Cytokinins in the ball gall of Solidago altissima and in the gall forming larvae of Eurosta solidaginis

Authors


Abstract

Summary

  •  The identities and concentrations of cytokinins in goldenrod ball galls, normal goldenrod stems, and in the gall-forming larvae of the tephritid fly Eurosta solidaginis were determined in order to gain insight into the mechanism of ball gall formation on Solidago altissima.
  •  Zeatin, zeatin riboside, isopentenyladenine, and isopentenyladenosine were identified and quantified by gas chromatography–mass spectrometry in ball galls, goldenrod stems, and in larvae of E. solidaginis.
  •  Concentrations of the four cytokinins were higher in gall tissues than in stem tissues when expressed on a weight per stem length basis but not when expressed on a weight per weight basis. Isopentenyladenine was the most abundant cytokinin in larvae from developing galls and in larvae from fully formed galls exhibiting ‘green islands’. In first instar larvae, isopentenyladenine was present at a much higher concentration (c. 50 fold) than in control stem tissues.
  •  The presence of an E. solidaginis larva results in higher cytokinin amounts in a given length of stem than are found in its absence. Larvae of E. solidaginis may act as point sources of isopentenyladenine in developing ball galls.

Abbreviations

BHT, butylated hydroxytoluene; DMSO, dimethylsulphoxide; DMSO, dimethylsulphoxide anion; GC-MS, gas chromatography mass spectrometry; HPLC, high pressure liquid chromatography; iP, isopentenyladenine; iPA, isopentenyladenosine; Me-dHZ, methyl-dihydrozeatin; Me-dHZR, methyl-dihydrozeatin riboside; Me-iP, methyl-isopentenyladenine; Me-iPA, methyl-isopentenyladenosine; Me-Z, methyl-zeatin; Me-ZOG, methyl-zeatin-O-glucoside; Me-ZR, methyl-zeatin riboside; Me-ZROG, methyl-zeatin riboside-O-glucoside; m/z, mass to charge ratio; PVPP, polyvinylpolypyrrolidone; SIM, selected ion monitoring; TEAB, triethylammonium bicarbonate; Z, zeatin; ZR, zeatin riboside.

Introduction

As is the case for insect-induced plant galls in general, the mechanism of ball gall formation by the larva of the tephritid fly Eurosta solidaginis on goldenrod (Solidago altissima) has not yet been determined. The ball gall of S. altissima is a stem gall that forms in response to the activities of a single larva of E. solidaginis. Adult females lay eggs in the folded leaves of terminal buds of S. altissima in the spring. Once a larva has hatched, it travels down through the folded leaves of the terminal bud and burrows into stem tissue right below (or sometimes at) the apical dome. Gall formation continues over a 3–4-wk period and results from cell division, enlargement, and differentiation (Beck, 1947).

Cytokinins are produced by a number of gall-forming bacterial phytopathogens including Agrobacterium tumefaciens, Erwinia herbicola pv. gysophilae, Pseudomonas savastonoi, and Rhodococcus fascians (Morris, 1995). Pathogenic strains of the gall forming bacterium E. herbicola pv. gysophilae have been shown to secrete significant amounts of cytokinins into culture media, while nonpathogenic strains did not (Lichter et al., 1995). Genes that encode for the production of cytokinins have been localized on the T-DNA of the Ti plasmid of A. tumefaciens (Barry et al., 1984), and a linear plasmid with a cytokinin synthesis gene has been found in R. fascians, a bacterium that induces leafy galls (Crespi et al., 1992).

Given the evidence for cytokinin production by gall forming bacteria and the well characterized role for cytokinins in the development of crown galls induced by A. tumefaciens (Morris, 1995), there has been interest in determining the role of cytokinins in insect gall development. Leitch (1994), using radioimmunoassay following high pressure liquid chromatography (HPLC), found that the levels of isopentenyladenine, isopentenyladenosine, isopentenyladenine ribotide, and isopentenyladenine-9-glucoside were higher on a per gram f. wt basis in galls induced by Pontania proxima during the first 2 wk of gall development compared with levels in leaf tissue. Levels of isopentenyladenosine were reported as 50-fold higher in hackberry (Celtis occidentalis) gall tissues than in control leaf tissues when detected by radioimmunoassay following immunoaffinity chromatography and HPLC (McDermott et al., 1996). By contrast, Engelbrecht (1971) found that the levels of cytokinin bioactivity in the tobacco callus bioassay were not elevated in galls formed by Mikiola fagi when compared with healthy leaves. Similarly, Van Staden & Davey (1978) reported that cytokinin concentrations in gall tissues formed by a chalcid wasp on Erythrina latissima were consistently lower than in surrounding leaf tissues.

There have been some reports of cytokinins associated with gall-inducing insects. Ohkawa (1974) isolated compounds active in the tobacco callus bioassay from the oriental chestnut gall wasp (Dryocosmus kuriphilus), and presented evidence for the presence of zeatin in the larvae. Larvae of a chalcid wasp that forms leaf galls on E. latissima were found to contain bioactivity coeluting with zeatin, zeatin riboside, and zeatin glucoside on HPLC (Van Staden & Davey, 1978). In a study of willow galls induced by Pontania pacifica, McCalla et al. (1962) associated growth promotion with two unidentified adenine derivatives in female accessory glands of the sawfly, but did not detect significant cytokinin bioactivity of gland extracts in the Xanthium leaf disc assay.

As the cytokinins in ball galls on S. altissima have not been previously identified, the purpose of these experiments was to determine which cytokinin nucleosides and free bases are present in developing ball galls by screening fractions eluting after HPLC for evidence of cytokinin bioactivity in the Amaranthus betacyanin bioassay (Biddington & Thomas, 1973) and further identifying the cytokinins by gas chromatography mass spectrometry (GC-MS). In addition, these experiments were undertaken to determine the levels of cytokinins in a developmental series of galls, in normal stem tissues, and to characterize and quantify the cytokinins found in the larvae of E. solidaginis.

Materials and Methods

All glassware that was used had been baked at 500°C to destroy any cytokinin contaminants prior to use. For the qualitative determination of cytokinins, 890 g of ball gall tissue from goldenrod (Solidago altissima L.) were extracted, while samples extracted for quantitative analyses consisted of 6–40 g of gall tissue and 2–8 g of normal goldenrod stem tissues, which were collected from fields near Ithaca, NY, USA, during the period of active gall growth. Leaves and lateral buds were removed from the tissues. Larvae were not removed from galls prior to extraction. Therefore, gall extracts contained the larvae that were associated with the extracted galls, typically at one larva per gall. The apical bud was removed from the stems, which were trimmed to 7 cm in length. The collected stem samples were further subdivided into the top 3.5 cm and the bottom 3.5 cm to act as separate controls for galls found at differing locations along the stem. In quantitative studies, the collected galls were processed as 7 different size classes and were trimmed to a length that was specific for each size class as follows: class A: 4–6 mm wide, 10 mm long; class B: 6–9 mm wide, 10 mm long; class C: 9–12 mm wide, 15 mm long; class D: 12–15 mm wide, 20 mm long; class E: 15–18 mm wide, 25 mm long; class F: 18–21 mm wide, 30 mm long; class G: 21–24 mm wide, 30 mm long. The top 3.5 cm of stem were the control tissues for class A–D galls, and the lower 3.5 cm of stem were the control tissues for class E–G galls.

Gall and stem tissues were processed and weighed as rapidly as possible, frozen in liquid N2 and stored at −80°C until they were extracted. The extraction procedure followed was essentially the procedure of Davies et al. (1986). The tissue samples were ground in five volumes of cold 80% methanol plus 20 mg l−1 butylated hydroxytoluene (BHT) in a Sorvall omni-mixer. In the quantitative studies, [2H5]-trans zeatin and [2H5]-trans zeatin riboside (98%[2H5], Apex Organics, Ltd, Honiton, UK) were added as internal standards to account for losses during extraction.

After the extract was stirred overnight at 4°C, it was filtered through cheesecloth and filter paper under vacuum. The residue was re-extracted with a half volume of 80% methanol +20 mg l−1 BHT and filtered. The methanol and some of the water of the combined filtrates were evaporated in vacuo at 40°C, and the pH of the aqueous residue was adjusted to 3.0 with 3N HCl. The extract was shaken against an equal volume of n-hexane, the hexane was evaporated in vacuo at 40°C, and the remaining solution was filtered through polyvinylpolypyrrolidone (PVPP). After the pH was raised to 8.0 with KOH, the extract was partitioned five times against equal volumes of water-saturated butanol. The extract in butanol was evaporated to dryness at 40°C. The residue was dissolved in water adjusted to pH 3.0 with glacial acetic acid and loaded onto an HCl/NH4OH washed P1 floc cellulose phosphate (Whatman) column. The column was washed with four column volumes of H2O that had been made pH 3.0 with glacial acetic acid, and the cytokinins were removed with four column volumes of 0.5 N NH4OH.

The eluate was evaporated down to 2 ml in vacuo at 40–45°C. Ten ml of methanol were added and the extract was centrifuged to remove the precipitate. After the supernatant was removed, the precipitate was washed with 5 ml methanol, centrifuged, and the combined supernatants were evaporated to dryness. The residue was dissolved, with mild warming, in 4 ml H2O adjusted to pH 6.9 with triethylammonium bicarbonate (TEAB) and centrifugally filtered through a 0.45-µm nylon-66 filter (Rainin, Woburn, MA, USA) prior to HPLC.

HPLC and GC-MS of the bulk gall extract

The bulk gall extract was loaded onto a 150 × 10 mm preparative C18 reverse phase column (Spherisorb ODS2, Phase Sep, Norwalk, CT, USA) on a Beckman HPLC with an increasing gradient of CH3CN in pH 6.9 H2O + TEAB at 5 ml min−1 (7–12% CH3CN over 25 min, 12–20% CH3CN over 5 min, and 20% CH3CN held for 10 min). Zeatin (Z) was shown to elute at 16.1 min; zeatin riboside (ZR) at 21.2 min; and isopentenyladenine (iP) and isopentenyladenosine (iPA) coeluted at 38.3 min (detected by UV absorption on a model 153 Beckman Analytical UV detector). Samples were collected at 1 min intervals in vials that had been siliconized in Aquasil (Pierce, Rockford, IL, USA). All glassware that was used in subsequent steps was siliconized prior to use. One tenth of each fraction collected, starting after 10 min, was tested for cytokinin bioactivity in the Amaranthus betacyanin bioassay (Biddington & Thomas, 1973).

Fractions containing peaks of cytokinin bioactivity were bulked into separate samples and were evaporated to dryness. Early peaks were re-dissolved in 1.0 ml of 12.5% CH3CN in pH 6.9 H2O + TEAB while late running peaks were dissolved in 22% CH3CN in pH 6.9 H2O + TEAB prior to centrifugal filtration through 0.45 µm nylon-66 filters (Rainin, Woburn, MA, USA). The cytokinin-containing samples were run on a 250 × 4.6 mm analytical HPLC column (Alltech Econosphere C18, 5 µm) at 1 ml min−1 in a H2O + TEAB/CH3CN gradient with a 2% change in the CH3CN concentration over 20 min. The gradient was specific for each cytokinin as follows (Z: 12–14%; ZR: 14–16%, and iP and IPA: 22–24%). One tenth of each one ml fraction collected was tested for cytokinin bioactivity in the Amaranthus betacyanin bioassay (Biddington & Thomas, 1973). Fractions containing peaks of bioactivity were bulked as individual samples.

The samples were evaporated to reduce the volume, transferred to 1 ml conical-bottomed vials, evaporated to dryness, dried over phosphorus pentoxide under vacuum, and permethylated with dimethylsulphoxide anion (DMSO) and methyl iodide (Martin et al., 1981). Fifty µL DMSO (prepared from silylation grade DMSO, Pierce, Rockford, IL, USA) and 10 µL methyl iodide were added to each sample. Fifty µL H2O were added to stop the reaction and to improve the two fold partitioning into 100 µL CHCl3. The samples were evaporated to dryness and dried over phosphorus pentoxide under vacuum.

Each sample was dissolved and sonicated in 20 µL CHCl3 followed by 80 µL n-hexane and centrifugally filtered through 0.45 µm nylon-66 filters (Rainin, Woburn, MA, USA). The samples were run on a 250 × 4.6 mm normal phase silica column (Alltech Econosphere Silica, 5 µm) at 2 ml min−1 with n-hexane: isopropanol: triethylamine at 97 : 3 : 0.1; 95 : 5 : 0.1; and/or 85 : 15 : 0.1 as needed, for sample purification and confirmation of peak identity by comparisons with the retention times of the following standards: methyl-dihydrozeatin riboside (Me-dHZR), methyl-zeatin riboside-O-glucoside (Me-ZROG), methyl-zeatin riboside (Me-ZR), methyl-dihydrozeatin (Me-dHZ), methyl-zeatin-O-glucoside (Me-ZOG), methyl-zeatin (Me-Z), methyl-isopentenyladenine (Me-iP) and methyl-isopentenyladenosine (Me-iPA). The UV spectra and absorption maxima of single peaks which were coincident with the retention time of the respective standards were determined. Fractions that showed a UV absorption maximum at the same wavelength as the methylated standard with the same retention time, were evaporated to dryness in 1 ml conical-bottomed vials under N2 and placed in a vacuum desiccator with phosphorus pentoxide overnight.

The samples were dissolved in 2–4 µL of ethyl acetate, and 1 µL was injected onto a bonded methyl silicone capillary column (HP-1, 12 m × 0.2 mm × 0.33 µm coating thickness, Hewlett Packard, Palo Alto, CA, USA) in a model 5890 A gas chromatograph connected to a model 5970B mass spectrometer (Hewlett Packard), with splitless injection at an injection port temperature of 250°C and a carrier gas flow rate of 25–30 cm s−1. The column was heated from 70°C to 230°C at 30°C min−1 and then 230°C to 290°C at 8°C min−1. The scan mode was used to obtain a complete mass spectrum of extracted compounds that chromatographed parallel to the cytokinin standards.

HPLC and GC-MS of the quantitative gall and stem extracts

Each quantitative extract that had been prepared for preparative HPLC was loaded onto a 150 × 10 mm preparative C18 reverse phase column (Spherisorb ODS2, Phase Sep, Norwalk, CT, USA) on a Beckman HPLC with an increasing gradient of CH3CN in pH 6.9 H2O + TEAB at 5 ml min−1 (6–14% CH3CN over 14 min, 14–20% CH3CN over 1 min, and 20% CH3CN held for 20 min). Samples were permethylated as described above. Preliminary normal phase runs of methylated samples were on a 250 × 4.6 mm normal phase silica column (Alltech Econosphere Silica, 5 µm) at 2 ml min−1 with n-hexane : isopropanol : triethylamine at 87 : 13 : 0.1, while final runs were with n-hexane : isopropanol : triethylamine at 97 : 3 : 0.1. The zone chromatographing at the retention time of Me-Z was collected as was the zone chromatographing at the retention time of Me-ZR. In extracts analysed for iP and iPA, the Me-iP and the Me-iPA peaks were bulked with the Me-ZR fraction.

The samples were dissolved in 2–10 µL of ethyl acetate, and 1 µL was run on GC-MS as above, except for the use of on-column injection and detection by selected ion monitoring (SIM). The 23 m column was heated from 75°C to 175°C at 40°C min−1; 175°C to 250°C at 8°C min−1; and 230°C to 290°C at 45°C min−1, where it was held for 7.79 min. The ions monitored for each of the four cytokinins were as follows: Me-iP, 107, 134, 162, 176, 188, 199, 216, 231; Me-Z and Me-[2H5]Z, 107, 134, 162, 188, 199, 216, 230, 235, 261, 266; Me-iPA, 162, 174, 202, 216, 217, 348, 376, 391; Me-ZR and Me-[2H5]ZR, 101, 174, 202, 216, 221, 246, 348, 390, 395, 421. Typical retention times were: Me-iP, 10.949; Me-Z, 13.038; Me-iPA, 17.580; Me-ZR, 19.995.

The ion at m/z 390 was used for the quantification of the levels of Me-ZR, while the ion at m/z 395 was used for quantification of Me-[2H5]ZR. For determination of the endogenous levels of ZR, the ratio of the area of the 395 ion of Me-[2H5]ZR to the area of the ion at m/z 390 of the endogenous Me-ZR was used to calculate the endogenous concentration of ZR in the tissue based on the known amount of [2H5]ZR initially added. The amount of endogenous zeatin was determined by use of the ratio of the area of the ion at m/z 235 of Me-[2H5]Z to area of the ion at m/z 230 of the endogenous Me-Z.

When Me-iP and Me-iPA levels were estimated in samples, the extracted Me-iPA and Me-iP were coinjected onto the GC-MS with the extracted Me-ZR so that the recovery of the Me-[2H5]ZR internal standard could be used to account for losses in the entire sample. The Me-[2H5]ZR internal standard was used to account for losses because deuterated standards of Me-iP and Me-iPA were not available. The amount of Me-iPA in the sample was assessed by comparison of the area of the base peak at m/z 391, found at the appropriate retention time, to a standard curve, while the amount of Me-iP was assessed by comparison of the peak area of the base peak ion at m/z 188, having the appropriate retention time, with a standard curve. Both values were adjusted for losses during extraction based on the recovery of the Me-[2H5]ZR internal standard which was determined by the comparison of the peak area of the ion at m/z 395 to a standard curve.

Extraction of larvae for cytokinins

Three hundred and forty larvae of Eurosta solidaginis Fitch were removed from developing class B-F galls that were collected from a field near Ithaca, NY, USA. The larvae were washed three times, frozen in liquid N2, and stored at −80°C until extraction. The larvae were extracted in cold 80% methanol plus 20 mg l−1 BHT in a 7-ml tissue grinder. [2H5]-trans Z and [2H5]-trans ZR (98%[2H5]), were added as internal standards to account for losses during extraction. The extraction procedure was as described for the quantitative plant tissue extracts, except that a cellulose phosphate column was not used for clean up of the sample.

One hundred and twenty seven larvae that were removed from fully formed galls exhibiting the ‘green island effect’ collected in the month of October, were frozen in liquid N2, and stored at −80°C until extraction as described above. The ‘green island effect’ is evident as a region of green tissue surrounding the larva in the centre of the gall, with the outer regions of the gall and neighbouring stem tissues achlorophyllous and ‘pithified’.

Results

Qualitative determination of cytokinins in ball galls

There were four zones of bioactivity following preparative HPLC of the bulk gall extract: samples A, B, C, and D. Samples A, B, and D were coincident with the elution positions of Z (16.1 min); ZR (21.2 min); iP + iPA (38.3 min), respectively, during the preparatory reverse phase HPLC run. The bioactive fractions collected as sample C did not coincide with the elution positions of tested standards. During analytical reverse phase HPLC, the bioactivity of samples A and B ran somewhat ahead of the positions of standard Z and ZR, respectively, detected by UV absorption. When sample C was re-analysed following analytical HPLC, the sample did not show evidence of cytokinin bioactivity when tested in the Amaranthus betacyanin bioassay and was not subjected to further analysis. The bioactivity of sample D coincided with the elution positions of the combined iP and iPA standards (12.2 min) on analytical HPLC. All fractions showing bioactivity were permethylated and run on normal phase HPLC.

Sample A, when run on 85 : 15 : 0.1 (n-hexane : isopropanol : triethylamine), yielded peaks at 4.5 min, 4.8 min, 6.4 min, 13.6 min, and 18.2 min. These peaks were close to the retention times of the following standards: Me-dHZR (4.5 min), Me-ZR (5.0 min), Me-ZROG (6.4 min), Me-Z (13.2 min), and Me-ZOG (18.0 min). Comparison of the UV absorbance of the putative Me-Z with that of a standard, indicated that approximately 30 ng were present. Quantitative information could not be obtained for the putative Me-ZOG and Me-ZROG fractions as standards of known concentrations were not available. No peaks coincided with Me-dHZ. A subsequent run on 95 : 5 : 0.1 (n-hexane : isopropanol : triethylamine) of the early fractions containing the putative Me-dHZR, Me-ZR, and Me-ZROG, contained no peaks close to the retention time of Me-dHZR (8.5 min), but contained a large peak at 10.0 min, a similar retention time to standard Me-ZR (9.9 min) and a small peak at 18.2 min, a retention time close to that of Me-ZROG (18.1 min). The amount of putative Me-ZROG in the sample was considerably less than what had been suggested by the results of the previous normal phase run. The ‘Me-ZR’ fraction had a UV absorption spectrum characteristic of a cytokinin with an absorption maximum at 274–276 nm. The positive identification of Me-Z and Me-ZR was subsequently obtained by GC-MS, with the mass spectrum of Me-Z obtained by SIM and Me-ZR obtained in the scan mode (Table 1).

Table 1. Gas chromatography mass spectrometry (GC-MS) identification of endogenous cytokinins in ball gall tissue
Permethylated cytokinin (GC-MS Rt)Characteristic ions m/z (relative abundance)
  1. Me-ZR, methyl-zeatin riboside; Me-Z, methyl-zeatin; Me-iPA, methyl-isopentenyladenosine; Me-iP, methyl-isopentenyladenine.

Me-ZR 
(16.3 min) 
Standard:421(M+,5), 390(94), 348(12), 246(5), 216(100), 202(9), 174(17), 101(9)
Gall extract:421(M+,5), 390(87), 348(10), 246(5), 216(100), 202(9), 174(16), 101(13)
Me-Z 
(13.9 min) 
Standard:261(M+,6), 230(100), 216(8), 199(10), 188(26), 162(10), 134(10), 107(9)
Gall extract:261(M+,6), 230(100), 216(4), 199(8), 188(35), 162(9), 134(8), 107(10)
Me-iPA 
(14.9 min) 
Standard:391(M+,100), 376(22), 348(54), 246(12), 217(54), 216(53), 202(74), 174(87), 162(14)
Gall extract:391(M+,100), 376(21), 348(52), 246(12), 217(52), 216(44), 202(72), 174(83), 162(12)
Me-iP 
(12.0 min) 
Standard:231(M+,33), 216(41), 199(19), 188(100), 176(12), 162(24)
Gall extract:231(M+,34), 216(44), 199(21), 188(100), 176(16), 162(26)

The normal phase HPLC run in 85 : 15 : 0.1 (n-hexane : isopropanol : triethylamine) of sample B, yielded peaks at 4.4 min and 5.0 min. These peaks were coincident with standard Me-dHZR (4.5 min) and Me-ZR (5.0 min), respectively. A peak at 13.6 min with a retention time close to that of Me-Z (13.2 min) contained approximately 40 ng of putative Me-Z. Following further purification on 95 : 15 : 0.1 (n-hexane : isopropanol : triethylamine), there was no evidence for the presence of Me-dHZR. The ‘Me-ZR’ fraction with an elution position that coincided with standard Me-ZR at 10.2 min, was shown to have a UV absorption peak of 274–277 nm compared with 274–277 nm for standard Me-ZR, with approximately 2 µg present. When this peak was run on the GC-MS it was shown to have the same retention time as Me-ZR and the mass spectrum of Me-ZR (Table 1).

The normal phase HPLC run in 97 : 3 : 0.1 (n-hexane : isopropanol : triethylamine) of sample D, yielded peaks at 7.7 min and 26.3 min. These peaks were coincident with standard Me-iPA (7.6 min) and Me-iP (26.3 min), respectively, and consisted of approximately 150 ng of ‘Me-iPA’ and 90 ng of ‘Me-iP’, judged by absorption. The ‘Me-iPA’ peak was shown to have a UV absorption peak of 272–273 nm compared with 273 for standard Me-iPA, and when run on the GC-MS it was shown to have the retention time and mass spectrum of Me-iPA (Table 1). The mass spectrum of Me-iP from extracted gall tissue was also identical to a standard (Table 1).

Quantitative determination of cytokinins in ball galls and stem tissue

Class A galls, the smallest galls, contained approximately half the amount of ZR on a weight/weight basis, measured by GC-MS with [2H5]ZR internal standard, compared with the top 3.5 cm of stem (up to the base of the apical bud), which is the region of the stem at which galls of this size are found (Table 2). The amount of ZR present in the galls remained relatively stable at a mean value of 3.5–4.9 ng g−1 f. wt during the early stages of gall growth, for galls up to 15 mm in diameter. An approximately two-fold lower level was found in class E and class F galls, with levels at one-half to two thirds the amount present in the corresponding region of the stem at which these stages are typically found. Class G galls, the largest galls in the developmental series, showed a five fold increase in the amount of ZR present compared with the preceding stage, with a mean of 10.28 ng g−1 f. wt, the highest level of all stages, and an amount that was three times higher than the amount found in the corresponding stem tissue. The stem tissue showed a three-fold higher amount of ZR, on a weight/weight basis, in the top 3.5 cm compared with the lower 3.5 cm (Table 2).

Table 2. Levels of extractable zeatin riboside (ZR) from seven developmental stages of ball galls and from the corresponding goldenrod (Solidago altissima) stem sections at which the stages are found, detected and quantified by gas chromatography mass spectrometry (GC-MS)
Stem location: (distance from apical bud)Size class of galls (diameter)Gall [ZR] (ng g−1 f. wt)Stem [ZR] (ng g−1 f. wt)Gall [ZR] (ng cm−1)Stem [ZR] (ng cm−1)Gall [ZR] (ng gall−1)
  1. Results are presented as means (± SE; n = 2 (except as noted)) and are reported as ng g−1 f. wt, ng cm−1, and ng gall−1. an = 1.

0 cm–3.5 cmA (4–6 mm) 4.48 ± 0.039.35 ± 3.00 0.71 ± 0.020.67 ± 0.18 0.71 ± 0.02
 B (6–9 mm) 4.90 ± 1.80  1.40 ± 0.53  1.40 ± 0.53
 C (9–12 mm) 3.50 ± 0.68  1.96 ± 0.31  2.93 ± 0.46
 D (12–15 mm) 3.79 ± 0.61  3.60 ± 0.81  7.19 ± 1.15
3.5 cm–7.0 cmE (15–18 mm) 1.41a3.05 ± 0.55 1.78a0.30 ± 0.08 4.44a
 F (18–21 mm) 2.13 ± 0.71  4.04 ± 1.57 12.13 ± 4.72
 G (21–24 mm)10.28 ± 2.51 32.47 ± 11.14 97.40 ± 23.64

The levels of Z, measured by GC-MS with [2H5]Z internal standard, expressed on a weight/weight basis were lower than ZR levels in all stages of galls and in the stem tissues, and lower than iP and iPA levels for all of the gall and stem tissues that were analysed (Tables 2, 3 & 4). Zeatin concentrations did not show an initial decline in class A galls compared with stem levels as was evident for ZR. Mean levels of Z for galls up to 12 mm in diameter were in the range of 0.9–1.02 ng g−1 f. wt compared with a mean of 0.87 ng g−1 f. wt in the corresponding stem tissue. Zeatin levels dropped two fold in class D galls, and dropped again in class E galls, rising to a level of 0.39–0.40 ng g−1 f. wt in class F and G galls, compared with 0.80 ng g−1 f. wt in the corresponding stem tissues. The level of zeatin was not appreciably higher in the top 3.5 cm of the stem compared with the lower 3.5 cm.

Table 3. Levels of extractable zeatin (Z) from seven developmental stages of ball galls and from the corresponding goldenrod (Solidago altissima) stem sections at which the stages are found, detected and quantified by gas chromatography mass spectrometry (GC-MS)
Stem location: (distance from apical bud)Size class of galls (diameter)Gall [Z] (ng g−1 f. wt)Stem [Z] (ng g−1 f. wt)Gall [Z] (ng cm−1)Stem [Z] (ng cm−1)Gall [Z] (ng gall−1)
  1. Results are presented as means (± SE; n = 2) and are reported as ng g−1 f. wt, ng cm−1, and ng gall−1.

0 cm–3.5 cmA (4–6 mm)0.98 ± 0.220.87 ± 0.150.16 ± 0.040.07 ± 0.020.16 ± 0.04
 B (6–9 mm)1.02 ± 0.32 0.30 ± 0.10 0.30 ± 0.10
 C (9–12 mm)0.90 ± 0.31 0.50 ± 0.16 0.75 ± 0.23
 D (12–15 mm)0.45 ± 0.09 0.42 ± 0.08 0.85 ± 0.17
3.5 cm–7.0 cmE (15–18 mm)0.28 ± 0.050.80 ± 0.070.37 ± 0.080.08 ± 0.010.93 ± 0.20
 F (18–21 mm)0.39 ± 0.19 0.74 ± 0.39 2.21 ± 1.17
 G (21–24 mm)0.40 ± 0.08 1.25 ± 0.24 3.75 ± 0.71
Table 4. Levels of extractable isopentenyladenine (iP) and isopentenyladenosine (iPA) from three developmental stages of ball galls and from goldenrod (Solidago altissima) stem sections, detected and quantified by gas chromatography mass spectrometry (GC-MS) using [2H5] zeatin riboside (ZR) as an internal standard
Tissue analysed:[iP] (ng g−1 f. wt)[iP] (ng cm−1)[iP] (ng gall−1)[iPA] (ng g−1 f. wt)[iPA] (ng cm−1)[iPA] (ng gall−1)
  1. Results are reported as ng g−1 f. wt, ng cm−1, and ng gall−1. na = not analysed.

Class A galls:6.14 1.01 1.011.890.31 0.31
Class D galls:2.98 2.82 5.641.311.25 2.50
Class G galls:3.8812.2836.841.404.4313.29
Top 3.5 cm of stem:6.58 0.49 2.780.21 
Top 7.0 cm of stem:nana 1.980.18 

Expressed on a weight/stem length basis, as ng cm−1, Z levels (Table 3) were higher in the gall tissues than in the stem controls in all gall classes, and ZR levels (Table 2) were higher in all classes except class A, indicating that the presence of the larva of E. solidaginis results in higher Z and ZR amounts in a given length of stem than are found in its absence. Class B-F galls contained 2–14 times more ZR than the corresponding stem controls, with class G containing 108 times the amount in the corresponding stem control. The gall classes contained from 2 to 16 times the amount of Z compared with the corresponding stem controls. Data are also presented on a ng gall−1 basis (Tables 2 & 3).

The data for iP and iPA levels determined using the recovery of [2H5]ZR to correct for losses during extraction, are shown in Table 4. About 6 ng g−1 f. wt iP was found in class A galls, making it the cytokinin in highest concentration in first visible galls. Levels of iP in the top 3.5 cm of stem were similar to levels in class A galls. Levels in class D and class G galls were 37–51% lower than levels in class A galls. Expressed on a weight/weight basis, iPA levels were one third lower in class A galls compared with the top 3.5 cm of stem. Levels in class D and G galls were about 25% lower than levels in class A galls. Comparison of the amount of iPA found in the top 3.5 cm of stem, to the amount found in 7.0 cm of stem, suggested that concentrations were higher in the top 3.5 cm of stem compared with the lower 3.5 cm of stem, as had also been noted for ZR levels in the stem. Expressed on a weight/stem length basis (Table 4), levels of iP and iPA were elevated in the gall tissues in all classes analysed compared with the corresponding stem controls. Results are also expressed on a ng gall−1 basis (Table 4).

Cytokinins in the larvae

Following extraction of 340 washed first instar larvae of E. solidaginis, which had been removed from developing class B-F galls, Z, ZR, iP, and iPA were detected in permethylated fractions by coincidence of the following ions for each compound at the standard retention time of the compound: Me-Z, ion at m/z 230; Me-[2H5]Z, ion at m/z 235; Me-ZR, ion at m/z 390; Me-[2H5]ZR, ion at 395; Me-iPA, ion at m/z 391; and Me-iP, ions at m/z 188 and 216. The amounts of each of the four cytokinins found in the larvae are listed in Table 5. iP was the most abundant cytokinin present in the larvae at about 350 ng g−1 f. wt of larvae or about 0.07 ng larva−1. The second most abundant cytokinin was iPA at about 43 ng g−1 f. wt or about 0.009 ng larva−1. Similar amounts of Z and ZR were present at 8.36 ng g−1 f. wt and 12.22 ng g−1 f. wt or 0.002 ng larva−1 and 0.003 ng larva−1, respectively.

Table 5. Levels of extractable zeatin (Z), zeatin riboside (ZR), isopentenyladenosine (iPA), and isopentenyladenine (iP) from larvae of Eurosta solidaginis removed from developing galls, and from E. solidaginis larvae removed from fully formed galls exhibiting the ‘green island effect’, detected and quantified by gas chromatography mass spectrometry (GC-MS) using [2H5]Z and [2H5]ZR as internal standards
CytokininLevel in larvae from developing gallsLevel in larvae from fully formed galls exhibiting ‘green islands’
(ng g−1 f. wt)(ng larva−1)(ng g−1 f. wt)(ng larva−1)
  1. Results are reported as ng g−1 f. wt and ng larva−1.

Z  8.360.002none detected 
ZR 12.220.0031.980.114
iPA 43.440.0091.620.093
iP349.610.0734.750.273

Extraction of 127 third instar larvae derived from fully formed galls exhibiting the ‘green island effect’ found in the field in October, yielded 3 cytokinins: ZR; iP, and iPA, detected as described for the first instar larvae. The most abundant cytokinin was iP at about 5 ng g−1 f. wt or about 0.3 ng larva−1. ZR was present at 1.98 ng g−1 f. wt or 0.114 ng larva−1; and iPA was present at about 1.6 ng g−1 f. wt or about 0.09 ng larva−1 (Table 5).

Discussion

Cytokinins in ball galls and goldenrod stem tissues

A tissue extract of developing ball galls was screened for the presence of cytokinin nucleosides and free bases exhibiting bioactivity in the Amaranthus betacyanin bioassay, and the positive identities of four cytokinins were subsequently determined by GC-MS, namely Z, ZR, iP, and iPA. There was no evidence to suggest the presence of novel cytokinins in the gall extract. The most abundant cytokinin in the gall extract was ZR and the least abundant was Z. Small amounts of Me-ZOG and Me-ZROG may have also been present based on retention time comparisons on normal phase HPLC, though these two cytokinins may be underestimated due to their unfavourable partitioning coefficients into butanol, with the problem being most pronounced for ZROG (Horgan & Scott, 1987).

The levels of the cytokinins Z, ZR, iP, and iPA were not elevated when analysed on a weight/weight basis in developing ball galls (4–21 mm) in comparison with the levels found in actively growing stem tissues. The levels of Z and iP in newly formed class A galls were similar to levels found in the corresponding stem controls, while the levels of iPA and ZR were 32%–52% lower in the youngest galls compared to stem controls. However, levels of the four cytokinins were higher in gall tissues compared with stem tissues when expressed on a weight/stem length basis, indicating that the presence of a Eurosta solidaginis larva results in higher cytokinin amounts in a given length of stem than are found in its absence. In galls, the greater width of the gall masks the increased cytokinin levels when the data are represented on a weight/weight basis.

The lack of elevated weight/weight concentrations of the four cytokinins in developing galls compared with stem tissues is not surprising given that the control stem tissues had been actively growing and already contained high levels of cytokinins. In addition, a dramatic increase in the concentration of cytokinins in whole ball galls compared with stem tissues would not be expected, as unlike crown galls, ball galls are not unorganized masses of callus-like growths or regions of abnormal shoot and root proliferation, morphologies suggestive of gross changes in hormonal levels. The ball gall is a very organized tumour that is found developing in a region of active stem growth. When the larva induces gall development, it enters a region of active cell division, and by its presence, it ‘captures’ the meristematic tissues, creating an island of meristematic tissue in a region that would otherwise differentiate into nondividing pith parenchyma cells. These factors suggest that major changes in hormonal levels may actually be found in the area surrounding the larva, and not throughout the entire gall. The increase in cytokinins in developing ball galls that is evident when levels are expressed on a weight/stem length basis, and the presence of a high concentration of cytokinins associated with the larvae of the developing galls could be reflective of a localized high concentration of cytokinins in the gall tissue surrounding the larva. The data obtained in this study of extracts of entire galls would not reflect the changes in cytokinin concentrations in the immediate vicinity of the larva.

With the exception of the finding of an elevated level of zeatin riboside in class G galls, the highest levels of zeatin and zeatin riboside in gall tissues were found in young developing galls. Levels of Z and ZR remained relatively high and stable during the early stages of gall growth, suggesting a role for cytokinins in gall development. Both cytokinins dropped by about 50% in class E galls. The levels of iP and iPA were also highest in the early stages of gall growth with levels in class A galls higher than levels in class D and G galls.

The levels of the bases Z and iP in class A galls were closer to the levels in equivalent stems than were the levels of the corresponding ribosides. The cytokinin bases are probably the active forms of cytokinins, with ribosides needing to be converted into bases before exerting an effect (Laloue & Pethe, 1982). The decline in ribosides seen in the earliest stage of gall formation may be a result of increased conversion of the ribosides to the active bases. The high level of ZR in class G galls may represent a return to lower metabolism of ZR to Z in galls that have ceased growing.

Given the growing evidence that cytokinin nucleotides may play a central role in regulating cytokinin metabolism and given the fact that high levels of the nucleotides may show better correlation with cytokinin activity than the ribosides and bases, an analysis of the cytokinin nucleotides in goldenrod stems and developing ball galls might provide further evidence in support of a role for cytokinins in the development of the ball gall. Cytokinin nucleotide levels have been shown to increase in response to exogenous application of cytokinins and in response to enhanced cytokinin biosynthesis without a concomitant rise in the levels of the ribosides and bases, despite evidence for cytokinin activity in the system studied (Laloue & Pethe, 1982; Horgan, 1985). The nucleotides may act as storage pools with high rates of conversions to the ribosides and bases in some tissues exhibiting cytokinin mediated responses (Laloue & Pethe, 1982).

Comparison of ball gall cytokinin levels with reports of cytokinins in other insect galls

The results presented here are similar to the results of cytokinin analyses of chalcid galls on E. latissima leaves by Van Staden & Davey (1978) and the results of bioassays for cytokinins in leaf galls formed by M. fagi (Engelbrecht, 1971), in that nongalled tissues contained higher levels of cytokinin activity and cytokinins, expressed on a f. wt basis, than the insect-induced gall tissues. However Van Staden & Davey (1978) found that the cytokinin activity increased in gall and leaf tissue throughout the growing season, whereas the results for the ball gall indicate that the levels of ribosides and bases are highest during the most active early stages of gall growth, with the exception of the large level of ZR found in the fully developed class G galls. The results presented here differ from those of Leitch (1994) and McDermott et al. (1996) who reported higher weight/weight concentrations of cytokinins in galls induced by P. proxima and in hackberry galls, respectively, compared with levels in control tissues.

Cytokinins in the larvae of Eurosta solidaginis

Z, ZR, iP, and iPA were detected and quantified by GC-MS in purified extracts of first instar larvae of E. solidaginis removed from developing galls. The cytokinin found in highest concentration in the larvae was iP, at about 350 ng g−1 f. wt, which was 53 times more than the concentration found in the control stem tissues; iP was also found in the highest concentration in larvae removed from fully developed galls exhibiting ‘green islands’. The elevated levels of iP and iPA in larvae from developing galls in comparison to ZR and Z levels, were not reflective of the relative amounts of cytokinins in the gall or stem tissues. In the larvae, iP was present at a concentration that was 29 times the amount of ZR, while the levels of ZR and iP in the gall tissues were approximately the same. Accumulation of iP from the plant tissues and storage in the fat body of E. solidaginis is not likely given the water solubility of iP. As preferential accumulation of iP by the larvae seems unlikely, the high concentration of iP in the larvae probably results from larval iP synthesis. As iP is the precursor of ZR in plants (McGaw & Burch, 1995), this iP may serve to provide the precursor for some of the ZR and Z in the plant tissue of the gall.

The high levels of iP, iPA, and ZR in the larvae from fully formed galls may be partially responsible for the ‘green island effect’ that is evidenced by the persistence of green tissue in the immediate vicinity of the larvae of E. solidaginis in a region of otherwise senesced plant tissues. These results are similar to those of Engelbrecht (1971), who found high levels of cytokinin bioactivity in the leaf mining larvae of Stigmella spp. and in the ‘green islands’ surrounding leaf mines on beech, birch, and poplar, as well as associated with the ‘green islands’ surrounding Hartigola annulipes galls.

It is interesting to note that iPA has been detected in the nongall-forming insect Locusta migratoria and has been shown to be conjugated to the insect hormone, ecdysone in newly laid eggs of L. migratoria with evidence obtained by NMR and GC-MS (Tsoupras, 1983). Ohkawa (1974) and Van Staden & Davey (1978) have presented evidence for cytokinins in the larvae of insect gall formers, with detection by bioassay and coincidence with standards following LH20 column chromatography, gas liquid chromatography, and thin layer chromatography; and by bioassay following fractionation on an LH20 column, respectively. There have been no previous reports of cytokinins detected in gall-forming insects with detection by GC-MS.

The results presented here differ from those of Van Staden & Davey (1978), who suggested that the larvae of the chalcid wasp that forms leaf galls on E. latissima, were accumulating cytokinins from the plant tissues. Larvae from mature galls on leaves of E. latissima had very high concentrations of cytokinins, containing 90% of the cytokinins found in the gall tissue. Since larvae from senesced galls contained no detectable cytokinin activity, Van Staden & Davey (1978) suggested that the cytokinins in the larvae from mature galls may have been accumulated from the tissues rather than synthesized by the larvae. For the ball gall, the total amount of each cytokinin per gall considerably exceeded the amount per larva (Tables 2, 3, 4, & 5). As the extracted galls contained larvae, and there is typically only one larva per gall, this means that most of the cytokinins in the gall extracts were associated with the plant tissue. These results suggest that larvae of E. solidaginis, unlike those of the gall former on E. latissima, are not accumulating cytokinins from the plant tissue, but instead may be secreting them.

Cytokinins have been shown to be associated with other plant tumour-forming organisms. Stevens & Berry (1988) presented evidence that the single major cytokinin released by cultures of Frankia, an actinomycete that forms nitrogen fixing nodules on Alnus glutinosa, was iPA. They detected iPA at the high level of 30 µg g−1 d. wt of Frankia, by GC-MS (SIM) in the culture medium. Wang et al. (1982) detected iP in the culture medium of the root nodule-forming bacterium Rhizobium strain LPR1105, by CI-GC-MS. Z, ZR, and iP have been shown to occur in the culture medium of pathogenic strains of the gall-forming bacterium E. herbicola pv. gysophilae with quantification by radioimmunoassay following HPLC separation of immunopurified cytokinins (Lichter et al., 1995). It has also been shown that one of the genes within the fas locus on the linear plasmid of the leafy gall-inducing bacterium, R. fascians, codes for an isopentenyl transferase (Crespi et al., 1992) which differs from the isopentenyl transferase gene of A. tumefaciens that is responsible for the formation of isopentenyladenine metabolites (Barry et al., 1984).

Evidence presented in this study shows a comparatively high level of only one of the cytokinins in E. solidaginis larvae, iP, which is found in a weight/weight concentration that is 53-fold greater than the amount found in the stem tissues. It is present in larvae, not only during stages of active gall growth, but also in larvae removed from mature galls, suggesting that the larvae may have the capability of iP synthesis. No previous studies have provided evidence for iP or iPA in gall-forming arthropods.

We have demonstrated that the E. solidaginis larva contains both cytokinins and high levels of IAA (Mapes & Davies, 2001), and is almost certainly acting as a source of these compounds to the goldenrod stem tissue that forms the ball gall. No applications of plant hormones or related substances have been able to duplicate the development or structure of insect galls. The failure to mimic gall development by application of compounds has almost certainly resulted from the inability to effectively simulate both the precise location of the insect within the plant tissue, and the continuing production of the gall-inducing stimulus by the insect source. In addition it is unlikely that one compound alone induces gall formation. For the ball gall, the ‘green island effect’ surrounding the larva is consistent with a central source of cytokinins, while the vascular tissue pattern of the gall can be accounted for by a central production of IAA. The tissue proliferation of the ball gall is most likely a response to both IAA and cytokinins. Our results suggest that larvae of E. solidaginis may act as point sources of IAA and cytokinins in developing ball galls.

Acknowledgements

This research was funded in part by Hatch funds administered by the New York State College of Agriculture and Life Sciences at Cornell University and by Grants-In-Aid of Research from Sigma Xi and from the Cornell Chapter of Sigma Xi. The GC-MS used in this research was funded by the National Science Foundation (grant DMB-8505974) and the New York State College of Agriculture and Life Sciences at Cornell University.

Ancillary