IV RESULTS

The following chapter is structured according to the themes of the original publications (I-VI) as indicated in the chapter Materials and Methods.

1 Biodiversity of tansy

1.1 Morphological variation (I, II)

The mean height of tansy genotypes varied from 60.5 cm (Tv 19) to 115.4 cm (Tv 15), the mean number of nodes from 14.6 (Tv 14) to 26.7 (Tv 19), the mean number of flower heads per stem from 17.6 (Tv 19) to 79.8 (Tv 16), and the mean length of the corymb from 4.9 cm (Tv 19) to 24.8 cm (Tv 16). The date of onset of flowering varied from the 20^th of July (Tv 3 and Tv 15) to the 4^th of August (Tv 12). Significant differences were found among the tansy genotypes in all of these traits (P=0.0001; for the date of flowering P=0.05). Pearson correlation coefficients between genotype's height and number of flower heads, and height of corymbs were 0.53, and 0.48 (P=0.0001), respectively. Pearson correlation coefficient between the number of flower heads and height of corymb, and date of beginning of flowering were 0.78 (P=0.0001), and 0.35 (P=0.1), respectively.

Principal components 1 and 2 of tansy genotypes based on the morphological data are illustrated in paper I, Fig. 2. Out of the 9 genotypes belonging to group I, 6 were positioned below the principal component 1-axis, whereas 8 out of the 11 genotypes of group II were above the principal component 1-axis.

1.2 Genetic variation (I)

The overall mean 2 C value of nuclear DNA content for tansy genotypes was 8.86 pg. The means of the different genotypes ranged from 7.84 pg (Tv 14) to 9.95 pg (Tv 4) (I, Table 4) and differed significantly (P=0.05).

The genetic distance of the tansy genotypes was analysed by RAPD-PCR. The smallest genetic distance observed was 0.153, between the genotypes Tv 18 and Tv 19, whereas the widest genetic distance, 0.476, was observed between the genotypes Tv 8 and Tv 16 (I, Table 5). The mean genetic distance was 0.294 among all of the 20 tansy genotypes included in this study.

Complete linkage cluster analysis initially separated the 20 genotypes to two groups (group I and II) that were further divided into six smaller sub-groups (I, Fig. 5). The widest and shortest normalised maximum distances were 1.62 (1.62 x 0.294=0.476) and 0.52 (0.52 x 0.294 = 0.153) between group I and II, and between genotypes Tv 18 and Tv 19, respectively.

The first and second principal components are illustrated in Fig. 6 (I), where most of the genotypes that belonged to group I and group II were positioned below and above the principal component -1 axis, respectively (I, Fig. 6).

Most genotypes in the same subgroup based on complete linkage analysis of RAPD patterns (I, Fig. 5) originated from close by geographical locations (I, Fig. 1). Seven out of the 9 genotypes of group I and 10 out of the 11 genotypes of group II originated south and north of latitude 60° 3'N, respectively (I, Table 1).

According to the S-N-K test, the mean number of nodes per stem, the mean number of flowers, the mean length of the corymb, and the date of flowering differed significantly between groups I and II (P=0.05) based on the data from RAPD patterns (II, Table 3).

To explore the potential association of RAPDs and moprhological data the following two calculations were carried out. Firstly, the means of principal components 1 and 2 of the genotypes based on RAPD patterns and belonging to group I and II, respectively, were calculated and plotted to a scatter diagram (I, Fig. 3). The data based on genetic differences positioned the groups nearly symmetrically to the origo: group I below the principal-1 axis, whereas group II was positioned above the principal-1 axis. The means of the principal components 1, 2, 3, and 4 differed significantly between groups I and II (P=0.05), which means that the position of the groups differed also significantly from each other. Distant position of the groups based on RAPD profiles indicated genetic differences. Secondly, the means of principal components 1 and 2 of the morphological data of tansy genotypes belonging to group I and II, respectively, were calculated and plotted on a scatter diagram. Similar to the data from RAPD patterns, the position of the groups was nearly symmetrical to the origo (I, Fig. 3). The t-test indicated that the means of the principal components 1, 2, 3, and 4 for group I and II differed significantly from each other which means that the position of the groups differed. This suggests that morphologically these two groups may be different. The position of groups (I and II) based on RAPD and morphological data on the same scatter diagram was distinct. The principal coordinates for group I based on RAPD and moprhological data were positioned opposite in relation to the origo as compared to the position for group II based on RAPD and morphological data.

1.3 Variation in volatile compounds (I, II)

A total of 55 aromatic volatile compounds were detected from the petroleum-ether extraction of dried flower heads of tansy. The concentrations of 47 of the 55 compounds detected varied highly significantly between the tansy genotypes (P<0.0001) and only the genotypic variation in artemiseole, trans sabinen hydrate, nerol, nerolidol, spathulenol, caryophyllene oxide and two unidentified compounds was not significant (II, Table 1).

Fifteen of the genotypes had a main chemical component, which consisted of at least 40% of the total peak area of the volatile compounds, and five genotypes contained at least two terpenes as main components. The two first principal components clustered the 20 genotypes into eight groups when plotted in the scatter diagram. Four of the groups contained more than 19.0% of camphor, and the other four groups less than 4.2% of camphor. The pure camphor chemotypes (>65.2%) formed the largest group. In two groups, camphor was detected in significant concentrations (35.7-50.8%) but in association with other compounds, the concentration of which exceeded 5% such as a-pinene, camphene, 1,8-cineole, borneol, pinocamphone, chrysanthenyl acetate, bornyl acetate, isobornyl acetate, and germagrene D. A group containing a high concentration of 1,8-cineole (47.1%) and a lower concentration of camphor (29.4%) was also identified. Four other groups consisted of genotypes which contained either artemisia ketone, ß-thujone, davanone, or tricyclene + myrcene as a major compound with or without artemisyl acetate and umbellulone (II; Table 1, Figs 1 and 2).

On the basis of absence or presence of a compound, complete linkage cluster analysis divided the genotypes into two groups. These were further divided into smaller groups which differed from each other more than on average tansy genotypes in this study (II, Fig.3). Some similarities on the grouping of the genotypes in the two analysis methods used could be observed. The two most distant groups (Tv 7, 8, 10 & 20 and Tv 14 & 16) based on the analysis of the absence or presence of the compounds were also distinguished from the other chemotypes in the principal component analysis (PCA). This was especially clear when principal components 1 and 2 based on the concentration of volatile compounds were plotted in a scatter-gram. In addition, the four small groups formed by the complete linkage cluster analysis (Tv 14 & 16; Tv 9 & 13; Tv 4 & 18; Tv 8 & 19) (II, Fig. 3) positioned close to each other also in the PCA (II, Fig. 1).

Six of the seven chemotypes which did not contain camphor as the main component originated from Southern Finland, and only the chemotype containing thujone was collected from Eastern Finland. Eight of the 13 chemotypes with camphor concentration exceeding 18.5% originated from Central Finland and five from Southern Finland. Artemisia ketone was found only from genotypes originating from Southwestern Finland.

The eight groups (II; Fig. 1 and Table 2) resulting from the principal component analysis of volatile compounds, could be joined to two clusters based on camphor concentration. The cluster consisting of groups the camphor concentration of which was between 1.4-4.2% had less nodes per stem, more flower heads and a taller corymb, than the other cluster consisting of groups with camphor concentration more than 19.0%.

Distance matrices of volatile compounds were calculated by the same method as the genetic dictance matrices based on RAPD-PCR patterns. Pair-wise comparisons of the matrices showed that Pearson correlation coefficient, 0.407 (P<0.0001), had a 41% analogy between relative genetic and chemical differences between the 20 tansy geno- and chemotypes studies.

The 20 tansy genotypes were arranged according to the groups formed as the result of the principal component analysis of the composition and concentration of volatile compounds. The variation of morphology was compared between the groups using the SAS CONTRAST procedure. The group containing 1,8-cineole had the tallest shoots and differed significantly from the others (P<0.01 or P<0.0001) whereas the two groups containing mixed chemotypes had the shortest shoots. The group having the shortest shoots had the highest number of nodes in the stems, the number of which differed significantly (P<0.05, and P<0.0001) from the number of nodes in the other groups. The number of flower heads and the height of the corymb in groups containing davanone or artemisia ketone were the highest, and the results differed significantly (P<0.05, number of flower heads; P<0.01, height of the corymb) from the other groups. The mixed group containing 1,8-cineole-camphor-chrysanthenyl acetate had the lowest number of flower heads per stem differed significantly (P<0.05) from the other groups, except for the group containing thujone. Flowering occurred significantly earlier (P<0.05) in the group containing thujone compared to the groups containing davanone or artemisia ketone. The difference in flowering time was significant (P<0.05) between the first and last flowering groups.

2 In vitro shoot tip culture for tansy and pyrethrum

2.1 Establishment of shoot cultures in vitro (IV)

In vitro cultures were successfully established from shoot tips using MS medium supplemented with low NAA (0.54 - 1.07 µ M) and BAP (0.44 - 0.89 µ M) concentrations. Strong apical growth of the main shoots under low illumination could be suppressed by increasing the concentration of BAP. Embryos excised from surface sterilised flower heads germinated and developed faster into plantlets compared to plantlet development from seeds.

2.2 Role of BAP in micropropagation of tansy (III)

The effect of BAP (2.66 µM) in the MS medium on micropropagation efficiency in tansy was studied simultaneously with the effect of antibiotics on the growth of in vitro cultured tansy (III). The initiation of shoots occurred 6.9 and 8.3 days after subculture for plants grown without and with BAP (2.66 µM), respectively. The number of shoots increased from 1.1 to 7.8 per explant in parallel with an increase in growth rate from 4.3 to 44.8 when BAP was added to the growth medium. Of the three genotypes, only Tv 5 produced roots consistently and thus the rooting data from only that genotype could be analysed. Plants growing in MS medium containing BAP initiated root growth significantly later than plants growing without growth regulators. The number of roots and the percentage of roots decreased from 10.1 to 4.0 and from 100% to 29.2%, respectively, when the medium was supplemented with BAP. The final pH in media containing BAP was significantly higher than pH in media without BAP.

2.3 Culture of in vitro plantlets for protoplast isolation (V)

Culture conditions had a significant effect on the yield of isolated protoplasts. The growth of tansy plants differed depending on the light intensity during the in vitro culture. In high light intensity (80-200 µM m^-2 s^-1) the growth of tansy was slow, the anthocyanin coloured leaves were thick, and only a few shoots were produced even in the BAP containing micropropagation medium. In contrast, leaves of tansy grown in low light intensity (20-40 µM m^-2 s^-1) grew larger and no anthocyanin could be detected. The yield of protoplasts isolated from donor tissue cultured under 20-40 µM m^-1 s^-1 was the highest (V; Table 1, condition 5). Tv 14 and Tv 142094 were the most frequently used genotypes for protoplast isolations due to their adventious shoot growth. Also, medium without BAP and with or without a low concentration of NAA was used for the growth of visually 'normal'-looking plantlets yielding a high number of protoplasts. The effect of other culture conditions of in vitro grown plantlets for the yield of protoplasts is summarised in Table 1 (V).

2.4 Problems of endophytic bacteria in in vitro cultures of tansy (III)

Occasional contamination, notably caused by bacteria, became visible around roots of in vitro plantlets of tansy about 3-5 months after transferring them to tissue culture. This problem led us to study the possibility of endophytic bacteria, and the eradication of these from in vitro cultures with antibiotics, and to observe the possible alterations in the growth of tansy caused by the antibiotics.

2.4.1 Characterization of bacteria

All isolated bacteria were Gram-negative. Five isolates were fermentative, lacked the cytochrome C oxidase activity and belonged to Enterobacteriaceae. Three isolates showed respiratory metabolism and belonged to the fluorescent Pseudomonas. Two aerobic isolates belonged to neither of these groups (III; Table 1).

2.4.2 Sensitivity of bacteria to antibiotics

No antibiotic alone was effective against all of the bacteria isolated from in vitro cultures of tansy (III; Table 2), but the growth of most, but not all, bacteria was prevented by rifampicin and gentamicin even at their lowest concentrations tested. Combinations of rifampicin and gentamicin at the lowest concentrations in the growth medium prevented the growth of all bacteria. Cefotaxime efficiently inhibited the growth of Enterobacteriaceae. In combination with gentamicin, cefotaxime was effective against all of the bacteria except for no.13. Streptomycin was only modestly effective and ampicillin did not effectively control any strain at the tested concentrations (III; Table 2). Thus, only gentamicin, rifampicin and cefotaxime were further tested for their effects on plant growth.

Increased concentrations of gentamicin, cefotaxime, and rifampicin delayed the initiation of shoot growth linearly from 8.3 days to 19.0, 22.1, and 20.5, respectively (III; Fig. 2a). Also, a linear genotype x gentamicin interaction was observed (III; Fig. 2b). When the medium contained the combination of gentamicin + cefotaxime or gentamicin + rifampicin, the shoots started to grow 21.7 days or 19.1 days after subculture, respectively. The number of shoots decreased linearly from 7.8 with increasing concentrations of gentamicin to 2.0, with cefotaxime to 2.2, and with rifampicin to 2.3 (III; Figs 1 and 3). The number of shoots was higher in the combination of gentamicin+rifampicin than in the combination of gentamicin+cefotaxime (III; Fig. 1).

The growth rate of shoots was 44.8 in media supplemented with only BAP. Increased antibiotic concentration decreased the growth rate of shoots (III; Figs 1 and 4). Increase of gentamicin concentration decreased the growth rate from 44.8 to 9.5 (III; Fig. 4), and the reduction was rate-quadratic. Increasing concentrations of cefotaxime and rifampicin decreased the growth rate linearly from 44.8 to 16.1, and to 23.3, respectively. A genotype x gentamicin-quadratic interaction was observed (III; Fig. 4). Gentamicin combined with rifampicin resulted in growth rate which was significantly higher than the growth rate when a combination of gentamicin and cefotaxime was used (III; Fig. 1). Gentamicin, cefotaxime, and rifampicin reduced the shoot height linearly from 4.7 cm to 1.4 cm, 1.8 cm, and 2.2 cm, respectively (III; Figs 1 and 5). A genotype x cefotaxime-linear interaction was observed. The genotypes exhibited different quadratic responses to differing antibiotic concentrations (III; Fig. 5). Shoots in the medium containing the combination of gentamicin and rifampicin were longer than those in the medium containing gentamicin and cefotaxime (III; Fig. 1).

Of the three genotypes, only Tv 5 produced roots consistently in most of the treatments, and therefore only Tv 5 could be analysed with analysis of variance. The number of roots per explant and the percentage of roots was 4.0 and 29.2%, respectively, in the medium supplemented with BAP. Increased concentrations of cefotaxime delayed the root initiation linearly from 24.0 days to 32.5 days. Root length had a quadratic response to gentamicin. Increased concentrations of rifampicin increased the percentage of rooted plants from 29.2% to 83.4% (III; Figs 1 and 6). The pairwise comparisons (t-test) of Tv 5 vs Tv 6 and Tv 5 vs Tv 142094 showed that the percentage of rooted plants of Tv 5 was significantly higher than in Tv 6 or Tv 142094 when plants were cultured in media containing gentamicin, cefotaxime or rifampicin. The root growth of Tv 6 was totally inhibited following treatment with gentamicin (III; Fig. 6).

In a few cases, the morphology of the explants was affected by treatment with antibiotics compared to controls (III; Fig. 1). Leaves were dark green at low concentrations of gentamicin and cefotaxime. Leaves turned curly and sometimes thickened becoming chlorotic at high concentration of gentamicin, cefotaxime, and combinations of gentamicin with cefotaxime or rifampicin. Higher concentrations of rifampicin increased the frequency of hyperhydricity of shoots.

Final pH of the media differed significantly between the tansy genotypes Tv 5, Tv 6 and Tv 142094 after 35 days of culture (III; Fig. 7). Final pH of the media exhibited a quadratic response to concentration of gentamicin and rifampicin, with differences among the genotypes (III; Fig. 7).

The linear regression coefficient of gentamicin differed highly significantly (P<0.001) from the linear regression curves of cefotaxime and rifampicin for initiation of shoot growth, number of shoots, and height of shoots as tested by t-test. The linear regression coefficient of cefotaxime and rifampicin differed significantly (P<0.05) from each other in the initiation of shoot growth, and shoot number.

3 Callus formation and regeneration from explants of tansy (IV, V)

3.1 Callus and suspension culture (V)

For several months the suspension culture derived from Tv 4 leaf callus contained globular structures and very few single cells. Culture medium with NAA and BAP promoted the formation of green and globular cell structures whereas 2,4-D resulted in production of light and small cell aggregates. Gradually the size of the globular structures decreased, the colour lightened and a suspension culture of small 4-6-cell aggregates was established. Small callus aggregates grown in suspension cultures were used for protoplast isolation.

3.2 Callus formation and regeneration from leaf explants (IV)

Dark green and compact callus was produced from the leaf and stem explants, the latter which produced also shoots. Callus formation, but not shoot or root regeneration was enhanced by the addition of CH to the medium. The first signs of callus formation were visible on explants grown in the dark on high NAA (16.11 - 24 17 µM) and BAP (13.32 - 19.98 µM) concentrations in two weeks. In cultures grown in the light callus was observed one to two weeks later than in the dark grown explants and the strong proliferation of callus growth that was initiated in darkness was visible after several subcultures. Without NAA the callus formation from leaf and petiole explants, and root and shoot regeneration from callus was very poor although the medium was supplemented with different concentrations of BAP. As the medium contained various combinations of both NAA and BAP the morphogenesis was enhanced. The best response was observed with the high concentration of NAA (16.11 - 24 17 µM) whereas the concentration of BAP varied (6.66 - 19.98 µM ) in the medium (IV; Table 3). Generally, high concentrations of both NAA and BAP increased the callus formation from the explants in the MS medium.

The best shoot regeneration was obtained on the highest concentration of NAA (24.17 µM) and BAP (19.98 µM ) in the MS medium, whereas the highest frequency of root regeneration occurred on the MS medium containing 8.06 µM NAA only (III; Table 3). Most of the regenerated shoots of tansy were hyperhydric. However, also vigorous shoots which had normal appearance and produced roots after transfer to hormone free MS medium were obtained.

4 Protoplast technique for tansy and pyrethrum

4.1 Culture conditions for protoplast isolation (V)

The effects of culture conditions on the yield of protoplasts from tansy and pyrethrum are summarised in paper V, Table 1. Culture conditions had a significant effect on the yield of isolated protoplasts. The yield of protoplasts isolated from donor tissue cultured under 20-40 µM m^-1 s^-1 and incubated at 29± 1° C in the low enzyme concentration (V; Table 1, condition 5) and supplemented with sucrose (+ a low concentration of CaCl₂) was the highest (V; Table 1, condition 5). In tansy, this combination of culturing conditions differed significantly from conditions 1 (P<0.05), 2 (P<0.001), and 4 (P<0.05) and in pyrethrum it differed significantly from conditions 1 (P<0.001) and 2 (P<0.05) (V; Table 1.). Overall, pyrethrum yielded less protoplasts than tansy. Visually, leaf tissue producing a low yield of protoplasts was usually brown after enzyme incubation. Ascorbic acid added to the enzyme solution decreased the browning but also decreased the yield of protoplasts isolated. Preconditioning the leaves for 2 days on MS medium decreased the browning but did not affect the viability or the protoplast yield. If preconditioned longer, viability and yield decreased.

4.2 Effect of season on protoplast isolation (V)

The percent of successful protoplast isolations and the mean protoplast yields per season are illustrated in paper V, Fig 1. The Pearson correlation coefficient between the mean percent of successful protoplast isolations and the mean yield of protoplasts of the experiments carried out in autumn (September-November), winter (December-February), spring (March-May) or in summer (June-August) was 0.98 for tansy and 0.53 for pyrethrum.

Protoplast isolations performed during the period from December to April yielded the highest number of protoplasts whereas the division of tansy protoplasts and the formation of callus were most successful in February, March, and April. During that time in total 32 successful protoplast isolations from tansy were performed of which 12 isolations from Tv 142094 and 3 isolations from Tv 50 resulted in callus formation. Up to 1.0% of the isolated protoplasts formed callus and more than 3000 calli per isolation could be obtained. Pyrethrum formed microcolonies which did not develop further.

4.3 Protoplast isolation (IV, V)

In total 488 protoplast isolations were performed of which 309 were from tansy, 158 from pyrethrum, and 15 from the tansy cell suspension culture. Of these isolations 187 (60.9%), 96 (58.5%), and 11 (73.3%) yielded protoplasts from tansy, pyrethrum and tansy cell suspension culture, respectively. The mean yields of the protoplasts isolated from tansy, pyrethrum, and the suspension culture were 6.0 x 10⁶, 3.0 x 10⁶, and 7.5 x 10⁴ g^-1 FW, respectively. The mean protoplast yield of tansy genotypes ranged from 2.5 x 10⁶ (Tv 16) to 10.5 x 10⁶ (Tv 11) g^-1 FW and among the pyrethrum lines from 1.0 x 10⁶ (Tc 24) to 5.0 x 10⁶ (Tc 18) g^-1 FW. From the 309 protoplast isolations carried out from tansy 232 consisted of genotypes Tv 5, 8, 14, or 142094 and from the 158 isolations from pyrethrum, 110 consisted from lines Tc 18, 21, or 22 (V).

The diameter of protoplasts ranged between 20-60µM (IV; Fig.1a). Some of the protoplasts contained anthocyanin pigments. The viability of protoplasts increased from 60% up to 95% when relative centrifugal force was reduced from 100 x g to 70 x g during the first spin and to 41 x g during the subsequent three centrifugations during washing of the protoplasts (IV).

4.4 Effect of media components on protoplast culture (IV, V)

Accumulation of brown pigmentation was frequently visible in protoplasts after 4-5 days of culture. However, dilution of the cell mixture decreased the intensity of the pigmentation and improved the growth of the cultures. Later, when protoplasts were plated in agarose, weekly replacement of the surrounding media also hindered the growth reduction by the brown metabolites produced by the cultured cells. Protoplasts grew well and stayed viable in the medium supplemented with 90 g l^-1 of glucose and a low concentration of ammonium nitrogen as the N-source (IV).

Protoplast viability was usually high (90%->) when a high yield of protoplasts (>5 x 10⁶ g^-1 FW) was obtained. In terms of the antibiotics studied the least reduction in viability was observed during the two week culture period in which rifampicin (12.5-25 mg l^-1), gentamicin (25-50 mg l^-1), or cefotaxime (100 mg l^-1) was added to the medium. At the end of the experiment the viability was the highest when the cells had been cultured in medium supplemented with gentamicin (25 mg l^-1) or rifampicin (12.5 mgl^-1) (V; Fig. 2) The increase in acidity of the culture medium from pH 5.8 to pH 3.5 - 4 during the first few days after protoplast isolation was prevented by adding 5 mM MES to the medium (V).

The growth hormones NAA and BAP were essential for protoplast division. High concentration of NAA (26.85 µM) promoted cell division with a wide range of zeatin (4.56-45.62 µM) or BAP (4.44 - 44.40 µM). Either symmetrical or asymmetrical protoplast division was observed. Without NAA, BAP or zeatin the protoplasts did not divide, and the division ceased if the protoplast medium was not solidified (V).

4.5 Effect of media components on protoplast-derived callus growth of tansy (IV, V)

In semi-solid agarose (0.35%) loose colonies continued to form microcalli, but in solid agarose (0.6%) the division was very slow or it discontinued (V). Protoplast-derived calli, 1-2 mm in diameter, were transferred from agarose plating medium to fresh medium for further proliferation, usually 2-3 months after isolation (IV; Fig. 1d). Most of the calli reached a diameter of 5-10 mm in 3-4 months after isolation, after which they were transferred to shoot regeneration medium (IV). Protoplast-derived calli grew rapidly in medium containing glucose or sucrose, but grew very slowly and turned brown on growth medium containing mannitol. Silver nitrate (>20 µM) in the growth medium decreased callus growth (V).

Of the various hormone combinations shown in paper IV, Table 2, modified MS medium (MSIII) with BAP (13.32 µM) + NAA (5.37 µM) resulted in production of green and compact callus with nodules (IV; Fig 1f). BAP (8.88 µM) + GA₃ (0.29 µM) in the MS medium resulted in many green nodules with some specialised arrangements on the surface. Production of anthocyanin was observed also in calli. Different types of callus growth were observed after 3-4 months of culture depending on the combination of hormones as shown in paper IV, Table 2. White, unorganised, and friable callus was induced on media, in which the concentration of auxins exceeded cytokinins. Green, compact, and nodulating callus with meristematic zones was produced on media, in which the concentration of cytokinins exceeded auxins, whereas white, friable and pre-embryogenic calli (IV; Fig. 1e) were observed in association with both of the above mentioned callus types (IV; Table 2). Protoplast-derived calli were subcultured 6-7 times, however, no shoot regeneration occurred.

Calli grown without hormones grew slowly, and no globular or meristematic structures formed. Both auxins and cytokinins had a significant effect on protoplast-derived callus growth. The callus growth was most abundant with NAA (1.07-17.18 µM) and zeatin (0.91-14.60 µM) (V; Tables 2 and 3). Formation of meristematic structures differed significantly only between the auxin treatments. The calli were visually scored 'the most meristematic' when NAA or IBA of the auxins were combined with BAP or zeatin (V; Tables 2 and 4). Formation of meristematic structures was observed especially when IBA (0.98-7.87 µM) or NAA (1.07-8.59 µM) were used in combination with BAP (0.89 - 14.21 µM).

At the end of the 30-day experiment the greenish and most compact protoplast-derived calli were further cultured up to 20 months subculturing every 30-40 days. The calli cultured on MS medium supplemented with BAP (56.84 µM) and/or NAA (68.74 µM) produced green, globular and meristematic formations whereas the calli cultured on high concentrations of BAP (up to 113.68 µM) and/or NAA (up to 137.48 µM) turned brown.

4.6 Sensitivity of protoplast-derived calli of tansy to antibiotics (V)

The increased concentrations of gentamicin decreased the growth rate of calli from 2.21 to 1.15 (V; Fig. 4). When cefotaxime and rifampicin were used alone they did not have an affect on the growth rate. When gentamicin and rifampicin were used in combination, a significant interaction was observed. The increased concentration of rifampicin from 12.5 to 25 mg l^-1 increased the growth rate of callus from 1.63 to 2.26 (P<0.05) at the basal rate of 25 mg l^-1 of gentamicin. The increased concentration of gentamicin from 25 to 50 mgl^-1 decreased the growth rate of calli from 2.6 to 1.53 (P<0.05) at the rate of 25 mg l^-1 of rifampicin.

4.7 Regeneration from protoplast-derived callus of tansy (IV, V)

Cell differentiation with root formation was observed on callus cultured on the MS medium with BAP (13.32 µM) + NAA (10.74 µM) supplemented with CH 250 mg l^-1 (IV; Fig. 1g). Spontaneous shoot formation occurred from protoplast derived callus of tansy cultured on MS medium supplemented with 2,4-D (3.62 µM) and zeatin (1.82 µM) (V; Fig. 3). The growth of the about 5 mm long shoots ended due to contamination. None of the other attempts resulted in shoot formation even if globular structures with trichomes, nodules and anthocyanin pigmentation were frequently observed on media with several hormone concentrations (V).

4.8 Protoplast fusion and characterization of the fusion calli (VI)

In total 48 chemical fusion experiments with PEG between tansy and pyrethrum protoplasts were carried out, of which three resulted in callus formation. The volume of the fusion solution was found critical for the fusion procedure. The optimum volume of the PEG solution was 1.5 x the volume of the protoplasts. If the volume was less (1x) no fusion occurred, and if it was higher (2x) the protoplasts collapsed. Protoplast viability before and after the fusion was 90-95% and 30-50%, respectively.

4.8.1 Culture of fusion-derived callus

Fused tansy and pyrethrum protoplasts [F46; (Tv 50 + Tv 142094) + Tc 22] began to divide in 2 days, whereas the protoplasts of the intraspecific tansy fusion (F43; Tv 50 + Tv 142094) started to divide 5 days after protoplast fusion. Two weeks after the fusion small microcolonies containing 30-50 cells were observed in F46, whereas the cells of the F43 had reached only the 5 cell stage. Calli formed from fusion F46 were 1 mm diameter in 20 days after fusion compared to F43 which needed twice that time to reach the same diameter. No divisions were observed in the control plates of tansy and pyrethrum protoplasts.

In total, 5 calli were obtained from the F43 fusion and 12 from the F46 fusion. Calli derived from F43 grew slowly and were friable, unorganised, albino or yellowish, and coloured with anthocyanin. In 6-10 months the first signs of organization such as globular structures were observed. Callus growth was better or similar in MS medium supplemented with 30 or 40 g l^-1 glucose instead of sucrose. Calli derived from the fusion F46 grew fast and were compact, green and formed globular and meristematic structures. Clear differences in the growth rate and structure between the calli clones derived from the fusion were observed during the 20 months of culture.

4.8.2 Nuclear DNA content

The nuclear DNA content of Tv 142094, Tv 50, protoplast-derived calli of Tv 142094 and pyrethrum were 6.41, 7.39, 8.09-8.15, and 13.16-14.76 pg, respectively, whereas the DNA content of five fusion-derived calli (F43A, B, C and F46A, B) ranged from 8.84 to 31.87 pg (VI; Table 1).

4.8.3 RAPD patters

The DNA yields from fresh calli and leaves were 0.7 - 0.8 µg of total DNA per 100 mg, respectively. Four RAPD primers were used, which produced a total of 56 reproducible bands (350 bp - 2350 bp) (VI; Fig. 1) of which 4 were common for all samples and 52 polymorphic. The genetic distance was the widest (0.714 and 0.762) between pyrethrum and the two tansy genotypes (Tv 142094 and Tv 50), and the smallest (0.031) between the two callus clones derived from the fusion F46 (F46B vs F46C). Fusion callus F46B [(Tv 142094 + Tv 50) + Tc 22] was similar to Tv 50 and to Tc 22 (0.472), but the artificial mix of tansy+pyrethrum DNA had more similarity to pyrethrum (distance 0.245 and 0.292) than to the tansy RAPD pattern (distance 0.373 and 0.417). Fusion callus F43A (Tv 50 + Tv 142094) had a higher similarity to Tv 50 (distance 0.269) than to Tv 142094 (distance 0.519) (VI; Table 2.). Average linkage cluster analysis separated the samples into two groups. Both of the tansy genotypes and the fusion F43A belonged to one group whereas pyrethrum, fusion F46B, F46C, and artificially mixed DNA from tansy and pyrethrum belonged to the other group (VI; Fig. 2).

4.8.4 Analysis of volatile compounds

Kovat's indices (K_I) (Retention indices, R_I), authentic compounds and mass spectral library were used to identify the composition of the dichloromethane extracts. In total 36 volatile organic compounds (K_I 1800) were quantified, of which 24 were either identified or tentatively identified. No compounds common for fusion-derived calli (F43A, F46B) and tansy (Tv 142094) or pyrethrum (Tc 22) were observed. Decadienal, artedouglasia oxide, syringaldehyde, heptadecane and coniferyl alcohol were identified only from protoplast-fusion derived calli (F43A, F46B). Coniferyl alcohol was found also from protoplast-derived callus of Tv 142094. Beta farnesene and germacrene D were volatiles present in both tansy (Tv 142094) and pyrethrum (Tc 22) (VI; Table 3).

Among the less volatile compounds (oven temperature 164-260 ° C) hexadecanoic and linoleic acid were present in both the protoplast fusion-derived calli (F43A, F46B) and tansy (Tv 142094). One unidentified compound (K_I 2264) was detected in the fusion callus (F46B) and the pyrethrum (Tc 22). In addition, several compounds were observed only from protoplast fusion-derived calli (F43A, F46B). Four peaks with mass spectra corresponding to the mass spectra of authentic pyrethrins I (K_I 2281 and 2321) and II (K_I 2264 and 2294) were detected from pyrethrum (Tc 22) extracts. Peaks with mass spectra corresponding to pyrethrins were not observed in the protoplast fusion-derived calli.