Expression and assembly of spectrally active recombinant ...

Proc. Natl. Acad. Sci. USA Vol. 88, pp. 10387-10391, December 1991 Biochemistry

Expression and assembly of spectrally active recombinant holophytochrome (plant photoreceptor/phytochrome biosynthesis/yeast/Eschenchia coli)

JILL A. WAHLEITHNER, LIMING LI, AND J. CLARK LAGARIAS Department of Biochemistry and Biophysics, University of California, Davis, CA 95616

Communicated by Winslow R. Briggs, August 9, 1991

translation fails to yield sufficient material for structural and/or spectrophotometric analyses (4). Here we show that the recombinant apophytochromes produced in both yeast and a bacterium are soluble and full-length and can assemble with bilins to produce spectrally active holophytochromes.

To develop an in vitro phytochrome assembly ABSTRACT system, we have expressed an oat phytochrome cDNA in both the yeast Saccharomyces cerevisiae and the bacterium Escherichia col. Analysis of soluble protein extracts showed that the recombinant apophytochromes were full-length and capable of covalently attaching the phytochrome chromophore analogue phycocyanobilin. Difference spectra indicated that in vitroassembled holophytochrome species were photoreversible; however, maxima and minima difference absorption values were blue-shifted relative to those of the native photoreceptor. Extracts containing the recombinant apophytochromes were also incubated with phytochromobilin, the natural chromophore synthesized from biliverdin by cucumber etioplast preparations. In these experiments, the difference spectrum obtained was identical to that of native oat holophytochrome. These results suggest that the recombinant apophytochromes adopt a structure similar to that of the apoprotein biosynthesized in vivo. ELISAs were used to quantitate phytochrome expression levels in both yeast and E. coil extracts. These measurements show that 62-75 % of the phytochrome apoprotein in the soluble protein extract was competent to assemble with bilins to form spectrally active holophytochrome.

MATERIALS AND METHODS Plasmids. Standard protocols utilizing Escherichia coli DH5a were used for all clone constructions (7). An E. coli expression plasmid containing the oat apophytochrome phyA3 coding region (8) was constructed as follows. Plasmid pPC3 (4), which contains the full-length oat phyA3 cDNA, was linearized with Pvu I. Plasmid pAQE58 (9), which contains the genes for the a and 83 subunits of phycocyanin from Synechococcus sp. PCC 7002, was linearized with BstEII. An equal molar ratio mixture of the two linearized plasmids was then treated with mung bean nuclease to remove single-stranded overhangs, digested with Kpn I, ligated, and transformed into E. coli. From the resulting transformants, clone pAQ3'PC-18 was selected. This clone contained a 3.2-kilobase Kpn I-Pvu I fragment of pPC3, comprising the 3' end of the phyA3 cDNA, which was inserted into the Kpn I-BstEII sites of pAQE58. In this construction, phyA3 sequences were inserted within the phycocyanin operon replacing most of the a and ( subunit genes. To construct a full-length phyA3 cDNA clone, the plasmid pAQ3'PC-18 was digested with the enzymes Sma I and Nsi I. Plasmid pPC3 was also digested with Xba I and the overhangs were removed with mung bean nuclease and then digested with Nsi I. A 7.8-kilobase-pair Sma I-Nsi I fragment from pAQ3'PC-18 and a 2492-base-pair Xba I-Nsi I fragment from pPC3 containing the 5' end of the phytochrome gene were gel purified and ligated together. After transformation, the plasmid pAQPC-5 (L27-5), which contained the complete phytochrome coding region, was isolated. For the final cloning step, a 5.0-kilobase HindIII-EcoRI fragment derived from plasmid pAQPC-5, which contained the entire phytochrome coding region, flanking cyanobacterial phycobiliprotein promoter, and transcription termination signals, was cloned into the HindIII-EcoRI sites of pGEM-4. The resulting oat phytochrome construct, pGphyA3 (Fig. 1A), should express apophytochrome as a full-length polypeptide, not a fusion protein, since a stop codon lies between the phycocyanin and phytochrome translation initiation sites. A control plasmid, pGphy-10, was also constructed. This construct lacks a 900-base-pair HindIII-BamHI fragment that contains the promoter sequences of the phycobiliprotein operon. A yeast expression plasmid containing the oat phyA3 sequences was prepared by insertion of the phytochrome coding region from plasmid pAQPC-5 (described above) into the yeast-E. coli shuttle vector pMAC105, a derivative of the yeast expression vector pAC1 (10). The plasmid pMAC105

The effect of light on many growth and developmental processes of plants is mediated by the photoreceptor phytochrome (1). The functionally active photoreceptor molecule is comprised of a large apoprotein of =1100 amino acids to which the linear tetrapyrrole (bilin) chromophore phytochromobilin (P46B) is covalently bound (2). Synthesis of the holoprotein, therefore, involves the convergence of two biosynthetic pathways-one for the apoprotein and the other for the chromophore. Holophytochrome can be assembled in vitro by the incubation of plant-derived apophytochrome preparations with purified bilins (3). In addition, an in vitro transcription and translation system has been used to show that holophytochrome assembly is autocatalytic, requiring only apophytochrome and a free bilin (4). In vitro synthesis of phytochrome is a useful tool for the structural and functional analysis of this important photoreceptor. For this reason, the objective of this study was to produce apophytochrome using recombinant expression in yeast and bacteria. Both of the presently available experimental systems for apophytochrome production have their limitations. Plant-derived apophytochrome preparations contain significant levels of spectrally active holophytochrome due to incomplete inhibition of chromophore synthesis by tetrapyrrole synthesis inhibitors (3, 5). Apophytochrome heterogeneity also arises from the expression of multiple phytochrome genes (6) and from posttranslational and artifactual posthomogenization modifications of the protein. Apophytochrome produced by in vitro transcription and The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: PCB, phycocyanobilin; P4B, phytochromobilin. 10387

10388 A

Biochemistry: Wahleithner et al.

Proc. Natl. Acad. Sci. USA 88 (1991)

Xn AIz-l

Hindldl

a

with phytochrome sense and antisense orientations were named pMphyA3 and pMphy-11, respectively. The structure of plasmid pMphyA3 is shown in Fig. 1B. Apophytochrome Expression and Extraction from E. coli Cells. Single colonies containing either the sense (pGphyA3) or the control (pGphy-10) plasmids were inoculated into 5 ml of 2YT medium [Bacto tryptone (16 g/liter)/Bacto yeast

i < < ~~~~~Kpril!

/~ ~~~

extract (10 g/liter)/NaCl (5 g/liter)] containing ampicillin (100 ,ug/ml) and grown for 5 hr at 37TC. The cultures were then added to 750 ml of 2YT medium plus ampicillin (100 ,ug/ml) and shaken vigorously at room temperature until the OD6w was 1.0. These cultures were chilled to 40C and centrifuged for 5 min at 2000 x g at 40C. The cell pellet was washed with 300 ml of 20 mM Tris'HCl (pH 8.0) containing

pGphyA3 7750 bp

>

l0-arll~l +'-~~~X~ DC"A~

~ ~/ ~

C

i/I-/

TNCA

---

Ball I

FIG. 1. E. coli (A) and yeast (B) expression plasmids containing the full-length oat phytochrome cDNA sequences. For both plasmids, phytochrome sequences are indicated with open boxes. Promoter initiation sites and direction of transcription are indicated with an arrow. ATG translation start and TGA stop codons for phytochrome are indicated. The shaded boxes on the E. coli plasmid pGphyA3 (A) represent sequences originating from the cyanobacterium Synechococcus sp. 7002 (9). Sequences on the yeast expression plasmid pMphyA3 (B) that are derived from the yeast enoI structural gene including the "TATA box" and transcription initiation and termination sites are labeled with shaded boxes. The solid black box on this plasmid denotes the yeast galactose-inducible upstream activating sequence (GAL UAS). Selectible marker genes, origins of plasmid replication, and selected restriction sites are also indicated on both plasmids.

chosen to place phytochrome expression under regulatory control of the yeast enoI promoter and a galactosewas

inducible upstream activator sequence (GAL UAS). The unique HindIII site in pMAC105 was first converted to a BamHI site by insertion of a 12-base-pair BamHI linker and the

resulting plasmid

was

named

pMAC105(H

-*

B). A

gel-purified 4.0-kilobase-pair BamHI fragment containing the phytochrome sequences was then excised from plasmid pAQPC-5, gel-purified, and ligated to BamHI-linearized pMAC1O5(H

--

B). Clones containing both the

sense

and

antisense orientations of apophytochrome sequences were identified by restriction enzyme analysis and both plasmids were subsequently isolated. The yeast expression vectors

20 mM NaCl and 1 mM EDTA. Washed cell pellets were frozen in liquid nitrogen and stored at -80°C until needed. All subsequent steps of the protein isolation were performed at 4°C or on ice. Frozen cells from a single 750-ml culture were thawed by resuspension in 1:3 (wt/vol) homogenization buffer [50 mM Tris HCl, pH 7.2/0.1 M NaCl/1% dimethyl sulfoxide/1 mM EDTA/1 mM EGTA/2 mM phenylmethylsulfonyl fluoride/1 mM benzamidine/leupeptin (1.5 ,ug/ ml)/10 mM dithiothreitol] and immediately lysed by two passages through a French pressure cell at 10,000 psi (1 psi = 6.9 kPa) using a cell with a 3/8-inch diameter (1 inch = 2.54 cm). The lysate was cleared by centrifugation for 15 min at 22,000 x g and 0.23 g of (NH4)2SO4 was added per ml of the resulting supernatant to precipitate the apophytochrome. After incubation for 10 min, the precipitate was collected by centrifugation for 30 min at 22,000 x g. The resulting pellet was dissolved in 0.25 ml of TEGE [25 mM Tris-HCl, pH 7.2/2 mM EDTA/25% (vol/vol) ethylene glycol/2 mM phenylmethylsulfonyl fluoride/1 mM dithiothreitol] per g (fresh cell weight). This protein solution was cleared by ultracentrifugation for 30 min at 200,000 x g and directly used for holophytochrome assembly experiments. Apophytochrome Expression and Extraction from Yeast Cells. Sense (pMphyA3) and antisense (pMphy-11) plasmids were transformed into lithium acetate-treated Saccharomyces cerevisiae 29A (MATa leu2-3 leu2-112 his3-AJ adel-101 trpl-289) using published protocols (11). Single colonies were inoculated into 3 ml of liquid medium containing 0.67% yeast nitrogen base without amino acids, 2% (wt/vol) galactose, and 0.02% amino acid "drop out" mixture (minus leucine) and grown for 2-3 days at 30°C. Two of these cultures were then combined, transferred to 1 liter of the same medium in a 2-liter Fernbach flask, and shaken vigorously at 30°C until the cells reached an OD580 of 0.5-1.5. This typically took 2-3 days. The cells were harvested by centrifugation for 5 min at 500 x g. All subsequent steps of the protein isolation were performed at 4°C or on ice. Cell pellets were resuspended in ice-cold distilled water, recentrifuged, and then resuspended in 1.25 ml of homogenization buffer per g (fresh cell weight). The cell suspension was transferred into chilled 2-ml polyethylene vials (Biospec Products, catalogue no. 1083, Bartlesville, OK) containing 0.5 vol of acid-washed 0.5-mm glass beads. Cell lysates were prepared using five 1-min pulses with a Mini-bead beater (Biospec Products) with cooling on ice between each pulse. The crude homogenate was cleared by ultracentrifugation for 1 hr at 100,000 x g and 0.23 g of (NH4)2SO4 was added per ml of the resulting supernatant. After a 15-min incubation, the precipitate was collected by centrifugation for 30 min at 16,000 x g. The resulting pellet was resuspended in 1.0-1.5 ml of TEGE per g (fresh cell weight). The apophytochrome preparation was either directly used for holophytochrome assembly experiments or frozen in liquid nitrogen and stored at -80°C. Holophytochrome Assembly. The following procedures were performed under a green safelight (12). Apophytochrome preparations from E. coli and yeast were divided

into three 200-,ul samples. One sample [termed + phycocyanobilin (+PCB)] was diluted with 400 /l of TEGE or assembly buffer (20 mM Tes/10 mM Hepes-NaOH, pH 7.7/500 mM sorbitol/1 mM phenylmethylsulfonyl fluoride/ 0.5 mM dithiothreitol/2 ,uM leupeptin) to which PCB was added from a 1.0 mM stock solution in dimethyl sulfoxide to give a final concentration of 4 ,uM PCB. A second sample (+P4B) was diluted with 400 Al of a P4)B-containing mixture in assembly buffer. The P4B solution was obtained after a 30-min incubation of cucumber etioplast preparations with biliverdin and a NADPH-regenerating system (13). The third sample (-PCB control) was diluted with 400 /l of TEGE or assembly buffer only. All three samples were then incubated at 280C for 30 min. After incubation, the suspensions were cooled to 40C and clarified by ultracentrifugation for 15 min at 200,000 x g prior to holophytochrome photoassay. SDS/PAGE, Zinc-Blot, and Immunoblot Analyses. SDS/ PAGE analyses were performed using 1- or 1.5-mm-thick minigels according to Laemmli (14). After electrophoresis, proteins were transferred to poly(vinylidene difluoride) membranes (Immobilon P, Millipore) for 1 hr at 100 V. After soaking the membranes in 100 ml of 1 M Zn(OAc)2 for 5-30 min, transblotted biliproteins were visualized with a Fotodyne UV transilluminator (model 3-3000) and photographed with Technical Pan film (Kodak type 4415) using a Schott RG-630 red cutoff filter and a 2-min exposure (15, 16). The same membranes were immunostained as specified by Birkett et al. (17) using the following sequence of antibodies: 10 antibody, affinity-purified oat phytochrome polyclonal rabbit antibody (52 ng/ml) (18); bridging antibody, goat anti-rabbit IgG fraction (Boehringer Mannheim, catalogue no. 605-200, 1:7000 dilution); 20 antibody, swine anti-goat IgG alkalinephosphatase conjugate (Boehringer Mannheim, catalogue no. 605-280, 1:7000 dilution). Blots were typically developed for 3-10 min. After immunostaining, blots were also stained with Coomassie blue according to instructions of the poly(vinylidene difluoride) membrane manufacturer. Holophytochrome Photoassay. Holophytochrome concentrations were estimated using an absorbance difference assay with a HP8450A UV-visible spectrophotometer (19). For PCB experiments, a 636-nm interference filter was substituted for the 660-nm filter because of the blue-shifted difference maxima of the PCB-apophytochrome adduct (3). The concentration estimate for the PCB-apophytochrome adduct was made with the assumption that the molar absorption coefficients and photoequilibrium values were identical with those of native oat phytochrome except for the position of the absorption maxima. Quantitative Immunoassays. Phytochrome concentrations were estimated using a double-antibody sandwich ELISA protocol (20) with the following modifications. A 96-well flat-bottom microtiter plate (Coming no. 25805-96) was coated with 50 jml of affinity-purified oat phytochrome polyclonal rabbit antibody (5 ,g/ml) (18) in 50 mM Na2CO3 (pH 9.6) for 2 hr at 4°C. Wells were then blocked with 200 ,l of 2% (wt/vol) bovine serum albumin in phosphate-buffered saline (PBS = 20 mM potassium phosphate, pH 7.4/150 mM NaCl) overnight at 4°C. Phytochrome-containing samples and all subsequent antibody incubation media were diluted with 1% bovine serum albumin in PBS/Tween (PBS/0.05% Tween 20). Samples of 50 ,l per well were- used for these incubations. Between each step, the wells were washed for three 10-min periods with PBS/Tween. Three dilutions of each sample were prepared, transferred to the appropriate antibody-coated well, and incubated for 2 hr at 4°C. Immunodevelopment involved incubation with Oat-25 or Pea-25 (1 ,g/ml), two phytochrome-directed mouse monoclonal antibodies (21), followed by incubation with a 1:7000 dilution of horseradish peroxidase-conjugated goat anti-mouse IgG (Boehringer Mannheim, catalogue no. 605-250) for 1 hr at

10389



room temperature. Color development utilized a freshly prepared solution (100 /l per well) of 0.01% 3,3',5,5'tetramethylbenzidine (added as a 1% stock solution in dimethyl sulfoxide) and 0.3% H202 in 0.1 M NaOAc adjusted to pH 6.0 with citric acid. After a 20-min incubation, the reaction was stopped by the addition of 25 /l of 0.2 M H2SO4 per well. Absorbance at 450 nm was read using a Molecular Devices kinetic microtiter plate reader. Unless otherwise indicated, measurements were performed in triplicate and each microtiter plate contained a standard dilution series of oat phytochrome of known concentration. Protein Assays. Protein assays were performed using the BCA method (Pierce) with bovine serum albumin as the protein standard (22).

RESULTS AND DISCUSSION Recombinant Apophytochrome Expression. Expression of full-length apophytochrome was the first objective of these studies. After constructing the expression plasmids pGphyA3 and pMphyA3 for E. coli and yeast, respectively (see Fig. 1), we examined protein extracts from transformed cell cultures for the presence of phytochrome polypeptides using two immunoassays, Western blot and quantitative ELISA. Western blot analyses were performed on whole-cell SDS extracts and on crude and (NH4)2SO4-fractionated soluble protein extracts. Fig. 2 A and B shows immunoblots of (NH4)2SO4-fractionated soluble protein extracts from yeast and E. coli cultures, respectively, that contain phyA3 sense plasmids (lane 6) or control plasmids (lane 8). These blots demonstrate that both recombinant systems produce apophytochromes with electrophoretic mobility similar to purified oat phytochrome (compare lanes 5 and 6). Similar results were obtained using whole-cell SDS extracts and crude soluble protein extracts from both organisms (data not shown). These results indicate that recombinant apophytochromes produced in both yeast and E. coli systems are predominantly full length. Zinc Blot A

1

2

3

Coomassie Blue

Immunoblot 5

6

7

8

9

10

11

12

180 116 - 84 58

48 36

56 7

8

9

0 1 1 12 ~~~~~~1

180 116 84 --58 48

FIG. 2. Zinc-blot and immunostaining analyses of phytochrome from recombinant strains of yeast (A) and E. coli (B). Soluble protein extracts were prepared from yeast and E. coli cultures containing sense (pMphyA3 and pGphyA3) and control (pMphy-11 and pGphy10) plasmids, respectively, and fractionated with (NH4)2S04. Two samples from each sense plasmid extract were removed and incubated with PCB or buffer only and analyzed by SDS/PAGE, zincblot, and immunostaining protocols. For the control plasmid experiments, only the PCB incubation was performed. Zinc-blot, immunostaining, and Coomassie blue-staining analyses for yeast (A) and E. coli (B) samples are as indicated. Lanes: 1, 5, and 9, purified oat holophytochrome; 2, 6, and 10, -PCB, sense construct; 3, 7, and 11, +PCB, sense construct; 4, 8, and 12, +PCB, control construct. Lanes 1 contain 98 ng (A) or 25 ng (B) of purified oat phytochrome. For yeast lysates 70 ,ug of protein or for E. coli lysates 87 ,ug of protein was added per lane. Molecular mass markers with sizes in kDa are indicated on the right. The arrow indicates the position of the 124-kDa phytochrome polypeptide.

10390



Results from the quantitative immunoassay analysis, summarized in Table 1, reveal that a significant amount of immunochemically cross-reactive apophytochrome is produced in both cell systems. Control experiments were performed to ascertain that the antigenicity of the recombinant apophytochromes was similar to that of phytochrome purified from etiolated oat seedlings, and that yeast and E. coli proteins did not interfere with immunoquantitation (data not shown). Under the respective experimental conditions used for cell culture, the level of apophytochrome expression in yeast cells is -3-fold greater than that found in E. coli cells on a per g fresh weight basis (Table 1). Immunochemical estimates of apophytochrome recoveries in soluble protein extracts after (NH4)2SO4 fractionation were 28% for E. coli and 33% for yeast. In both systems, these results were obtained under conditions of constitutive expression. In the vector pGphyA3, apophytochrome transcription is under the control of the cyanobacterial phycocyanin promoter that also functions in E. coli (23). Growth of the yeast cultures in galactose leads to constitutive expression of apophytochrome in cells containing pMphyA3. Although the yeast expression vector pMphyA3 contains regulatory sequences for induction of apophytochrome expression, such growth conditions did not lead to enhanced yields or improved recoveries of apophytochrome in the soluble protein fraction (data not shown). Holophytochrome Assembly. To determine if recombinant apophytochromes could covalently attach bilins, zinc-blot analyses were performed. In these analyses, polypeptides with covalently bound bilins can be visualized on blots in the presence of Zn2+ ions by UV transillumination (15, 16). These experiments were performed with the phytochrome chromophore analogue PCB, whose assembly with apophytochrome in vitro has been demonstrated (3, 4). Fig. 2 shows that a single orange-fluorescing band at 124 kDa appears in protein extracts prepared from both yeast and E. coli cultures that are expressing phyA3 sequences only when PCB was added [compare lanes 2 with lanes 3 for yeast (Fig. 2A) and E. coli (Fig. 2B) samples]. For both yeast and E. coli samples, the molecular mass of the zinc-dependent fluorescent band is identical to that of phytochrome isolated from oats (Fig. 2, lanes 1). This fluorescent band was not detected in the soluble protein fraction obtained from either yeast or E. coli cultures that were transformed with control plasmids and subsequently incubated with PCB (Fig. 2, lanes 4). These results demonstrate that both yeast and E. coli apophytochrome covalently attach to the bilin pigment PCB under our experimental conditions. After zinc-blot analysis, immunostaining of the -same membrane was performed. As was described above, these analyses confirm that apophytochrome extracted from both yeast and E. coli cells is predominantly full length (Fig. 2, lanes 6 and 7) and that the 124-kDa phyTable 1. Recombinant apophytochrome and

tochrome polypeptide is not found in the control extracts (Fig. 2, lanes 8). In addition, the lack of other zinc-dependent fluorescent bands in the control samples indicates that PCB attachment is specific for the recombinant apophytochrome. Indeed, the- Coomassie blue-stained blots reveal that both extracts contain many other proteins that do not covalently bind to this bilin (Fig. 2, lanes 10-12). To determine whether the interactions between both recombinant apophytochromes and bilins yield photoreversible holoproteins, the +PCB and -PCB samples described above were spectrophotometrically assayed for holophytochrome. The resulting difference spectra, shown in Fig. 3, indicate that the addition of PCB to the soluble-protein extracts from both the recombinant yeast and E. coli cell lines leads to the production of photoreversible holophytochrome adducts. For both systems, the maximum and minimum absorbance regions of the difference spectra, 650-652 nm and 716-720 nm, respectively, are blue-shifted from those of the native 0.008

0.004

0.000

-0.004

CD a) 1.4

-.4

-0.008

a)

0)

0.002 0

,0 0.001

0.000

-0.001

-0.002

holophytochrome yields Yield,

;Lg/g (fresh weight)

E. coli Yeast Source Apophytochrome 9.51 ± 0.% 28.54 ± 4.83 Whole cells 7.74 ± 0.67 11.01 ± 0.93 Soluble extract 9.50 ± 0.75 2.73 ± 0.19 (NH4)2SO4 fraction Holophytochrome 1.62 ± 0.35 7.1 ± 0.3 Spectral estimate 74.7 ± 7.3 62.3 ± 3.51 % ligatable Phytochrome recoveries were determined by ELISA. Each value represents an average of three experiments with at least three replicas of each data point. Holophytochrome recoveries were determined by spectrophotometric assay of the +PCB samples. Each value represents an average of three experiments.

500

600

700

800

wavelength (nm) FIG. 3. Holophytochrome photoassay of in vitro-assembled holophytochromes. Apophytochrome-containing protein extracts from yeast (A) and E. coli (B) cells containing sense plasmids pMphyA3 and pGphyA3, respectively, were prepared and fractionated with (NH4)2SO4. Spectrophotometric difference assays of samples incubated with PCB (solid line), P#B (dashed lines), and buffer only (dotted line) are indicated. Difference maxima and minima (in nm) were taken from mathematically smoothed plots. Data points for the four absorbance values from 652 to 658 nm were omitted due to a light-scattering artifact arising from masking of these diodes in the HP8450A spectrophotometer. The absorbance values at these wavelengths represent interpolated values from the measured values at 650 and 660 nm.

Biochemistry: Wahleithner et al. chromoprotein, which lie at 668 and 730 nm (19). These blue-shifted spectra are similar to that of the PCBapophytochrome adduct obtained from tetrapyrrole-deficient oat seedlings (3, 18). The control spectra, shown in Fig. 3, also demonstrate that the apophytochrome present in both yeast and E. coli extracts is not photochromic in the, absence of added bilin. Based on immunochemical estimates of the amount of apophytochrome present in these extracts and the assumptions that the molar absorption coefficients and quantum yield of the PCB adduct are similar to those of native oat phytochrome, an estimate of percent bilin-apophytochrome ligation yield was obtained. Table 1 shows that the majority of the apoprotein in the final soluble-protein extract, 62-75%, is competent to form photoreversible holophytochrome for both yeast and E. coli expression systems. It is significant to note that the E. coli holophytochrome assembly experiments described above were performed using apophytochrome preparations obtained from cells grown at temperatures below 30'C. Although apophytochrome is expressed from pGphyA3 in E. coli cultures grown at 37TC, this apophytochrome failed to assemble with PCB to form a photoactive holoprotein (data not shown). In this regard, improved yields of functional proteins that are expressed in E. coli grown at lower temperatures have been well documented (24). The observed blue-shifted difference spectra of the recombinant apophytochrome-PCB adducts (described above) could arise from the attachment of a nonnatural bilin chromophore or from structural alterations of the apoprotein. To distinguish between these two possibilities, we performed similar assembly experiments with both recombinant apophytochromes using the natural phytochrome chromophore P4B. For these experiments, P4B was obtained by the enzymatic conversion of the -bilin precursor biliverdin in isolated plastids (13). Difference spectra produced by the incubation of recombinant yeast and E. coli apophytochrome preparations with P4B are shown in Fig. 3. In both cases, the difference maxima at 666-668 nm and difference minima at 730 nm are nearly indistinguishable from those of native oat phytochrome preparations (19). In addition, the AAm/AAm ratios of 1.15 (yeast) and 1.04 (E. coli) are also similar with the native photoreceptor from oats. Based on these results, we conclude that both recombinant apophytochromes adopt protein structures that are similar to the natural phytochrome polypeptide synthesized in planta. These observations reaffirm the conclusions of our earlier in vitro transcriptiontranslation studies (4) that plant-specific posttranslational modifications are not required for proper folding of the apophytochrome polypeptide, for subsequent bilin attachment, or for photoreversibility of the newly assembled holophytochrome. Concluding Remarks. The development of E. coli and yeast experimental systems to express and assemble photoactive holophytochrome represents a significant advance in the study of this important plant photoreceptor. Recently, the expression of high levels of pea apophytochrome in E. coli was reported; however, this protein was both truncated and insoluble (25). The ability of this recombinant apophytochrome to assemble with bilins was not described. We believe that the lower constitutive level expression of our constructions contributed to the ability to produce soluble ligation-competent apophytochromes. In conjunction with the holophytochrome assembly systems described here, site-specific alterations can now be readily introduced into the apophytochrome polypeptide


10391

using well-established recombinant DNA methodologies. In vitro mutagenesis experiments will be especially useful to elucidate the structural basis for bilin-apophytochrome interaction and to address the structural requirements for the molecule's photoactivity. This approach, in conjunction with in vitro or in vivo assays for recombinant holophytochrome function, will also enable experimental -dissection of the molecular basis of phytochrome action. We thank Dr. M. A. Innis (Cetus, Emeryville, CA) for the gift of the pMAC105 plasmid, Dr. M. J. Holland (University of California, Davis, CA) for S. cerevisiae 29A, Drs. M.-M. Cordonnier and L. H. Pratt (University of Georgia, Athens) for monoclonal antibodies Pea-25 and Oat-25, Dr. A. N. Glazer for PCB, and Dr. Matthew Terry for providing P4B samples. This work was funded by U.S. Department of Agriculture Grant GAM89-001162. 1. Furuya, M., ed. (1987) Phytochrome and Photoregulation in

Plants (Academic, Tokyo). 2. Vierstra, R. D. & Quail, P. H. (1986) in Photomorphogenesis in Plants, eds. Kendrick, R. E. & Kronenberg, G. H. M. (Nijhoff, Dordrecht, The Netherlands), pp. 35-60. 3. Elich, T. D. & Lagarias, J. C. (1989) J. Biol. Chem. 264, 12902-12908. 4. Lagarias, J. C. & Lagarias, D. M. (1989) Proc. Natl. Acad. Sci. USA 86, 5778-5780.5. Elich, T. D. & Lagarias, J. C. (1987) Plant Physiol. 84, 304310. 6. Tomizawa, K.-I., Nagatani, A. & Furuya, M. (1990) Photochem. Photobiol. 52, 265-275. 7. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY), 2nd Ed. 8. Hershey, H. P., Barker, R. F., Idler, K. B., Murray, M. G. & Quail, P. H. (1987) Gene 61, 339-348. 9. de Lorimier, R., Wang, Y.-J. & Yeh, M.-L. (1988) in Light Energy Transduction in Photosynthesis: Higher Plant and Bacterial Models, eds. Stevens, S. E., Jr., & Bryant, D. A. (Am. Soc. Plant Physiol., Rockville, MD), pp. 332-336. 10. Innis, M. A., Holland, M. J., McCabe, P. C., Cole, G. E., Wittman, V. P., Tal, R., Watt, K. W. K., Gelfand, D. H., Holland, J. P. & Meade, J. H. (1985) Science 228, 21-26. 11. Rose, M. D., Winston, F. & Hieter, P. (1989) Laboratory Course Manual for Methods in Yeast Genetics (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 12. Litts, J. C., Kelly, J. M. & Lagarias, J. C. (1983) J. Biol. Chem. 258, 11025-11031. 13. Terry, M. J. & Lagarias, J. C. (1991) J. Biol. Chem., in press. 14. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 15. Berkelman, T. R. & Lagarias, J. C. (1986) Anal. Biochem. 156, 194-201. 16. Jones, A. M. & Quail, P. H. (1989) Planta 178, 147-156. 17. Birkett, C. R., Foster, K. E., Johnson, L. & Gull, K. (1985) FEBS Lett. 187, 211-218. 18. Elich, T. D., McDonagh, A. F., Palma, L. A. & Lagarias, J. C. (1989) J. Biol. Chem. 264, 183-189. 19. Lagarias, J. C., Kelly, J. M., Cyr, K. L. & Smith, W. O., Jr. (1987) Photochem. Photobiol. 46, 5-13. 20. Thomas, B., Crook, N. E. & Penn, S. E. (1984) Physiol. Plant.

60, 409-415. 21. Cordonnier, M.-M. (1989) Photochem. Photobiol. 49, 821-831. 22. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Oson, B. J. & Klenk, D. C. (1985) Anal. Biochem. 150, 76-85. 23. Bryant, D. A., Dubbs, J. M., Field, P. I., Porter, R. D. & de Lorimier, R. (1985) FEMS Microbiol. Lett. 29, 343-349. 24. Schein, C. H. (1989) BiolTechnology 7, 1141-1147. 25. Tomizawa, K., Ito, N., Komeda, Y., Uyeda, T. Q. P., Takio, K. & Furuya, M. (1991) Plant Cell Physiol. 32, 95-102.