presence of arginine.

-y-Aminobutyric acid (GABA) In the mammalian brain, y-Aminobutyric acid (GABA) is said to be the major inhibitory neurotransmitter (Stephenson, 1995). There is substantial evidence proving GABA to be an excitory neurotransmitter (Hahner, 1991; Colwell, 1997). GABA mediates its effects through receptors known as GABAA and GABAB. These receptors send messages through to the brain via ion channels; GABAA receptors are the fastest (Hahner, 1991; Colwell, 1997). Wagner et al. (1997) reported that mature neurons in the suprachismatic nucleus (SCN) of the hypothalamus can be excited by GABA. The smoking of Sutherlandia seeds has been recorded (Dr Nigel Gericke, pers. comm.) and the plant may well have some psycoactive properties as a result of the high levels of GABA.

6.4 Pinitol Pinitol is one of the major cyclitols in plants (Kuo et a/., 1997) and has been found in various plant organs. Although pinitol has been found to be the major cyclitol in soybean seeds, other plant seeds accumulate myo-inositol (Kuo et al., 1997). Pinitol has been reported to have many physiological functions. In soybean leaves it has been shown to accumulate during stressful periods especially at high temperatures and was also found to have anti-growth activity for various insects. There is relatively little information regarding the general functions of pinitol, but it is known to have hypoglycaemic and antidiabetic activity (Buckingham, 1994).

70 CHAPTER 6 MEDICINAL VALUE

Conclusion The three above-mentioned amino acids were detected in high yields in leaves of Sutherlandia and they are known to be responsible for numerous biochemical activities and for curing many diseases in the human body. Considering all the diverse metabolic functions that these amino acids perform and given the traditional anticancer use of Sutherlandia, there is undoubtedly a link between the two. Thus, the amino acids, triterpenoids and pinitol found in Sutherlandia leaves provide a rationale behind the medicinal uses of the plant which would be worth exploring in more detailed studies.

71 CHAPTER 7

PHENETIC ANALYSIS

Introduction

The aim of this study was to investigate and summarize infrageneric relationships between the closely related species and regional forms of Sutherlandia. The data was analysed phenetically to quantify and conceptualise overall similarities among the species. The computer programme NTSYS-PC 2.0 (Rohlf, 1997) was used to construct phenograms. A total of 51 provenances (populations) was included as Operational Taxonomic Units (OTU's). Each population is represented by one or more herbarium specimens, as shown in Table 7.1. Three of the populations (OTU's 13 & 14, 15 & 16, 46 & 47) were subsampled to study the effect of subtle differences between individual plants from the same population.

A data matrix of 25 characters and 51 OTU's was constructed. Character and character states included in the analysis are given in Table 7.2. The data matrix is given in Table 7.3. Four different coefficients of similarity were used to generate phenograms. In the first three analyses, the SIMNT algorithm in the SAHN option of the programme was used with UPGMA as clustering method. In these three analyses a distance coefficient (analysis 1), a correlation coefficient (analysis 2) and a Manhattan coefficient (analysis 3) were used (Fig. 7.1 to 7.3 respectively). In the fourth analysis, the SIMQUAL option of the programme was used and a simple matching coefficient was calculated (Fig. 7.4) (Rohlf, 1997). Various other methods were also tested but only the above-mentioned four analyses are reported here.

72 •.• 10111- I ,111 PHENETIC ANALYSIS

Table 7.1 List of populations and specimens of Sutherlandia 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) Ghamsberg 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) Graaf-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, Thlken & Venter 388 (PRE) Kodakspiek 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) VVitsand S. tomentosa 4 Van Wyk 3669 (JRAU) Still Bay S. tomentosa 5 Moshe, Van Wyk & De Castro 1 (JRAU) Koeberg Nature Reserve

73 VII 110,1 • LAI • • PHENETIC ANALYSIS

Table 7.2 Characters and character states used for phenetic analysis. 1, Maximum leaf length (mm). 2, Maximum number of leaflet pairs. 3, Minimum number of leaflet pairs. 4, Maximum petiole length (mm). 5, Maximum leaflet width (mm). 6, Maximum leaflet length (mm). 7, Maximum leaflet length : maximum leaflet width. 8, Leaflet apex (1 = acute; 2 = rounded; 3 = emarginate). 9, Leaflet pubescence abaxial (1 = sparsely pubescent; 2 = pubescent ). 10, Leaflet pubescence adaxial (1 = glabrous; 2 = densely pubescent). 11, Maximum petiolule length (mm). 12, Minimum petiolule length (mm). 13, Mean petiolule length (mm). 14, Twig pubescence (1 = sparsely pubescent; 2 = densely pubescent). 15, Maximum flower length (mm). 16, Maximum pod length (mm). 17, Maximum pod width (mm). 18, Maximum pod breadth (mm). 19, Maximum pod length : maximum pod width. 20, Maximum pod length : maximum pod breadth. 21, Maximum pod stipe length (mm). 22, Maximum pedicel length (mm). 23, Pod stipe orientation (1 = in line with the pod; 2 = closer to upper suture; 3 = closer to lower suture). 24, Habit (1 = erect; 2 = erect to procumbent; 3 = procumbent to prostrate). 25, Size (1 = large, > 1.5 m; 2 = medium, 0.3-1.2 m; 3 = small, < 0.3 m).

74 CHAPTER 7 PHENETIC ANALYSIS

Table 7.3 Matrix of 51 OTU's and 25 characters used in constructing the phenograms of the genus Suthedandia in Figures 7.1 to 7.4 (OTU's as in Table 7.1); characters and character states as in Table 7.2).

Character number OTU 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 incanal 60 10 8 9 4 11 3 2 2 1 0.8 0.6 0.7 2 32 40 24 23 1.7 1.7 5 7 2 2 2 incana2 57 7 5 11 3 11 4 2 2 1 0.8 0.7 0.75 2 32 57 26 28 2.0 2.0 6 6 2 2 2 incana3 63 10 8 7 4 8 2 2 2 1 0.7 0.7 0.7 2 33 55 25 26 2.0 2.1 3 7 2 2 2 incana4 56 9 8 8 3 10 4 2 2 1 0.4 0.3 0.35 2 33 51 27 28 1.9 1.8 4 7 2 2 2 incana5 68 8 7 7 3 11 4 2 2 1 0.7 0.4 0.55 2 32 55 25 24 2.0 2.3 3 7 2 2 2 incana6 62 9 7 7 3 9 3 2 2 1 0.4 0.3 0.35 2 30 54 29 34 1.9 1.6 3 5 2 2 2 incana7 55 8 7 6 6 8 4 2 2 1 0.6 0.6 0.6 2 30 40 20 25 2.0 1.6 3 7 2 2 2 frutescens1 82 9 7 13 3 15 5 1 1 1 0.9 0.7 0.8 1 25 45 20 26 1.8 1.7 4 6 2 1 1 frutescens2 65 9 6 12 3 13 4 1 1 1 0.8 0.6 0.7 1 30 45 20 24 1.9 1.9 4 7 2 1 1 frutescens3 75 8 6 13 3 14 5 1 1 1 1.0 0.6 0.8 1 25 48 25 30 1.9 1.6 2 7 2 1 1 frutescens4 83 8 7 14 3 12 4 1 1 1 1.0 0.7 0.85 1 27 45 25 23 1.8 1.9 2 7 2 1 1 frutescens5 65 9 5 12 3 12 4 1 1 1 1.3 0.7 1.0 1 28 42 22 20 1.9 2.1 5 8 2 1 1 frutescens6 70 6 4 13 3 13 4 1 1 1 0.7 0.7 0.7 1 25 43 21 23 2.0 1.9 5 7 2 1 1 frutescens7 65 9 6 11 4 13 3 1 1 1 0.7 0.5 0.6 1 30 50 29 23 1.7 2.1 3 8 2 2 3 frutescens8 82 8 6 14 3 14 4 1 1 1 0.7 0.7 0.7 1 32 40 25 21 1.6 1.9 5 9 2 2 3 frutescens9 75 8 5 12 3 10 3 11 1 0.7 0.7 0.7 1 32 53 24 25 2.0 2.1 4 8 2 1 1 frutescens10 75 9 7 12 2 11 5 1 1 1 1.0 1.0 1.0 1 32 55 25 20 2.0 2.7 7 6 2 2 3 frutescens11 70 8 5 12 3 14 5 1 1 1 0.4 0.3 0.35 1 32 60 19 15 3.1 4 5 6 2 2 3 3 frutescens12 65 6 4 14 2 12 6 1 1 1 0.6 0.5 0.55 1 25 55 23 16 2.0 3.4 5 6 2 2 3 humilis1 73 6 4 14 3 10 3 2 1 1 0.6 0.6 0.6 1 29 40 24 26 1.7 1.5 3 5 2 2 3 humilis1 51 6 5 10 3 10 4 2 1 1 0.8 0.5 0.65 1 25 40 21 15 1.9 2.6 3 5 2 2 3 humilis1 83 6 5 14 3 13 4 2 1 1 1.0 0.7 0.85 1 28 42 20 19 2.0 2.2 4 5 2 2 3 humilis1 62 6 5 12 3 11 4 2 1 1 0.8 0.7 0.75 1 29 43 22 17 1.9 2.5 2 6 2 2 3 humilisl 50 5 3 10 2 8 4 2 1 1 0.3 0.2 0.25 1 26 38 15 21 2.0 1.8 5 6 2 2 3 microphyllal 70 7 5 13 3 12 4 1 1 1 1.0 1.0 1.0 1 32 50 19 16 2.6 3.1 10 8 1 1 1 microphylla2 81 9 7 10 3 10 3 1 1 1 0.9 0.7 0.8 1 30 55 12 19 4.5 2.8 6 7 1 1 1 microphylla3 76 8 5 6 3 16 6 1 1 1 1.0 0.8 0.9 1 31 55 10 17 5.5 3.2 9 6 1 1 1 microphylla4 72 7 4 6 3 15 5 1 1 1 0.8 0.5 0.65 1 30 55 12 17 4.5 3.2 9 7 1 1 1 microphylla5 53 9 6 10 2 9 4 1 1 1 1.0 0.6 0.8 1 35 57 13 19 4.3 3.0 10 8 1 1, 1 microphylla6 79 9 7 15 3 11 3 1 1 1 1.0 0.7 0.75 1 30 46 9 15 5.1 3.0 9 8 1 1 1 microphylla7 65 6 6 15 4 16 4 1 1 1 0.9 0.4 0.65 1 27 45 8 12 5.6 3.7 8 8 1 1 1 microphylla8 60 10 8 12 3 6 2 1 1 1 0.4 0.3 0.35 1 30 45 15 21 3.0 2.1 3 9 1 1 1 microphylla9 40 7 6 10 3 6 3 1 1 1 0.5 0.4 0.45 1 32 45 10 15 4.5 3.0 9. 9 1 1 '1 microphylla10 70 10 8 10 3 8 3 1 1 1 1.0 0.9 0.95 1 35 45 7 14 6.4 3.2 9 8 1 1 1 microphylla11 60 9 8 12 3 8 3 1 1 1 1.0 1.0 1.0 1 32 58 17 10 3.4 5.8 10 8 1 1 1 montanal 74 9 7 11 4 10 3 2 1 1 0.6 0.6 0.6 1 40 55 16 24 3.4 2.2 15 8 2 2 2 montana2 90 9 10 15 5 14 3 2 1 1 0.6 0.3 0.45 1 35 50 17 22 2.9 2.2 108 2 2 2 montana3 65 8 6 5 3 11 4 2 1 1 0.7 0.5 0.6 1 47 55 19 22 2.8 2.5 11 8 2 2 2 montana4 90 9 6 11 6 11 2 2 1 1 0.5 0.4 0.45 1 42 55 16 25 3.4 2.2 9 8 2 2 2 montana5 65 10 6 10 5 7 2 2 1 1 0.7 0.4 0.55 1 41 50 19 15 2.6 3.3 9 8 2 2 2 montana6 40 8 6 10 5 10 3 2 1 1 0.7 0.3 0.5 1 38 50 16 17 3.1 2.9 9 8 2 2 2 montana7 40 8 7 8 5 9 4 2 1 1 0.6 0.5 0.55 1 39 49 14 18 3.5 2.7 9 8 2 2 2

Table 7.3 (continued) Matrix of 51 OTU's and 25 characters used in constructing the phenograms of the genus Suthedandia in Figures 7.1 to 7.4 (OTU's as in Table 7.1; characters and character states as in Table 7.2)

Character number OTU 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 speciosa3 65 6 5 9 6 15 2 2 1 1 1.0 0.3 0.65 1 52 50 23 31 2.0 1.6 15 8 3 3 2 speciosa4 65 8 7 10 3 17 6 2 1 1 1.0 0.6 0.8 1 45 50 26 35 1.9 1.4 15 8 3 3 2 tomentosal 65 10 9 5 5 8 2 3 2 2 0.6 0.4 0.5 2 30 50 24 37 2.0 1.3 2 6 2 2 2 tomentosa2 75 10 9 8 5 8 2 3 2 2 0.6 0.5 0.55 2 32 55 28 37 1.9 1.4 3 6 2 2 2 tomentosa3 80 10 8 8 8 9 1 3 2 2 0.5 0.3 0.4 2 32 60 24 28 2.0 2.1 36 2 22 tomentosa4 70 11 8 12 6 10 2 3 2 2 1.0 0.4 0.7 2 32 60 29 44 2.Q 1.3 6 6 2 2 2 tomentosa5 60 11 10 6 4 5 1 3 2 2 0.3 0.1 0.2 2 32 50 20 22 2.0 2.2 36 2 22

75 _••- ■ • • -• • • PHENETIC ANALYSIS

Results and Discussion The results of the analyses did not always group populations in the expected way, i.e., populations from different taxa were sometimes closely associated in the phenograms. The choice of coefficient has a marked effect on the structure of the phenograms, not only in the degree of similarity but also in the inclusion (grouping) of the populations.

Analysis 1 (Distance coefficient)

This coefficient measures the distance between OTU' s in a space defined in various ways. It actually measures dissimilarity. No resolution was found with this coefficient as obviously related populations did not form groups as would be expected (Fig. 7.1). Note, for example, that two practically identical subpopulations od Sutheriandia frutescens var. incana from Aurora ( ■ = incana 1 and incana 2), grouped into two different phenons. The two subpopulations of S. tomentosa (• = tomentosa 1 and tomentosa 2) and S. microphylla (*= microphylla 4 and microphylla 5) were grouped even wider apart.

Analysis 2 (Correlation coefficient)

The coefficient measures proportionality and independence between pairs of OTU vectors. It is among the most frequently employed in numerical taxonomy and in plants (Soria & Heiser, 1961 in Rohlf, 1997; Morishima, 1969b in Rohlf, 1997). Although this coefficient did give some satisfying groupings, it isolated related populations into smaller clusters, for example, clusters 1 and 3 and clusters 2 and 4. Sutheilandia speciosa (cluster 5) seems to form a convincing group and this reflects the significance of its unique pod stipe orientation (character 23) (Fig. 7.2).

Analysis 3 (Manhattan distance coefficient)

This coefficient is commonly used as a measure of dissimilarity in numerical taxonomy (Rohlf, 1997). The Manhattan distance coefficient gave better groupings of some related populations than the above-mentioned two coefficients. The expected grouping of S. tomentosa with S. frutescens var. incana (cluster 1), S. frutescens with S. humilis (cluster 2), and S. speciosa with S. montana (cluster 3) and the various populations of S. microphylla (cluster 4) are positive aspects of this phenogram. It however does not accurately reflect relationships, as S. microphylla 1 is, for example, found in the S. frutescens group (cluster 2) (Fig. 7.3).

76 PHENETIC ANALYSIS

Incanal incana7 microphylla8 tomentosa5 frulescens2 frutescens5 frutescens6 humilis4 humilisl Incana2 incana4 Incana3 incana5 frutescens7 Incana6 tomentosal frutescens1 frutescens4 frutescens8 humilis3 frutescens3 frutescens9 frutescens1 0 tomentosa3 frutescens11 frutescens12 mlcrophylla 1 mlcrophylla2 microphylla3* microphylla4 microphyIla6 microphyllal 0 microphylla7 montanal montana3 montana5 speciosal speciosa2 speciosa3 speclosa4 tomentosa2 • tomentosa4 montana2 montana4 humilis2 humilis5 microphylla9 L montana6 montana7 mIcrophylla5 microphylla11 • 0.77 2.13 3.30 4.86 6.23 Coefficient

Figure 7.1 Phenogram of 51 populations and 25 characters based on Distance coefficient to illustrate phenetic similarity among the species and subspecies of Sutherlandia.

77 PHENETIC ANALYSIS

Incanal incana7 ncana3

umilis4 frutescens7 humilis5 incana2 incana4 incana6 i—tomentosa 1 t—tomentosa2 tomentosa4 frutescens1 frutescens4 frutescens6 humilis1 hfrutescens8 umilis3 frutescens3 frutescens11 frutescensl2 microphylla11 microphyllal microphylla8 montanal montana3 montana5 microphylla2 fnicrophylla3 trnicrophylla4 microphylla7 --microphylla6 r montana2 montana4 microphylla1 0 speciosal rspeciosa2 t-speciosa3 speciosa4 microphylla5 microphylla9 ontana6 Lm ontana7 0.94 0.95 0.97 0.98 1.00 Coefficient

Figure 7.2 Phenogram of 51 populations and 25 characters based on Correlation coefficient, to illustrate phenetic similarity among the species and subspecies of Sutherlandia.

78 - • • - - • - - • PHENETIC ANALYSIS

incanal Incana7 tomentosa5 Incana2 incana4 incana3 Incana5 incana6 tomentosal tomentosa2 tomentosa4 tomentosa3 frutescens1 frutescens4 frutescens3 frutescens8 humilis3 humilis1 frutescens2 frutescens5 frutescens6 frutescens7 frutescens9 frutescens10 frutescens11 microphyllal frutescens12 humilis2 humilis4 humilis5 montanal montana4 montana2 montana3 montana5 speciosal speciosa2 speciosa3 speciosa4 microphylla2 microphylla6 microphylla3 microphylla4 microphylla7 microphylla10 microphylla5 microphylla11 microphylla8 microphylla9 montana6 montana7 I T 0.48 110 1.72 2.34 2.96 Coefficient

Figure 7.3 Phenogram of 51 populations and 25 characters based on Manhatan coefficient, to illustrate phenetic similarity among the species and subspecies of Sutherlandia.

79 PHENETIC ANALYSIS

incanal Inca na2 incana3 incana5 Inca na6 Incana7 incana4 omentosa 1 Itomentosa2 tomentosa3 tomentosa4 -tomentosa5 frutescens1 frutesce ns2 frutesce ns4 frutescens6 frutesce ns3 frutesce ns5 frutescens9 -microphylIa11 microphyllal microphylla2 microphylla6 microphylla1 0 microphylla3 microphylla4 microphylia9 microphylla8 microphylla 5 microphylla7 frut esce ns7 frutescens8 . frutescens1 0 frutescens11 frutescens 12 umilis1 humilis2 humilis3 humilis4 humilis5 montanal montana4 montanal montanal rnontana3 montana6 montana5 speciosal speciosa3 speciosa4 speciosal I 0.18 0.29 0.41 0.52 0.64 Coefficient

Figure 7.4 Phenogram of 51 populations and 25 characters based on Simple Matching coefficient, to illustrate phenetic similarity among the species and subspecies of Sutherlandia.

80 l.rIkkr I Cr( / PHENETIC ANALYSIS

Analysis 4 (Simple Matching coefficient)

This analysis gave the best results and grouped related populations and species according to the outcome of the morphological studies and enzyme electrophoresis study. Thus this analysis agrees most closely with my ideas about taxonomic patterns and infrageneric relationships. The phenogram has two groups (clusters 1 and 2). The first group comprises Sutherlandia tomentosa and S. frutescens var. incana (i.e. the two hairy taxa) and the second groups comprises S. frutescens in the broader sense. These two basic groups are congruent with the proposed change of circumscription and rank (see Chapter 8), but the subspecies and forms in the S. frutescens group are not satisfactorily represented (Fig. 7.4). Note that the taxa previously regarded as species group almost perfectly into clusters. These include the hairy coastal form of S. frutescens, hirtherto known as S. frutescens var. incana (cluster 3), the hairy coastal species S. tomentosa (cluster 4, a particularly convincing cluster), the high-altitude S. montana (cluster 5) and the Namaqualand species S. speciosa (cluster 6). Note the partial grouping of populations of S. frutescens, S. microphylla and S. humilis in cluster 7, where there is considerable overlap.

The most obvious inconsistencies are the S. frutescens var. incana cluster which groups with S. tomentosa rather than S. frutescens. A similar result was obtained in enzyme electrophoresis and I ascribe this to possible introgression between S. frutescens var. incana and S. tomentosa. Similarly S. montana (cluster 5) appears to be a very distinct Glade related to S. speciosa rather than S. frutescens. This taxon is undoubtedly associated with S. frutescens. Important characters such as stipe orientation and fruit shape are in conflict with associating S. montana with S. speciosa and the similarities may well be superficial only.

Conclusions

Ideally, one would like to reflect overall similarity in the final taxonomic hierarchy. Such a mechanical way of dealing with the complexicity in the Sutherlandia species complex would certainly be an oversimplification. The hierarchy would vary depending on which character are thought to be more significant than others and it is clear that no one single character (or combination of characters) lead to any logical hierarchy. It is interesting to note that similar results were obtained in the electrophoresis study where allozyme patterns did not agree with diagnostically significant morphological characters. The proposed hierarchy in Chapter 8 should therefore perhaps not be taken too seriously, as I merely tried to reflect the overall patterns of similarity to present a practical solution for identifying and naming the various regional forms within the species. A dogmatic approach to species and infraspecific circumscriptions would give a false impression that character evolution in Sutheriandia proceeded in a hierarchial fashion.

81 CHAPTER 8

GENERAL CONCLUSIONS

The habit of Sutherlandia is quite variable. The small and prostrate habit is diagnostic for S. humilis but the overall variation limits the use of this character to distinguishing other species (S. tomentosa, S. frutescens, S. montana and S. microphylla). Some forms of S. frutescens are included in S. humilis purely on the basis of their small size.

Leaflet shape and pubescence can be regarded as reliable characters for distinguishing 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. The densely tomentose adaxial surface and the obcordate shape of the leaflets satisfactorily distinguishes S. tomentosa from the rest of the species.

Flower length proved to be of diagnostic value 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 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. The fruit length to width ratio differentiates S. microphylla from the rest of the species. The stipe orientation appears to have been overlooked in the past and is a useful character to differentiate S. speciosa from the rest of the species.

Enzyme electophoresis has, in agreement with morphology, shown that there is a highly complex gene pool and that the species are closely related. This suggested that the current classification of the genus as species is not well supported. The subspecies rank therefore seems to be more appropriate in dealing with regional variation. The results also revealed the products of introgression between the two taxa (S. tomentosa and S. frutescens var. incana). These two taxa are known to co-occur and intermediates were also observed.

Although secondary metabolites have proved to be of taxonomic value in most genera and species, this was not the case in Sutherlandia. Of all the chemical compounds extracted, none were of taxonomic significance. In view of the medicinal value of the plant, large concentrations of amino acids were detected and identified. Some major compounds identified in Sutherlandia

82 1,11A1-' I GENERAL CONCLUSIONS

have known anticancer activity, and provide a rationale behind the traditional use of the plant.

The morphology and allozyme data support the notion that the species are very closely related. The Simple Matching coefficient reflected the expcted infrageneric grouping.

The results from this multidisciplinary study suggest that a more conservative taxonomic classification system for the genus Sutherlandia is called for and therefore I propose the following changes to be made in the circumscription of taxa (Table 8.1).

Table 8.1. Proposed changes in the classification of Sutherlandia.

Current classification Proposed classification

S. frutescens (L.) R. Br. ■ • ■ ■ ■ ■ ■ ■ ■ S. frutescens subsp. frutescens (typical form)

S. frutescens var. incana E. Mey. Linnaea 01. (hairy form)

S. humilis Phillips & Dyer Ol• (dwarf form)

S. montana Phillips & Dyer 0 (high altitude form)

S. microphylla Burchell ex DC. S. frutescens subsp. microphylla

S. speciosa Phillips & Dyer S. frutescens subsp. speciosa

S. tomentosa Eckl. & Zeyh. 11111. "" •. * S. tomentosa

83 CHAPTER 9

TAXONOMY

TAXONOMY OF THE GENUS SUTHERLANDIA

9.1 Historical overview The South African plant for which Robert Brown proposed the name Sutherlandia had previously been named by Linnaeus, Colutea frutescens (Sp. Pl., 1753). The name Colutea stood up to 1812, until Brown, recognising differences between the plant and the other species of Colutea, suggested the name Sutherlandia frutescens. Linnaeus' plant was accepted as the type species.

The genus Sutherlandia was named by Robert Brown in 1812 (Aiton, Hort. Kew. ed 2. IV. 327). Prior to the publication of this name, Gmelin in 1791 (Syst 1027) used the name for a plant belonging to the family Sterculiaceae. Later Sutherlandia Gmelin was sunk under Heritiera Dryand which gave motivation for conserving Brown's name.

Both Harvey (1862) and Meyer (1836) considered all South African plants to constitute one variable species, Sutherlandia frutescens. Phillips and Dyer's (1934) were inclined to believe that they were dealing with a large complex which separated into definite species in different geological areas. Phillips and Dyer (1934) observed that morphological characters of the stem and leaves rarely exhibit any outstanding or constant features which could be used for grouping the specimens into distinct species, and floral features were found to be of even less service in this respect, but in the shape of the pod or the fruit, three groups were recognised. They also described three new species of Sutherlandia, namely S. humilis, S. speciosa and S. montana. Phillips and Dyer (1934) further discussed S. frutescens var. incana of the coast belt, proposed by Meyer (1836), which they found to be similar to S. tomentosa but differed in that S. tomentosa had bilobed leaflets of a narrow-elliptic ovate shape. They also pointed out the possibility of Sutherlandia frutescens var. incana being a hybrid of S. frutescens and S. tomentosa.

When studying morphological characters I found three groups based on the fruit shape, the stipe orientation and the stipe length. These findings are in congruence with those of Phillips & Dyer (1934). I propose that these groups be referred to as subspecies of S. frutescens rather than distinct species.

84 I CK TAXONOMY

9.2 The genus Sutherlandia R. Br.

Sutherlandia R. Br. ex. Ait. f., Hort. Kew. (2) 4: 327 (1812) nom. cons.; DC., Prodr. 2: 273 (1825); DC.; Mem. leg: 293 (1826); Harv., Fl. Cap. 2: 212 (1862); PhiII., Gen. Fl. Pl.: 328 (1926); Phill. & Dyer in Rev. Sudamer. Bot. 1:69-90 (1934); Hutch., Gen. Fl. Pl. 1: 406 (1964); Merxm., Prodr. FWSA: 212 (1970). Type species: Colutea frutescens L. [now S. frutescens (L. ) R. Br. ex. Ait. f.]

Shrubs, variable in habit, small prostrate, procumbent to large and erect, 0.1 m to 2,5 m high. Branches prostrate to erect, woody, young ones densely to sparsely pubescent. Leaves (40—)50- 82(-90) mm long, pinnate with (5—)7-9(-11) pairs of leaflets. Stipules small, narrowly lanceolate. Petiole adaxially grooved (5—)6-12(-15) mm long. Leaflets alternate, lanceolate to linear or elliptical, sometimes ovate or obovate, (3—)8-12(-17) mm long, apices rounded to emarginate, abaxial surface densely to sparsely pubescent to tomentose, adaxial surfaces densely tomentose, sparsely hairy or more often glabrous. Petiolule (0—)0.1-1 mm long. Inflorescences axillary and or terminal racemes usually few-flowered, with (3—)5-6(-8) flowers. Flowers tubular, large and showy, bright red, rarely white. Pedicels (5—)6-8(-9) mm long. Calyx campanulate (5.5—)8-14(- 17) mm long, 11-20 mm broad, densely tomentose or pubescent, the tube much longer than the lobes; lobes triangular (2.5—)4-12(-12) mm long and 10-13 mm broad, the two vexillary lobes longer and/ or wider than the lower three lobes. Standard oblanceolate to obovate, reflexed at the apex, 23-52 mm long, 10-13 mm broad, narrowed to the base, bright red and blotched or streaked with white. Wings minute, (5—)6-10(-20) mm long, claw 2-3(-5) mm long. Keel large, boat shaped, (20—)21-24(-38) mm long, 6-8(-15) mm broad; claw (5—)6-8(-15) mm long. Stamens fused 0.75 of their length, tube glabrous, (14—)16-20(-36) mm long; anthers uniform in size and shape. Ovary, oblong 10-14(-20) mm long, semi—translucent, glabrous, stipitate; stipe 3-6mm long; style 10-12(-17) mm long, curved upwards, apex laterally bearded; stigma small. Fruit large, inflated and oblong to ovate, 38-60 mm long and (7—)14-24(-29) mm broad, length to width ratio (1.6—)1.9-4.6(-6.4), stipitate; stipe (2—)3-10(-15) mm long, in line with fruit, directed upwards (i.e. towards the seed—bearing suture) or downwards. Seeds numerous ca. 28-32, kidney shaped, 2.5-3 mm long, pale brown to dark brown, surface foveolate to smooth.

Diagnostic features: Shrubs or shrublets with large, tubular, bright red flowers. Fruits large, inflated and bladdery.

Distribution: Widely distributed in South Africa, Namibia, Botswana and Lesotho. Found at low and high altitudes and common on disturbed areas, especially along roadsides.

85 UMIAK I tK TAXONOMY

9.3 Key to species and subspecies of Sutherlandia Plant densely tomentose; leaf lamina (adaxial surface) not visible between hairs; leaflets rounded to emarginate; fruit ovoid; restricted to the coast 2. S. tomentosa

;-- Plant glabrescent to sparsely tomentose; adaxial surface glabrescent, leaf lamina visible between the hairs; leaflets oblong to linear, usually not emarginate; fruit oblong to ovoid; widespread in southern Africa; found on the coast or inland 1. S. frutescens

Fruit ovoid, less than twice as long as wide (I:w ratio less than 2); stipe directed upwards or downwards: Flowers less than 35 mm long; stipe directed upwards (towards the upper seed-bearing suture of the pod); habit variable, prostate to erect, up to 1 m high and usually less than 0.8 m wide la. subsp. frutescens (typical form) Shrubs more than 0.2 m high (usually 0.3 to 0.8 m), coastal distribution (never far from the sea) la. subsp. frutescens (hairy form) Shrubs less than 0.2 m high, inland distribution la. subsp. frutescens (dwarf form) Flowers more than 35 mm long; habit erect to procumbent, up to 0.8 m high and 1m wide: Stipe directed downwards (towards the lower suture of the pod); habit rounded, leaflets oblong lb. subsp. speciosa Stipe directed upwards; leaflets oblong to ovate, often notched la. subsp. frutescens (high altitude form) Fruit oblong, more than twice as long as wide (I:w ratio more than 2); stipe more or less in line with the pod: Flower less than 35 mm long; leaflet linear to elliptic, often small: Flowers red; widely distributed in the interior of southern Africa lc. subsp. microphylla (typical form) Flowers white, known only from cultivation 1c. subsp. microphylla (albino form) Flowers more than 35 mm long; leaflets rounded and relatively large 1a. subsp. frutescen: (high-attitude form)

86 CHAPTER 9 TAXONOMY

9.4 The species and subspecies of Sutherlandia

1. Sutherlandia frutescens (L.) R. Br. ex. Ait. f., Hort. Kew. (2) 4: 327 (1812); DC., Prodr. 2: 273 (1825); DC., Mem. leg: 293 (1826); Eckl. & Zeyh., Enum. 2: 251 (1836); Harv., Fl. Cap. 2: 212 (1862); Phill., Gen. Fl. PI.: 328 (1926); Phill. & Dyer, in Rev. Sudamer. Bot. 1: 73-75 (1934); Hutch., Gen. Fl. Pl. 1: 406 (1964); Merxm., Prodr. FSWA: 212 (1970). Type: Herb. Linn. No. 914.4 (LINN, lecto., designated by Schrire (1997), photo!).

= Colutea frutescens L., Sp. Pl. 2: 723 (1753); Sp. Pl. Veg.: 1045 (1759); Burm., Prodr. Pl. Cap.:22 (1768); Thunb., Prodr. Pl. Cap.: 134 (1800); R. Br. ex. Ait. f., Hort. Kew. (2) 4: 327 (1812); Thunb., Fl. Cap.: 603 (1823); DC., Prodr. 2: 273 (1825). Type as above.

la. Sutherlandia frutescens (L.) R. Br. subsp. frutescens

A shrub variable in size, from small and prostrate to large and erect, 0,3 m to 1,5 m high. Leaves imparipinnate, (40—)50-83(-90) mm long, pinnate with (5—)6-9(-10) pairs of leaflets. Stipules small, narrowly lanceolate. Petiole (5—)8-10(-15) mm long. Leaflets alternate, lanceolate to linear sometimes ovate to obovate (6—)8-14(-15) mm long, (2—)3-4(-6) mm wide, apices acute or narrowly elliptic, rounded or emarginate, abaxial surface sparsely hairy and adaxial surface more often glabrous. Petiolule 0.3-1 mm long. Inflorescence axillary and or terminal racemes usually few—flowered with (3—)5-6(-8) flowers. Flowers large, 25-47 mm long, bright red or scarlet with white streaks. Pedicels (5—)6-8(-9) mm long. Calyx campanulate 7-9(-15) mm long and (10- )11-12(-20) mm broad; lobes triangular, 2.5-3 mm long; the tube 4.5-10 mm long. Standard oblanceolate to obovate narrowed to the base, bright red and blotched or streaked with white, reflexed at the apex 22-45 mm long, 10 mm broad. Wings minute, 6 mm long, claw 2 mm long. Keel large, boat-shaped, 21 mm long, 6 mm broad, claw 6 mm long. Stamens fused 0.75 of their length, stamina! tube 16 mm long. Ovary oblong, 12 mm long, semi-translucent, glabrous, stipitate; stipe 2-8(-15) mm long; style, 12 mm long, curved upwards and laterally bearded at apex, stigma small. Fruit large, inflated and ovoid to ovate (38—)40-55(-60) mm long, (15—)17-25 (-29) mm broad, length to width ratio (1.6—)1.7-2.0(-2.5), stipitate; stipe (2—)3-9(-15) mm long, directed upwards (i.e towards the seed—bearing suture). Seeds numerous (28-32), kidney shaped, 2.5-3 mm long, pale brown to dark brown, surface smooth or rugose.

Diagnostic features: Procumbent to erect shrubs with small to large flowers. Fruits inflated, ovoid or ovate, with stipe directed upwards (towards seed—bearing suture) (Fig. 9.4.1).

Distribution: Widely distributed in southern Africa, also occurs at high altitudes and common on t. disturbed areas especially along roadsides (Maps. 9.4.1 to 9.4.3).

87 CHAP I tti 9 TAXONOMY

Sutherlandia frutescens subspecies frutescens can be divide into four more or less recognisable forms.

Form 1 ( typical form)

= Suthertandia frutescens var. communis Harv., Fl. Cap. 2: 212 (1862); nom. illeg. Type as for S. frutescens

= Sutherlandia frutescens var. garipensis E. Mey., Comm. Pl. Afr. Austr. 1(1): 121-122 (1836)

Type: South Africa, "Garip", Drege 3354 (P!, lecto., designated here; P!, isolecto.).

[Note: The original Drege material on which Meyer based his descriptions was located in P. The locality details of specimen labelled 3354 exactly agrees with that of the original description]

Diagnostic features: Procumbent glabrescent shrub reaching up to 0.3 m high (Fig. 9.4.1), flowers small, fruits ovoid and much inflated. Stipe directed upwards.

Distribution: Widely distributed in South Africa extending to Botswana and Namibia (Map. 9.4.1).

Specimens examined —2114 (Uis): Brandberg, Numasplato (—BA), Oilver s.n. (PRE); main camp near Numas valley (—BD), Wiss 1469 (PRE). —2216 (Otjimbingwe): Farm Haris WIN 367 (—DD), Giess 13472 (PRE). 2217 (Windhoek ): Between Avis and Gobabis (—CA), Seydel 4428 (PRE); Farm Finkestein (—CB), Seydel 3586 (PRE); 1 mile east of Windhoek (—CB), Schelpe 153 (PRE); Windhoek (—CB), Ban s.n. (PRE); Farm Lichtenstein (—CC), Merxmueller & Giess 3595 (PRE). —2316 (Nauchas): Gamsberg Pass (—AD), Jensen 338 (PRE) 2421 (Tshane):1 km north east of Tshane (—BA), Skarpe 188 (PRE). —2426 (Mochudi): Mochudi (—AC), Harbor 14071 (PRE). —2616 (Aus): Farm Frisgewaagd (—BA), Giess 12801 (PRE). 2715 (Bogenfels) Klinghardt Mountains (—BC), Muller 665 (PRE). 2716 (Witputz): Swartpunt (—BC), Merxmueller & Giess 28470 (PRE). —2718 (Granau): Klein karas (—CA), Dinter 4968 (PRE). —2722 (Olifantshoek): 10 miles from Sishen to Olifantshoek (—DD), Van der Schjiff 8031 (PRE). —2723 (Kuruman): 30 miles north west of Kuruman (—AA), Louw 859 (PRE); McCarthys—Rest, RSA, Botswana border (—AD), Jordaan 21 (PRE). —2822 (Glen Lyon ): 121 km from Upington to Kuruman (—AB), Palmer 1 (JRAU). 2823 (Griekwastad): Witsand turnoff road towards Griekwastad and Groblershoop (—CD) Zietsman 615 (PRE); Herbert division, "Brakken", P.O. Campbell (—DC), Van Ecklon 3 (PRE). —2917 (Springbok): 18 km from Springbok to Kamieskroon (—DD), Klapwijk 20 (PRE). —2918 (Gamoep): Springbok, 48 km from Springbok to Pofadder (—AD), Moshe, Van Wyk & De Castro

88 l.r1/4r I TAXONOMY

5 (JRAU); Aggeneys district, 2 km west at turnoff to Blomhoek, north of Ghamsberg (—BD), Palmer 2 (JRAU). —2924 (Hopetown): 5 km from Luckoff on road to Koffiefontein (—DB), Muller 1375 (PRE). —2925 (Jagersfontein): 1,5 km from Jagersfontein to Fauresmith (—CB), Moshe, Van Wyk & De Castro 20 (JRAU); Fauresmith Nature Reserve garden next to information office (—CB), Moshe, Van Wyk & De Castro 21 (JRAU). —3019 (Loeriesfontein): Boesmanland, 6 km south east of Witputs and 18 km north west of main road, between Loeriesfontein and Brandvlei near gate (—DB), Perold 2282 (PRE); 1 km north of Lekkeroog on the road to Brandvlei (—DB), Crosby 890 (PRE). —3118 (Vanrhynsdorp): 12,8 km north of Vanrhynsdorp (—BA), Palmer 6 (JRAU). 3119 (Calvinia): Farm Kareeboom, Loeriesfontein road (—AB), Burger & Louw 241 (PRE); Nieuwoudtville, Oorlogskloof (—AC), Van Wyk & Viljoen 3709 (JRAU). 3218 (Clanwilliam): 18,2 km west of Aurora, along roadside from Piquetberg to Velddrift (—DC), Moshe, Van Wyk & De Castro 14 (JRAU). 3219 (Wuppertal): Theerivier, Citrusdal (—CA), Hanekom 2053 (PRE); 10 km north of Leeuwshof on Calvinia to Ceres road (—DD), C. M. Van Wyk 497 (PRE). —3220 (Sutherland): Houthoek (—CA), Hanekom 1103 (PRE). —3221 (Merweville): Layton, Kloofspitskop (—BB), Shearing 1306 (PRE). —3222 (Beaufort West): 15 km north of Beaufort West, Molteno Pass (—BA), Shearing 1066 (PRE). —3318 (Cape Town): Saldanha, west coast road, 110 km from Cape Town to Saldanha (—AA), Moshe, Van Wyk & De Castro 15 (JRAU); Camps Bay (—CD), Moshe, Van Wyk & De Castro 18 (JRAU); Sea point, along Bantry Bay (—CD), Smith 2930 (PRE). 3319 (Worcester): Tulbaghweg, Waterfall, forestry station (—AC), Fellingham 177 (PRE); De Mond, Hermanus (—AD), Williams 910 (PRE); 16 miles from Worcester to Robertson (—DA), Marsh 1016 (PRE); Droogeriviersberg (—DC), Moshe, Van Wyk & De Castro 4 (JRAU). —3321 (Ladismith): Farm Rouxpos below Klein Swartberg mountains (—AC), Stirton 9494 (PRE). 3322 (Oudtshoorn): Kleinsleutelfontein (—AB), Palmer 23 (JRAU); Swartbergpas (—AC), Palmer 21 (JRAU); Hartebeesrivier, south of VVillowmore (—CB), Hugo 1469 (PRE); George (—CD), Mogg 11850 (PRE); Farm Doornkraal near De Rust (—DA), Dahlstrand 2121 (PRE). 3418 (Simonstown): Chapman's peak (—AB), Palmer 15 (JRAU); 2,9 km from Bakoven to Chapman's Peak (—AB), Van Wyk & De Castro 3671 (JRAU). —3419 (Caledon): 15 km from Napier to Caledon (—BD), Grobbelaar 2745 (PRE); 1 km from Gansbaai on road to Stanford (—CB), Germishuizen 4135 (PRE). —3420 (Riversdale): 4 km from Stormvlei to Swellendam (—AA), Stirton 6139 (PRE); 22 km from Swellendam to Bredasdorp (—CD), Van Wyk & De Castro 3667 (JRAU).

89 CHAPTER 9 TAXONOMY

L

Figure 9.4.1 The typical form of Sutherlandia frutescens subsp. frutescens showing the small habit with bright red flowers and inflated pods.

• 12 14 16 18 24 26 28 30 32 arldnajavgancLanyadufaiiporigrimii. sitaliair sressarisaluiamm. 661116.r trail 18 l1171111.11111"1111.111 166116:11Atii01 416511WEWIIMiliiilliiiiiK6116mgMa1111.11MillEralliffilitnillal reisiiiallithilkiaiiiaiiiiiiidaW - entilikUNIVINtilliallWattikaiamozsraird 'ia111111MIANIAMMIliiiiiiiiiialaNtiaMiOlLitilLitatilliMilintallii5W118"1"1:1lklailm JIIMMWRIMMUMMIMMMI2z?. M-2225M LIMP. MR_Meliftm& 1'. regllatriti illIMIENIUMEENI- rt"Mitit'''' 11'"1ania ''' sn'eruasisatiraminS"1"6"11.1.11141111 tWliitilakataiLiafr-MikiaMil.11.11111rall".11 2423433 7 : .. . ..7111111111Mill lei lialliallMINEEME11.111.1lt-iiihliM41106- .2526 _,-, ..,. _1 ilifelfilipumr.m 'Maid6111M1 r talialUl ,IIIIMMM*1111"nEMEN5501 .1111151.111111411111012625 1262e - -.12623 2630 2E01 ;t1111111111VIAMIUMEM IL•LAiW lir 1112.11111MMIRMIIIIIIthilablillalialir. inallinalltalalWagagimmo" na27715 launipZ•76 V27 mirna 273.W11111111111 ■i61U11 1ikiji;;1 2877 2S2 2821 asp 21e, 111.110110 111I11 allaagnaliallinleiriaingtikaZaMIMENIESMINIMMINPMeina 111111111111011111WW■Aii i164611167,41 1;;ErallinUilillin 101111111111 11111111•11111111111011Erid111111.111. el■— 11181110111111•1 --ingslill01111111,1 Lareir--Wamialleaffiron„._11111111111.1,, 4 Illirmm ai usiorra_Lisit_.:itiggromilisiam .. t;:apshi itatiAaaVilillialtaiMil3,27 3128 rrillintlailiminti 1111111111111ffsrragIMMUUMMIIMMIMMUMMIMIEWWWILIMMU a"gra .=:Qii;:iilarem- •miall 1.---"WilantailiiemItMin ill -----imiNRIIIU- er vallIALLIMMInitnalemillina 81rUililliragrar (° 1111 ■ iumm11111116114 iriarri llet111H,Iwkii Aridcaiztallinli;awarrft_ -imgusimiln l ninyammummaMLI IM c asoiRRIIIVEMen MEMMalil 1 1 10 12 14 16 18 20 22 24 26 28 30 32

Map 9.4.1 The known geographical distnbution of S. frutescens subsp. frutescens (typical form). 90 UMW I t ► TAXONOMY

Form 2 ( hairy form) = Sutherlandia frutescens var. incana E. Meyer in Linnaea. 7: 170 (1832); E. Mey., Comm. Pl. Afr. Austr. 1(1): 121-122 (1836); Phill. & Dyer, in Rev. Sudamer. Bot. 1: 75 (1934). Type: South Africa, "Greenpoint". Distr. Caledon", Ecklon s.n. ("Dec. 1826"), piece on far right hand side of sheet, (S!, lecto, designated here; S!, specimen with original Meyer label, isolecto).

[Note: The specimen in S is chosen as lectotype because it has an original label in Ecklon's hand and the name "fl incana" was added by Meyer himself. There is no original specimen of this variety in P]

This form is similar to the typical form of Sutherlandia frutescens subspecies frutescens (form la) except for the densely pubescent twigs and leaves (Fig. 9.4.2).

Distribution: This form is distributed only along the coast (Map 9.4.2) and co—occurs with Sutherlandia tomentosa but differs in the leaflet shape. It has slightly notched leaflets. as opposed to the deeply emarginate leaflets in the latter and the less densely pubescent upper surfaces of the leaflets (with the lamina visible between the hairs) is a further distinction (see key).

Specimens examined 2916 (Springbok): Port Nolloth, 500 m to 2 km from sea (—BD), Plowes 7262 (PRE); Mc Dougall's Bay, 2 miles south of Port Nolloth (—BD), Hardy 653 (PRE). —3318 (Cape Town): Bloubergstrand (—AD), Palmer 12 (JRAU); Bloubergstrand, 5.7 km N of Potterfield/ Parks Drive road, west coast road (—AD), Moshe, Van Wyk & De Castro 16 (JRAU); Aurora (—DC), Palmer 9 &10 (JRAU). 3322 (Oudtshoorn): George, on sand dunes (—CD), Mogg 11850 (PRE); Wilderness, near George (—DC), Mogg s.n (PRE). 3418 (Simonstown): Hout Bay (—AB), Palmer 14 (JRAU). 3419 (Caledon): De Mond, Voelklip, Hermanus (—AD), Williams 910 (PRE); Between Franskraal and Pearly Beach, roadside (—CB), Van Wyk 3162 (JRAU); 3,8 km from Pearly Beach to Quin Point (—DC), Van Wyk & De Castro 3668 (PRE). —3420 (Bredasdorp): Brandfontein, amongst sand dunes (—CA), Smith 3118 (PRE). 3421 (Riversdale): Struisbaai, Cape Agulhas (—AC), Palmer 18 (JRAU); Muir s.n. sub 4608 (PRE).

91 CHAPTER 9 TAXONOMY

Figure 9.4.2 The hairy coastal form of Sutherlandia frutescens subsp. frutescens showing the erect habit with pubescent twigs and leaves. Note the sandy habitat in which this form occurs. 12 14 16 18 20 22 24 26 28 30 32 34 aragaWianumm#!41ilimmramemgaRAALtrImmulmimaiummacuicaula lMiaririntil4816MOIM 8 1lallarararararariail EMLIILERen 8 legrailiragrallaillinirialaiiiliMILIMMItertrati ilia10111AMMIIIIIIIMINIMMIMINUMIUMUNIIIIMMIL-9 1aallimitiiiiailibilltbiatia 6niciathnike.16i16:16•416416 J646•Milikr- - ,--- 11111sawasnammonnumuumummonunumumi """'' MhilWiiiililhillhilfilanitiliiiniillMiiiiilibeiN616“Mitial61LiiiIV Ilia"' MMOia Ilii • IMMI INIVUIRIMPAS URISMSMIMMINIORIMIilwAVAJW1661161J6h16i16411.16iM.IWLainiirakinp-,miaaw-1111111IA"aliiirUto," illualwMUMUMOIUMMEMEMMEREMMIMMNWIMMinundPaaCW' IM tratiaffail6iii66 -111810016ilwratiMuliMMAiligingli IiralkaWJW/10.1111111411 iffif:117:222:±711 6.4776:77w11/1:5:artili, t IIIIIIRFIIMMILI annallialIMUUMIor Millnialli '4?.5:1 lillISSUMUSIMMAWJWAAJWAWAWNWIWAWWWWW4g4ereallertuft -,,--111111mini IZEsin. OIMI1131111111WWW.M.Altgliwama tti it M lairth I6171 71 75li t: 6 67 671Z11MILntna t tis afIs Ir t■ ■ ■ i sl iiaaao i ra 5ua ri 77;7Z:lrh a l g

JAJAJ4R466 IMi ESSOSSREUWAJAJAJ LAWJWWWINUMIIREP RO IIMIRNMInmwu da ■-- l r MMOMUMa.rammu a udirJ aazgaiii.umai l Q gium A umeis 8-----a MMIIIIn 1 lreassumusuracannu1 sm mwomusiunn"'""i ul u ntair .6w b lmuukvivamennanuem mmnis-u1e 1V. :a 1 1.-1rg " la w.vg= .ak--,zain 1 mamil lll !!! 1. '. lAiiiim minsim mi mu1i1n1 in l m-mM-_- u- r b1 10 12 14 16 18 20 22 24 26 28 30 32 34 36 Map 9.4.2 The known geographical distribution of the hairy form of Sutherlandia frutescens subsp. frutescens.

92 CHAF' I tK TAXONOMY

Form 3 (dwarf form)

= Sutherlandia humilis Phill. & Dyer, in Rev. Sudamer. Bot. 1: 79-80 (1934). Type: South Africa,

Orange Free State, Fauresmith district, Fauresmith, Smith 429 (PRE!, lecto; designated here);

Smith 442 (PRE!), Smith 968 (PRE!), isosyntypes.

[Note: Smith 429 is chosen as lectotype merely because it shows the fruits. All the syntype

specimens were annotated by Phillips]

1 . This form is similar to form 1 but the habit is more prostrate and the plant grows up to 0.2 m high r (Fig. 9.4.3). The fruits tend to be somewhat more globose and smaller than in the typical form.

Distribution: Southern Cape, South—eastern Cape Province and Fauresmith in the Free State

(Map 9.4.3).

Specimens examined —2827 (Senekal): Excelsior, Vegkop (—CC), Du Preez 1742 (PRE). 2925 (Koffiefontein): Fauresmith, Nature Reserve (—CB), Henrici 1809 (PRE); Hill slope side of Botanical Reserve on hill above town (—CB), Smith 968, 442 , 429 (PRE); Koppie on north slope (—CB), Verdoorn 2370 (PRE). —3024 (De Aar): East of Naauwpoort station (—DA), Liebenberg 871 (PRE); Nooitgedacht, Colesberg district (—DB), Watt & Brandwyk 1715 (PRE). 3025 (Colesberg): Middelburg (—CB), Gill 61 (PRE). —3122 (Loxton): "Alarmskraal", south east of Carnarvon (—AB), Hugo 329 (PRE). —3124 (Hanover): Wapadsberg Pass (—DD), Retief & Reid 570 (PRE). —3220 (Sutherland): Kanariesfontein (—BC), Van der Walt s. n. (PRE). 3225 (Somerset East): Cradock, on Babilon's Tower, Mountain Zebra National Park (—AB), Liebenberg 7224 (PRE); Mountain Zebra National Park (—AB), Zietsman 1305 (PRE). 3319 (Worcester): 3 km from Tulbagh to Ceres, Malmesbury shales (—AC), Stirton 5880 (PRE); Worcester (—CB), Bayliss 1665 (PRE); 16 miles from Worcester to Robertson (—DA), Marsh 1016 (PRE). 3320 (Montagu): Barrydale, 9 km from Montagu to Barrydale (—DC), Palmer 17 (JRAU). 3321 (Ladismith): Farm Rouxpos, below Klein Swartberg mountains (—AC), Stirton 9494 (PRE). —3322 (Oudtsthoorn): Meiringspoort, 1 km form the turnoff to Oudtshoorn, between Oudtshoorn and Beaufort West (—BC), Palmer 22 (JRAU). —3323 (Willowmore): Uniondale, 3 km from turnoff to Uniondale (—CA), Palmer 25, Moshe, Van Wyk & De Castro 19 (JRAU); Hartebees river, south of Willowmore (—CB), Hugo 1469 (PRE). —3325 (Port Elizabeth): Uitenhage, Kromkloof, Mannetjie catchment basin (—CB), Scharf 1568 (PRE).

93 CHAPTER 9 TAXONOMY

Figure 9.4.3 The dwarf form of Sutherlandia frutescens subsp. frutescens, showing the prostrate habit (only reaching 0.2 m high) and fruits which tend to be somewhat more globose and smaller than in the typical form.

12 14 16 18 20 24 26 28 30 32 iltaraggagialliMmop rforgagaimaciaw_i- ultimattmizillecijit‘Jirialkiiiki ■ liir"l aiiiiiiInigiliiiiiiiiiillithlialirial emilliiMilailallaillialiLlilligilLiranagelailla 18 18 ligaiiig 1111111 14)41/141111r10:1aililgalliidilialli 1111111111:1111121; 6ilitia661641 641.216111ilijkiiiMia 20E" Lir - gal 9 ' 11111011alliiiiM266161111661mmummulmilummiummummummutsie'"i Isssomusalsawavicimaisimmirawarainimataiggpitairiamatatiwinaimarm" r~ rirrnirairapriereatti:Et%Imir:t25:211 ::11 111111111111111WhiMilailiMil6616616i161116.116ariallkiiiiVatailiZa eagama,.111.. a MUmmummalmmumm MM Wittig Mal . ' ' uswoinimmatuatfiatamamummuummummumssaminviammurnp issIsaalinnatitaitainuctaavagagmar.v.:-.46zawmatiiiimagli■sir 25:30 tagari9-.12.411"/11 61111111.2111M11111111111.WIFFAMMINAIMENIMMINIPMErAiraln ' illrearnarai 71111:2:11rrillWriaral"Wiaifitalff liVrillidi .711:111111:111:2823 11651::: llitil ° Illialliallilatiat*WdEririailai 2527:9:41 284° =Wall "IC 11:11711111"111-M.61111e1 441°11;1 railigl:111111",::::::+11:1:1111 11111111111111111111111111111611MliiMilaiiiaiglailklMatiiii6NAVAII:WA 733P12 VAIMIIIIIII lartirap6:41111 926%.c1616. iAvallAilr itTl allardiria ■numir c 1111111111111111111111111111111111111' 104Zillitaaa rejaptigiainaimiar mom 411111111111.1111111111111111MBariaLMIllgliasirmrsoli:Lp 32.22 . 11124;111:41.711111711:11111111111181:4 . - Wasinlir'ill11111111111111MUll■NM "I Ilarligillinierlitainiiim tugimataisna " aaikatzaallumiggliimil...di 1111111111...... ipm c llialiiiiiirmlizamaiNinsiumas EM10121111111111M11111111/11 as --..•ril • 10 12 14 16 18 22 24 26 28 30 32 36

Map 9.4.3 The known geographical distribution of the dwarf form of Sutherlandia frutescens subsp. frutescens.

94 CHAPTER 9 TAXONOMY

Form 4 (high altitude form)

= Sutherlandia montana Phill. & Dyer, in Rev. Sudamer. Bot. 1: 78-79 (1934). Type: South Africa, Ladismith district,: Mount-Aux Sources, Doidge s.n. sub PRE 8720 (PRE!, holo.).

[Note: The Doidge specimen has no fruits and rather atypical leaflets, but the locality leaves no doubt that it must be this form. Phillips & Dyer explicitly designated Doidge s.n. as the type.]

Similar to form 1 but the habit is often more erect and the plants may be larger. The fruits have the stipe directed upwards, but the shape may be somewhat oblong as in is S. frutescens subsp. microphylla.

Diagnostic features. The flowers and fruits are large and the pods are long-stipitate (Fig. 9.4.4).

Distribution: This form occurs at high altitudes in the Drakensberg, Magaliesberg, Free State and in the Cape mountains (Map 9.4.4).

Specimens examined

—2429 (Zebediela): Wolkberg; Boyne,Talus slopes (-BB), Van Vuuren 1818 (PRE). 2728 (Frankfort): Free State, 24 km from Reitz to Tweeling (-CB), Van Wyk 3800 (JRAU); Reitz, 22 km from Reitz to Tweeling (-CB), Moshe, Van Wyk & De Castro 23 (JRAU). —2729 (Frankfort): Majuba near Volksrust (-BD), Van der Merwe 8 (PRE). 2828 (Bethlehem): Golden Gate (-DA), Van Wyk 2771 (JRAU); path below Sentinel Peak before chain ladder (-DB), Arnold 855 (PRE); Mnweni cut-back area, Ledger's Cave (-DD), De Castro 138 (JRAU); Mont Aux Sources (-DD), Doidge 8720 (PRE). 2829 (Harrismith): Platberg (-AC), Jacobsz 3068 (PRE); Plaas Rensburgkop, Manyenyeza mountain (-. AB), Jacobsz 605 (PRE). 2926 (Bloemfontein): Thabanchu mountain (-BB), Peters, Gericke & Burelli 368 (PRE). —2927 (Maseru): Letlapeng district (-CB), Dieterlen 643 9PRE). 2928 (Marakabei): Semonkong, western aspect slope to river (-CC), Davidson 3046 (PRE). —2929 (Underberg): Escourt district, near Champagne Castle Hotel, 58 miles west of Escourt (-AB), Codd 2474 (PRE); Drakensberg Botanical Gardens (-AC), Jacobsz 98/72 (PRE); Giant's Castle (-AD), Wood 10550 & 10560 (PRE); Loteni Nature Reserve, near Giant's Castle, along river, west of Camp (-BC), Jacobsz 3980 (PRE); Garden Castle Forest Reserve (-CD), Burtt 13448 (PRE); Sehlabathebe National Park, east side of road between the top of Kubutsane Nek and Phula (-CC), Hoener 1633 (PRE); Tarn cave above Bushman's Nek, below Devil's Knuckles (-CC), Hilliard & Burtt 16891 (PRE). —3027 (Lady Grey): Near Barkly East, west of Naude's Nek, on road to Rhodes (-DD), Phillipson 723 (PRE). 3119 (Calvinia): Oorlogskloof, south of Niewoudtville (-AC), Powrie 84 (PRE) (possibly S. speciosa, mature fruit required for positive identification).

95 CHAPTER 9 TAXONOMY

—3126 (Queenstown): Broughton, near Molteno (—BC), Flanagan 1892 (PRE); Hangklip mountain (—DD), Galpin 1619 (PRE).

—3127 (Lady Frere): Between Dordrecht and Barkly East (—BA), Werdemann & Oberdieck 1084 (PRE).

—3218 (Clanwilliam): Pakhuis summit (—BB), Williams 1250 (PRE); West of Patrysberg (—CD), Bayer 5922 (PRE); Piquetberg (—DC), Godfrey 1246 (PRE); Piquetberg, De Hoek, 12,8 km W of Piquetberg to Velddrift (—DD), Palmer 8, Moshe, Van Wyk & De Castro 13 (JRAU). 3219 (Wuppertal): Biedouw valley, top of Uitkyk Pass (—AA), FeHingham 1132 (PRE); 76 km from Ceres to Sederberg near Rosendal (—AA), Stirton 5893 (PRE). 3220 (Sutherland): Jakkalsvlei, south of Sutherland (—BC), Oliver 8973 (PRE). —3222 (Beaufort West): Molteno Pass, 15 km north of Beaufort West (—BA), Shearing 1066 (PRE). 3226 (Fort Beaufort): Stockenstrom division (—DD), Hilliard & Burtt 13268 (PRE). 3227 (Stutterheim): Izeleni (—CD), Sim 19443 (PRE). 3319 (Worcester): Groot Winterhoek State forest, Driebosch (—AA), Haynes 856 (PRE); 3 km from Tulbagh to Ceres (—AC), Stirton 5880 (PRE); Tulbaghweg, Waterfall Station (—AC), Fellingham 177 (PRE).

96 CHAPTER 9 TAXONOMY

Figure 9.4.4 The high altitude form of Sutherlandia frutescens subsp. frutescens, showing the more erect and larger habit and large flowers.

. ifs...;4,..„4.6 ,,,,,„.14,..iir..iii.a_ „ila4 14w6 limi8 iainittivaucciiiirlara, mis 18 13111111111111enatal.11&frilikatliiilial161416111iadarlitiaillIMI liariliViiidriarikil69.ll6a21.iilliliii0iiiiimiliiii 611:41,111160111411iiiiallili1111Vial inikildraisitlitainiA ' .1111110Sitlirarilliala"illnarallalhareahlatlig711: Liciwrintarrrirearaiiii 1:11:4:11 111MMUM6"1 :162 -AZIWrialliagaltiiiMiliWiiMiEWAQT'l%=.11F7-,-- '-;---••:4:81141■11:114711111111111 0• - -- 1111111111161iniliniiiiHMEMMMUMMiraraialL7aLV ''. reinn- - illinitiiiiiiiIittat0116tattaiiiiiiiiiiiViinall614'641LiMiaLiii123P run allimeR1 rairstilliaLisivair 4 allignire , gai 2 atill la=113"1 1 ruv gregaratar 1LIMI8612s33;efrliti 6 illIMISPIIVELIMIL;, 2736 EV i al ' , ruarraairigir1411PASC1 IPEret=n 2/217 1196ii.i 1111MiliainiEllat 111111111186111111111111.1glidellarline9StillAkaliZatr 921181121811111111181181 jra6119.IIM3°2516"26 1 lrrillinlininnlnailialarAeriINEIMMEMEMILinairallifitrAl111111111111118116111.111111111 4 I .------rillial liinumastaigarimmarattagamuativrairaiwgign----umanall11111111111S inalltil" arainsumunumnimsadra,rmmui .----- 1111111111111111111111M16614 inanti 4 nissumundic mumanorwrimmeouse""ur 2 L 1"""On"I unsigimAi araiaimaaw6-awawfaunniuund"""ma ilienurounrsuattimaithienuessuirainEllrneiniFiiiiiiiiiifinintinlira2MMIUM1111111811.1 , • 1 10 12 14 16 18 20 22 24 26 28 30 32-- 34 38 Map 9.4.4 The known geographical distribution of the high altitude form of Sutherlandia frutescens subsp. frutescens.

97 CHAPTER 9 TAXONOMY

1 b. Sutherlandia frutescens subsp. speciosa (Phill. & Dyer) Moshe & Van Wyk comb. et stat. nov. Type: South Africa, Little Namaqualand, Waterklip [3118 AC], Schlechter 11162 (PRE!, holo., P! , iso.).

= Sutherlandia speciosa Phill. & Dyer, in Rev. Sudamer. Bot. 1: 75 (1934). Type as above.

This subspecies is closely similar to subsp. frutescens (typical form). Procumbent shrub, 0.5 m high and 1m wide. Flowers large, 43-52 mm long, bright red or scarlet with white streaks. Pedicels 8 mm long. Calyx campanulate, 14-15 mm long and 20-21 mm broad, the tube 10-11 mm long; lobes 3.5-4 mm long. Standard 40-52 mm long, 22-23 mm broad. Wings 15-16 mm long, claw 2.5-3 mm long. Keel 35-37 mm long, 12-15 mm broad, oblong-laceolate; claw 13-14 mm long. Staminal tube 28-30 mm long. Ovary oblong 19-20 mm long, semi-translucent, glabrous, stipitate; stipe 12-13 mm long; style 15-17 mm long. Fruit ovate, large and inflated and stipitate 40-50 mm long, 16-26 mm broad; stipe 14-15 mm long, uniquely directed downwards (closer to lower suture).

Diagnostic features: The large flowers and large fruits with its unique stipe orientation characteristic of this subspecies (Fig. 9.4.5).

Distribution: The subspecies is restricted to Namaqualand (Northern Cape and Namibia) (Map 9.4.5).

Specimens examined —2616 (Aus): Farm Klein Aus, West of Aus'(–CA), Giess & Van Vuuren 781 (PRE); Farm Kubub near Aus (–CB), Giess 14674 (PRE). 2816: (Oranjemund): Kodaspiek, main ridge south east of Beacon and up to the summit (–BB), Oliver, Toelken & Venter 388 (PRE). —2817 (Vioolsdrift): 5 km from Eksteenfontein on road to Khubus (–AC), Germishuizen 4724 (PRE). 2917(Springbok): Mesklip (–DD), Grant 4796 (PRE); Namaqualand (–DD), Schlechter 11162 (PRE); 24 km from Kamieskroon to Springbok (–DD), Stirton 6006 (PRE); —3017 (Honderklipbaai): 22 miles south of Kamieskroon on Vanrhynsdorp road (–DB), Hardy & Bayliss 1097 (PRE). —3018 (Kamiesberg): Kamiesberg (–AC), Palmer 4, Moshe, Van Wyk & De Castro 11 (JRAU).

98

CHAPTER 9 TAXONOMY

Figure 9.4.5 The large flowers and inflated and long-stipitate fruits of S. frutescens subsp. speciosa. 12 14 16 .iaw..1ii iae _...... 01= 18 ...14_ A.mma28w m, aum30nu mu il32w ria 3a ai,.ka _ _wa init1ar t n nA iimiialai isp. _ ciaLia 18 aalalall rsaffzM& lla . g gui iu ni_ tiimmi a ffia7 18 airlrli91"u igar m6 6. .bimMMer.-P i 1a11 11111111M1_1111•11 1 . 111a . IIIn- a -m ,mnd . ,_ liiiiiii _ , .. . ., „., , , . ha i U6ism:A111 lall tiii 1 MMEMMIIM11r__Ea#m •■•- Mf IWIll • fiumul11116illWa6IiiiiWIL W 1 2a29 223o zona 11111 sousurum.....simm...z.monsimirismumm)71.1,::: 111111.111511 maismurimumnimummeimm unsullim li°16i'l'' ."11"1"Ii"164' liariaassimmouni.A 4 1,65".1 01010:11iwinalsikatamalrow- , , _-_-• .tar-- , : tiormis::111.11""71 711111 " rseXpism Isuar":7: 28 airanurArr461.5=7:.' • Ilraii7171"1".11:11:111 riam""asionmzrunimmsalim..rmo mum mu8:: ImimumAisti 1.ngarsani litiralkalial maltanniuma,■mast a Mitainni lailfirialliwillinwil Ei 111111110111111111=11111111 111111M11111 1 --- ',LrdrrrrirZa ruff61aw%. 7".,a..a5---Alnqe:rt-- ilV 11.11----- 1111111119M11111111 adEMIMINIMinmuMun . ausassumiezeatLi0immuli,15 P'2° ainig==11111111muurtnim4u,-- 10 12 14 16 18 20 22 24 26 ZS 30 32 34 36 Map 9.4.5 The known geographical distribution of S. frutescens subsp. speciosa.

99 CHAPTER 9 TAXONOMY

lc. Sutherlandia frutescens subsp. microphylla (Burch. ex DC.) Moshe & Van Wyk comb. et . stat. nov. Type: South Africa, Burchell 1510 (G?, K!)

= S. microphylla Burchell ex DC., Phill. & Dyer, in Rev. Sudamer. Bot. 1: 76-78 (1934).

= S. frutescens (L.) R.Br. var. microphylla (Burch. ex DC.) Harv., Fl. Cap. 2: 212 (1862), nom. superfl. Type as above.

= S. frutescens (L.) R. Br. var. angustifolia E. Mey., Comm. Pl. Afr. Austr. 1(1): 121-122 (1836). Type: South Africa, dry places near Schiloh [3226 BB] Drege s.n. (P!, lecto., designated here).

[Note: The Schiloh specimen is chosen because it is the only one of the three syntypes which are beyond doubt this taxon. The other two ("Zilverfontein" and "Orogap", PO belong to subsp. speciosa]

A large and erect shrub, 0.8 m to 2,5 m high (Fig. 9.4.6). Leaves 53-81 mm long; pinnae small, alternate 6-10 on each side of rachis, 8-16 mm long, 2-4 mm broad, oblong or linear-oblong to narrowly elliptic. Petiole (6–)10-13 mm long, petiolule (0.1–)0.5-1 mm long. Fruit narrow and oblong (Fig. 9.1.6), 45-58 mm long, (8–)10-15(-19) mm broad, length to width ratio more than 2; stipe (3–)6-9(-10) mm long and in line with the pod.

An albino form of this subspecies has been collected (Conradie s.n. sub Moshe, Van Wyk & De Castro 3 (JRAU).

Distribution: Interior of southern Africa and extends to Gauteng Province (Map 9.4.6). Particularly common in the Karoo region, where it often forms dense stands along roadsides.

Specimens examined 2526 (Zeerust): Swartruggens, farm Brakfontein about 2 km south east of town (–DA), Van Hoepen 1736 (PRE); Sutton 951 (PRE). 2527 (Rustenburg ): Rustenburg (–CA), Moore 13 (PRE); Krugersdorp, Olifantsnek, Magaliesberg (–CD), Obermeyer s.n. (PRE). —2528 (Pretoria): Fountains (–CC), Leendertz 8384 (PRE). —2727 (Kroonstad): Between Kroonstad and Heilbron at Rietspruit, crossing near Amerika (–CB), Joffe 429 (PRE). —2824 (Kimberley): Barkley West, on the islands in the Vaal at Warrenton (–BB), Acocks 1276 (PRE). —2917 (Springbok): Bottom end of Kamiesberg Pass, outside Kamieskroon (–BB), Palmer 3 (JRAU). 2927 (Lady Grey): 42,2 km from Zastron to Wepener (–CC), Van Wyk 3801 (JRAU). 3017 (Hondeklipbaa): Kotzesrus (–DD), Stirton 6060 (PRE); —3018 (Kamiesberg): Kamiesberg, north–east of Lelikebankiestoppies, Warm Viool farm (–CC), Welman 00053 (PRE). —3024 (De Aar): Leeuwberg Pass, bottom of the pass (–BC), Moshe, Van Wyk & De Castro 20 (JRAU); 21,3 km from Hanover to Colesberg (–DC), Van Wyk 3671 (JRAU).

100 CHAPTER 9 TAXONOMY

—3025 (Colesberg): 4 km before Colesberg along N1 (—CA), Van Wyk C. M. 2673 (PRE). 3118 (Vanrhynsdorp): 12,8 km north of Vanrhynsdorp (—BC), Palmer 7, Moshe, Van Wyk & De Castro 12 (JRAU); 37 km north of Bitterfontein (—DB), Palmer 5 (JRAU); along roadside from Van Rhys Pass to Vanrhynsdorp opposite Kobee (—DB), Van Wyk, C. M. 2585 (PRE); 1 km from Vanrhynsdorp (—DA), Stirton 5975 (PRE). 3119 (Calvinia): Kareeboom, Loeriesfontein road (—AB), Burger & Louw 241 (PRE); Lokenburg (—CA ), Acocks 17350 (PRE). r 3121(Fraserburg): Williston, Walkraal (—AC), Foley 169 (PRE). —3124 (Hanover): Middelburg (—DC), Gill 62 (PRE). 3125 (Steynburg): Cradock, 8 km from Cradock to Middleburg (—BA), Palmer 27 (JRAU). —3126 (Queenstown): 37 km from lndwe to Elliot (—BC), Van Wyk 3804 (JRAU). —3219 (Wupperthal): Citrusdal (—CA), Hanekom 2053 (PRE). —3226 (Fort Beaufort): near Oxton, road from Queenstown to Katberg (—BA), Story 2808 (PRE). 3319 (Worcester): 20 km from Ceres on road to Calvinia, Theronberg Pass (—BA), Germishuizen 4056 (PRE). 3320 (Montagu): Montagu district, south of Touws river (—AC), Bayer 5937 (PRE); on road between Touwsriver and Montagu(—AC), Palmer 16 (JRAU). —3322 (Oudtshoorn): Kleinsleutelfontein, 3 km from turnoff to Kleinsleutelfontein (—AB), Palmer 24 (JRAU); Farm Doornkraal, 3 miles from De Rust (—DA), Dahlstrand 1473 (PRE). 3323 (Willowmore): about 3 km before Nelspoort on Murraysburg road (—AA), Van Wyk C. M. 2734 (PRE); Unionpoort (—CA), Palmer 26 (JRAU).

101 V11,1 1 L-1-1 0 TAXONOMY

Figure 9.4.6 The large and erect S. frutescens subsp. microphylla, with small leaflets, narrow oblong pods and a stipe in line with the pod.

12 14 16 18 20 22 24 26 28 30num_ 32 01I forarsimaimunaurr_.,55„,„___Ismaingugua 1830 •Z1 18 1 9 1830 Lamillumna 18 vgasgaullawkailiammikabillabii 18 1930 1931 InalManigaVialtailikaigalialVrat :==11::Mrni ,0 ii.M11111L023 • 207201111111125 2027 111.11.11.1.1.1111.111111 lmrps lUari 2028 2029 2030 1161112031 120,33120 20 Arniatiar 20 1111111111 IMAM I isasisomissummiammillonsi imantimuta warsinviminttimagalL061116 i .: 2224 11 222.. r 11116117111111111;11111111 1; 2226 12222 mummer.110 iiiiiall. liti 34' rlignimarinalli223 az lc Ignailliatialtatlindinallttagainja e 1 1111111111111111111511111111 < -v.2 222, __ _ e_2 1 2234 328 2329 12330 12 111111111111111111110111111116VARAMMIMInallUll I I i 11111111111111111•11 420 244s 114 i244.z 2424 ■ 242n 2z..6 27 2meinsimr 1111111111 2 428 12429 2430223:322 245!" 2.23:1C14111233:au9rai '1111111111111106111WitilagiMatiiiiill 202-.5- . ,.. 11111111111/1111111L I ' I inn1110101 21 12522 ; !Zi aL24 :202 eCZVS-i 2 9 1 2530 2531 12 12521 622 12523 2524 ; cu,..,..cc6 2:6316-124112629 2630 Al' ar Nal I . ' Illne 26 11111nni11.1101111ummillirerkillat rille"i" 11111.1114111111.7112721 12722 lje--2724 ; 2722 2"..2E 12727 l272: 9 2r21114mr 11011311111.1.1111111111 0 . I ,.1 , r I grommitimaumme,_11/111111alarnalliala gmtakiiiirailigiatal. : )1 ; 28 illM16.,___7111111"- _,A.L....L 1 mizionaimmillinni; ilaillliallan 2 02921i2922 lam oii, 1 2_ :2926 23Allarr 9 1293:milk inallnialliMIHMaraigirail immTimaililltaiiiin 1111.11.111111 I ; * I I 111111111111 3019 3020 3021 3022 poza :30 413025 :3C25 27 6 ISINIIKMOMMISSIMMAIWA 111111111121•41111111111111 Ira 120.12, 1 Ri?:::. pr-• 111111111111111 1 3127 3128 13129 130 111111111111111111111111111111111all airilliiiIIIIM1111/711.111.111.1 ini 1111111111: isijus_111111111ram2111111111214111.1111111111 .,..e. lialimalnurlinilli , Milli:7ahall9 : r 3325 I3M5 nsinni; ill2: 721 11"11173:1211.1.11111 Eareallililli 1 •''' I 11" ..m...11111171111mIMMITI 1111191111111111111111 =1117:8:1111112: • ausiallOnliennomonmsaisiN11111111111 rij'are.mmillsininn . 10 12 14 16 18 20 24 26 28 30 32

Map 9.4.6 The known geographical distribution of S. frutescens subsp. microphylla.

102 TAXONOMY

2. Sutherlandia tomentosa Eckl. & Zeyh., Enum. 2: 251 (1836); PhiII. & Dyer in Rev. Sudamer. Bot. 1:75 (1934). Type: South Africa, Cape Agulhas, Ecklon & Zeyher 1659 (S!, specimens with original label, lecto., designated here; S!, isolecto.).

= S. frutescens var. tomentosa (Eck. & Zeyh.) Harv., Fl. Cap. 2: 212 (1862). Type as above.

[Note: There are two identical specimens in S, but the chosen lectotype has, in addition to the Enumeratio label, also a handwritten note, probably in Zeyher's hand.]

An erect to procumbent shrub, 0.6 m high. Leaves imparipinnate; 60-80 mrri long, pinnae alternate, obcordate and deeply emarginate; densely tomentose beneath and above (Fig. 9.1.7 a), 8-10 on each side of rachis, (5–)6-9(-10) mm long, 4-8 mm broad. Petiole 5-8(-12) mm long, petiolule (0–)0.1-0.5 mm long. Inflorescence a few–flowered raceme, usually about 5-6- flowered. Flowers 30-32 mm long, bright red or scarlet with white streaks on the standard petal. Pedicels 6 mm long. Calyx 8 mm long and 12 mm broad, tube 5.5 mm long; lobes 2.5 mm long. Standard narrowed to the base, with the margins usually reflexed above, 21 mm long, 10 mm broad. Alae 9 mm long, claw 3 mm long. Keel oblong-lanceolate, 21 mm long, 8 mm wide; claw 6 mm long. Stamina) tube 20 mm long. Ovary oblong 11 mm long, semi-translucent, glabrous, stipitate; stipe 4 mm long; style laterally bearded above, 12 mm long. Fruit membraneous, inflated, hairy, 50-60 mm long, 20-29 mm broad; stipe 2-3(-6) mm long and closer to upper (seed–bearing) suture (Fig. 9.4.7 a).

Distribution and habitat. Restricted to coastal areas (Map 9.4.7) on sand dunes (Fig. 9.4.7 b).

Specimens examined 3318 (Cape Town): Koeberg Nature Reserve (–AC), Moshe, Van Wyk & De Castro 1 (JRAU). Bloubergstrand (–AD), Palmer 13 (JRAU); 5,7 km N of Potterfield/Park Drive road, West Coast road (–AD), Moshe, Van Wyk & De Castro 17 (JRAU). —3418 (Simonstown): Witsand (–AB), Palmer 19 (JRAU); Karbonkelberg, Houtbay (–AB), Salter 7867 (PRE). 3420 (Bredasdorp): Waenhuiskrans (–CA), Van der Westhuizen 145 (PRE). —3421 (Riversdale): Still Bay (–AD), Van Wyk 3669, Moshe, Van Wyk & De Castro 6 (JRAU).

103

TAXONOMY

(a) (b)

Figure 9.4.7 The coastal species, Sutherlandia tomentosa, showing the densely tomentose obcordate and deeply emarginate leaflets with the fruit similar to that of S. frutescens (a) and the sand dunes where this species occurs (b).

12 14 16 18 20 22 24 26 28 30 32 34 raragaitTimussufralumucaural ersimangdaisamiammuniagralellailltIl 18 kunaglialli"11111117"mirm*Haifi"liawigissmawatuimualuaramminuna l 18 milarmarailaailislilirsail ftgi ima:11::: dliCa"1111"1:11"1.11161126191141111 11111111.141811111 , • rararadvirg 1511ill1aulls"61111611;20 2 aireami EmplImiluilltull01112: amiwimaravkaiiiii""1"amillirealW1111111.1 IllikWUrmaimilaalgifigau 1111111111111111111111MIM MI sumws ' seasimatgastatausimmitaa01111111111111 tuairfardia...i.xidualialatiiitainuumiul 11111111111111111011101111111111 h___._1 _r_ arkimumilmmummuni 111111111111allitgliWIMIttillittigl Eiral _7_ 1 :m.23iam:::: 21111111111111111M1 11111111111111 AilliiMiliiraMill LW '4-44 44 ; J24 24 .:11111„ Fizq :7: 111411:( 1111111111111111M ! ' 24 1111111111111111:11111111111111111 19 23 ! 624 2-: 2!25 e. :v.122 ilnialliienalitingiiatla iw. loin 1wain 11111111111111111111MM111111111 Ill, NM. :. 4 6 3 2624 525 2925 .r27. ;2 28 11111111111111111111WMIUMMilla • 11111111111111111111 i 1m :2728. 1111111111111111111168A111111111 • ttili ‘. ia.kj;;;WlrdMTITlidl 1111111111111111914111111111 - 61.1261j '28 3 2i1 ;925 2878 2827- 2820... . illiallagrill 111111111111111111111111A ' • Maim4,11N ms -mg To 11111111111111111M SE 1 23 /§25 mt- 1.111111111111111.111W I! - 1 302-5-h-2, c . rrniallearill itear lill lataigirAtial 322 IlliallaninlinallW ltiiiiiiitiail 312: :65411; 111111111111111111a1110111111111111MIM 1 11111111111111:1111111 •mmaiminimammEraniummaaralitarpumazu 4 CallaliffillialiMMEI NIMIMPallininnala Mil iffiri... VIIMMIIIIIINWAVONAWKalliallimar■ . iEDLeglaammiliminalineix lira mil Li Effillasomm IMNIMILIMILxvirailaria- ram 4 ' 1111111110-10EnrairginiMilliffilin.inglininii.11111111111111U111M21114 10 12 14 16 18 20 24 26 30 32 34 36

Map 9.4.7 The known geographical distribution of S. tomentosa.

104 Acknowledgements

Sincere thanks goes to the following persons and institutions for contributing to the success

of this study:

My supervisor, Prof. Ben-Erik Van Wyk for his guidance, encouragement, constructive critisism and believing in my potential.

My co-supervisor Michelle and Prof. Herman. van der Bank for guidance and encouragement and assistance with enzyme electrophoresis.

The Foundation for Research Development, the Rand Afrikaans University and South African Druggists for financial support.

Professor Fanie van Heerden of the Chemistry Department for identifying most of the leaf compounds and for supplying chemical standards, support and encouragement.

Mr Pat Smith from Medunsa (Chemical Pathology Department) for identifying amino acids with great enthusiasm.

Alvaro Viljoen and his wife Aimy for all their assistance with chemistry, computer packages and support throughout my study

The Botany department personnel for their laboratory equipment and assistance.

My family and friends for inspiration, encouragement and moral support.

My collegues at the Botany department; Gael Campbell, Tony de Castro, Nozuko Makhuvha for encouragement and support. •

105 CHAPTER 10

REFERENCES

ADAMS, M. R., JESSUP, W., MCCREDIE, R., ROBINSON, J., SULLIVAN, D. & CLERMAJER,

D. S. 1997. Oral L-arginine improves endothelium-dependent dilation and reduces

monocyte adhesion to endothelial-cells in young men with coronary-artery disease. J. Am.

College Card. 29: 7956-7956.

ARCHER, F. M. 1990. Planning with people—Ethnobotany and African uses of plants in

Namaqualand (South Africa), Proceedings of the twelfth plenary meeting of Aetfat, Mitt.

Inst. AIlg. Bot. Hamburg 23: 959-972.

ARROYO, M. T. KALIN. 1981. Breeding systems and pollination biology in leguminosae. In R.

M. Polhill & P. H. Raven (eds), Advances in Legume Systematics, Part 2, pp. 723-

769. Royal Botanic Gardens, Kew.

ATTIAS, J. M., SCHLESINGER, J. & SCHLESINGER, S. 1969. The effect of amino acid

analogues on alkaline phosphatase formation in Escherichia coll. J. Biol. Chem. 244:

3810-3817.

BAILEY, L. H. 1976. Hortus Third, p. 1085. Macmillan, New York.

BARNEBY, R. C. 1964. Atlas of North American Astragalus. Mem. New York. Bot. Gard.

13: 1-1188. is BELL,'D. 1974. The effect of canavanine on herpes simplex virus replication. J. gen. Virol. 22:

319-330.

BELL, E. A. 1958. Canavanine and related compounds in the Leguminosae. Biochem. J. 70:

617-619.

BELL, E. A. 1960. Canavanine in the Leguminosae. Biochem. J. 75: 618.

106 BELL, E. A. 1971. Comparative biochemistry of non-protein amino acids. In J. B. Harborne, D.

Boulter, & B. L. Turner (eds), Chemotaxonomy of the Leguminosae, pp. 179-204.

Academic Press, London.

BELL, E. A. 1980. The non-protein amino acids of higher plants. Endeavour 4: 102-104.

BELL, E. A. 1981. Non-protein amino acids in the Leguminosae. In R. M. Polhill & P. H. (eds),

Advances in Legume Systematics, Part 2, pp. 489-499. Royal Botanic Gardens, Kew.

BELL, E. A. 1981. Non-protein amino acids occuring in plants. In L. Reinhold, J. B. Harborne

& Swain T. (eds), Progress in Phytochemistry, Vol. .7, pp. 171-196. Pergamon Press, .

BELL, E. A., LACKERY, J. A. & POLHILL, R. M. 1978. Systematic significance of canavanine

in Papilionoideae (Faboideae). Biochem. Syst. Ecol. 6: 201-212.

BENTHAM, G. 1865. Leguminosae. In: G. Bentham, & J. D. Hooker (eds), Genera

Plantarum Vol. 1, pp. 434-600. Lovell Reeve, London.

BIRDSONG, B. A., ALSTON, R. & TURNER, B. L. 1960. Distribution of canavanine in the family

Leguminosae as related to phyletic grouping. Can. J. Bot. 38: 499.

BOGER, R. H., BODE-BOGER, S. M. & FROLICH, J. C. 1996. L-arginine-nitric oxide pathway:

role in atherosclerosis and therapeutic implications. Atherosclerosis 127: 1-11.

BREDT, D. S., HWANG, P. M. & SNYDER, S. H. 1991. Localization of nitric oxide synthase

indicating a neural role for nitric oxide. Nature 347: 768-770.

BROWICZ, K. 1963. The genus Colutea L. A monograph. Mono. Bot. 14: 1-136.

BROWN, R. 1812. Sutherlandia ex. Ait. f. In Hort. Kew. ed. 2, Vol. 4, p. 327.

BRUMMERHOFF, S. W. D. 1969. Sommige inhoudstowwe van Sutherlandia microphylla.

Unpublished D. Sc. thesis, University of the Free State.

BRUNETON, I. 1995. Pharmacognosy, Phytochemistry, Medicinal Plants, p. 692. Intercept, L: Hampshire.

BUCKINGHAM, J. 1994. Dictionary of Natural Products, Vol. 3, p. 3296. Chapman & Hall,

London.

107 BUSSE, R. & MULSCH, A. 1990. Induction of nitric oxide synthase by cytokines in vascular

smooth muscle cells. FEBS Letters 275: 87-90.

CENDAN, J. C., TOPPING, D. L., PRUITT, J. SNOWDY, S., COPELAND, E. M. III. & LIND, D.

S., 1996. Inflammatory mediators stimulate arginine transport & arginine-derived nitric

oxide production in a murine breast cancer cell line. J. Surg. Res. 60: 284-288.

CHO-CHUNG, Y. S., CLAIR, T. BODWIN, J. S. & HILL, D. M. 1980. Arrest of mammary tumor

growth in vivo by L-arginine: stimulation of NAD-dependent activation of adenylate

cyclase. Biochem. Biophys. Res. Commun. 95: 1306-1313.

COLWELL, C. S. 1997. Circadian rhythms—Time to get excited by GABA. Nature 387: 554-555.

COOKE, J. P. & TSAO, P. S. 1997. Arginine: A new perspective for the atherosclerosis?

Circulation 95: 311-312.

CORNER, E. J. H. 1951. The Leguminosae seed. Phytomorphology 1: 117-150.

CURTIS, W. 1792. Botanical Magazine p. 6, t.181. London.

DAS, I. & KHAN, N. S. 1995. Inhibition of nitric-oxide synthase by L-arginine metabolites.

Biochem. Soc. Trans. 23: 324S.

DAS, I. & KHAN, N. S. 1996. Differing effects of polyamines on nitric oxide synthase. Biochem.

Soc. Trans. 24: 484S.

DE CANDOLLE, A.P. 1825. Prodromus systematis naturalis regni vegetabilis 2, p.273. Treuttel

& Wurtz, Paris.

DE CANDOLLE, A.P. 1826. Memoires sur la famille des Legumineuses, p. 293. A. Belin, Paris.

DORMER, K. J. 1945. An investigation of the taxonomic value of shoot structure in angiosperms

with special reference to Leguminosae. Annals of Bot. Lond. 9: 141-153.

DORMER, K. J. 1946. Vegetative morphology as a guide to the classification of the

Papilionoideae. New Phytol. 45: 145-161.

DUNHILL, P. M., & FOWDEN, L. 1965. The amino acids of seeds of the Cucurbitaceae.

Phytochemistry 4: 933-944.

DUNHILL, P. M., & FOWDEN, L. 1967. The amino acids of the genus Astragalus.

108 Phytochemistry 6: 1659-1663.

DYKMAN, E. J . 1908. Kook-, Koek- en Resepte Boek. p. 145. Paarlse drukpers Maatskappy,

Paarl.

ECKLON, C. F. & ZEYHER, K. L. P. 1836. Enumeratio plantarum Africae Australis extratropicae

2, p. 251. Perthes & Besser, Hamburg.

FAEGRI, K. & VAN DER PIJL, L. 1979. The Principles of Pollination Ecology, 3rd ed. pp.110-115,

146. Pergamon, Oxford.

FEARON, W. R. & BELL, E. A. 1955. Canavanine: Detection and occurrence in Colutea

arborescens. Biochem. J. 59: 221-224.

FOWDEN, L., LEWIS, D. & TRISTRAM, H. 1967. Toxic amino acids: As antimetabolites. Advan.

Enzymol. 29: 89-163.

FURCHGOTT, R. F. & ZAWADZKI, J. V. 1980. The obligatory role of endothelial cells in the

relaxation of arterial smooth muscle by acetylcholine. Nature 288: 373-376.

GABRIELSE, V. S. 1996. Pharmacological studies on rooperol, rooperol derivatives and

Sutherlandia extracts. Unpublished M.Sc. thesis (Medical Sciences-Pharmacology),

University of Stellenbosch.

GARTHWAITE, J., CHARLES, S. & CHESS-WILLIAMS, R. 1988. Endothelium-derived

relaxing factor release on activation of NMDA receptors suggests a role as intercellular

messenger in the brain. Nature 336: 385-388.

GHOSAL, S. D., GOSH & DUTIA, S. K. 1970. Occurrence of erythrosine and other alkaloids in

Erythrina variegata. Phytochem. 9: 2379-2398.

GMELIN, 1791. Linnaeus's Systema Naturae, 13th ed., Vol. 2, p. 1027. Leipzig.

GOLDBERG, A. L. 1972. Degradation of abnormal proteins in Escherichia coli. Proc. Natl. Acad.

Sci. USA. 69: 422-426.

GOMES, C. M. R., GOTTLIEB, 0. R. & SALANTINO, A. 1981. Phytochemistry in perspective:

Chemosystematics of the Papilionoideae. In R. M. Polhill & P. H. Raven (eds), Advances

in Legume Systematics, Part 2, pp. 465-488. Royal Botanical Gardens, Kew.

109 GORBUNOV, N. V., TYURINA, Y. Y., SALAMA, G., DAY, B. W., CLAYCAMP, H. G., ARGYROS,

G., ELSAYED, N. M., & KAGAN, V. E. 1998. Nitric oxide protects cardiomyocytes against

tent-butyl hydroperoxide-induced formation of alkoxyl and peroxyl radicals and peroxidation

of phosphostidylserine. Biochem. Biophys. Res. Commun. 244: 647-651.

GOTTLIEB, L. D. 1982. Conservation and duplication of isozymes in plants. Science 216: 373-

380. L GREEN, M. H. & WARD, J. F. 1983. Enhancement of human tumor cell killing by L-canavanine

in combination with y-radiation. Cancer Res. 43: 4180-4182.

GREEN, M. H., BROOKS, T. L., MENDELSON, J. & HOWELL, S. B. 1980. Antitumor activity of

L-canavanine against L1210 murine leukemia. Cancer Res. 40: 535-537.

GRIFFITHS, S. M. 1992. The New Royal Horticultural Society Dictionary of Gardening, p. 1129.

McMillan, London.

HAHNER, L., MCOUILKIN, S & HARRIS, R. A. 1991. Cerebellar GABAB receptors modulate

function of GABAA receptors. FASEB J. 5:2466-2472.

HAMRICK, J. L. 1979. Genetic variation & longevity. In 0. T. Solbrig, S. Jain, G. B. Johnson &

P. H. Raven (eds), Topics in Plant Population Biology, pp. 83-113. Columbia

University Press, New York.

HARBORNE, J. B. 1971a. Distribution of flavonoids in the Leguminosae. In J. B. Harborne, D.

Boulter, & B. L. Turner (eds), Chemotaxonomy of the Leguminosae, pp. 257-283.

Academic Press, London.

HARBORNE, J. B. 1971b. Triterpenoids and other low molecular weight substances of

systematic interest in the Leguminosae. In J. B. Harborne, D. Boulter, & B. L. Turner

(eds), Chemotaxonomy of the Leguminosae, pp. 257-283. Academic Press, London.

HARBORNE, J. B. 1973. Phytochemical methods, pp. 1-32,182-192. Chapman & Hall,

London.

HARVEY, W. H. 1862. Leguminosae. In W. H Harvey & 0. W. Sonder (eds), Flora Capensis

110 Vol. 2, p. 212. Hodges, Smith & Co., Dublin.

HENDIL, K. B. 1975. Degradation of abnormal proteins in Hela cells. J. Cell Physiol. 87: 289-

296.

HIBBS, J. B., TAINTOR, R. R., VAVRIN, Z. RACHLIN, E. M. 1988. Nitric oxide: a cytotoxic

activated macrophage effector molecule. Biochem. Biophys. Res. Commun. 157: 87—

94.

HILLIS, D. M. & MORITZ, C. 1990. Molecular Systematics, p. 45. Sinauer Associates Inc.,

Massachusetts.

HORNERO, J. & PEREZ, C. 1997. Evaluation of genetic variability in Iberian Colutea spp.

using isozyme electrophoresis. Biochem. Syst. Ecol. 25: 13-20.

HOROWITZ, H. N. & SRB, A. M. 1948. Growth inhibition of Neurospora by canavanine and

its reversal. J. Biol. Chem. 174: 371-378.

HUTCHINSON, J. 1964. The Genera of Flowering Plants, Vol. 1, p. 406. Oxford University

Press, Oxford.

HWANG, I. D., KIM, S-G. & KWON, Y. M. 1996. Canavanine metabolism in tissue cultures of

Canavalia lineata. Plant Cell, Tissue and Organ Culture 45 :17-23.

IGNARRO, L.J., BUGO, G. M., BYRNS, R. E. & CHAUDHUN, G. 1987. Endothelium-derived

relaxing factor produced and released from artery and vein is nitric oxide. Proc. Natl.

Acad. Sci. USA. 84: 9265-9269.

JOSHI, M. 1997. the importance of L-arginine metabolism in melanoma: an hypothesis for the role

of nitric oxide and polyamines in tumor angiogenesis. Free Radical Biology & Medicine

22: 573-578.

KAY, Q. 0. N. 1987. Ultraviolet patterning and ultraviolet-absorbing pigments in flowers of the

Leguminosae. In C. H. Stirton (ed.), Advances in Legume Systematics, Part 3, pp. 317-

354, Royal Botanic Gardens, Kew.

KELLY, A., MORRIS, S. M. & BILLIAR, T. R. 1995. Nitric oxide, sepsis, and arginine metabolism.

JPEN 19: 234-238.

111 KEPHART, S.R. 1990. Starch gel electrophoresis of plant isozymes: a comparative analysis of

techniques. Am. J. Bot. 77: 693-712.

KHAN, F., LITCHFIELD, S. J., MCLAREN, M., VEALE, D. J., LITTLEFORD, R. C. & BELCH, J.

J. F. 1997. Oral L-arginine supplementation and cutaneous vascular responses in patients

with primary Rynaud's phenomenon. Arthritis & Rheumatism, 40: 352-357.

KINGHORN, A. D. & SMOLENSKI, S. J. 1981. Alkaloids of Papilionoideae. In: R. M. Polhill & P.

H. Raven (eds), Advances in Legume Systematics, Part 2, pp. 585-598. Royal Botanic

Gardens, Kew.

KNOWLES, S. E., GUNN, J. M., HANSON, R. W. & BALLARD, F. J. 1975. Increased

degradation rates of protein synthesized in hepatoma cells in the presence of amino acid

analogues. Biochem. J. 146: 595-600.

KONTUREK, S. K. & KONTUREK, P. CH. 1995. Role of nitric oxide in the digestive system.

Digestion 56: 1-13.

KRUSE, P. F., WHITE, P. B., CARTER, H. A. & MCCOY, T. A. 1959. Incorporation of canavanine

into protein of Walker Carcinosarcoma 256 cells cultured in vitro. Cancer Res. 19: 122—

125.

KUO, T. M., LOWELL, C. A. & NELSEN, T. C. 1997. Occurrence of pinitol in developing soybean

seed tissues. Phytochemistry 45: 29-35.

LERSTEN, N. R. 1979. A distinctive seed coat pattern in the Viceae (Papilionoideae:

Leguminosae). Proc. Iowa Acad. Sci. 86: 102-104.

LERSTEN, N. R. 1981. Testa topography in Leguminosae, subfamily Papilionoideae. Proc. Iowa

Acad. Sci. 88: 180-191.

LERSTEN, N. R. & GUNN, C. R. 1981. Testa characters in tribe Viceae, with notes about tribes

Abreae, Cicereae, and Trifoleae (Fabaceae). U.S. Dept. of Agric. Tech. Bull.

LINNAEUS, C. 1753. Species plantarum. 2nd ed. Vol. 2, pp. 994-995 & 1004-1005, Stockholm.

LINNAEUS, C. 1759. Species plantarum (Veg.). p. 1045.

LISTON, A. 1992. Isozyme Systematics of Astragalus sect. Leptocarpi subsect. Californici

112 (Fabaceae). Syst. Bot. 17: 367-379

MA, Q., HOPER, M., ANDERSON, N. & ROWLANDS, B. J. 1996. Effects of supplemental L-

Arginine in a Chemical-Induced Model of Colorectal Cancer. World J. Surg. 20: 1087–

1' 1091.

MABBERLEY, D. J. 1987. The Plant-Book, p. 563. Cambridge University Press, Cambridge.

MAHATO, S. B., ASHOKE, K. N. & GITA, R. 1992. Triterpenoids. Phytochemistry 7: 2199-

2249.

MANNING, J. C. & VAN STADEN, J. 1987. The systematic significance of testa anatomy in the

Leguminosae—an illustrated survey. S. Afr. J. Bot. 53: 210-230.

MARKET, C. L. & FAULHABER, I. 1965. Lactate dehydrogenase isozyme patterns of fish. J.

Exp. Zool. 159: 319-332.

MEARS, J. A. & MABRY, T. J. 1971. Alkaloids in the Leguminosae. In J. B. Harborne, D. Boulter,

& B. L. Turner (eds), Chemotaxonomy of the Leguminosae, pp. 73-178, Academic Press,

London.

MEISTER, A. 1965. Biochemistry of the Amino Acids, Vol. 1, pp. 231-268. Academic Press, New

York.

MERCK 1989. The Merck Index. p. 263, 11th ed. Merck, Rahway.

MERXMOLLER, H. 1970. Prodromus einer flora von SOdwest Africa, pp. 112-113.

MEYER, E. H. F. 1836. Commentariorum de plantis Africa& Australions 1(1), pp. 121-122.

Leopoldum Voss, Leipzig.

MEYER, E. H. F. 1832. Plantae Ecklonianae. Linnaea 7: 170.

MILLER, E. J. & HARRISON, J. S. 1950. Growth inhibition of a yeast by uracil and its reversal

by arginine. Nature 166: 1035.

MONCADA, S. & HIGGS, A. 1993. The L-arginine-nitric oxide pathway. N. EngL J. Med. 329:

2002-2012.

MONCADA, S., PALMER, R. M. J. & HIGGS, A. 1991. Nitric oxide: Physiology, pathophysiology

and pharmacology. Pharmacol. Rev. 43: 109-142.

113 MORGAN, D. M. L. 1995. Polyamines, arginine and nitric oxide. Biochem. Soc. Trans. 22: 879—

883.

MOURAD, F. H., O'DONNELL, L. J. D., ANDRE, E. A., BEARCROFT, C. P., OWEN, R. A.,

CLARK, M. L. & FARTHING, M. J. G. 1996. L-arginine, nitric oxide, and intestinal

secretion: studies in rat jejunum in vivo. Gut 39: 539-544.

NAKANISHI, K. 1969. Studies on tumor growth inhibition of arginine imbalanced diet. Osaka

Univ. Med. J. 21: 193-204.

NATHAN, C. 1992. Nitric oxide as a secretory product of mammalian cells. FASEB J. 6:

3051-3064.

NEI, M. 1973. Analysis of gene diversity in subdivided populations. Proc Nat. Acad. Sci. USA

70: 3321-3323

NEI, M. 1978. Estimation of average heterozygosity and genetic distance from a small number

of individuals. Genetics 89: 583-590.

NEI, M. 1986. Definition and estimation of fixation indices. Evolution 40: 643-645.

NEURATH, A. R., WEINER, F. P., RUBEN, B. A. & WARTZELL, R. W. 1970. Inhibition of

Adenovirus replication by canavanine. Biochem. Biophys. Res. Commun. 41: 1509-1517.

OTA, J. 1993. DISPAN. Pennsylvania University, USA.

PALMER, R. M. J., FERRIGE, A. G. & MONCADA, S. 1987. Nitric oxide accounts for the

biological activity of endothelium-derived relaxing factor. Nature 327: 524-526.

PALMER, R. M. J., REES, D. D., ASHTON, D. S. & MONCADA, S. 1988. L- Arginine is the

physiological precursor for the formation of nitric oxide in endothelium-dependent

relaxation. Biochem. Biophys. Res. Commun. 153: 1251-1256.

PALMER, R. M. J., ASHTON, D. S. & MONCADA, S. 1988. Vascular endothelial cells synthesize

nitric oxide from L-arginine. Nature 333: 664-666.

PHILLIPS , R. & DYER, R. A. 1934. The Genus Sutherlandia R.Br. Rev. Sudam. Bot. 1 (3):

69-90.

PHILLIPS, E. P. 1926. Genera of Flowering Plants, p. 328. Dept. Agriculture Technical Services,

114 Pretoria.

PINE, M. J. 1967. Response of intracellular proteolysis to alteration of bacterial protein and the

implications in metabolic regulation. J. Bacteriol. 93: 1527-1533.

POLHILL, R. M. 1981a. Papilionoideae. In R. M. Polhill & P. H. Raven (eds), Advances in

Legume Systematics, Part 2, pp. 191-298. Royal Botanic Gardens, Kew.

POLHILL, R. M. 1981b. Galegeae. In R. M. Polhill & P. H. Raven (eds), Advances in Legume

Systematics, Part 2, pp. 357-363, Royal Botanic Gardens, Kew.

POLTE, T., OBERLE, S. & SCHRODER, H. 1997. Nitric oxide protects endothelial cells from

tumor necrosis factor-a-mediated cytotoxicity: possible invonement of cyclic GMP. FEBS

Letters 409: 46-48.

PROUTY, W. F., KARNOVSKY, M. J. & GOLDBERG, A. L. 1975. Degradation of abnormal

proteins in Escherichia coll. J. Biol. Chem. 250: 1112-1122.

RADOMSKI, M. W., JENKINS, D. C., HOLMES, L. & MONCADA, S. 1991. Human colorectal

adenocarcinoma cells: differential nitric oxide synthesis determine their ability to

aggregate platelets. Cancer Res. 51: 6073

RAMCHAND, C. N., DAS, I., GLIDDON, A. & HIRSCH, R. 1994. Schizophr. Res. 13: 249-

253.

RANKI, M. & KAARIAINEN, L. 1969. Canavanine as an inhibitor of Semiliki Forest virus

growth. Ann. Med. Exp. Fenn. 47: 65-72.

REBELO, A. G. 1987. Bird pollination in the Cape flora. In A. G., REBELO (ed.), A

preliminary synthesis of pollination biology in the Cape flora, pp. 83-104. S. Afr.

Nat. Sci. Prog. Rep No. 141, FRD, Pretoria.

REHR, S. S., BELL, E. A. & JANZEN, D. H. & FEENY, P. P. 1973. Insecticidal amino acids in

legume seeds. Biochem. Syst. Ecol. 1: 63-67.

RIDGWAY, G., SHERBURNE, S. W. & LEWIS, R. D. 1970. Polymorphism in the esterases

of Atlantic herring. Trans Am. Fish. Soc. 99: 2-15.

ROHLF, F. J. 1997. Numerical Taxonomy and Multivariate Analysis Systems. NTSYS-pc

115 2.01d. New York.

ROOT, R. K. & JACOBS, R. 1991. In: J. D. Wilson, E. Braunwald, K. J. Isselbacher, R. G.

Peterdorf, J. B. Martin, A. S, Fauci, & R. K. Root (eds), The Principles of Internal

Medicine, 12th ed., pp. 502-507. Mc Graw-Hill, New York.

) ROSENTHAL, G. A. 1977. The biological effects and mode of action of L-canavanine, a

structural analog of arginine. Quart. Rev. Biol. 52: 155-178.

ROSENTHAL, G. A. 1991. The biochemical basis for the deleterious effects of L-canavanine.

Phytochemistry 30: 1055-1058.

ROSENTHAL, G. A. 1992. L-Canavanine and chemical defense in higher plants. In: K. Takai,

(ed.) , Frontiers and New horizons in Amino Acid Research, pp. 109-113. Elsevier, t. New York.

ROSENTHAL, G. A. & RHODES, D. 1984. L-Canavanine transport and utilization in

developing jack bean, Canavalia ensiformis (L.) DC. J. Agric. Food Chem. 36: 1159—

1163.

ROSENTHAL, G. A., LAMBERT, J. & HOFFMAN, D. 1989. L-Canavanine incorporation into

proteins can impair macromolecule function. J. Biol. Chem. 264: 9768-9771.

ROSENTHAL, G. A. & DAHLMANN, D. L. 1991. Incorporation of L-canavanine into proteins

and the expression of its antimetabolic effects. J. Food Agr. Chem. 39: 987-990.

ROSENTHAL, G. A., SWAFFAR, D. S. & ANG, C. Y. 1995. Inhibition of the growth of human

pancreatic—cancer cells by the arginine antimetabolite L-canavanine. Cancer Res. 55: t_ 4486.

ROSENTHAL, G. A., DAHLMAN, D. L., CROOKS, P. A., PHUKET, S. N. & TRIFONOV, L. S.

1995. Insecticidal properties of some derivatives of L-canavanine. J. Agric. Food

Chem. 43: 2728-2734.

ROSENTHAL, G. A. & HARPER, L. 1996. L-homoarginine studies provide insight into the

antimetabolic properties of L-canavanine. Insect Biochem. Molec. Biol. 26: 389-394.

ROSS, R. 1993. Thepathogenesis of atherosclerosis: a perspective for the 1990's. Nature.

116 362:801-809.

SASS, J. E. 1958. Botanical Microtechnique, 3rd ed. pp. 15-18, Iowa State University Press,

Iowa.

SCHACHTELE, C. F. & ROGERS, P. 1965. Canavanine death in Escherichia coll. J. MoL Biol.

14: 474-489.

SCHRIRE, B. D. & ANDREWS, S. 1992. Sutherlandia: gansies or balloon peas: part 1. The

Plantsman 14: 65-69.

SCHWARTZ, M., ALTMAN, A., COHEN, Y. & ARZEE, T. 1996. Inhibition of polyamine

biosynthesis by L-canavanine and its effects on meristematic activity, growth, and

development of Zea mays roots. Israel J. Plant Sci. 45: 23-30.

SHACKLEE, J. B., ALLENDORF, F. W., MORITZ, D. C. & WHITT, G. S. 1990. Gene

nomenclature for protein-coding loci in fish. Trans. Am. Fish Soc. 119: 2-15.

SKEAD, C. J. 1967. The Sunbirds of Southern Africa, p.85. Balkema, Cape Town.

SHQUEIR, A. A., BROWN, D. L. & KLASING, K. C. 1989. Canavanine content and toxicity of

Sesbania leaf meal for growing chicks. Anim. Feed Sci. Technol. 25: 137-147.

SMITH, A. 1895. South African Materia Medica. pp. 60-62,66. Juta & Co.

SNYDERS, J. H. 1965. Chemiese ondersoek van Sutherlandia microphylla Burch. MSc,

Universiteit van die Oranje-Vrystaat.

SOLTIS, D. E. & SOLTIS, P. S. 1989. Isozymes in Plant Ecology, pp. 241-258. Chapman &

Hall, London.

SOLTIS, D. E., HAUFLER, D. C., DARROW, D. C. & GASTONY, G. J. 1983. Starch gel

electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and

staining schedules. Am. Fern J. 73: 9-27.

SOUTHON, I. W. 1994. Phytochemical Dictionary of the Leguminosae, pp. cxii—cxviii. Chapman

& Hall, London.

STACE, A. C. 1980. Plant Taxonomy and Biosystematics, 2nd ed., p. 135. Edward Arnold,

London.

117 STAMLER, J. S., SINGEL, D. J. & LOCALZO, J. 1992. Biochemistry of nitric oxide and its

redox-activated forms. Science 258: 1898-1902.

STARK, M. E. & SZURSZEWSKI, J. H. 1992. Role of nitric oxide in gastrointestinal and hepatic

function and disease. Gastroenterology. 103: 1928.

STEPHENSON, F. A. 1995. GABA A receptors. Biochem. J. 310: 13-21.

STUEHR, D. J., KWON, N. S., NATHAN, C. F., GRIFFITH, 0. W., FELDMAN, P. L. &

WISEMAN, J. 1991. N"'-Hydroxy-L-arginine is an intermediate in the biosynthesis of

nitric oxide from L-arginine. J. Biol. Chem. 266: 6259-6263.

SWOFFORD, D. L. & SELANDER, R. B. 1981. BIOSYS-1: A FORTRAN program for the

comprehensive analysis of electrophoretic data in population genetics and systematics.

J. Hered. 72: 281-283.

TAKEDA, Y., TOMINAGA, T., TEI, N., KITAMURA, M., TAGA, S., MURASE, J., TAGUCHI, T.

& MITAWANI, T. 1975. Inhibition effect of L-arginine on growth of rat mammary tumors

induced by 7,12-dimethylbenz(a)anthracene. Cancer. Res. 35: 2390-2393.

THOMAS, D. A., ROSENTHAL, G. A., GOLD, D. V. & DICKEY, K. 1986. Growth inhibition of a

rat colon tumor by L-canavanine. Cancer Res. 46: 2898-2903.

THUNBERG, C. P. 1800. Prodromus plantarum capensium 2, p. 134. Uppsala.

THUNBERG, C. P. 1823. Flora Capensis, p. 603. Stuttgart.

TUCKER, S. C. 1987. Pseudoracemes in papilionoid legumes: their nature, development and

variation. Bot. J. Linn. Soc. 95: 181-206.

TURNER, B. L. & HARBORNE, J. B. 1967. Distribution of canavanine in the Plant Kingdom.

Phytochemistry 6: 863-866.

VAN DER BANK, H., VAN DER BANK, M. & VAN WYK, B-E. 1995. Allozyme variation in Virgilia

oroboides (tribe Podalyrieae, Fabaceae). Biochem. Syst. Ecol. 23: 47-52.

VAN WYK, B.-E., VAN OUDTSHOORN, B. & GERICKE, N. 1997. Medicinal Plants of South

Africa, p. 246. Briza Publications, Pretoria.

VILJOEN, P. T. 1969. Die oxidasie van pinitol en gedeeltelike identifikasie van 'n triterpeen

118 glikosied uit Sutherlandia microphylla Burch. Unpublished M.Sc., University of the Free

State.

WAGNER, S., CASTEL, M., GAINER, H. & YAROM, Y. 1997. GABA in the mammalian

suprachiasmatic nucleus and its role in diurnal rhythmicity. Nature 387: 598-603.

WASCHER, T. C., POSCH, K., WALLNER, S., HERMETTER, A., KOSTNER, G. M. & GRATER,

W. F. 1997. Vascular effects of L-arginine: anything beyond a substrate for the NO-

synthase?. Biochem. Biophys. Res. Commun. 234: 35-38.

WATT, J. M. & BREYER-BRANDWIJK, M. G. 1962. The Medicinal and Poisonous Plants of

Southern and Eastern Africa, p. 649. Livingstone, London.

WEBERLING, F. 1989. Morphology of flowers and inflorescences, p. 337-338. Cambridge

University Press, Cambridge.

WEISBURGER, J. H., YAMAMOTO, R. S., GLASS, R. M. & FRANKEL, H. H. 1969. Prevention

by arginine glutamate of the carcinogenicity of acetamide in rats. Toxicol. Appl.

Pharmacol. 14: 163-175.

WINK, D. A., VODOVOTZ, Y., LAVAL, J., LAVAL, F., DEWHIRST, M. W. & MITCHELL, J. B.

1998. The multifaceted roles of nitric oxide in cancer. Carcinogenesis 19: 711-721.

WOLF, A ZALPOUR, C., THEILMEIER, G., WANG., B.-Y., MA, A., ANDERSON, B., TSAO, P.

S. & COOKE, J. P. 1997. Dietary L-arginine supplementation normalizes platelet

aggregation in hypercholesterolemic humans. JACC. 29:479-485.

WU, C.-C., HONG, H.-J., CHOU, T.-C. DING, Y.-A. & YEN, M.-H. 1996. Evidence of inducible

nitric oxide synthase in spontaneously hypertensive rats. Biochem. Biophys. Res.

Commun. 228: 459-466.

WRIGHT, S. 1978. Variability within and among natural populations. In Evolution and the

Genetics of Populations. pp. 82-89. Vol. 4. University of Chicago Press, Chicago IL.

XIE, Q. W., CHO, H. J., CALACAY, J., MUMFORD, R. A., SWIDEREK, K. M., LEE, T. D.,

DING, A., TROSO, T. & NATHAN, C. 1992. Cloning and characterization of inducible

119 nitric oxide synthase from mouse macrophages. Science 256: 225-228.

120