Chemical Diversity of the Natural Populations of Dalmatian Pyrethrum(Tanacetum cinerariifolium (Trevir.) Sch.Bip.) in Croatia

by Martina Grdisa*a), Sandra Babicb), Martina Perisab), Klaudija Carovic-Stankoa), Ivan Kolaka),Zlatko Liberc), Marija Jug-Dujakovicd), and Zlatko Satovica)

a) Department of Seed Science and Technology, Faculty of Agriculture, University of Zagreb,Svetosimunska 25, HR-10000 Zagreb, Croatia

(phone: þ385-1-2393623; fax: þ385-1-2393330; e-mail:mgrdisa@agr.hr)b) Department of Analytical Chemistry, Faculty of Chemical Engineering and Technology, University of

Zagreb, Marulicev trg 19, HR-10000 Zagreb, Croatiac) University of Zagreb, Faculty of Science, Division of Biology, Department of Botany, Marulicev trg 9,

HR-10000 Zagreb, Croatiad) Institute for Adriatic Crops and Karst Reclamation, Put Duilova 11, p.p. 288, HR-21000 Split, Croatia

Dalmatian pyrethrum (Tanacetum cinerariifolium (Trevir. ) Sch.Bip.) is a plant species endemic tothe east Adriatic coast. The bioactive substance of Dalmatian pyrethrum is a natural insecticide,pyrethrin, a mixture of six active components (pyrethrins I and II, cinerins I and II, and jasmolins I andII). The insecticidal potential of pyrethrin was recognized decades ago, and dried and ground flowershave traditionally been used in Croatian agriculture and households.

A total of 25 Dalmatian pyrethrum populations from Croatia were studied to determine thepyrethrin content and composition, and to identify distinct chemotypes. The total pyrethrin contentranged from 0.36 to 1.30% (dry flower weight; DW) and the pyrethrin I/pyrethrin II ratio ranged from0.64 to 3.33%. The statistical analyses revealed that the correlations between the percentage of pyrethrinI and of all the other components were significant and negative. The total pyrethrin content waspositively correlated with the percentage of pyrethrin I and negatively correlated with cinerin II. Themultivariate analysis of the chemical variability enabled the identification of five chemotypes among 25Dalmatian pyrethrum populations. The chemical characterization of indigenous Dalmatian pyrethrumpopulations may serve as a good background for future breeding and agricultural exploitation.

Introduction. – Dalmatian pyrethrum (Tanacetum cinerariifolium (Trevir.) Sch.Bip.) is a perennial plant species of the Asteraceae (Compositae) family, endemic tothe Adriatic coast and its islands. Dalmatian pyrethrum plants contain a potent naturalinsecticide, pyrethrin. This secondary metabolite is highly biodegradable when exposedto light, water, and air, and does not accumulate in food chains and ground water [1 –3].It is, therefore, an efficient and environmentally safe method of insect control [1].Pyrethrin acts as a contact poison, penetrating into the nervous system of insects,causing knock-down effect and death [4]. The insecticidal activity of pyrethrin is aresult of the synergistic action of six active pyrethrin components: pyrethrins I and II,cinerins I and II, and jasmolins I and II, with pyrethrins I and II being the most active(Fig. 1). A considerable difference in the insect toxicity of pyrethrins I and II has beendocumented; pyrethrin II is associated with a faster knock-down effect and pyrethrin Iwith a greater killing power [5]. Pyrethrin I alone is toxic, while the insects easily

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)460

� 2013 Verlag Helvetica Chimica Acta AG, Z�rich

metabolize pyrethrin II and recover in a few hours. Nevertheless, the combination ofpyrethrins I and II has an outstanding effect on a wide range of insect species [6].Therefore, the ratio of pyrethrin I to pyrethrin II is of utmost importance as itdetermines the quality of the pyrethrum extract. The ratio typically varies between 0.90and 1.30 [7], although a wider range (0.50 –3.50) has been reported in some breedinglines [8].

The main source of pyrethrin is the flower head, a compound inflorescenceconsisting of yellow disc florets and white ray florets [9]. Pyrethrin content varies indifferent flower parts; i.e., ca. 94% of pyrethrin is accumulated in the achenes, andminor quantities in the disc florets (2.0%), ray florets (2.6%), and receptacles (2.6%)[10].

The harvesting period, genotype, geographical source, and climate [11], as well asthe drying method and storage conditions are the factors which influence the pyrethrincontent [12]. The most appropriate harvesting period for pyrethrum flowers is when 3/4of the disc florets are opened. At that phase, the highest pyrethrin content has beendetected, and, beyond that stage, the yield of pyrethrin does not increase appreciably[10] [13] [14].

Pyrethrum powder has a long history as an insect control agent in many Croatianhouseholds and traditional agricultural systems. The cultivation of Dalmatian

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013) 461

Fig. 1. Chemical structures of pyrethrins I and II, cinerins I and II, and jasmolins I and II

pyrethrum along the Dalmatian coastal region began around 1854 [15]. From 1930 tothe beginning of the World War II, the production rapidly decreased, and the discoveryand usage of the synthetic pesticide, DDT (�Dichloro-Diphenyl-Trichloroethane�)further reduced its production. In the meantime, it was introduced to Japan, Kenya, andIndia where the first breeding programs have been initiated [16– 20]. Today, the majorproduction areas of pyrethrum are located in East Africa (Kenya, Rwanda, andTanzania), Tasmania, China, and Papua New Guinea [21]. Commercial varieties withca. 1.80 – 2.50% of pyrethrin [12] and high production clones with the highest pyrethrincontent of 3.0% have been reported [22].

Only a few reports regarding the total pyrethrin content of natural Dalmatianpyrethrum populations are available [14] [23]; thus, a detailed assessment of thechemotypical variations has not been conducted. This study represents the first attemptof such systematic evaluation with the following aims: i) to determine the pyrethrincontent and composition of 25 Dalmatian pyrethrum populations, ii) to determine thecorrelations between total pyrethrin content and six pyrethrin components (pyrethrinsI and II, cinerins I and II, and jasmolins I and II), and iii) to identify distinctchemotypes. For this purpose, an optimized ultrasound-assisted extraction has beendeveloped and applied, ensuring the optimal extraction conditions of pyrethrin [24].

The reason of the selection of indigenous Dalmatian pyrethrum populations for thepresent screening is based on its economic importance as a source of natural insecticide.Addressing the chemical diversity of natural Dalmatian pyrethrum populations will notonly contribute to conservation efforts of this indigenous plant species but will alsoprovide a solid foundation for future plant-breeding programs and agriculturalexploitation.

Results and Discussion. – The secondary metabolites of plants have historicallybeen an important source of insecticidal activity used for pest management with nonegative side-effects on human health, animals, and environment. Unfortunately, theadvent of numerous synthetic insecticides displaced their use. Synthetic insecticidesbrought new means of insect control, but also many risks with respect to human healthand environment quality, namely the pollution of water and air, the reduction of speciesdiversity and pest resistance [25] [26], and the alternation of food chains and energynutrient cycling patterns [27]. The concern about these possible adverse effects onceagain renewed the interest in naturally occurring plant compounds as alternativesources of pest control [28]. Research is now focused on the evaluation of differentplant chemicals, both novel and established, for the purpose of finding alternatives tosynthetic insecticides. Among them is a well-known natural insecticide, pyrethrin, theactive substance of Dalmatian pyrethrum.

Herein, we report the first detailed systematic evaluation of the chemical diversityof natural Dalmatian pyrethrum populations from Croatia. Geographical locations andbioclimatic data for the sampling sites of Dalmatian pyrethrum populations arepresented in Table 1 and Fig. 2. The populations were grown in a controlled field trial,and the flower samples were harvested at the same stage, when they were 3/4 opened,i.e., when they contained the highest amount of pyrethrin [10] [13] [14].

The evaluation of pyrethrin content and composition was performed by univariateanalysis of variance of six pyrethrin components, the total pyrethrin content, and

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)462

pyrethrin I/pyrethrin II ratio. The back-transformed least squares means of sixpyrethrin components (% of total pyrethrin), total pyrethrin content (% of the dryflower weight; DW), and pyrethrin I/pyrethrin II ratio are presented in Table 2. Totalpyrethrin content ranged between 0.36% DW for the population from the island Ugljan(L08) and up to 1.30% DW for the population from the island Lastovo (L19), with anaverage value of 0.86% DW. The value of pyrethrin content of the analyzed populationsis in accordance with the reports of other authors on pyrethrin content in naturalDalmatian pyrethrum populations [14] [23]. The highest pyrethrin I content wasrecorded for the population from Osor (L02; 67.46% of total pyrethrin), while thelowest value was recorded for the population from Premantura (L01; 31.02%), with anaverage value of 51.66%. Because of the different types of action of the pyrethrin I andpyrethrin II against insect species, the ratio between these two components is of utmostimportance for the quality of pyrethrum extract [29]. In analyzed populations,variations in quality of pyrethrins, determined as the pyrethrin I/pyrethrin II ratio, havealso been detected. The values ranged from 0.64 for the population from Premantura

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013) 463

Table 1. Geographical Locations and Bioclimatic Data for Dalmatian Pyrethrum Sampling Sites

No. Accession No.a) Population Latitude(N)b) [8]

Longitude(E)b) [8]

Altitude[m]

Annual meantemp. [8]c)

Annual precipi-tation [mm]c)

L01 MAP02137 Premantura 44.81 13.89 42 14.40 922L02 MAP02143 Osor 44.70 14.39 2 13.60 1107L03 MAP02158 Cres 44.96 14.41 72 14.60 1160L04 MAP02139 Mali Losinj 44.53 14.47 42 14.20 1037L05 MAP02156 Krk 45.08 14.67 24 14.00 1260L06 MAP02148 Gornja Klada 44.82 14.90 589 12.30 1245L07 MAP02138 Senj 44.93 14.92 6 13.60 1202L08 MAP02150 Ugljan 44.13 15.10 113 14.10 969L09 MAP02180 Pasman 44.00 15.29 17 14.20 935L10 MAP02171 Zlarin 43.70 15.84 45 15.30 853L11 MAP02166 Primosten 43.58 16.05 190 14.60 855L12 MAP02181 Vis 43.06 16.10 401 14.00 784L13 MAP02144 Ciovo 43.51 16.29 143 15.80 811L14 MAP02153 Solta 43.38 16.29 114 15.60 802L15 MAP02145 Kozjak 43.58 16.41 516 7.20 1126L16 MAP02155 Brac 43.38 16.52 116 16.30 824L17 MAP02173 Hvar 43.17 16.53 333 14.30 861L18 MAP02170 Omis 43.45 16.70 2 15.90 891L19 MAP02151 Lastovo 42.77 16.90 37 16.30 864L20 MAP02142 Kotiski Stanovi 43.32 17.06 1335 6.40 1132L21 MAP02140 La�ena 43.30 17.07 1295 6.50 1129L22 MAP02141 Ravna Vlaska 43.29 17.08 1228 8.30 1114L23 MAP02184 Peljesac 42.92 17.40 18 15.50 1139L24 MAP02152 Mljet 42.77 17.45 172 15.60 1117L25 MAP02146 Konavle 42.54 18.33 82 15.80 1304

a) Accession No from the Collection of Medicinal and Aromatic Plants, Zagreb, Croatia as available atthe CPGRD (http://cpgr.zsr.hr). b) N, North; E, East. Coordinates are in degree decimal format.c) Bioclimatic data for 25 sampling sites from the WorldClim-Global Climate Data, available atwww.worldclim.org. The data were extracted using DIVA-GIS ver. 5.2 (www.diva-gis.org).

(L01) up to 3.33 for the population from Osor (L02), and they were noticeably higherthan those reported by Maciver et al. [7] (0.90– 1.30) and Ambrozic Dolinsek et al.(1.45) [14], but very similar to the value of some breeding lines (0.50 – 3.30) [8]. Theanalysis of variance revealed highly significant (P<0.001) differences amongpopulations in all of the variables. The highest coefficient of variation was recordedfor jasmolin II (58.60), while the lowest was for pyrethrin I (17.47).

To determine the relationship between the total pyrethrin content and six pyrethrincomponents in the extracts, Pearson correlation coefficients were calculated. Thecorrelations between the percentage of pyrethrin I and the percentage of all of theother components were negative and significant except for jasmolin I, where thecorrelations were non-significant. On the other hand, total pyrethrin content was

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)464

Fig. 2. Geographical distribution of Dalmatian pyrethrum chemotypes. For populations L01–L25, seeTable 1.

positively correlated with the percentage of pyrethrin I and negatively correlated withcinerin II (Table 3).

On the basis of six pyrethrin components, 25 natural Dalmatian pyrethrumpopulations were further subjected to principal component analysis (PCA). The firstthree principal components (PCs) counted for 98.39% of the total variation (Table 4).The gap was observed between the second and third eigenvalue in the SCREE plot, andthe eigenvalues were greater than one for the first two principal components. The firstprincipal component (PC1) counting for 56.08% of the total variation separatedpopulations with a high percentage of pyrethrin I (the component that was highly andnegatively correlated with PC1) from populations characterized by a higher percentageof cinerin I, cinerin II, and jasmolin II (the components that were highly and positively

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013) 465

Table 2. Pyrethrin Components (in % of total pyrethrin) , Total Pyrethrin Content (in % of the dry flowerweight) , and Pyrethrin I/Pyrethrin II Ratio in 25 Dalmatian Pyrethrum Populations in Croatia

No. Population Pyrethrin components Totalpyrethrin

Pyr. I/pyr. IIa)Cinerin I Jasmolin I Pyrethrin I Cinerin II Jasmolin II Pyrethrin II

L01 Premantura 10.27 3.25 31.02 13.57 5.90 35.38 0.67 0.64L02 Osor 2.28 3.68 67.46 1.12 1.92 23.61 1.29 3.33L03 Cres 4.16 2.30 60.51 2.67 1.18 29.69 1.05 2.04L04 Mali Losinj 3.65 0.66 61.46 3.19 0.98 31.77 1.09 1.56L05 Krk 4.19 3.97 49.61 4.27 2.81 35.47 1.16 1.79L06 Gornja Klada 6.30 5.06 53.70 4.29 2.77 28.17 1.00 1.75L07 Senj 4.10 4.58 65.66 1.94 1.42 21.05 1.07 3.09L08 Ugljan 10.73 4.28 35.54 10.67 4.37 34.31 0.36 2.14L09 Pasman 6.38 4.72 60.13 3.66 2.70 21.79 0.84 1.16L10 Zlarin 6.35 0.89 54.53 4.91 0.86 31.05 1.17 2.18L11 Primosten 7.04 4.03 45.05 6.29 3.65 33.66 0.88 2.39L12 Vis 4.44 2.63 61.14 2.41 1.10 27.87 1.20 1.96L13 Ciovo 9.40 3.81 48.69 6.74 3.08 23.95 0.65 1.53L14 Solta 6.63 4.03 48.49 5.88 3.07 30.14 0.91 1.16L15 Kozjak 2.52 1.30 48.39 3.87 1.72 42.94 0.58 2.33L16 Brac 4.35 4.13 37.51 7.08 5.03 40.22 0.70 1.88L17 Hvar 7.60 6.88 51.81 4.82 3.87 23.31 0.62 2.34L18 Omis 8.43 7.46 49.63 6.13 3.99 23.10 0.92 1.61L19 Lastovo 3.80 4.28 56.61 2.70 2.27 29.20 1.30 2.18L20 Kotiski Stanovi 4.90 1.71 53.03 5.30 0.95 34.20 0.40 0.90L21 La�ena 9.59 1.69 46.48 8.50 1.23 32.41 0.66 1.39L22 Ravna Vlaska 3.71 1.96 53.40 3.69 1.26 35.26 0.53 1.99L23 Peljesac 7.21 5.76 41.48 6.85 5.36 32.19 0.62 2.72L24 Mljet 7.21 2.82 57.21 5.01 1.31 25.77 0.90 2.37L25 Konavle 5.42 1.56 52.98 4.58 1.52 32.97 0.87 1.18

Average 6.03 3.50 51.66 5.21 2.57 30.38 0.86 1.91CVb) 39.62 50.52 17.47 52.99 58.60 18.67 31.63 33.94Minimum 2.28 0.66 31.02 1.12 0.86 21.05 0.36 0.64Maximum 10.73 7.46 67.46 13.57 5.90 42.94 1.30 3.33

a) Pyr. I/pyr. II, pyrethrin I to pyrethrin II ratio. b) CV, Coefficient of variation.

correlated with PC1). The components highly correlated with the second axis werepyrethrin II (negatively) and jasmolin I (positively) (Fig. 3).

Cluster analysis (CA), using the first two PCs was applied to analyze therelationship among the populations and to group the populations according topyrethrin content. The highest values of both Pseudo F (PSF) statistics and CubicClustering Criterion (CCC) were obtained for five clusters; thus, the classification ofanalyzed Dalmatian pyrethrum populations into five chemotypes turned out to be theoptimal solution (Fig. 4). This is the first attempt at such a classification of naturalDalmatian pyrethrum populations.

Chemotype A was characterized with the highest percentage of pyrethrin I, and itincluded six populations: Osor (L02), Senj (L07), Cres (L03), Vis (L12), Lastovo(L19), and Mljet (L24). Populations from Mali Losinj (L04), Zlarin (L10), KotiskiStanovi (L20), Ravna Vlaska (L22), Kozjak (L15), and Konavle (L25) were assignedto chemotype B, characterized with a higher percentage of pyrethrin I and pyrethrin II,and a lower percentage of all the other components. The most abundant was chemotypeC. It included eight populations: Krk (L05), Gornja Klada (L06), Primosten (L11),Ciovo (L13), Solta (L14), Brac (L16), La�ena (L21), and Peljesac (L23), and wascharacterized by a lower percentage of pyrethrin I, and higher percentages of cinerin I,

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)466

Table 3. Pearson Correlation Coefficients between Six Pyrethrin Components (in % of total pyrethrin)and a Total Pyrethrin Content (in % of the dry flower weight)a)

Variable Jasmolin I Pyrethrin I Cinerin II Jasmolin II Pyrethrin II Total pyrethrin

Cinerin I 0.336 ns �0.614 ** 0.823 *** 0.503 * �0.152 ns �0.436 *Jasmolin I �0.190 ns 0.127 ns 0.724 *** �0.468 * �0.016 nsPyrethrin I �0.925 *** �0.734 *** �0.585 ** 0.629 **Cinerin II 0.617 ** 0.383 ns �0.634 **Jasmolin II 0.087 ns �0.312 nsPyrethrin II �0.392 ns

a) ns, Non-significant; *, significant at P< 0.05; **, significant at P< 0.01; ***, significant at P<0.001.

Table 4. Pearson Correlation Coefficients between Six Pyrethrin Components, and a Total PyrethrinContent and Scores of the First Three Principal Components

Traits Principal componentsa)

PC1 PC2 PC3

Cinerin I 0.793 *** 0.209 ns 0.565 **Jasmolin I 0.443 * 0.812 *** �0.341 nsPyrethrin I �0.937 *** 0.326 ns 0.106 nsCinerin II 0.932 *** �0.245 ns 0.252 nsJasmolin II 0.848 *** 0.299 ns �0.399 *Pyrethrin II 0.275 ns �0.891 *** �0.342 nsEigenvalue 3.365 1.752 0.786% of variance 56.08 29.20 13.10

a) ns, Non-siginificant; *, significant at P<0.05; **, significant at P<0.01; ***, significant at P< 0.00.

cinerin II, and jasmolin II). Populations from Pasman (L09), Hvar (L17), and Omis(L18) were grouped into chemotype D, characterized with a lower percentage of bothpyrethrins I and II, and a higher percentage of jasmolin I. Two populations, Premantura(L01) and Ugljan (L08), were grouped into chemotype E, characterized with a lowerpercentage of pyrethrins I and II, and a higher percentage of jasmolin II (Fig. 2).

Back-transformed least-squares means of six pyrethrin components (% of totalpyrethrin) and total pyrethrin content (% of the dry flower weight) in five chemotypesare compiled in Table 5. The analysis of variance revealed highly significant (P<0.001)differences among chemotypes in all of the variables. In five chemotypes, the totalpyrethrin content varied from 0.49 (% DW) for chemotype E to 1.13 (% DW) forchemotype A. The pyrethrin I/pyrethrin II ratio ranged from 0.86 for chemotype E to2.57 for chemotype A. On the basis of these results, it can be concluded that chemotypeA is the most promising for future breeding programs, due to both the highest pyrethrinI content and pyrethrin I/pyrethrin II ratio, the traits which are mostly responsible forthe bioactivity of the extract.

The obtained results revealed a high chemical variability among Dalmatianpyrethrum natural populations. The total pyrethrin content was not as high as incommercial varieties from main pyrethrum production regions where it was domes-ticated and systematically bred to achieve higher pyrethrin content and quality. Thebenefits, namely increased production, are the main reasons for initiating breeding

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013) 467

Fig. 3. Biplot of principal component analysis (PCA) based on six pyrethrin components in 25 pyrethrumpopulations form Croatia

programs, but, at the same time, continuous breeding efforts narrow the genetic base ofthe species, and increase genetic uniformity and consequently genetic vulnerability[30]. In East African countries, pyrethrum production is concentrated on small farmsand is entirely dependent on the availability of cheap labor for propagation andharvesting. Propagation is mainly conducted using vegetative transplants, and theexisting plants are split into multiple plantlets and then transplanted by hand [31].Vegetative propagation leads to genetic uniformity which can offer many advantages inagricultural practices, but over time clonal populations become more susceptible todiseases, pathogens, and unexpected environmental stresses due to exhaustion ofgenetic variability [32].

Contrary to cultivated material, the natural Dalmatian pyrethrum populationsmaintain a broad genetic variability, an important factor for the adaptability to diverseenvironmental conditions, and the resistance to diseases and pathogens. These plantgenetic resources could be used to expand the genetic base of the commercial varietiesand, therefore, alleviate their genetic vulnerability risk to biotic and biotic stresses, andensure their sustainability. Some attempts of broadening the genetic base of the Kenyanpyrethrum breeding populations has been reported [23]. The pollen of naturalDalmatian pyrethrum populations from Croatia was used to pollinate three selected

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)468

Fig. 4. UPGMA Dendrogram of cluster analysis on 25 Dalmatian pyrethrum populations form Croatiausing the first two principal components. The five major clusters, A–E, are indicated.

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013) 469

Tabl

e5.

Pyr

ethr

inC

ompo

nent

s(i

n%

ofto

talp

yret

hrin

),T

otal

Pyr

ethr

inC

onte

nt(i

n%

ofth

edr

yfl

ower

wei

ght)

,and

Pyr

ethr

inI/

Pyr

ethr

inII

Rat

ioin

Fiv

eC

hem

otyp

esof

Dal

mat

ian

Pyr

ethr

um

Che

mot

ype

na )

Pyr

ethr

inco

mpo

nent

sTo

tal

pyre

thri

nP

yr.I

/pyr

.IIb

)

Cin

erin

IJa

smol

inI

Pyr

ethr

inI

Cin

erin

IIJa

smol

inII

Pyr

ethr

inII

A6

4.24

bb)

3.31

b61

.60

a2.

51c

1.48

b26

.17

bc1.

13a

2.57

aB

64.

23b

1.39

c53

.94

ab4.

10bc

1.26

b34

.51

a0.

76ab

1.83

abC

86.

68ab

3.98

ab46

.24

b6.

16b

3.30

a31

.99

ab0.

82ab

1.53

bD

37.

49ab

6.29

a53

.90

ab4.

85b

3.47

a22

.63

c0.

78ab

2.43

aE

210

.64

a3.

85b

32.8

4c

12.1

1a

5.09

a35

.32

a0.

49b

0.86

b

a )n,

Num

ber

ofpo

pula

tion

sas

sign

edto

ach

emot

ype.

b)

Pyr

.I/

pyr.

II,

pyre

thri

nI

topy

reth

rin

IIra

tio.

b)

Val

ues

follo

wed

byth

esa

me

lett

erar

eno

tsi

gnif

ican

tly

diff

eren

tat

P<

0.05

.

clones, showing a high combining ability for flower yield and pyrethrin content. Theperformance of cultivars and crosses between the Kenyan varieties and the naturalDalmatian pyrethrum plants showed that selection on the Kenya pyrethrum was veryeffective and thereby significantly improved pyrethrum in Kenya.

Conclusions. – Considerable chemical variability has been detected, whereas fivechemotypes differing in insecticidal potential have been identified. Chemotype A,which was characterized with the highest percentage of pyrethrin I, as well as apyrethrin I and pyrethrin II ratio, appears to be the most promising for future plant-breeding programs. It can be concluded that natural Dalmatian pyrethrum populationsrepresent valuable resources of chemical variability, and there are highly favorableprospects for future plant-breeding programs for agricultural utilization.

This study was supported by the Ministry of Science, Education and Sports of the Republic of Croatia,within the framework of Projects Nos. 178-1191193-0212 and 125-1253008-1350. Special thanks to Ms.Valerie-Ann Cherneski for editorial help.

Experimental Part

Plant Material. Seed samples of 25 natural Dalmatian pyrethrum populations were collected alongthe Dalmatian coastal region and its islands. The geographical locations and bioclimatic data for thesampling sites are presented in Table 1. The seed samples are preserved ex situ as a part of the Medicinaland Aromatic Plants Collection at the Department of Seed Science and Technology, Faculty ofAgriculture, University of Zagreb and are available upon request (http://cpgrd.zsr.hr).

The field trial was set up at the experimental field of the Department of Seed Science andTechnology, Faculty of Agriculture, University of Zagreb. A row-column exper. design (5�5) with threereplications was used. During June 2009, pyrethrum flowers were harvested when ca. 3/4 of the rows ofdisc florets had opened. The flowers were air-dried to 10 to 12% moisture content in a dark, ventilatedroom. The dried flower heads were placed in sealed glass jars, and stored in a dark and cool place until theanalysis was performed. The plant material was pulverized prior to the extraction procedure. Each of thethree replications of 25 populations was included in the chemical analysis, corresponding to a total of 75samples.

Ultrasound-Assisted Extraction of Pyrethrin. Dry and finely ground pyrethrum flowers (0.25 g) wereextracted for 60 min at 508 with 5 ml of acetone in an ultrasonic bath Sonorex Digital 10 P (Bandelin, D-Berlin) with an ultrasound power of 1200 W and a frequency of 35 kHz. The extracts were filteredthrough a 0.45-mm nylon syringe filter and stored in a dark place at 48 until they were analyzed. Prior tothe chromatographic determination, 0.1 ml of internal standard 4’-metoxyflavanone (Sigma�Aldrich, D-Steinheim) was added to 1 ml of extract [24].

HPLC-DAD Analysis. The determination and quantification of extracted pyrethrin from pyrethrumflowers was performed by HPLC using a diode array detector (DAD). A Varian ProStar 500 (WalnutCreek, California, USA) equipped with a ProStar autosampler, ProStar 230 tertiary pump system,ProStar 330 DAD, and thermostatted column compartment was used. The chromatographic separationswere carried out with a ChromSep OmniSphere C18 column (5 mm particle size, 250�4.6 mm; Varian,NL-Middelburg).

To achieve a good separation, a two-component mobile phase consisting of MeCN and H2O with0.1% H3PO4 was used for gradient elution. The flow rate was kept constant at 1.4 ml/min. The detectorwavelength was set at 225 nm, and the sample injection volume was 20 ml. The pyrethrum technicalmixture Pestanal (Sigma�Aldrich, D-Steinheim) was used as a standard.

The pyrethrin components in the samples were identified by comparing the retention times ofstandard mixtures with those of peaks obtained in the analysis of pyrethrum extracts. The DAD providedUV spectra for each of the measured esters, and the peak identity was additionally confirmed by

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)470

comparing the UV spectra of peaks obtained from the pyrethrin extract and from the pyrethrin standard.The quantity of pyrethrin was expressed as mg of analyte per 1 g of flower sample.

Data Analysis. Statistical analysis of the data was performed using the SAS statistical software [33].The univariate analyses of variance using PROC MIXED in SAS were conducted for six pyrethrincomponents, total pyrethrin content, and the pyrethrin I/pyrethrin II ratio. Percentages were normalizedby arcsine transformation. Pearson correlation coefficients for six pyrethrin components and totalpyrethrin content in 25 populations were calculated using PROC CORR in SAS. The principalcomponent analysis (PCA) based on six pyrethrin components was performed using the PROCPRINCOMP procedure in SAS. The number of PCs was determined by checking the eigenvalues of thePCs (using the Kaiser criterion that retains components with eigenvalues greater than one and theSCREE plot) and the cumulative proportion of variance explained. The biplot was constructed by twoprincipal PCs showing populations and pyrethrin components (as vectors).

The standardized scores of the first two PCs were multiplied by the root of their eigenvalues and usedin cluster analysis (CA). The average linkage method (i.e., UPGMA) of PROC CLUSTER was applied,and, to determine the optimal number of clusters, Cubic Clustering Criterion (CCC) statistics and PseudoF (PSF) statistics were calculated. The populations were classified into groups representing distinctchemotypes.

The variance analysis using PROC MIXED was then repeated by introducing chemotype as a fixedeffect and the population nested within chemotypes as a random effect. Post hoc comparisons ofchemotype means were carried out using Tukey�Kramer�s test.

REFERENCES

[1] S. Jovetic, C. D. de Gooijer, Crit. Rev. Biochem. 1995, 15, 125.[2] D. G. Cochran, in �Pyrethrum Flowers: Production, Chemistry, Toxicology and Uses�, Eds. J. E.

Casida, G. B. Quistad, Oxford University Press, New York, 1995, p. 234.[3] G. D. Todd, D. Wohlers, M. Citra, �Toxicology Profile for Pyrethrins and Pyrethroids�, Department

of Health and Human Services, Agency for Toxic Substances and Disease Registry, Atlanta, 2003.[4] A. S. Gunasekara, �Environmental Fate of Pyrethrins�, Environmental Monitoring Branch, Depart-

ment of Pesticide Regulation, Sacramento, 2004.[5] R. M. Sawicki, E. M. Thain, J. Sci. Food Agric. 1962, 13, 292.[6] R. Winney, Pyrethrum Post 1979, 13, 132.[7] D. R. Maciver, in �Pyrethrum Flowers: Production, Chemistry, Toxicology and Uses�, Eds. J. E.

Casida, G. B. Quistad, Oxford University Press, New York, 1995, p. 108.[8] B. K. Bhat, in �Pyrethrum Flowers: Production, Chemistry, Toxicology and Uses�, Eds. J. E. Casida,

G. B. Quistad, Oxford University Press, New York, 1995.[9] J. G. Brewer, Euphytica 1974, 23, 45.

[10] S. W. Head, Pyrethrum Post 1966, 8, 32.[11] A. Glynne-Jones, �Pesticide Outlook-Biopesticides�, 2001, p. 195.[12] S. E. Morris, N. W. Davies, P. H. Brown, T. Groom, Ind. Crop. Prod. 2006, 23, 9.[13] V. A. Beckley, C. B. Gnadinger, F. Ireland, Ind. Eng. Chem. 1938, 30, 385.[14] J. Ambrozic Dolinsek, M. Kovac, J. Zel, M. Camloh, Ann. Ser. Hist. Nat. 2007, 17, 39.[15] P. Bakaric, Gospodarski List 2005, 17, 41.[16] A. R. Tuikong, Pyrethrum Post 1984, 15, 113.[17] S. E. Chandler, Pyrethrum Post 1948, 1, 10.[18] H. C. Thorpe, E. A. Agric. J. 1940, 265.[19] R. B. Contant, Proc. East Afr. Acad. 1963, 93.[20] S. P. Singh, J. R. Sharma, B. R. Rajeswara Rao, S. K. Sharma, Pyrethrum Post 1988, 17, 8.[21] J. M. G. Wainaina, in �Pyrethrum Flowers: Production, Chemistry, Toxicology and Uses�, Eds. J. E.

Casida, G. B. Quistad, Oxford University Press, New York, 1995, p. 50.[22] J. E. Casida, G. B. Quistad, �Pyrethrum Flowers: Production, Chemistry, Toxicology and Uses�,

Oxford University Press, New York, 1995.

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013) 471

[23] J. E. Parlevliet, J. G. Brewer, W. G. M. Ottaro, in �Broadening the Genetic Base of Crops�, Eds. A. C.Zeven, A. M. van Harten, Pudoc, Wageningen, 1979, p. 91.

[24] S. Babic, M. Grdisa, M. Perisa, D. Asperger, Z. Satovic, M. Kastelan-Macan, Agrochimica 2012,accepted.

[25] D. Pimentel, � Handbook of Energy Utilization in Agriculture�, Boca Raton (FL), CRC Press, 1980.[26] J. R. Coats, Ann. Rev. Entomol. 1994, 39, 489.[27] D. Pimentel, C. A. Edwards, Bioscience 1982, 32, 595.[28] J. H. Perkins, in � Handbook of Natural Pesticides: Methods. Vol. 1: Theory, Practice and Detection�,

Ed. N. B. Mandava, CRC, Boca, FL, 1985, p. 297.[29] B. Marongiu, A. Piras, S. Porcedda, E. Tuveri, S. Laconi, D. Deidda, A. Maxia, Nat. Prod. Res. 2009,

23, 190.[30] S. Rauf, J. A. Teixeira da Silva, A. A. Khan, A. Naveed, Int. J. Plant Breeding 2010, 4, 1.[31] M. Greenhill, Chron. Hortic. 2007, 47, 5.[32] S. A. Lei, J. Arizona-Nevada Acad. Sci. 2010, 42, 9.[33] SAS/STAT � 9.1 User�s Guide, SAS Institute Inc., Cary, NC, 2004.

Received January 13, 2012

CHEMISTRY & BIODIVERSITY – Vol. 10 (2013)472