Eur. J . Biochem. 97. 119-126 (1979)

Hydrodynamic Parameters of the Detergent-Solubilised Hydrogenase from Paracoccus denitrificans Edith SIM and Robert B. SIM Laboratoire de Biochimie, Departement de Recherche Fondamentale ( Laboratoire Associe du Centre National de la Recherche Scientifique N” 312), Centre d’Etudes Nuclkaires-Grenoble, Commissariat i I’Energie Atomique (Received December 13, 1978)

The hydrogenase from Paracoccus denitrlficans, which is an intrinsic membrane protein, has been solubilised from membranes by Triton X-100. The partial specific volume of the solubilised protein has been determined using sucrose density gradient centrifugation in H 2 0 and 2H20. The values of the specific volumes of hydrogenase, measured in the presence or absence of Trifon X-100, are 0.73 and 0.74 ml . g-’, respectively, indicating that hydrogenase binds much less than one micelle of Triton X-100. The sedimentation coefficient of hydrogenase is increased from 10.4 S to 15.9 S on removal of detergent. The Stokes’ radius of hydrogenase, determined by gel filtration on Sepharose 6B, is 5.5 nm in the presence of Triton X-100 compared to 6.7 nm in the absence of detergent. The apparent molecular weight therefore increases from 242 500 to 466000 on removal of detergent. In the presence of urea and sodium dodecylsulphate, the hydrogenase has an apparent molecular weight of 63000. The enzyme therefore behaves as a non-covalently linked tetramer in the presence of Triton X-100. Removal of Triton X-100 results in association of tetramers to form octamers.

Paracoccus denitrijicans, like other members of the class of aerobic hydrogen bacteria, grows autotrophically when supplied with H2, 0 2 and C02 [l]. Under these conditions, hydrogenase activity is induced. Since HZ provides the sole source of reducing equivalents for the cell, hydrogenase, which catalyses the cleavage of H2 to protons and electrons [Eqn (l)] is a key enzyme for cell growth.

H2 s 2H’

+ 2e-.

(1)

Certain aerobic hydrogen bacteria have both cytoplasmic and membrane-bound hydrogenase enzymes [2] but P. denitrficans has only one hydrogenase, which is membrane-bound [2,3]. The hydrogenase from P. denitrficans is an intrinsic membrane protein in that it can be solubilised from the membrane only by treatment with detergents and the non-ionic detergent, Triton X-100, has proved to be particularly effective in solubilising the enzyme [4]. Before the hydrodynamic properties of membrane proteins solubilised by detergent can be determined, it is necessary to measure the amount of detergent bound to the protein. Binding of non-ionic detergents has been estimated for a variety of membrane proteins, . . - ..

G7;j~rnc,.s.

Catalase (EC 1.1 1.1.6); hydrogenase or hydrogen dehydrogenase (EC 1.12.-.-); phosphorylase a (EC 2.4.1 .l); alcohol dehydrogenase (EC 1.1.1.1).

e.g. the cholinergic receptor protein from Electrophorus electricus IS], adenylate cyclase from bovine brain [6] and rat renal medulla [7] and the prostaglandin E binding protein from rat liver [S], by determining the partial specific volume of the protein-detergent complex using sucrose density gradient centrifugation in H 2 0 and 2 H 2 0 as initially described by Meunier et al. [5]. There are no published data on the detergent binding properties of hydrogenases solubilised from bacterial membranes although estimates of molecular weights which d o not take account of possible detergent binding have been reported for the detergentsolubilised hydrogenases from Rhodospirillum rubrum [9], Thiocupsa roseopersincina [lo, 111, Alicaligenes eutrophus H16 [12] and Chromatium vinosum [13,14]. However, there is evidence to suggest that the solubilised hydrogenases from A . eutrophus [15], and Chromatium [14] d o bind detergent. T o permit structural characterisation of the hydrogenase from P. denitrijccans, hydrodynamic parameters of the enzyme have been determined in the presence and absence of TritonX-100. The apparent molecular weight of the polypeptide chain corresponding to hydrogenase has been estimated by polyacrylamide gel electrophoresis in sodium dodecylsulphate and the subunit composition of the enzyme is discussed.

120

MATERIALS AND METHODS Preparation of' Membranes and Detergent Extraction Paracoccus denitrificans DSM 65 strain 381 (a gift from Prof. H. G. Schlegel, Gottingen, F.R.G.) was grown under autotrophic conditions, harvested and membranes prepared. after sonication as previously described [3]. Membranes were extracted under H2 (20 min, 30°C) with 1 mg Triton X-lOO/mg membrane protein and the Triton X-100 extract was obtained as the supernatant after centrifugation at 20 "C at 140000 x g (rav= 10.93cm) for 90 min. Phenylmethylsulphonyl fluoride was added to a final concentration of 1 mM to inhibit proteolysis. The Triton X-100 extract was stored at - 20 "C under H2 until use. Bacteria were also grown as described above but with the addition of 1 mCi of [32P]orthophosphate (175 yCi . mmol-')/l of growth medium. Cells were washed four times after harvest but otherwise were treated identically to non-radioactive material.

Hydrodynamic Parameters of Hydrogenase from P. dmitr[/r'cans

Catalase was detected by measuring H 2 0 2 consumption at 240 nm [17] and yeast alcohol dehydrogenase was assayed by following the reduction of NAD' by ethanol at 340 nm [18]. Other proteins were detected by measuring protein content by the method of Lowry et al. [19]. Proteins used as standards do not bind significant quantities of detergent [20]. Molecular weight markers used for polyacrylamide gel electrophoresis in the presence of dodecylsulphate were reduced and alkylated preparations of rabbit phosphorylase a ( M , 92 000), bovine liver catalase (60000), bovine serum albumin (68 000), ovalbumin (42000), ClF(58000and 36000) and ClS(58000 and 29000). The last two proteins were prepared as described by Sim et al. [21]. Values for the Stokes' radius, sedimentation coefficient, partial specific volume (u2) and apparent molecular weight of subunits of standard proteins are taken from published tables [21,22]. Gel Filtration

Sucrose Density Gradient Centrifugation

The Triton X-100 extract (0.5 ml containing 3-5 mg protein) was loaded onto a linear gradient of sucrose (5-20"/, w/v) in 10 mM Tris-HC1, pH 8.0 or in 10 mM Tris-HC1, 0.05 (w/v) Triton X-100, pH 8.0. The sucrose gradient was made up in either 100 H 2 0 or H ~ 0 / ~ H 2(5/95, 0 v/v). 2H20 (99.8 %) was used to make up solutions and was kept in sealed glass vials until use to avoid exchange with atmospheric H20. Gradients were run in a Kontron 50 ultracentrifuge using a Beckman SW41 rotor. Gradients were pre-equilibrated to the run temperature (20°C) by centrifugation at 10000 rev./min for 3 h. A temperature of 20°C was used since hydrogenase has been shown to be cold labile [4]. Gradients were fractionated into 25 equal fractions by peristaltic pumping from the bottom of the gradient. Sucrose concentrations across the gradient were measured with a Zeiss refractometer. For gradients run in 95% 2H20, a standard curve was established to permit correction for the lower refractive index of 2H20 [16]. Protein Standards

For sucrose density gradient centrifugation bovine liver catalase ( S X I . = ~ 11.3 S, u2 = 0.73 ml . g-'), yeast alcohol dehydrogenase (s20,w= 7.61 S, u2 = 0.73 ml . g-I), porcine thyroglobulin ~ 2 0 =, ~19.2 S, z12 = 0.72 ml . g-') and hen ovalbumin (sZo,,,, = 3.66 S, r2 = 0.75 ml . g-I) were used as standards. Standard proteins for gel filtration chromatography were bovine liver catalase (Stokes' radius = 5.22 nm), porcine thyroglobulin (Stokes' radius = 8.1 nm) and yeast alcohol dehydrogenase (Stokes' radius = 4.55 nm).

Samples ( 5 ml, 30-50 mg protein) of the Triton X-100 extract of membranes of P. denitrificans were applied to a column of Sepharose 6B (80 x 2.5 cm), equilibrated with 10 mM Tris-HC1, pH 8.0, containing 0.05 "/, (w/v) Triton X-100, where indicated. The void volume (Yo)was determined from the elution position of protein aggregates in human serum. The total permeable space was measured with 3 H 2 0 . The Stokes' radius of hydrogenase was calculated using the average pore radius determined both in the presence and absence of detergent, by the method of Ackers [23], using standard proteins. Molecular weights and frictional ratios were calculated from experimentally determined Stokes' radii, partial specific volume measurements and sedimentation coefficients as described by Siege1 and Monty [24].

(v

Gel Electrophoresis

Polyacrylamide gel electrophoresis in the presence of dodecylsulphate was performed either according to the method of Weber and Osborn [25], using 7.7 (w/v) polyacrylamide gels or by the method of Fairbanks et al. [26] using 5.6 % (w/v) polyacrylamide gels. Samples were reduced in capped vials with 1,4-dithiothreitol (3 mg . ml-', 1 h at 37°C) and then alkylated with iodoacetamide (20 mg . ml-', 20 min at 37"C) before application to gels, unless otherwise indicated. Gel electrophoresis in the absence of dodecylsulphate was carried out as described by Ornstein and Davis [27]. Gels were stained for hydrogenase activity with benzyl viologen in an atmosphere of Hz, as previously described [3]. Alternatively, gels were stained for protein with Coomassie brilliant blue R250 and scanned at 550 nm in a Beckman Acta 3 spectropho-

E. Sim and R. B. Sim

121

tometer fitted with a scanning attachment. In cases where quantification was required from gel scans, gels from any one series were stained with equal volumes of stain solution and destained in equal volumes of destaining solution. Each gel was scanned at least twice and the gel was rotated through 90" between scans to minimise possible artefacts due to inhomogeneity in staining. The average area under each peak was measured. Comparisons were made only with gels run in the same set. Calculation of Partial Specific Volume and Sedimentation Coefficient

The partial specific volume and sedimentation coefficient of hydrogenase in the presence and absence of Triton X-100 were determined after sucrose density gradient centrifugation in 'H20 and H20 by the method of Clarke [I71 except that the viscosities of sucrose solutions at half-migration distance were taken from published tables [28]. The use of information from published tables to construct a viscosity curve was found to be more satisfactory since it provides more data points than the graphical method of Clarke [17]. Both methods gave acceptable values for protein standards compared with published values (i57J. The densities and viscosities of sucrose in 2 H 2 0 were calculated from values for the equivalent solutions in H20 as described by O'Brien et al. [29]. Analyrical-Scale Purification of Hydrogenase

A sample of a Triton X-100 extract of cell membranes of P. denitrijkans was centrifuged on a linear w/v) in H 2 0 / 2 H 2 0(5/95, sucrose gradient ( 5 - 20 v/v) in 10 mM Tris-HC1, 0.05 o/, Triton X-100 pH 8.0, for 20 h at 35000rev./min (Fig.1B). The peak of hydrogenase activity was pooled and samples were then subjected to a polyacrylamide gel electrophoresis in the absence of dodecylsulphate. The gels were stained for hydrogenase activity [3]. A single purple band developed on each gel. One gel from the set was rinsed in water and stained for protein with Coomassie blue, and then scanned. The other gels were treated as follows. The band of hydrogenase activity was excised from each of the gels and each slice was incubated with 100 pl of 2% (w/v) sodium dodecylsulphate in 8M urea, 0.2 M Tris-HC1, pH 8.0 either with or without 1,4-dithiothreitol for 1 h at 37°C followed by 3 min at 100 "C. Iodoacetamide was added to a final concentration of 20 mg . ml-' and samples were incubated for a further 2 min at 100 "C. Gel slices and the incubation fluid were then loaded onto 5.6 2)(w/v) polyacrylamide gels containing dodecylsulphate and subjected to electrophoresis. The gels containing dodecylsulphate were stained with Coomassie blue and scanned.

x,

Other Methods

Hydrogenase activity was determined by following the increase in absorbance at 555 nm on reduction of benzyl viologen by H2, as previously described [3]. T o permit calculation of the specific radioactivity of phospholipids in membranes of P. denitrifieans grown in the presence of [32P]orthophosphate, lipids were extracted by treatment with chloroform/methanol (2/1, v/v) [30] and phospholipid phosphorus was determined by the method of Bartlett [31]. Liquid scintillation counting was done in an Intertechnique CC30 counter using the scintillation fluid described by Patterson and Greene [32]. Under these conditions efficiency of counting 32P was greater than 98 %. Materials

Deuterated water and 3 H 2 0 were obtained from C.E.A. (Saclay,' Paris, France). Inorganic [32P]orthophosphate was from Radiochemicals (Amersham, U.K.). Sucrose was purchased from Mallinckrodt (St Louis, Mo., U.S.A.) and Sepharose 6B was from Pharmacia Fine Chemicals (Uppsala, Sweden). Yeast alcohol dehydrogenase and phosphorylase a were from Boehringer (Mannheim, F.R.G.). Thyroglobulin, catalase, ovalbumin, bovine serum albumin and Triton X-100 were supplied by the Sigma Chem. Co. (St Louis, Mo., U.S.A.). Iodoacetamide was from B.D.H. Chemicals Ltd (Poole, England) and 1,4-dithiothreitol was from Calbiochem Inc. (La Jolla, Calif., U.S.A.). Sodium dodecylsulphate was from E. Merck (Darmstadt, F.R.G.). All other reagents were from Prolabo (Rh6ne-Poulenc Industries, Paris, France). RESULTS Behaviour of Hydrogenase in the Presence of Detergent

After sucrose density gradient centrifugation of a Triton X-100 extract of membranes of P. denitrificans containing hydrogenase activity, the hydrogenase activity profile shows a peak at a position between catalase and yeast alcohol dehydrogenase marker proteins (Fig. 1). Hydrogenase migrates to the same position relative to standard proteins in gradients containing 0.05'x Triton X-100 in either H 2 0 (Fig. 1A) or 2 H 2 0(Fig. 1 B). Calculation of the partial specific volume of hydrogenase from sucrose density gradient centrifugation in both solvents [17] gives a value of 0.73 ml . g-', as shown in Table 1. The sedimentation coefficient of hydrogenase calculated from centrifugation in the presence of detergent (Fig. 1A and B) is 10.4 S (Table 1). When a Triton X-100 extract of membranes of P. denitrijicans grown in the presence of [32P]orthophosphate and containing phospholipids labelled with

Hydrodynamic Parameters of Hydrogenase from P. tlenirr;fic.an.s

the molecular weight of hydrogenase in the presence of Triton X-100 is 242500 and the frictional ratio is 1.2 (Table 1). Behuviour of Hydrogenase in the Absence qf'Detergent

Fraction number

Fraction number

Fig. I . Suc~roseckrisity grudient cmtriJugation in the presence ofdetergcwt. Samples (0.5 ml) of a Triton X-100 extract of membranes of P. ricwitrifi'cons were loaded onto linear gradients of sucrose (5-20",, w/v) in 0.05:!, Triton X-100, 10 m M Tris-HC1, pH 8.0 in ( A ) HzO and centrifuged for 15 h a t 35000 rev./min or in ( B ) HzO/'HzO (5195, v/v) and centrifuged for 20 h a t 35000 rev./min. Centrifugation was done at 20°C using a Beckman SW41 rotor. Gradients were fractionated into 25 fractions and hydrogenase activity ( 0 ) expressed as the rate of reduction of benzyl viologen, was measured in each fraction. Sucrose concentrations across the gradient (- ~ - - )were measured in each fraction. (B) A sample of a Triton X-100extract containing [3zP]phospholipids(5x 104counts . min-' pmol-') was centrifuged as in (A). Radioactivity (0) was measured in each fraction. The migration positions of standard proteins, centrifuged as described for hydrogenase, are shown ~

32P is centrifuged in a sucrose gradient containing 0.05'?:, Triton X-100 (Fig. 1 A) there is no peak of radioactivity associated with hydrogenase and greater than 98 of all radioactivity is found at the top of the gradient. This indicates that there is no phospholipid associated with the hydrogenase. The Stokes' radius of hydrogenase in the presence of Triton X-100 was determined by gel filtration using Sepharose 6B. Hydrogenase was eluted with buffer containing 0.05% Triton X-100, as shown in Fig.2A and the value calculated for the Stokes' radius is 5.52 nm (Table 1). Using the calculated values for partial specific volume, sedimentation coefficient and Stokes' radius,

Hydrogenase solubilised in Triton X-100 was centrifuged in a sucrose density gradient with either HzO or ' H 2 0 as solvent and no detergent in either gradient. From the profiles of hydrogenase activity in HzO and 'HzO gradients (Fig.3A and B) the partial specific volume and sedimentation coefficient were calculated [17] to be 0.74 m l . g-' and 15.9 S, respectively (Table 1). On centrifugation of membranes solubilised with Triton X-100 and labelled with [32P]phospholipidsin a gradient of sucrose containing no Triton X-100, the radioactivity associated with [32P]phospholipids is found at the top of the gradient (Fig.3A) and no significant radioactivity is associated with hydrogenase. The profile of solubilised hydrogenase on elution from Sepharose 6B with 10 mM Tris-HC1, pH 8.0 and no detergent, is shown in Fig. 2 B. The Stokes' radius of the main hydrogenase peak is 6.72 nm (Table 1). The minor peak is likely to be due to incomplete removal of detergent. The molecular weight and frictional ratio of hydrogenase in the absence of detergent have been computed from the values of partial specific volume, sedimentation coefficient and Stokes' radius. The molecular weight is 466000 and the frictional ratio is 1.2 (Table 1). Molecular Weight of Subunits of Hydrogenuse

Samples of each fraction over the hydrogenase activity peak from sucrose density gradient centrifugation in the presence of Triton X-100 (e.g. Fig.1A) have been subjected to electrophoresis on polyacrylamide gels in the presence of dodecylsulphate. A scan of a gel of the peak activity fraction is shown in Fig. 4A. The intensity of Coomassie blue staining of each of the bands over the region of hydrogenase activity has been compared with the profile of hydrogenase activity. As shown in Fig.4B, the peak of staining intensity of a band of molecular weight 63400 (designated as band 2) coincides with the peak of hydrogenase activity (Fig. 4A). Another band, of apparent molecular weight 30400 (band 4), also has a peak of staining intensity close to, but not matching exactly, that of hydrogenase activity. Hydrogenase from P. denitr(jicuns has been purified on an analytical scale by extraction of membranes with Triton X-100, sucrose density gradient centrifugation and polyacrylamide gel electrophoresis in nondissociating conditions followed by polyacrylamide gel electrophoresis in the presence of dodecylsulphate, as described in Methods.

E. Sim and R. B. Sim

123

Fig.? Gc~lfilrr-utionof'liydr-ogenuse.Samples (5 ml containing 30- 50 mg protein) of a Triton X-100 extract of membranes of P. d~,7itr-ifi'c~rrns were applied to a column of Sepharose 6B equilibrated at 20°C with either (A) 0.05 '2)(w/v) Triton X-100 in 10 m M Tris-HCI pH 8.0 or ( B ) 10 m M Tris-HCI pH 8.0 and eluted with the equilibration buffer. Fractions of 3 ml were collected and hydrogenase activity. measured as the rate of benzyl viologen reduction, was determined in each fraction ( 0 ) .The absorbance at 230 nm is shown by the broken line. The void volume ( VO)and total permeable space ( V J were determined as described in the text. The elution positions of standard proteins under both sets of conditions are indicated by arrows

- 25 10

-20

._ L

.

-

E

I

15

Y)

a "7 ,

c c 3

5

-10;

s

e

a 0 N

- 5

?

0

0

10

15

20

25

v

Fraction number

Fig. 3 . SII(~I.O.SL, tle/7sitv gradient ~ m / r - i / u p / i o in n tlzr uh.wnc,e o/rle/c~rgent.Samples (0.5 ml, 3-5 nig protein) of a Triton X-100 extract of membranes of P. denitr~ficonswere loaded onto linear gradients of sucrose (5-20x;. wiv) in 10 m M Tris-HCI. pH 8.0 in ( A ) HzO and centrifuged for 14 h at 30000 rev./min or (B) HzO/'HzO (5195, v/v) and centrifuged for 16 h at 30000 rev./min. Centrifugation was done at 20 C using a Beckman SW41 rotor. Gradients were fractionated into 25 fractions and hydrogenase activity ( 0 ) .expressed as rate of reduction of benzyl viologen, was measured in each fraction. Sucrose concentrations across the gradient were measured in each fraction and are shown by the broken line. A sample of a Triton X-100 extract containing [3ZP]phospholipids( 5 x lo4 counts . min pmo1-l) was centrifuged as in (A). Radioactivity (0)was measured and is expressed per fraction. The migration positions of standard proteins under both sets of conditions are indicated by arrows

After electrophoresis in non-dissociating conditions, a single band is found on staining for hydrogenase activity. The hydrogenase activity band has an R F value of 0.28 T 0.03 (SD of five determinations) compared to the tracking dye (bromphenol blue). Subsequent staining of the gel with Coomassie blue has permitted identification of the protein band which corresponds to hydrogenase. A scan of the gel after staining for protein is shown in Fig.5 and the band which coincides with the hydrogenase activity stain is indicated. When the band of hydrogenase activity is excised and itself subjected to electrophoresis on polyacrylamide gels in the presence of dodecylsulphate a single band of apparent molecular weight 63000 is seen on

staining with Coomassie blue (Fig. 6). The scan shown in Fig. 6 is of a gel loaded with a reduced and alkylated gel slice containing hydrogenase. No change was found if the gel slices containing hydrogenase were only alkylated prior to electrophoresis in the presence of dodecylsulphate. DISCUSSION

As summarised in Table 1, in the presence of Triton X-100, the sedimentation coefficient of hydrogenase from P. denitvificans is 10.4 S and the Stokes' radius is 5.52 nm. The molecular weight is calculated to be 242 500. It has been demonstrated [33] that centrifugation of a complex between a protein and a non-ionic

124

Hydrodynamic Parameters of Hydrogenase from P. denirrifi'cuns A

1.5

t

6 10

lo

r

vi

1

m

A

110

B

I

I

c

-

L 3 I

-

E

L

m ._ a

-

3.

-

x

0.5

c .._

z

L

m

5-

.,..';L"",.

.A

/ p a \

m

m Ln

- 5;.

. \.

.

c c a

c ._

m m

e

0 Migration

I

1

1

I

m L .c ._ m

n

-

5-

I

-

Ln ._

L

z

0

I

of benzyl viologen ~,.10-3

100 75

0.4 0.07:

i

E

c

x m

I

I

I

50

25

I

I

0

m

~

0.2

u

c

0.05C

J2 m

5: m

i

0

4 J2

L

m

a u c m

e

Lo

0

2

4

6

8

10

9

0.025

Migration distance (cm)

Fig. 5 . ~ ' o l ~ u c ~ i gel ~ ~ rl~ctroplio~c~.sis i i ~ n ~ ~ l ~ ~of lijdrogmuse. A sample (200 pl) of a pool of hydrogenase from sucrose density gradient centrifugation (under conditions of Fig. 1 B) was subjected to polyacrylamide gel electrophoresis without dodecylsulphate and stained for hydrogenase activity with benzyl viologen and subsequently stained for protein with Coomassie blue. A scan of the gel stained for protein is shown and the position of the single band of hydrogenase activity stain is indicated by the arrow

detergent through a sucrose gradient containing no detergent results in removal of detergent from the protein. Under similar conditions, hydrogenase has a sedimentation coefficient of 15.9 S. The Stokes' radius of hydrogenase measured from gel filtration in the absence of Triton X-100, is 6.72 nm and the molecular weight is 466000 (Table 1). Therefore, on removal of detergent, hydrogenase undergoes a molecular weight increase of approximately twofold.

0 0

2

4

6

8

10

Migration distance (crn)

Fig. 6. D O d ~ ~ C y ~ . S u ~ l l U I l ' i / ) O l J ' U ( . I . ) ' /g1.l ~llm el~~ct~o/Jhore.~;.S ;~/~~ Ofpurifi'cV/ hydrogenase. The region of a polyacrylamide gel which stained for hydrogenase activity, as indicated by the arrow in Fig.5, was reduced and alkylated and subjected to electrophoresis on 5.6'?;, polyacrylamide gels with dodecylsulphate. This gel was then stained for protein with Coomassie blue and a scan at 550 nm is shown. The molecular weight scale was determined from standard proteins run in parallel

An estimate of the amount of non-ionic detergent bound to a protein-detergent complex can be made from knowledge of the partial specific volume of the

E. Sim and R. B. Sim

I25

Table 1. Sumnzury of molerulur properties of hydrogenuse in the presence and absence of Triton X - 100 Results are expressed as the mean value T standard deviation. The number of independent estimates is shown in parenthesis. Where no error limits are given results are the average of two determinations. Partial specific volumes, 1'2, and sedimentation coefficients, .s20.u. were calculated according to Clarke [I71 using Eqns (14) and ( 13) therein, respectively. Stokes' radii were calculated according to Ackers [23]. Molecular weights, M,,were calculated using the . ~ N A (xi method of Siege1 and Monty [24] i.e. M , = ~ 2 0 6nqzo.,., ( 1 - 1 . 2 ~ 2 ( , . ~where ) N A is Avogadro's number, z is the Stokes' ~ density radius, rlzo.w is the viscosity of water a t 20°C and ~ 2 0 is. the of water at 20 C. Frictional coefficients I f i f i ) ) were calculated from the equation given by Neer [7] i.e. 4n

.fif;, = cl I 3 M r (!,2

NA

+

)I

I .J

where 6 = 0.2 g of solvent per g of protein [7] ~~

~

~

Parameter

1'2

.Szo.u

Stokes' radius Mr ,/;/;I

Apparent M , of subunits Apparent number of subunits

Unit

ml . g-' S nm -

Value with Triton X-100

without Triton X-100

0.73 k 0.02 (4) 10.4 i O . 2 (4) 5.52 i 0.16(4) 242 500 i 8000 1.2

0.74 15.9 6.72 k 0.19 ( 3 ) 466000 k 15000 1.2

63000 f 500 (8)

-

4

7-8

complex (51. This method relies on the detergent having a lower density, i.e. higher partial specific volume, than a typical protein. The partial specific volume of Triton X-100 is 0.91 ml . g-' [33] whilst the range for most proteins is between 0.71 and 0.76 ml . g-I. More precise measurements of detergent binding can be obtained by using radioactive detergent and pure protein. However, such an approach was not possible in this study since pure hydrogenase was not available in sufficient quantity. All hydrogenases which have been described are non-haem-iron proteins (see [34] for review) and it could be argued that the metal content is likely to alter the partial specific volume compared with other proteins. Although this is the case for ferredoxin from Clostridium pasteurianum which has eight iron atoms per polypeptide chain of molecular weight 6000 and a partial specific volume of 0.63 ml . g-' [35], this is not so for larger iron-sulphur proteins. For example, the hydrogenase from Desuljbvibrio gigus has 12 iron atoms per molecule of molecular weight 89000 and the partial specific volume is 0.73 ml . g-' [36]. As shown in Table 1, the partial specific volume of hydrogenase from P. denitrijicans, determined by sucrose gradient centrifugation in 'H20 and H20 is 0.73 ml . g-' in the presence of detergent and so is

within the range for a 'typical' protein. Measurement of the partial specific volume of hydrogenase in the absence of Triton X-100 (Fig.3 and Table 1) gives a value of 0.74 ml . g - * . Thus whether in the presence or absence of detergent, the values found for the partial specific volume of hydrogenase are essentially identical, and within the range for most proteins. Removal of detergent, however, does cause a profound change in the association properties of the enzyme, resulting in dimerisation of the 242500-M, species. This change in molecular weight on removal of detergent indicates that hydrogenase interacts with Triton X-100 but the measurements of partial specific volume show that the quantity of Triton X-100 bound is small and is likely to represent only a few Triton X-100 monomers. The fact that the molecular weight of hydrogenase increases by a factor of 1.93 rather than 2, on removal of detergent, although within experimental error, may be accounted for by the binding of Triton X-100 monomers to the 242 500-M, species. The differences between the protein in the presence and absence of Triton X-100 are not due to phospholipid binding since there is negligible phospholipid associated with hydrogenase on sucrose density gradient centrifugation either in the presence (Fig. 1 A) or in the absence (Fig. 3 A) of detergent. The formation of aggregates of specific size on removal of detergents has been observed for other membrane proteins [37,38] and it has been proposed that this is a general property of lipid-binding membrane proteins 1391. It has been shown that the enzymic properties of hydrogenase are not significantly altered on solubilisation by Triton X-100 [40]. Therefore the solubilised form of the enzyme is likely to have a similar structure to the membrane-bound form. We propose that hydrogenase exists in the membrane with an apparent molecular weight of 242500 and that, on solubilisation, detergent replaces lipid. When detergent is removed, two such molecules of hydrogenase associate to form a water-soluble aggregate presumably via interaction of restricted hydrophobic areas previously masked by detergent. Since hydrogenase is almost spherical, both in the presence of detergent and after aggregation, in as far as this can be determined from the frictional ratio (Table l), no proposal can be made as to the relative orientation of the molecules in the aggregate. It has been demonstrated by two methods that hydrogenase activity is associated with a polypeptide of apparent molecular weight 63 000 as determined by electrophoresis on polyacrylamide gels in the presence of dodecylsulphate. The Coomassie blue staining intensity of this protein band exactly parallels the profile of hydrogenase activity after sucrose density gradient centrifugation (Fig. 4B). The possible contribution of a band of lower molecular weight (Fig.4, band 4) has been eliminated by purification of hydro-

126

E. Sim and R. B. Sim: Hydrodynamic Parameters of Hydrogenase from P. denitrtfi'cans

genase on an analytical scale which results in a single band of molecular weight 63000 as seen on polyacrylamide gel electrophoresis in the presence of dodecylsulphate, both with and without reduction of the purified hydrogenase with dithiothreitol. The species of hydrogenase of molecular weight 242 500 is likely to consist of four non-covalently linked subunits, each of apparent molecular weight 63000. These results are summarised in Table 1. Many soluble hydrogenases have been shown to have molecular weights around 60000, including the soluble hydrogenase from Clostridium pasteurianiurn [41] and D. vulgaris strain Hildenborough [42]. It has also been reported that membrane-bound hydrogenases from Chromatiurn [12]. T. roseopersincina [lo], A . c.urrophus HI6 [12], D. gigus [36] and Proteus mirahilis [43]have subunits ofmolecular weight arpund 60000. For the last three enzymes the existence of other subunits of molecular weight around 30000 have been indicated. The hydrogenase from P. miruhilis, which was extracted from bacterial membranes with amyl alcohol at - 4 "C and subsequently chardcterised in the absence of detergent, behaves as a multimer of molecular weight 200000. However, determination of the evolutionary significance of the occurrence of a subunit of molecular weight around 60000 in hydrogenases from diverse bacteria must await amino acid sequence data. We are grateful to D r B. Osborne for a gift of 'HzO and for his interest and helpful discussion, to D r P. M. Vignais in whose laboratory the work was carried out and to Drs G. S. Schoenmaker, B. Schink and Prof. H. G. Schlegel for supplying results prior to publication. We also thank Mrs G. Kelley for excellent technical assistance. E. Sim was in receipt of a Royal Society European Exchange Fellowship and R. B. Sim was the holder of Medical Research Council/Institut National de la Sante et de la Recherche Mc~tlicalFrench Exchange Fellowship. This research was supported in part by grants from the C.N.R.S. (A.T.P. PhotosyntllPse) and the E.E.C. Solar Energy Research and Dcvelopemcnt Programme (Contract N". 538-78 ESF).

REFERENCES 1. Kluyver, A. J . & Verhoeven, W. (1954) Antonie van Leeuwen-

hoek J . Micwihiol. Serol. 20, 337- 358. 2. Schneider. K . & Schlegel, H. G. (1977) Arch. Microhiol. 112, 229-238. 3. Sim, E. & Vignais, P. M. (1978) Biochimie (Paris) 60,307-314. 4. Sim, E., Colbeau, A. & Vignais, P. M. (1978) Hydrogenases: Their. Catalj,tic, Activit-v, Structure and Function (Schlegel, H. G. & Schneider, K., eds) pp. 269-280, E. Goltze, Gottingen. 5 . Meunier, J. C., Olsen, R . W. & Changeux, J. P. (1972) FEBS L.ctt. 24, 63 - 68. 6. Neer. E. J. (1974) J . Bid. Cliem. 249, 6527-6531. 7. Neer, E. J . (1978) J . Biol. Chem. 253, 1498- 1502. 8. Smigel, M. & Fleischer, S. (1977) J . Biol. Chrm. 252, 36893696. 9. Adams. M. W. W. & Hall, D. 0. (1977) Bioclirm. Biophvs. RPS. Commun. 77,730- 137. 10. Serebryakova, L. T., Zorin, N. A. & Gogotov, I. N. (1976) Biokhimiya, 42, 571 - 575.

11. Gogotov, 1. N., Zorin, N . A,, Serebriakova, L. T. & Kindrdtieva, E. N. (1978) Biohim. Biophys. Ac,ta, 523, 335- 343. 12. Schink. B. & Schlegel. H. G. (1979) Biochim. Bioplq's. Acta, 567, 315-324. 13. Gitlitz, P. H. & Krdsna, A. I. (1975) Biochemistry, 14, 2561 2567. 14. Kakuno, T., Kaplan, N. 0. & Kamen, M. D. (1977) Proc,. Natl Acud. Sci. U.S.A. 74, 861 - 863. 15. Schink, B. & Schlegel, H. G. (1978) Biochimie (Paris) 60. 297 - 306. 16. Kirshenbaum, I. (1951) Pliysical Propertic,s and Analysis of Heavy Water, chap. 1, McGraw-Hill, New York. 17. Clarke, S. (1975) J . B i d . Cliem. 250, 5459-5467. 18. Racker, E. (1 955) method.^ Enzymol. I , 500 501. 19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J . (1951) J . Bid. Chrm. 193, 265-275. 20. Makino, S., Reynolds, J. A. & Tanford, C . (1973) J . Biol. Chrm. 248.4126-4132. 21. Sim, R. B., Porter, R. R., Reid, K. B. M . & Gigli, I . (1977) B i o ~ l i ~ Jf i. ~163, . 21 9 - 227. 22. Damn, H. C., Besch, P. K., Couri, D. & Goldwyn, A. J . (eds) (1974) The Handbook of BiochrmistryL.,pp. 67 -69. World Publishing Co., New York. 23. Ackers, G . K. (1 964) Biochmzistry, 3, 723 730. 24. Siegel, L. M. & Monty, K . J. (1966) Biochim. Biophys. Acra. 112, 346 - 362. 25. Weber, K. & Osborn, M. (1969) J . Biol. Cliem. 244,4406-4412. 26. Fairbanks, G., Steck, T. L. & Wallach, D. F . H. (1971) Biochemistry, 10, 2602 - 261 7. 27. Omstein, L. & Davis, B. J. (1964) Ann. N . Y . A m d . Sci. 121. 321 - 349. 28. Anderson, N. G. (1972) Handbook of Bioclicmistry, 3rd edn, pp. 3248 - 5251, Chemical Rubber Company, Cleveland, Ohio. 29. O'Brien, R. D., Timpone, C. A. & Gibson, R. E. (1978) Anal. Biochcw. 86, 602 - 61 5 . 30. Folch, J., Lees, M. & Sloane-Stanley, G. H. (1957) J . Biol. Cliem. 226,497 509. 31. Bartlett, G. R. (1959) J . Biol. Chem. 234, 466-468. 32. Patterson, M. S. & Greene, R. C. (1965) Anal. Chem. 37, 854- 857. 33. Helenius, A. & Simons, K. (1975) Biocliim. Biophys. Acta, 415, 29 - 79. 34. Chen, J.-S. & Mortenson, L. E. (1974) Microhiul Iron Metaholism (Nielands, J. B., ed.) pp. 231 -280, Academic Press, New York. 35. Buchanan, B. B. & Arnon, D. I . (1970) Adv. En:ymol. 33. 119- 176. 36. Hatchikian, E. C., Bruschi, M. & Le Gall, J . (1978) Biodirm. Biophys. Res. Commun. 82, 451 -461. 37. Helenius, A. & von Bonsdorff, C.-H. (1976) Biochim. Biopl7y.s. Acta, 436, 895 - 899. 38. Kuchel, P. W., Campbell, D. G., Barclay, A. N. & Williams, A. F . (1978) Biochem. J . 169, 411 -417. 39. Tanford, C. & Reynolds, J. (1976) Biochim. Biophys. Actu, 457, 133- 170. 40. Sim, E. & Vignais, P. M. (1979) Biochim. Biopl7j.s. Acta, in the press. 41. Chen, J.-S. & Mortenson, L. E. (1974) Biochim. Biophys. Acta, 371, 283-298. 42. Van der Westen, H. M., Mayhew, S. G. & Veeger, C. (1978) FEBS Lett. 86, 122 - 126. 43. Schoenmaker, G. J., Oltman, L. F.& Stouthamer, A. H. (1978) Hyrirogenases: Their Cata/.vtic Activity, Structurc and Func.tion (Schlegel, H. G. & Schneider, K., eds) pp. 209-219, E. Goltze, Gottingen. -

-

-

-

E. Sim and R. B. Sim, Department of Biochemistry, University of Oxford, South Parks Road, Oxford, Great Britain, OX1 3QU