Planta

Planta (1986) 167: 9-18

9 Springer-Verlag1986

The partial purification and characterisation of gibberellin 2p-hydroxylases from seeds of Pisum sativum V.A. Smith and J. MacMillan Agricultural and Food Research Council Research Group, School of Chemistry, University of Bristol, Bristol BS8 1TS, UK

Abstract. The gibberellin (GA) 2fl-hydroxylases in mature and immature seeds of Pisum sativum have been partially purified and characterised. The enzymes are unstable when stored below pH 7.0 or in the absence of a thiol reagent. The optimum assay pH is between 7.4 and 7.8 and activity is dependent upon the presence of ~-ketoglutarate, Fe z + and ascorbate. The 2fl-hydroxylase activities for GA1, GA4, GA9 and G A 2 o are chromatographically inseparable and correspond to a protein of Mr 44 000. The rate of GA 2fl-hydroxylation varies according to substrate and some evidence indicates that the 2fi-hydroxylase activities for GA1 and GA~ and for GA9. and GA2o may reside in different proteins. During pea seed maturation, the specific activity of the enzyme(s) increases dramatically and reaches a maximum at a time when e n d o g e n o u s G A g , G A / o , G A 2 9 and GAs 1 are also at their greatest concentration. This correlation is not the result of substrate induction of enzyme activity. Since the GA 2fl-hydroxylases operate at maximal rate at low substrate concentrations they are incapable of rapidly 2fl-hydroxylating excessive quantities of (exogenously applied) GA1 o r G A 2 o . On the basis of the kinetic parameters of the GA 2fl-hydroxylase activities, a generalised model is discussed for the control of the steady-state levels of bioactive hormone under normal physiological conditions. Key words: Gibberellin (2fl-hydroxylases) - Oxidative enzyme (plant) - Pisum (gibberellin) - Seed (gibberellin metabolism). DTE = dithioerythritol; EDTA = ethylenediaminetetraacetic acid; GAn = gibberellin An ; HPLC _high-performance liquid chromatography; HSS =high-speed supernatant; LSS=low-speed supernatant; PMSF=phenylmethane sulphonyl fluoride

Introduction

The in-vivo metabolism and localisation of gibberellins (GAs) in maturing and germinating seeds of Pisum sativurn have been researched extensively (Frydman et al. 1974; Frydman and MacMillan 1975; Sponsel and MacMillan 1977, 1978, 1980; Sponsel 1983). Furthermore, using cell-free extracts of maturing pea seeds, Kamiya and Graebe (1983) have determined details of the GA metabolic pathway. As yet, however, none of the enzymes catalysing these transformations have been isolated and examined. The present communication is concerned with the purification and characterisation of the enzyme(s) that are present in mature and maturing pea seeds and which catalyse 2fl-hydroxylation of endogenous C 19-gibberellins. The products formed (Scheme 1) are biologically inactive. Material and methods Labelled GA substrates. [1,2-3Hz]Gibberellin A1 (1.57.1015 Bq.

mol- 1) was a gift from Dr. J.L. Stoddart (Welsh Plant Breeding Station, Aberystwyth, UK). [1,2-3H2]Gibberellin A 4 of specific activity 1.79.1015 Bq.mol 1 was a gift from Dr. L.M. Srivastava (Simon Frazer Univeristy of British Columbia, Canada). [ 2 , 3 - 3 H z ] G i b b e r e l l i n A 9 and [2,3-3Hz]GAzo with specific activities of 1.72.1015 and 1.21 "1014 Bq.mo1-1, respectively, were gifts from Dr. A. Crozier (University of Glasgow, UK) and Dr. R.P. Pharis (University of Calgary, Alberta, Canada). These substrates were analysed by reverse-phase high-performance liquid chromatography (HPLC). The estimated radiopurity of the [1,2-3H2)GA1, [1,2-3H2]GA4, [2,3-3H2]GA9 and [2,3-3Hz]GAz0 was found to be 73%, 62%, 77% and 90%, respectively.

Abbreviations:

Plant material. Seeds of Pisum sativum cv. Progress No. 9 were

obtained from Nutting and Thoday, Longstanton, Cambridge, UK. Seeds used for enzyme extraction were surface sterilised in a 5% (v/v) hypochlorite solution, extensively washed in ster-

10

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin

GA

12

- aldehyde

s

GA 9

J

"~

GA 51

GA29

HO

H

"l" ~

"COOH

GA1 0 m, i!

r~/".,,, ~..O H ,

I ~

GA80 HO

"COOH

ile water and imbibed for 21 h at 25~ in sterile Petri dishes (9 crn diameter), lined with Whatman (Springfield Mill, Kent, UK) No. 1 filter paper and containing 15 ml sterile water. Cotyledons were then separated from testas and embryonic axes and grated into liquid nitrogen. Immature pea seeds of known ages were harvested from plants maintained in an unheated greenhouse under natural photoperiod between April and September. These seeds were dissected immediately after harvesting and the cotyledons, testas and embryonic axes were stored separately in liquid N2 until required for extraction. For experiments in which developing pea seeds were treated with G A prior to harvesting, the plants were maintained in constant-environment growth chambers (Conviron S10 H; Controlled Environments, Winnipeg, Canada) at 20 ~ C day temperature, 15 ~ C night temperature and a light regime of 18 h day length, 450 g m o l - m - 2 " s -1 photon fluence rate.

Procedure for enzyme extraction from pea cotyledons. Frozen tissue (from approx. 180 seeds) was pulverised in liquid N2 using a mortar and pestle, allowed to warm up to 4 ~ C and handhomogenised in 150 ml 0.1 M 2-amino-2-(hydroxymethyl)-l,3propanediol (Tris)-HC1 buffer, pH 7.6, containing 2 mM dithioerythritol (DTE). The homogenised tissue was then filtered through muslin and centrifuged at 18000 g for 45 rain at 4 ~ C. The low-speed supernatant (LSS) was decanted. For some experiments, this material was then ultracentrifuged at 120000 g for 2 h at 4 ~ C to obtain a soluble-protein fraction (HSS). For enzyme purification the LSS was placed in a conical flask standing in an ice bath. Redistilled methanol was added slowly with stirring to give a final concentration of 20% (v/v). After 1 h, precipitated material was removed by centrifugation. The clarified supernatant was then dialysed for 2 h against 0.05 M TrisHC1, pH 7.6, to remove the methanol and concentrated by further dialysis against 40% (w/v) polyethylene glycol (PEG) dissolved in 0.05 M Tris-HC1, pH 7.6, containing 2 mM DTE.

Enzyme assays. Gibberellin 2/?-hydroxylase activity was measured routinely at 25 ~ C by determining the rate of liberation of tritiated water using 2/?-tritiated gibberellins as substrates. The assay protocol was similar to that previously described (Smith and MacMillan, 1984) except that 0.1 M Tris-HC1, pH 7.6, was used as the buffer and catalase was omitted. The G A substrates were assayed at the concentrations spe-

'!

~,~OH

~',,,~OH

Scheme 1. Gibberellin metabolism in Pisum sativum. Although the GA1 2/~-hydroxylase is present in both the seed and vegetative tissue, GA~ is synthesised only in vegetative tissue

cified in the text. These concentrations were sub-saturating and maximal reaction velocities (Vmax) were thus never obtained. Unless otherwise stated [1,2-3H2]GAI was used at a concentration of 5.8 gM (9.15.103 Bq'ml-1).

High-performance liquid chromatography. Gibberrellin products formed during incubation of G A I , GA9 or GA2o with a partpurified 2fl-hydroxylase enzyme preparation were analysed and quantified by reverse-phase HPLC using a stainless-steel column (250 mm long, 8 mm internal diameter) packed with ODS Hypersil (5 gM) (Shandon Southern Products, Runcorn, Cheshire, UK), fitted to an LDC HPLC apparatus (Riviera Beach, FI., USA). Separation of GA1, GAs, GA2o and GA29 was achieved by loading prepared samples onto a column pre-equilibrated with 20% methanol in water containing 0.8 % (v/v) phosphoric acid at a flow rate of 1.2 ml. rain - 1. Samples were eluted by applying an exponential methanol gradient of 20-70% (v/v, methanol: H20) over 30 min. Under these conditions the retention volumes of GA1, GAs, GA2o and GA29 were 15.8 ml, 6.8 ml, 27.5 ml and 8.9 ml, respectively. Separation of GAg, GAzo and GAs1 was achieved by pre-equilibrating the column with 50% (v/v) methanol in water containing 0.8% (v/v) phosphoric acid at a flow-rate of 1.2 m l ' m i n -1, and eluting with an exponential methanol gradient of 50-85% (v/v, methanol:H20) over 20 min. The retention volumes of GAg, GA2o and G A s , were found to be 30.2 ml, 11.5 ml, and 17.3 ml, respectively.

Protein estimation. The protein content of various enzyme preparations was estimated by the Biuret method (Gornall et al. 1949) or by measuring peptide absorption at 225nm 1 cm (A225 1 m g - m l - 1 9,17, Hall and Hard 1975).

Sodium dodecyl sulfate-polyaerylamide gel eleetrophoresis. Enzyme preparations were analysed on 15% (w/v) acrylamide slab gels, prepared and run using the procedure described by Laemmli (1970). Estimation of molecular weight. The apparent molecular weight (Mr) of the gibberellin 2fl-hydroxylase(s) was determined by gel filtration on a column (90 cm long, 2,2 cm diameter) of Sephadex G-100 equilibrated and eluted with 0.1 M Tris-HC1 buffer, pH 7.6, containing 2 m M DTE at a flow-rate of 4 ml. h - 1 The column was calibrated with bovine serum albumin (Mr

V.A. Smith and J. MacMillan : The partial purification and eharacterisation of gibberellin

:

"8.0

~

m

~ EO

oeo -7.0 ~ ~I

-g

11

25 I

~r m

~

2o

~.:

0 0

- 5.0

~

15

-4,0

10-

3.0 20, % -2.0

/

e

1.0

., ..,~

~ooOO .o*~ . . .

~

i%

J

"-.~

--o~ -

2's

30

3'5 Fraction

40 Time

no,

(h)

Fig. 1. Sephadex G-100 gel filtration elution profile, o - o 9 Enzyme activity; ........ protein elution profile

lrig. 2. Stability of the HSS enzyme preparation at 4 ~ in 0.05 M Tris-HCI, pH 7.2, in the presence (o-o) or absence (o-e) of DTE (2 raM). Samples were assayed for a fixed incubation period of i h. The protein concentration was 8.8 mg-ml-1

68000), ovalbumin (M, 45000), lactoglobulin (Mr 38000) and chymotrypsinogen (Mr 25000).

Results

Enzyme purification. Unless otherwise stated, all of the following operations were performed at 4 ~ C.

Diethylaminoethyl (DEAE)-cellulose chromatography. A column (30 cm long, 2.2 cm diameter) of DEAE-cellulose (Whatman, DE-52) was equilibrated in 0.05 M Tris-HCl, pH 7.6, containing 2 mM DTE. The concentrated soluble extract of the pea cotyledons (40 ml) was pumped onto the column and eluted with 0.05 M Tris-HC1, pH 7.6, containing 2 mM DTE at a flowrate of 20 rot. h - 1 Fractions (15 ml) were collected and assayed against [1,2-3Hz]GA1 for 2fl-hydroxylase activity. Those that contained enzyme activity were pooled and concentrated by dialysis against 40% PEG in 0.1 M Tris-HC1, pH 7.6, containing 2 mM DTE.

Gel filtration. A column (60 cm long, 3.2 cm diameter) of Sephadex G-100 was equilibrated with 0.1 M Tris-HC1, pH 7.6, containing 2 mM DTE. The concentrated DEAE-cellulose enzyme preparation (18 ml) was loaded onto the column and eluted with equilibration buffer at a flow rate of 17 ml-h -1. Fractions (8.5 ml) were collected and assayed against [I,2-3H2]GA~. Fractions containing enzyme activity were pooled and concentrated by dialysis against 40% PEG in 0.1 M Tris-HC1, pH 7.6, containing 2 mM DTE to a final volume of around 5.0 ml. A typical elution profile is shown in Fig. 1. No further purification could be achieved by rechromatographing the enzyme preparation on an analytical (90 cm long, 2.2 cm diameter) column of Sephadex G-100. Analysis of the eluted material using [I,2-3H2]GA1 and [1,2-3H2]GA9 as enzyme substrates indicated that the A~ and A9 2/%hydroxylase activities co-eluted and corresponded to the only protein peak resolved.

Enzyme stability. In preliminary experiments, cotyledons from imbibed mature pea seed were homogenised in 0.05 M Tris-HC1, pH 7.2, and sequentially centrifuged at 18000g (LSS) and 120000 g (HSS). This impure HSS enzyme preparation loses its GA 2/~-hydroxylase activity with time (Fig. 2). When dialysed against 0.05 M Tris-HC1, pH 7.2, and assayed in the same buffer, the preparation retains enzyme activity. However, as shown in Table 1, when MgSO4 (10 mM) is included in the dialysis and assay buffer, almost all enzyme activity is lost. Magnesium does not inhibit enzyme activity directly. The observed results may be explained by the removal, during dialysis, of a low-molecular-weight compound which is not a cofactor for the GA 2/%hydroxylase(s) but which may inhibit the activity of an Mg 2+-dependent protease present in the HSS protein fraction. Attempts were made to stabilise the GA 2/?hydroxylase(s) using phenylmethane sulphonyl fluoride (PMSF), a potent inhibitor of proteases possessing serine residues at their active sites. The effects of ethylenediaminetetraacetic acid (EDTA) and dithioerythritol (DTE) were also examined. The addition of PMSF did not enhance enzyme stability at 4 ~ C, but EDTA had some beneficial efect on enzyme stability if M g 2+ w a s present in

12

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin

Table 1. Effect of Mg 2+ on the stability of the HSS enzyme preparation before and after dialysis. Assays were performed in the presence and absence of E D T A Treatment

g 30

a T20 liberated from [I,2-aH2]GA1 ( B q . m g - 1) 0

-- MgSO,

Control 4~ C Dialysed ( - EDTA) Dialysed ( + EDTA)

+ MgSO4

--EDTA

+EDTA

--EDTA

+EDTA

34 27 35

33 24 34

30 27 -

34 33 0.7

27

26

3

0.8

a Measured after a fixed incubation period of 1 h

the storage buffer, but not otherwise. The activity of the HSS enzyme preparations stored at 4 ~ (Fig. 2) was enhanced by DTE. This activity enhancement was slight but consistent and not observed when DTE was added to the HSS enzyme preparation immediately before assaying for enzyme activity. Subsequently it was found that DTE is an absolute requirement for retaining activity during column chromatography. It was therefore incorporated routinely in all buffers, at a concentration of 2 mM.

Enzyme activity and buffers. The activity of the GA1 2fl-hydroxylase was assayed in phosphate buffer (pH range 5.3-8.2), maleate-NaOH (pH range 5.2-6.8), sodium acetate (pH range 5.0-6.5) and Tris-HC1 (pH range 7.2-7.8). For a given pH value, similar activities were obtained independent of the chemical nature of the buffers. Activity was highest at pH 7.0-pH 8.0 and decreased rapidly below pH 7.0 (see, for example, Fig. 3). Activity was slightly increased by raising the buffer concentration from 0.05 M to 0.1 M. Both Tris-HC1 and phosphate have reasonably high buffering capacities between pH 7.0 and pH 8.0. However, 0.1 M Tris-HC1, pH 7.6, was chosen for routine use because Fe 2 + is precipitated from phosphate buffer, resulting in an apparent loss of enzyme activity. The pH stability of the GA 2fl-hydroxylase(s) was examined by dialysing an HSS enzyme preparation against 0.05 M potassium-phosphate buffer, pH 5.8, 6.8 or 7.8 for 16 h at 4 ~ and assaying for activity in 0.1 M Tris-HC1 buffer, pH 7.2. The results obtained (Table 2) confirm that the enzyme is very acid-labile and indicate that the observed inhibition of enzyme activity below pH 7.0 (Fig. 3) is a consequence of enzyme denaturation.

lo,

/ t / *

g5

/ 620

6]5

720

Z'5 pH

8[0

Fig. 3. Gibberellin A~ 2p-hydroxylase activity in 0.05 M phosphate buffer. Samples were assayed for a fixed incubation period of 1 h. The protein concentration was 8.3 m g . m l - x

Table 2. Dependence of the stability of GA1 2fl-hydroxylase activity on pH pH of dialysis buffer

a TzO liberated from [1,2-3H2]GA1 ( B q . m g - 1)

5.8 6.8 7.8

4 78 89

Measured after a fixed incubation period of 1 h

Enzyme purification. The GA1 2fl-hydroxylase does not bind to DEAE-cellulose (DE-52) but sufficient material is retained on the column to warrant incorporating this step in the purification procedure. Attempts to bind the enzyme to the cationic exchange CM-cellulose (CM-52; Whatman, Maidstone, Kent, UK) were also unsuccessful. Application of the enzyme to a CM-52 column equilibrated in either acetate or phosphate buffer, pH 6.0, resulted in total loss of activity, presumably because of the acid-lability of the enzyme. When applied to CM-cellulose columns equilibrated in Tris-HC1 or phosphate buffer, pH 7.0, almost all the applied protein, including the enzyme, eluted in the column flow-through. Thus the enzyme appears to be ionically neutral or to carry a very weak positive charge at neutral pH. On the basis of these results, attempts were made to purify the enzyme using phenyl-sepharose CL-4B (Pharmacia Fine Chemicals, Uppsala, Sweden). The chromatographic behaviour of the enzyme indicated that it was an extremely hydrophobic protein. However, a suitable balance between enzyme activity yield and protein purification could not be obtained. The procedure adopted for the purification of

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin Table 3. Enzyme purification

Step

Total units

13

06.

Specific activity (pmol 9h- 1. rag- 1 protein)

Recovery Purl(%) fication (fold)

LSS

670

0.18

100

1

Methanol preparation

716

0.35

105

1.9

DEAE

450

0.74

67

4.1

G100

246

1.47

37

8.2

0 , 5 84

E

0,4

0.3

: " ........ J ": 0,2. ,D

the GA1 2fl-hydroxylase activity achieved only an eight-fold increase in specific activity (Table 3). The molecular weight of this enzyme was estimated to be 44000 by gel filtration on Sephadex G-100. Examination of the G-100 enzyme preparation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed it to contain one major protein of Mr 45 000 and one minor protein of Mr 43 000. Moreover, as shown in Fig. 4, neither of these proteins were major components of the LSS enzyme preparation.

Cofactor requirements. The GA1 2fl-hydroxylase present in the cotyledons of mature seeds of Phaseolus vuIgaris is believed to be a dioxygenase that requires 0~-ketoglutarate, ascorbate and Fe 2+ for activity (Smith and MacMillan 1984). Although these cofactors were routinely added when assaying 2fl-hydroxylase activity in the HSS preparation of pea seed cotyledons, no reduction in activity was observed in their absence. Dialysis of the pea cotyledon HSS preparation against 0.05 M TrisHC1, containing 1 m M EDTA, or elution through DEAE-cellulose both yielded preparations that exhibited 45 and 62% of maximum activity when assayed for G A 1 2fl-hydroxylase activity in the absence of ascorbate and 0r respectively. However, total dependence of enzyme activity upon ~-ketoglutarate and ascorbate was shown by the Sephadex G-100 enzyme preparation. The concentrations of ~-ketoglutarate and ascorbate yielding maximum enzyme activity were found to be around 2 m M and 1 0 m M respectively. These values are similar to those obtained for the GA1 2fl-hydroxylase from seeds of P. vulgaris (Smith and MacMillan 1984). After dialysis against 0.05 M Tris-HC1, pH 7.2, containing 1 m M EDTA, the 2fl-hydroxylase activity of the pea seed HSS preparation exhibits total dependence upon the addition of Fe 2 § Maximum activity is obtained when Fe z § is present at concentrations around 1 mM.

0.1.

I

? o

x

100 e5.

70c~ @ r 5O

~

t

I

rylaseS (Mr 98000)

~oooJ

"~LOvalbumin { Mr 45000)

m

I;%%7.n )

o

30

Relative mobility (cm)

Fig. 4. Sodium dodecyl sulfate-po]yacrylamide gel electrophore-

sis of the gibberellin 2fl-hydroxylasepreparations. Five gg of protein was applied to each slot in the gel....... Impure LSS preparation; , purified enzymepreparation after gel filtration on Sephadex G-100 Enzyme activity was not detectable when Ca 2 +, Zn 2+, Mn 2+ or Mg 2 + (1 mM) were substituted for Fe z+. In the presence of Cu 2 + (1 raM), however, the enzyme exhibited some 45% of the activity measured in the presence of Fe z + (1 raM). The enzyme was also assayed in the presence of Fe 2 + (1 mM) together with varying concentrations of Ca 2 +, Cu 2 +, Zn 2 +, Mg 2 + and Mn 2 + (Fig. 5). Zinc was found to be a potent inhibitor of enzyme activity. The observed effect of copper is possibly caused by an ability to effectively displace Fe z+ from the enzyme surface and, albeit less efficiently fulfil the mechanistic role of Fe 2 + in G A 2fl-hydroxylation. The other metal ions had no effect on enzyme activity.

Analysis of the products formed from [3H]GA1. The H P L C analysis of samples recovered from an incubation mixture containing [3H]GA1 and an

14

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin Table 4. Product analysis using [3H]GA1 as substrate for HSS enzyme preparation

:~ ._> ~o 9

Time (min)

100"

0 40 80 120

80

E Ill

60'

y

40

e

1.0

GA8

0.107 0.159 0.210

0.106 0.164 0.220

0.24 0.124 0.079 0.05

Table 5. Specific activities of GA1, GA4, GAzo and G A 9 2flhydroxylation in the LSS and Sephadex G100-enzyme preparation Substrate

0,5

T20

Substrate remaining (pmol) GA1

9 "-~O

e

0

Product formation (pmol)

1,5 Metal

2.0

2,5

ion concentration

(raM)

Fig. 5. Gibberellin A1 2/~-hydroxylase activity in the presence of Fe 2 + (1 mM) and various other metal ions. m-m, Mn z+; rn-rn, Mg2+; T-v, Ca2+; e - e , Cu2+; o - o , Zn 2+

HSS preparation of mature pea cotyledons indicated that the only radioactive products formed during the course of the reaction were tritiated water and tritiated GAs. These products were generated in a 1 : 1 molar ratio and entirely accounted for the disappearance of [3H]GA1 (Table 4). The results are based on specific activities of 1.57.1015 Bq.mo1-1, 1.02-1015 Bq-mo1-1 and 5.5-1014 Bq" mol-1 for GA1, GAs and TzO, respectively (Nadeau and Rappaport 1974).

GA1 GA 4 GA2o GA 9

Specific activity (pmol 9h - 1 . m g - 1)

Purification (fold)

LSS enzyme preparation

G100 enzyme preparation

0.19 0.22 0.13 0.41

1.54 1.64 0.66 2.61

8.1 7.5 5.0 6.3

v

2~-Hydroxylase substrate specificity. The results shown in Table 5 were obtained from experiments in which the GA 2fl-hydroxylase activity was measured when equimolar concentrations (0.033 gM) of [1,2-3H2]GA1, [I,2-3H2]GA4, [2,3-3H2]GAzo or [2,3-3Hz]GA9 were incubated with the LSS or G-100 enzyme preparations. Although both preparations catalyse 2fl-hydroxylation of the four substrates, after gel filtration, the overall purification of the GAs and GA4 2/%hydroxylase activities was consistently higher than that of the GA2o and GA9 2/%hydroxylase activities. Comparison of the thermal-denaturation kinetics of the GA1, GA2o and G A 9 2fl-hydroxylase activities at 41~ indicated that the GA/o and G A 9 2fl-hydroxylase activities decayed at the same rate with half-lives of 60 min. By contrast, however, the half-life of the GA1 2/~-hydroxylase activity was estimated to be 83 min (Fig. 6). The above data indicate the presence of at least

N

c

bd

30

30

60

90

120

150 Time ( r a i n )

Fig. 6. Decay of G A 2fl-hydroxylase activity at 41 ~ C. e - o , Substrate [3H]GA1 ; m-m, substrate [3H]GA9; t~-D, substrate [3H]GA2o. The [2,3-3H2]GA2o used in this experiment was of specific activity 1.1.1015 Bq-mol 1 and 84~ radiopurity

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin

two discrete 2fl-hydroxylase activities in the cotyledons of mature pea seeds. Attempts were made to separate these activities by gel-filtration HPLC using an 1-125 protein analysis column (Waters Associates (Inst.), Northwich, Cheshire, UK). A sample of the Sephadex G-100 purified enzyme was injected onto the column, eluted with 0.1 M potassium phosphate, pH 7.2, at a flow rate of 0.5 ml. min- ~. Fractions of the eluant were collected and assayed against [1,2-3Hz]GA~, [2,3-3H2]GA2o and [ 2 , 3 - 3 H z ] G A 9 as substrate. The results obtained indicated that the three activities co-eluted and that HPLC could not enhance the resolution obtained by conventional gel filtration procedures.

Table 6. Kinetic parameters for 2fl-hydroxylase activities in ma-

ture pea seeds Substrate

Km (laM)

Vm.x (pmol" h - 1 .mg- 1)

GA1 GA2o GA9

0.069 _+0.004 1.55 __.0.3 0.299 +_0.03

6.95 24.80 22.90

on the initial rates of T 2 0 release in enzyme assay mixtures containing [3H]GA1, or [3H]GA2o, both at 0.005 gM, were in line with the Km values shown in Table 6 for the [3H]GAs. From [3H]GA2o the initial rate of release of T 2 0 w a s progressively reduced by GAI (and GA4) in the concentration range 0.1-0.25 gM (Fig. 7a) but G A 9 and G A 2 o were effective only at higher concentration (0.4-10.0 gM) (Fig. 7a, b). From [3H]GA1 the initial rate of release of T20 was reduced by GA~ (and GA4), but not by G A 9 and G A 2 o , in the concentration range 0.005-0.05 gM. Analysis of the inhibition kinetics of GA~ upon [3H]GA2o 2fl-hydroxylation at varying concentration of GA1 and substrate indicates that GA~ is a competitive inhibitor. The Ki was calculated to be 0.013 +0.001 gM. This value is lower than the K m (0.069 gM) determined for [3H]GA~ 2fl-hydroxylation. This observed discrepancy cannot be the result of a primary isotope effect since the Ki for unlabelled GA~ is lower than the Km for [3H]GA1. Therefore, in addition to showing that

Kinetic parameters. The rates of transformation of the labelled GA substrates used in the experiments reported here may be around an order of magnitude lower than those expected for the corresponding unlabelled substrates. This is a consequence of the presence of the tritium label at the site of catalytic attack. A Sephadex G-100 enzyme preparation was used for determining the Km and Vm,x values for the enzyme(s) catalysing 2fl-hydroxylation of [3H]GA~, [3H]GAzo and [3H]GA9. The results presented in Table 6 were obtained from EadieHofstee plots of v versus v/s. Competitive studies with cold gibberellins. The effects of adding cold GAa, GA4, GA9 and GA/o

'~176

15

.

""--"- ~

,o

60

c

40-

g

B

20-

olo,

oho

o:,~

oi~o

o:~,

o

i

~.o

i

,,.o

i

~.o

do

i

,o.o

Unlabelled GA (pM) Unlabelled GA I,uM) Fig. 7a, b. Inhibition of the gibberellin 2fl-hydroxylase activities, a, b Inhibition of the [3H]GA2o 2fl-hydroxylase activity by unlabelled GA1 (,, i), GA4 (m-D), GAzo (o-o) and GA9 (o-o). b Inhibition of the [3H]GA1 2fl-hydroxylase activity by unlabelled GA2o (A--A)and GA 9 (A--A)

16

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin

GA~ is an efficient inhibitor of GA2o 2fl-hydroxylation, these data may add further support to the suggestion made previously, that at least two discrete 2fl-hydroxylase activities exist in the cotyledons of mature pea seeds. 2fl-Hydroxylation in immature pea seeds. All experiments were carried out using soluble preparations of immature pea seeds homogenised in 0.1 M Tris HC1, pH 7.6, containing 2 mM DTE and sequentially centrifuged at 18000 and 120000 g as described previously. The characteristics of the 2flhydroxylase activities located in soluble-protein preparations from cotyledons of immature seeds were identical to those of the enzyme(s) present in the HSS preparation of cotyledons from mature seeds. Localisation of enzyme activity. Immature pea seeds were harvested between day 22 and 36 postanthesis. The cotyledons, embryonic axes and testas were separated by dissection and frozen immediately in liquid nitrogen. The soluble-protein preparation obtained from these materials were assayed against similar concentrations of [3H]GAt, [3HIGA4, [3H]GA2o and [3H]GA9 (0.1 txM). No 2fl-hydroxylase activity was detected in the soluble-protein preparation of the testa. The relative specific activities of the 2fl-hydroxylase activities in the cotyledons and embryonic axes are given in Table 7. The ratio of the 2fl-hydroxylase activities in the cotyledons and embryonic axes are different for the pairs GA1/GA4 and GA9/GA2o. This observation indicates the possibility of the existence of at least two discrete enzymes. Product analysis. Samples taken fiom enzyme-assay incubation mixtures were analysed by HPLC. The soluble-protein preparation from cotyledons catalysed the conversion of [ 2 , 3 - 3 H 2 ] G A 9 t o [3-3H]GA51. Two radioactive peaks, corresponding to authentic reference samples of G A 9 and GAs 1, were obtained. No other radio-active product (i.e. GA2o) was detectable. Similarly, HPLC data also indicated that [ 3 - 3 H ] G A 2 9 w a s the only radioactive product formed during the incubation of [2,3-3H]GA2o with either the soluble-protein preparation from cotyledons, or a mixture of the soluble-protein preparations of cotyledon and testa. No radioactive compounds corresponding to GA1 or to GAz9-catabolite were detected in the elution profiles. The expected molar ratio of T 2 0 liberated to [3-3H]GA51 and [ 3 - 3 H ] G A 2 9 produced from [2,3-3H2]GA9 and [2,3-3H2]GA2o, respectively, is

Table 7. Specific activities of the enzyme(s) catalysing 2fl-hydroxylation of GA~, GA4, GA2o and GA9 in immature pea seeds. Numbers in ( ) are the ratios of the enzyme specific activities in the cotyledons and embryonic axes Gibberellin substrate

Specific activity (pmol- h - t. rag- 1) Cotyledons Embryonic axes

GA1 GA 4 GAzo GA9

2.1

0.08

2.3 0.9 2.1

0.085 (27) 0.0725 (12.4) 0.1625 (12.9)

(26)

Table 8. Quantitative analysis of the products found during incubation of [2,3-3H2]GA9 and [2,3-3I-I2] GA20with a soluble-protein extract of immature pea seed cotyledons Time (rain)

0 30 60 90

[3Hz]GA9 substrate

[3Hz]GAzo substrate

T20 (pmol)

[3H]GA51 (pmol)

T20

[3H]GAz9

(pmol)

(pmol)

0 0.08 0.174 0.273

0 0.079 0.172 0.270

0 ND ND 0.472

0 ND ND 0.467

1:1. Thus, the specific activity of the label at the C-2 positions of G A 9 and GA2o was calculated from the relative amounts of radioactivity associated with the respective GA products formed and T20 liberated. The data presented in Table 8 are based on the values obtained. Given a total specific activity of 1.72" 10 is Bq. tool-1 for [2,3-3Hz]GA9, the estimated specific activity at the C-2 position is 7.33"1014 Bq'mo1-1. Similarly, given a specific activity of 1.21"10 t4 Bq'mo1-1 for [2,3-3H2]GAzo, the estimated specific activity at the C-2 position is 4.1013 Bq "mo1-1 Time course of 2fl-hydroxylase activity during pea seed maturation. Pea seeds were harvested at particular time points during the maturation period. The developing cotyledons were dissected out and immediately frozen in liquid nitrogen. Soluble-protein preparations were obtained from the frozen material using standard procedures and extensively dialysed at 4 ~ against 0.1 M Tris-HC1, 2 mM DTE, pH 7.6. Samples were then assayed for 2fl-hydroxylase activity using [2,3-3H]GA2o, [2,3-3H]GA9, [1,2-3H2]GA1 and [I,2-3H2]GA~ as substrate. For all the kinetic assays, the protein concentration used was 4 rag. m l - t and the substrate concentration 0.1 gM. The results shown in Fig. 8 indicate that the 2fl-hydroxylase activities for GA1, GA4, G A 9 and GA2o are approximately

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin

-3.5-

17

Table 9. Specific activity ( m o l - h - l . m g - l x 10 -15) of the gibberellin 2/~-hydroxylase(s) present in immature pea seed cotyledons

o

q~ 3.0-

Treatment E 2.5cc

~llI '~

2.0 -

0

1.5-

~1

o

/III II

GA1

GAzo

GA9

6.57 7.29 6.10

1.94 1.90 1.88

4.76 4.45 4.10

~ ~.

"~

1.0-

Assay substrates

9

9

None + GA2o + Solvent

9

enzyme(s) appear to be present throughout seed maturation, albeit at a very low level for most of the time.

0.5-

,s

20

2'2

;,

2'e

2'8

~o

a'2

~,

Days a f t e r antheeis

Fig. 8. Gibberellin 2fl-hydroxylase activity during pea seed maturation. 9- 9 Substrate [3H]GA1 ; zx-A, substrate [3H]GA4 ; e-o, substrale [3H]GA2o ; o-o, substrate [3H]GA9

parallel over the time period considered. Maximum levels of enzyme activity occur between day 22 and day 30 post-anthesis. The specific activities of the enzyme(s) during this period are approximately ten-fold higher than in the mature pea seed. This increase in specific activity of the enzyme(s) correlates with the net increase in seed weight and also with the accumulation of endogenous GA2o and G A 9 (Frydman et al. 1974). Enzyme induction. This correlation between the appearance of GA2o and G A 9 and the increase in specific activity of the 2/Lhydroxylase(s) may be explained in two ways: (i) substrate induction of enzyme activity; (ii) developmentally triggered synchronous expression of (all) enzymes involved in gibberellin metabolism. These possibilities were tested. Ten days after flowering, seeds of developing fruit were either injected with GA2o (1 gg/seed) or with an equivalent volume of solvent (1 gl ethanol, 50% v/v) or left intact. Seeds from each treatment were harvested after 48 h (average fresh weight 0.21 g). Soluble-protein preparations were obtained from the cotyledons and assayed against GA~, GA2o and G A 9 (0.1 laM). The specific activities for GA~, GA2o and GA9 2/~-hydroxylation are very similar in the cotyledons of intact seeds and those injected with G A 2 o o r solvent (Table 9). Similar data were also obtained for treated and untreated seeds of mean fresh weight 0.06 g (10 d post-anthesis) or of mean fresh weight 0.t 1 g and 0.17 g. Clearly, the gibberellin 2/?-hydroxylase activity is not subject to induction by GAzo. The

Discussion

The GA 2fl-hydroxylase activities present in the cotyledons of immature and mature pea seeds are probably identical. The enzymes from both sources are small proteins (Mr 44000) extremely hydrophobic and have an optimal assay pH of 7.4-7.8. Their activity is stabilised by high-ionic-strength buffer and more particularly by the presence of DTE. The cofactor requirements of these enzymes are similar to those reported previously for the GA1 2•-hydroxylase present in the cotyledons of mature seed of Phaseolus vulgaris cv. Canadian Wonder (Smith and MacMillan 1984). This latter enzyme, however, did not require DTE or high-ionicstrength buffer for stability. It was also found to be slightly smaller (Mr 37000) and had an optimum assay pH of 5.8 (Smith and MacMillan 1984). The specific activity of the GA1 2/?-hydroxylase in soluble-protein preparations of immature pea seed cotyledons is similar to that recorded for the GA~ 2/~-hydroxylase in soluble-protein preparations of the cotyledons of mature seeds of P. vulgaris. With respect to [1,2-3Hz]GA1, the Km values of these enzymes are also very similar (0.069 gM and 0.085 I-tM, respectively). By contrast, however, the specific activity of the GA2o 2p-hydroxylase is approx. 20-fold higher in the immature pea seed preparation than in that of the mature bean seed. Assuming that a single protein catalyses gibberellin 2/~-hydroxylation, then it would appear that the bean seed enzyme more effectively discriminates between the 3fl-hydroxy and non-3-hydroxy gibberellins than does the enzyme in the pea seed. On the other hand, some evidence does indicate that discrete proteins catalyse 2/~-hydroxylation of GAx/GA4 and 2fl-hydroxylation of GAg/GA2o in the pea seed. If this is indeed the case, then it may be that the GAzo/GA 9 2~-hydroxylase is a poor

18

V.A. Smith and J. MacMillan: The partial purification and characterisation of gibberellin

catalyst for GA1 and GA~ 2fl-hydroxylation and that the GAI/GA4 2j3-hydroxylase is, in all respects, kinetically similar to that of the bean seed. Gibberellin 2fi-hydroxylase may have a physiological role in regulating the turn-over of the bioactive gibberellin(s). Recently, attention has been directed to GA1 which appears to be the sole gibberellin for stem elongation in pea (Ingram et al. 1984 and references therein), maize (Spray et al. 1984 and references therein) and other plant species (Phinney 1984). Data on the 3fl-hydroxylase which catalyses the step from GA2o to GA1 (Scheme 1) is not yet available. However, this enzyme does appear to have kinetic properties similar to the GAa 2fl-hydroxylase from the following observations (T.J. Ingram School of Chemistry, University of Bristol, personal communication). When GA2o is applied in low doses (0.01-0.1 gg) to vegetative tissue of pea plants, the amounts of the metabolites GA~, GA8 and GA29 vary with the dosage. At high dosages (approx. 10 gg) the relative amounts of GA1 and GAs formed are approximately constant with dosage but the amount of GA29 formed increases with increasing amounts of GA2o supplied. From the information and the kinetic data obtained in the present study for the 2fl-hydroxylase from pea seed, it is interesting to speculate on the steady-state levels of GA~. Since the 2fl-hydroxylase has a relatively high Km for GA2o and since GA~ suppresses the formation of G A 2 9 , the formation of GA1 should be favoured over the formation of GA29 at low steady-state concentrations of GA2o. Conversely, at steadystate concentrations of GAzo approaching the Km for GAzo, the formation of GAz9 from GA20 should be favoured over the formation of GA8 from GA~ since the maximal rate of the former reaction is three to four times that of the latter conversion. The GA 2fl-hydroxylases do not appear to be abundant enzymes in any of the plant tissues examined here. The GA1 2fl-hydroxylase, in particular, also operates at maximal velocity at relatively low substrate concentrations. During pea seed maturation, however, the specific activities of the enzyme(s) increase appreciably between day 19 and day 30 post-anthesis and correlate with the levels of endogenous GA2o, GA29, GA9 and GAs1 (Frydman et al. 1974). This observation indicated that, in addition to playing a role in the regulation of steady-state levels of bioactive hormone, the enzymes may, by substrate induction, rapidly control excessive fluxes of bioactive hormone in GA-sensitive tissue. However, this possibility proved not to be the case. The enzymes are non-inducible.

Variations in enzyme concentration appear to be a consequence of a general, genetically determined increase in gibberellin metabolic activity. We thank Dr. Valerie M. Sponsel for providing samples of immature pea seeds and also for critically reading the manuscript.

References Frydman, V.M., Gaskin, P., MacMillan, J. (1974) Qualitative and quantitative analyses of gibberellins throughout seed maturation in Pisum sativum cv. Progress No. 9. Planta 118, 123 132 Frydman, V.M., MacMillan, J. (1975) The metabolism of gibberellins A9, A2o and A29 in immature seeds of Pisum sat# rum cv. Progress No. 9. Planta 125, 181-195 Gornall, A.G., Bardawill, C.J., David, M.M. (1949) Determination of serum proteins by means of the biuret reaction. J. Biol. Chem. 177, 751-766 Hall, B.G., Hartl, D.L. (1975) Regulation of newly evolved enzymes. 11. The Ebg represser. Genetics 81, 427435 Ingrain, T.J., Reid, J.B., Muffet, I.C., Gaskin, P., Willis, C.L., MacMillan, J, (1984) Internode length in Pisum. The Le gene controls the 3/~-hydroxylation of gibberellin A2o to gibberellin A1. Planta 160, 455~463 Kamiya, Y., Graebe, J.E. (1983) The biosynthesis of all major pea gibberellins in a cell-free system from Pisum sativum. Pbytochemistry 22, 681-689 Laemmti, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685 Nadeau, R., Rappaport, L. (1974) The synthesis of [3H]gibberellin A3 and [3H]gibberellin A1 by the palladium-catalysed actions of carrier free tritium on gibberellin A3. Phytochemistry 13, 1537-1545 Phinney, B.O. (1984) Gibberellin A1 dwarfism and control of shoot elongation in higher plants. In: The biosynthesis and metabolism of plant hormones. S.E.B. Seminar series No. 23, pp. 1%45, Crozier, A., Hilhnan, J.R., eds. Cambridge University Press, UK Smith, V.A., MacMillan, J. (1984) Purification and partial characterisation of a gibberellin 2fl-hydroxylase from Phaseolus vulgaris. J. Plant Growth Regul. 2, 251-264 Sponsel, V.M. (1983) The localisation metabolism and biological activity of gibberellins in maturing and germinating seeds of Pisum sativum cv. Progress No. 9. Planta 159, 454468 Sponsel, V.M., MacMillan, J. (1977) Further studies on the metabolism of gibberellins (GAs) Ag, A2o and A29 in immature seeds of Pisum sativum cv. Progress No. 9. Planta 135, 129-136 Sponsel, V.M., MacMillan, J. (1978) Metabolism of gibberellin A29 in seeds of Pisum sativurn cv. Progress No. 9; use of [2H] and [3H]GAs, and the identification of a new GAcatabolite. Planta 144, 69-78 Sponsel, V.M., MacMillan, J. (1980) Metabolism of [13C1]gibberellin A29 to [13C1]gibberellin catabolite in maturing seeds of Pisum sativurn cv. Progress No. 9. Planta 150, 46-52 Spray, C., Phinney, B.O., Gaskin, P., Gilmour, S.J., MacMillan, J. (1984) Internode length in Zea mays (L). The dwarf-I mutation controls the 3fl-hydroxylation of gibberellin A2o to gibberellin At. Planta 160, 464-468 Received 15 November 1984; accepted 25 July 1985