brp and blh are required for synthesis of the retinal cofactor of bacteriorhodopsin in Halobacterium salinarum.

Bacteriorhodopsin, the light-driven proton pump of Halobacterium salinarum, consists of the membrane apoprotein bacterioopsin and a covalently bound retinal cofactor. The mechanism by which retinal is synthesized and bound to bacterioopsin in vivo is unknown. As a step toward identifying cellular factors involved in this process, we constructed an in-frame deletion of brp, a gene implicated in bacteriorhodopsin biogenesis. In the Deltabrp strain, bacteriorhodopsin levels are decreased approximately 4.0-fold compared with wild type, whereas bacterioopsin levels are normal. The probable precursor of retinal, beta-carotene, is increased approximately 3.8-fold, whereas retinal is decreased by approximately 3.7-fold. These results suggest that brp is involved in retinal synthesis. Additional cellular factors may substitute for brp function in the Deltabrp strain because retinal production is not abolished. The in-frame deletion of blh, a brp paralog identified by analysis of the Halobacterium sp. NRC-1 genome, reduced bacteriorhodopsin accumulation on solid medium but not in liquid. However, deletion of both brp and blh abolished bacteriorhodopsin and retinal production in liquid medium, again without affecting bacterioopsin accumulation. The level of beta-carotene increased approximately 5.3-fold. The simplest interpretation of these results is that brp and blh encode similar proteins that catalyze or regulate the conversion of beta-carotene to retinal.

Rhodopsins are integral membrane proteins containing seven transmembrane ␣-helices and a covalently bound molecule of retinal. Two distinct rhodopsin families are known: the visual rhodopsins, which bind 11-cis retinal or related compounds and function as photoreceptors in vertebrates (1) and invertebrates (2), and the archaeal rhodopsins, which bind all-trans-retinal and function as light-driven ion pumps and phototaxis receptors in archaea (3). Archaeal rhodopsin orthologs have been found recently in bacteria (4) and fungi (5), suggesting that retinal-based pigments are of widespread significance. Despite their importance, the biogenesis of these molecules is not fully understood. In particular, relatively little is known about how retinal is assembled with the opsin apoprotein in vivo. Thus, a goal in elucidating rhodopsin biogenesis is to identify the cellular factors that mediate the biosyn-thesis or uptake of retinal, the transport of retinal in the cell, and the binding of retinal to the corresponding opsin.
To this end, we have studied the biogenesis of bacteriorhodopsin (BR), 1 a light-driven proton pump in the archaeon Halobacterium salinarum. BR consists of the membrane protein bacterioopsin (BO) and all-trans-retinal. Under microaerobic conditions, BR is induced ϳ50-fold (6) and forms a twodimensional crystal known as the purple membrane. This system has served as a model for studying key steps in membrane protein biogenesis, including protein insertion into the membrane (7,8) and the assembly of protein-lipid complexes (9,10). H. salinarum is genetically tractable, and the genome sequence of a closely related organism, Halobacterium sp. NRC-1, has been determined (11). Thus, the prospect of identifying the cellular factors that mediate retinal assembly and other steps in BR biogenesis in H. salinarum is excellent.
Retinal is synthesized de novo in H. salinarum (12) and eventually binds BO to form BR. A pathway for retinal biosynthesis has been proposed from studies of cell-free preparations and by comparison with other carotenoid biosynthetic pathways (12). Intermediates in the pathway from the universal C 40 -carotenoid precursor phytoene to ␤-carotene have been identified (12). ␤-Carotene is thought to be the immediate precursor of retinal, although there is no direct evidence for its conversion to retinal in H. salinarum. Furthermore, the cellular factors that catalyze this conversion are unknown. The addition of retinal to BO to form BR has long been supposed to occur without the participation of cellular factors because BR can be regenerated from retinal and purified BO in vitro (13,14). However, cellular factors may be required to prevent the photooxidation or photoisomerization of the cofactor during its transport or binding to BO. These functions may be mediated by a retinal-specific chaperone or by a multifunctional enzyme that converts ␤-carotene to retinal and transports or binds retinal to BO.
As a first step to identify cellular factors that mediate these processes, we chose to study the brp gene, which encodes a putative membrane protein (Brp) implicated in BR biogenesis (15)(16)(17)(18). The brp gene is part of a gene cluster (Fig. 1) that includes bop, which encodes BO; bat, which encodes a transacting factor (Bat) that contains a region homologous to the PAS domain of the oxygen sensor NifL (19,20) and activates bop expression under microaerobic conditions (21,22); and blp, a gene of unknown function (23). Insertions in brp greatly decrease BR and bop mRNA levels (15)(16)(17)(18). This result can be interpreted to imply that Brp modulates bop expression (18,21). However, it has also been recognized that the effect of brp insertions may be attributable to an indirect effect on bat expression (18,21). The bat gene is immediately downstream of brp, and the termination and initiation codons of the two genes overlap (Fig. 1). Northern blot analysis of brp and bat mRNAs was interpreted as evidence for separate transcripts, although cotranscription of the two genes was not excluded (21). Thus, brp and bat may constitute an operon, and insertions in brp may reduce bop expression indirectly by a polar effect on bat transcription. In this case, Brp may play a role in BR biogenesis other than regulating bop expression, such as the biosynthesis, transport, or binding of the retinal cofactor of BR.
To examine the role of Brp, we created an in-frame deletion of brp using a recently developed ura3-based gene knockout strategy (24). Analysis of BR, BO, and carotenoid accumulation in the deletion strain suggests that Brp is essential for the production of BR but not BO. Instead, brp appears to be required for the synthesis of retinal from ␤-carotene. Parallel studies with blh, a paralog of brp identified from the Halobacterium sp. NRC-1 genome sequence (11), indicate that the blh gene product partially substitutes for Brp. The implications of these findings for BR biogenesis and retinal biosynthesis in H. salinarum are discussed.

EXPERIMENTAL PROCEDURES
Materials-Oligonucleotides were obtained from Operon (Alameda, CA), Taq polymerase was obtained from Promega (Madison, WI), and restriction endonucleases and ligase were obtained from New England Biolabs (Beverly, MA). All other reagents were obtained from Sigma.
Plasmid Construction-Plasmids were propagated in the Escherichia coli strain DH5␣ except where noted. A two-step polymerase chain reaction (PCR) was performed to construct the ⌬brp plasmid containing DNA homologous to brp with an in-frame deletion of codons 33-308 and an 18-bp insertion at the deletion site. In the first step, reactions were carried out with the primer pairs GTGACGGACTTCACGGTTG and CTGCTCGATCTCGATCTCAAGCGCGAGCAGTGACAG or GAGATC-GAGATCGAGCAGGTGCTCTGGCTGCTGTCC and GTGCGGCTACG-AAATCAC using Halobacterium sp. NRC-1 genomic DNA as template. The PCR products were used as a template in a second PCR step with the primers GTGACGGACTTCACGGTTG and GTGCGGCTACGAAA-TCAC. The 293-bp BamHI fragment of the PCR product was combined with the 4.3-kbp BamHI fragment of pMPK414, which was constructed by combining the 2.4-kbp KpnI-PstI fragment containing brp from pMPK14 (25) with the 2.8-kbp KpnI-PstI fragment of Litmus 28 (New England Biolabs). The resulting plasmid, pMPK417, contains the ⌬brp construct with 0.6-kbp 5Ј-flanking DNA and 1.0-kbp 3Ј-flanking DNA. The 1.8-kbp KpnI-PstI fragment of pMPK417 was combined with the 5.1-kbp KpnI-HindIII fragment of pMPK85 (9) and the 1.4-kbp NsiI-HindIII ura3 cassette fragment of pMPK408 (24) to make pMPK421, which contains a mevinolin-resistance determinant, the ura3 cassette, and the brp deletion.
To create a complementation strain containing brp at the ura3 locus, the oligonucleotides TCTAGAGTAGATCT and AGATCTACTCTAGA were annealed to create a duplex linker containing XbaI and BglII restriction sites. The linker was ligated into a partial PshAI digest of the ⌬ura3 plasmid pMPK404 (24) to make pMPK423. The 2.0-kbp AgeI-EcoRI fragment of pMPK423 was combined with the 5.1-kbp HindIII-XmaI fragment of pMPK85 and the 1.4-kbp HindIII-EcoRI fragment of pMPK408 to make pMPK424, which contains the XbaI-BglII linker in place of ura3, the mevinolin-resistance determinant and the ura3 cassette. The primers TCTAGATCTAGAGTGACGGACTTCACGGTTG and AGATCTAGATCTAAAAGCCGCGCCGGTTCATGGGACGTACCA-GATGCC were used to make a PCR product containing XbaI and BglII restriction sites flanking brp. This PCR product was digested with XbaI and BglII and ligated into the XbaI-BglII fragment of pMPK424 (prepared from the strain CSH26 dam Ϫ ) to make pMPK425. In addition to the brp open reading frame, pMPK425 contains 186 bp of the brp upstream region to allow possible regulation similar to the native locus and a 21-bp sequence derived from the bop transcription terminator (26) to ensure the transcription termination of brp.
The ⌬blh plasmid was constructed using the same PCR strategy as described for making the brp deletion, except that the primer pairs GTCGACGCGACGTTCTACAT and CTGCTCGATCTCGATCTCACCG-AGCACGAGGTAGAGG or GAGATCGAGATCGAGCAGGTGCTGTGG-TGGGCGGTA and GCCCATGATGTTCATCGACT were used with MPK1 genomic DNA as a template. The resulting PCR products were used as a template in PCR with the primer pair GTCGACGCGACGT-TCTACAT and GCCCATGATGTTCATCGACT. The 0.8-kbp BamHI-NcoI fragment of the ⌬blh PCR was combined with the 5.1-kbp BamHI-EcoRI fragment of pMPK85 and the 1.2-kbp EcoRI-NcoI fragment of pMPK408 to make pMPK427, which contains the ⌬blh construct, the mevinolin-resistance determinant, and the ura3 cassette.
H. salinarum Strain Construction-The brp::ISH27 strain (MPK8) is a spontaneous mutant derived from MPK1 (6) on the basis of the loss of the purple colony color, which was detected by illuminating colonies with 40-watt daylight fluorescent lamps for 2-3 days. All other H. salinarum strains were derived by using the ura3-based gene replacement method (24). Briefly, strains were transformed as described (6) with the plasmids described above and mevinolin-resistant transformants were replated on media containing 5-fluoroorotic acid to obtain recombinants. Colonies resistant to 5-fluoroorotic acid were screened by PCR to identify the desired recombinants. The ⌬brp strain MPK417 was isolated by transforming pMPK421 into the ⌬ura3 strain MPK407 (24), the brp complementation strain MPK420 was isolated by transforming pMPK425 into MPK417, the ⌬blh strain MPK424 was isolated by transforming MPK407 with pMPK427, and the ⌬brp⌬blh strain MPK423 was isolated by transforming MPK417 with pMPK427. The structures of the brp, blh, and ura3 loci were confirmed in all strains by PCR and Southern blot analysis as described (6). The sequences of the entire brp open reading frame and at least 120 bp of the flanking sequence of the ⌬brp strain MPK417 and the blh locus of the ⌬blh strain MPK424 were confirmed by ABI PRISM Big Dye Primer Cycle sequencing (Applied Biosystems, Inc., Foster City, CA) using the primers GTGACGGACTTCACGGTTG and GTGCGGCTACGAAATCAC plus GTCGACGCGACGTTCTACAT and GCCCATGATGTTCATCGACT, respect-ively. Sequencing reactions were analyzed with a 337XL automated DNA sequencer (Applied Biosystems, Inc.) at the University of Wisconsin Biotechnology Center.
Induction and Preparation of Cell Lysates-To induce BR synthesis, 120 ml of peptone medium (27) in a 125-ml Erlenmeyer flask were inoculated with 1.2 ml of saturated H. salinarum culture and grown in the dark at 40°C with shaking at 250 rpm for 96 -100 h. Where noted, 30 l of 10 mM retinal in isopropyl alcohol were added at ϳ14, 24, 38, 48, 62, and 72 h after culturing. At an OD 660 of 0.70 -0.85, the cultures were harvested by centrifuging at 8000 ϫ g for 30 min at 4°C. The cell pellet was resuspended in 100 ml of basal salts and centrifuged again at 8000 ϫ g for 30 min, followed by a brief spin to remove all traces of the supernatant. Cells were lysed in 3.5 ml of 4 g of DNase/ml, 0.5 mM phenylmethylsulfonyl fluoride, and 0.025% sodium azide in water and shaken for ϳ1 h at room temperature. Samples were purged with nitrogen gas to minimize carotenoid oxidation.
Quantification of BR and BO-To assess BR levels, cell lysates were diluted 1:4 in 30 mM sodium phosphate buffer (pH 6.9) containing 0.1% sodium azide and scanned in a Perkin-Elmer 2 spectrophotometer. The BR levels were determined by light-dark difference spectroscopy as described (6) with a standard curve obtained with purified BR added to a lysate from the ⌬bop strain MPK412 (24). The BR levels were expressed as a percentage of total cell protein as measured by the BCA assay (Pierce). The BO levels were determined by immunoblotting with BR-114 monoclonal antibody generously provided by Dr. H. G. Khorana. The blots were subsequently incubated with fluorescein-conjugated ␣-mouse IgG secondary antibody (Amersham Pharmacia Biotech), and the BO levels were quantified on a Hitachi FMBIOII Multi-View Fluoroimager. Purified BR was used to generate a standard curve.
Extraction and Characterization of Carotenoids-Total carotenoid was extracted from cell lysates as described (28). Lysates were illuminated for 3 min with a Ͼ520-nm light prior to extraction to convert all retinal to the all-trans isomer. The extracts were evaporated to dryness under nitrogen gas. To identify the major carotenoid in ⌬brp strains, the extracts were fractionated by HPLC on an HPLX solvent delivery system (Rainin Instrument Co., Inc., Emeryville, CA) coupled with a reverse phase Econosphere C18 column (250 ϫ 4.6 mm, 5-m particle size) (Alltech Associates, Inc., Deerfield, IL) and an Alltech Econosphere C18 5-m guard column. The mobile phase was a gradient of solvent A (95% methanol, 5% water) and solvent B (dichloromethane) eluting at 1 ml/min. The solvent change over a 25-min sample run was programmed as follows: elution with 100% solvent A, 3 min; gradient to 32% solvent B, 6 min; isocratic elution with 32% solvent B, 11 min; gradient to 100% solvent A, 1 min; and re-equilibration with 100% solvent A, 4 min. The eluate was monitored at 474 nm with a Dynamax UV-1 variable wavelength UV/visible absorbance detector (Rainin In-strument Co., Inc.). HPLC fractions eluting at ϳ18 min were collected and evaporated under nitrogen gas. The sample was resuspended in acetone to a ␤-carotene concentration of 10 M. Mass spectrometry was performed at the University of Wisconsin Biotechnology Center on a Bruker Biflex III MALDI-TOF instrument (Bruker Analytical Systems, Billerica, MA) using 2,5-dihydroxybenzoic acid as a matrix.
For simultaneous measurement of ␤-carotene and retinal, an identical HPLC system was employed except that an Alltech Altima C18 column (250 ϫ 4.6 mm, 5-m particle size) was used and the mobile phase consisted of a gradient of solvent A (95% acetonitrile, 5% methanol) and solvent B (dichloromethane). The solvent change over a 21min sample run was programmed as follows: elution with 100% solvent A, 8 min; gradient to 65% solvent B, 1 min; isocratic elution at 65% solvent B, 7 min; gradient to 100% solvent A, 1 min; and re-equilibration with 100% solvent A, 4 min. The eluate was monitored at 380 nm for the first 12 min and at 450 nm for the remainder of each run. Standard curves were generated with commercial ␤-carotene and all-trans-retinal. (Fig.  1). Because insertions in brp might reduce bop expression indirectly through a polar effect on bat, we reexamined the role of brp in BR biogenesis. As a first step, we confirmed the phenotype of brp insertions by isolating a spontaneous mutant from a laboratory strain of H. salinarum, MPK1 (6). Unlike MPK1, which forms purple colonies, the mutant strain yielded pale yellow colonies. PCR, Southern blot, and DNA sequence analysis of the mutant strain (data not shown) revealed the presence of the insertion element ISH27 (16) at the third nucleotide position of codon 177 in the brp open reading frame (Fig. 1). This strain was designated brp::ISH27.

An Insertion in brp Eliminates BO Expression-Earlier studies of brp function relied on spontaneous insertions in brp
To quantify the effects of the brp insertion on BR accumulation, MPK1 and brp::ISH27 were grown microaerobically to induce BR synthesis (27). Cells were grown in the dark to prevent the differences in BR levels from affecting cell energy states. Cell lysates from brp::ISH27 were defective in the accumulation of BR, as evidenced by the loss of the 570 nm peak corresponding to BR (Fig. 2A). As quantified by light-dark difference spectroscopy (6), BR was ϳ4% of the total cell protein in the wild-type strain but was undetectable in the brp::ISH27 strain (Fig. 2B). Immunoblotting of cell lysates with a BO C-terminal antibody revealed that brp::ISH27 lacked BO (Fig.  2, inset, lane 2), indicating that the loss of BR was attributable to the absence of BO. These results are consistent with the earlier findings that insertions in brp abolish BR and bop mRNA production (15,16,21). As acknowledged previously (21), insertions in brp may prematurely abort an mRNA transcript that encodes both Brp and Bat, thereby reducing bat transcription and preventing bop transcriptional activation.
In-frame Deletion of brp Reduces BR but Not BO-To study brp without the polar effects on bat, a strain containing an in-frame deletion of brp was constructed. The ⌬brp strain lacks codons 33-308 of the brp open reading frame, which encodes a polypeptide of 359 amino acids. After incubation under lights, ⌬brp colonies had a dark orange color that was distinct from the purple color of wild type and the pale yellow color of brp::ISH27. This result suggested that the in-frame deletion reduced but did not abolish BR accumulation. The UV/visible spectrum of cell lysates from the ⌬brp strain revealed a re-duced peak at 570 nm, confirming that BR levels are decreased compared with wild type (Fig. 3, solid and dotted lines). By light-dark difference spectroscopy, the BR levels were found to be ϳ4.0-fold lower than in the wild type (Fig. 4, samples 1 and  2). Unlike the brp::ISH27 strain, however, the ⌬brp strain had normal BO levels as determined by quantitative immunoblotting (Fig. 3, inset, lanes 1 and 2, and Fig. 4, samples 1 and 2). These results suggest that Brp does not regulate bop expression and is instead involved in BR biogenesis at a step between BO synthesis and BR formation, such as retinal biosynthesis, transport, or binding.
brp Deletion Increases ␤-Carotene and Decreases Retinal Levels-In addition to the reduced peak at 570 nm, the UV/ visible spectra of ⌬brp cell lysates revealed three prominent peaks between 400 and 500 nm that were absent from the wild type (Fig. 3, solid and dotted lines). This suggested that ⌬brp had higher levels of ␤-carotene, which is thought to be the precursor of retinal in H. salinarum (12). To test this possibility, carotenoid was extracted from the wild-type and ⌬brp cell lysates using a method that recovers Ͼ80% of the retinal bound to BR (28). The extracts were fractionated with HPLC to quantify ␤-carotene and retinal (Fig. 5). The species eluting at 15 min comigrated with commercial ␤-carotene and was present at ϳ3.8-fold higher levels in the ⌬brp strain than in the wild type (Fig. 4, samples 1 and 2, and Fig. 5, traces 1 and 2). Independent experiments were carried out to confirm that the major carotenoid from the ⌬brp strain is ␤-carotene. Under different HPLC conditions (see "Experimental Procedures"), the ⌬brp extract and commercial ␤-carotene yielded both a prominent species absorbing at 474 nm that eluted at 17.5 min and a minor species that eluted at 17.8 min, presumably attributable to cis-isomers of ␤-carotene (data not shown). The HPLC peak fractions obtained from the ⌬brp extract and commercial ␤-carotene had mass ion values of 536.441 and 536.437 Da and similar UV/visible spectra with absorption maxima at 449 and 451 nm, respectively (data not shown). The slight difference in absorption maximum may be attributable to a difference in the ratio of ␤-carotene isomers (29). Thus, the major carotenoid that accumulates in the ⌬brp strain is ␤-carotene.  2. A, UV/visible spectra of cell lysates from wild-type (dotted line) and brp::ISH27 (solid line) strains. The peak at 570 nm corresponds to BR. Cells were grown to stationary phase to induce BR synthesis. Spectra of lysates were obtained at a total cell protein concentration of ϳ3 mg/ml and normalized for slight differences in protein concentration. B, light-dark difference spectroscopy of wild-type (dotted line) and brp::ISH27 (solid line) lysates. Spectra were obtained from lysates dark-adapted for Ն12 h or light-adapted for 5 min. A light-dark difference spectrum was calculated, and the BR level was determined from the value at 587 nm. Inset, immunoblot of wild-type (lane 1) and brp::ISH27 (lane 2) cell lysates. Equal amounts of total cell protein (4.0 g) were electrophoresed and immunoblotted with the C-terminal BR antibody, BR-114. The blot was analyzed by incubating with a fluorescein-conjugated secondary antibody and scanning with a fluoroimager.
The ϳ3.8-fold increase in ␤-carotene accumulation in the ⌬brp strain was accompanied by a corresponding ϳ3.7-fold decrease in retinal accumulation (Fig. 4, samples 1 and 2, and   Fig. 5, traces 1 and 2). The simplest interpretation of these results is that Brp catalyzes or regulates the conversion of ␤-carotene to retinal.
⌬brp Is Complemented by an Intact Copy of brp-A complementation strain was constructed to confirm that the ⌬brp phenotype is caused by the loss of brp and not by a reduction in bat expression or a second-site mutation in an unknown gene. The brp open reading frame, flanked by 186 bp of the upstream sequence to allow normal expression of the gene, was integrated at the ura3 locus. The resulting complementation strain, ⌬brp ura3::brp, yielded purple colonies identical to the wild type. When this strain was grown under BR induction conditions, BR, ␤-carotene, and retinal levels were restored to the wild-type levels (Fig. 3, dashed line, Fig. 4, sample 3, and Fig. 5, trace 3). These results confirm that the ⌬brp phenotype is caused solely by the loss of brp and that the in-frame deletion of brp has no detectable effect on bat expression.
A brp Paralog May Function Similarly to brp-The ⌬brp mutation reduced BR levels by ϳ4.0-fold but did not completely eliminate BR synthesis, raising the possibility that other factors partially substitute for brp in BR biogenesis. To examine this possibility, the Halobacterium sp. NRC-1 genome sequence (11) was searched for genes encoding proteins homologous to Brp, and a single gene, blh, was identified. The blh open reading frame begins with a GUG start codon and encodes a putative 345-amino acid integral membrane protein (Blh) with 28% identity to Brp over its entire length. (blh was initially predicted to encode a 284-amino acid protein (11). However, our analysis suggests that the protein may be 61 amino acids longer at the N terminus based on GC bias in the third position of codons and a better match of the length and predicted topography of the blh and brp gene products.) To determine whether Blh plays a role in BR biogenesis, an in-frame deletion of blh was constructed in the wild-type and ⌬brp backgrounds. Codons 97-300 of the blh open reading frame were deleted. The ⌬blh colonies had a dark orange color similar to the ⌬brp colonies, but the ⌬brp⌬blh colonies had a yellow color, suggesting an accumulation of ␤-carotene without the expression of BR. When grown microaerobically in liquid media to induce BR expression, the ⌬blh strain had wild-type levels of BR, ␤-carotene, and retinal (Fig. 4, sample 4, and Fig.  6, dashed line). Given that ⌬blh colonies were clearly altered, the lack of an observable effect under liquid growth conditions in the dark is surprising but suggests that blh function or expression is sensitive to growth conditions. Significantly, in the ⌬brp⌬blh strain, no BR or retinal was detected, and ␤-carotene levels were ϳ5.3-fold higher than wild type (Fig. 4, sample 5, and Fig. 6, solid line). The BO levels were normal in both ⌬blh and ⌬brp⌬blh (Fig. 4, samples 4 and 5, and Fig. 6, inset). These results suggest that Blh acts similarly to Brp in converting ␤-carotene to retinal.
Addition of Retinal in Vivo Restores BR Accumulation in ⌬brp and ⌬brp⌬blh Strains-One model for Brp and Blh func- Carotenoids were extracted from cells as described (28) and analyzed by reverse-phase HPLC as described in Experimental Procedures. At the time indicated by the arrow, the detector was switched from 380 nm, the wavelength at which retinal was quantified, to 450 nm to quantify ␤-carotene. Traces are shown normalized to total cell protein and offset an equal distance along both axes for clarity. tion is that they aid BO folding to permit retinal binding. To test this possibility, retinal was added periodically to cultures of the ⌬brp and ⌬brp⌬blh strains during growth under conditions to induce BR. The addition of all-trans-retinal restored BR accumulation to the wild-type levels (Fig. 7), confirming that the BO produced by these strains is competent to bind retinal and suggesting that Brp and Blh are not required for the correct folding of BO. DISCUSSION We have identified two related H. salinarum genes, brp and blh, that are required for BR biogenesis. The in-frame deletion of brp alone results in decreased BR and retinal levels and a corresponding increase in ␤-carotene levels. These effects are enhanced by the deletion of both brp and blh. The deletion of these genes has no significant effect on BO levels. These results indicate that brp and blh are not involved in regulating bop gene expression. Instead, the genes are needed for the synthesis of the retinal cofactor of BR or for its transport or binding to BO.
The simplest model of brp and blh function is that they encode the proteins that catalyze or regulate the catalysis of the conversion of ␤-carotene to retinal. The concomitant decrease in retinal and increase in ␤-carotene levels in the ⌬brp strain strongly support this model. Furthermore, the only defect in strains lacking brp and blh is the inability to synthesize retinal, because exogenous retinal restores BR levels. However, Brp and Blh have no obvious primary structural features that indicate they catalyze the conversion of ␤-carotene to retinal. Significantly, Brp and Blh are unrelated to the recently described 15,15Ј-␤-carotene dioxygenase of Drosophila melanogaster (30), a soluble protein that catalyzes the oxidative cleavage of ␤-carotene to two molecules of retinal. Because we have been unable to find orthologs of this enzyme in the Halobacterium sp. NRC-1 proteome, Brp and Blh may be part of a novel retinal biosynthetic pathway unique to haloarchaea.
An alternative model is that Brp and Blh are involved in the transport of retinal in the cell or the binding of retinal to BO. If the proteins were involved exclusively in transport or binding, at least low levels of retinal would be expected to accumulate in the ⌬brp⌬blh strain. The failure to detect retinal in this strain argues against a role of the proteins in retinal transport or binding. However, we cannot exclude the possibility that the proteins are multifunctional enzymes that catalyze the conversion of ␤-carotene to retinal and also mediate retinal transport and binding.
Another model is that Brp and Blh encode proteins that regulate the expression of enzymes that convert ␤-carotene to retinal. Although the proteins lack features typical of transcriptional regulators, they may interact with transcriptional regulators to modulate transcription. This type of regulation was suggested previously in a model whereby the brp gene product acts as a light sensor (31) and activates Bat to modulate bop transcription. We showed that brp has important functions without light because the differences in BR and carotenoid levels between wild-type and deletion strains were observed in cultures grown in the dark. Moreover, Brp does not appear to regulate bop transcription because BO levels were normal in the ⌬brp strain grown either in the dark or in the light (Fig. 4). 2 Thus, Brp is unlikely to be a light-sensing regulator of Bat. Nevertheless, Brp may modulate the activity of Bat at genes other than bop, such as those that are required for retinal metabolism.
Our results support the pathway of retinal synthesis in H. salinarum proposed previously (12). This pathway was based on the in vitro reconstitution of ␤-carotene formation from mevalonate (32) and on the accumulation of C 40 -isoprenoid intermediates in colorless mutant strains that lacked ␤-carotene and retinal (33). However, direct biochemical or genetic evidence for the conversion of ␤-carotene to retinal has not been obtained. In our experiments, we have shown that the deletion of a single gene (brp) simultaneously results in decreased retinal accumulation and increased ␤-carotene accumulation. When brp and blh are both deleted, ␤-carotene accumulation increases further and no retinal is detectable. Thus, ␤-carotene is likely to be the precursor to retinal in H. salinarum and is not converted spontaneously to retinal.
Brp and Blh appear to have redundant functions. The redundancy may be needed to allow retinal production under both aerobic and anaerobic growth conditions. Of the four rhodopsins produced by H. salinarum, three are induced microaerobically (BR, halorhodopsin, and sensory rhodopsin I), whereas the fourth (sensory rhodopsin II) is suppressed under these conditions (34). The immunoblotting of cell lysates with an antibody directed against epitope-tagged Brp indicates that Brp is present only in cells grown microaerobically. 3 This finding is consistent with a model in which Brp is induced microaerobically to provide the retinal needed for the formation of BR, halorhodopsin, and sensory rhodopsin I, and Blh is expressed aerobically to provide retinal to sensory rhodopsin II.
The results presented here have implications for the regulation of BR biogenesis. BR biogenesis is regulated partly by bat, which is required for bop and brp expression as shown by the virtual absence of bop and brp mRNAs in bat deletion or insertion strains (18,21). If Brp is required for retinal synthesis, as suggested by our data, then Bat may be responsible for the coordinate regulation of polypeptide and cofactor synthesis. Our results also suggest that a further level of control may occur because of the cotranscription of brp and bat. These genes were previously suggested to yield separate mRNA transcripts by Northern blot analysis (21). However, we demonstrated that a spontaneous insertion in brp eliminates BO synthesis, whereas an in-frame brp deletion has no effect. These results indicate that brp insertions have a polar effect on bat expression and suggest that brp and bat are cotranscribed. Thus, Bat may positively regulate its own synthesis as well as that of Brp and BO. This positive feedback may account for the rapid increase in the levels of both BO and retinal that are required for BR biogenesis.
The factors identified in this study play a key role in the synthesis of the retinal cofactor that is essential for the biogenesis of BR in H. salinarum. Further studies are needed to confirm that Brp and Blh catalyze the conversion of ␤-carotene to retinal and to test whether these proteins play a role in transporting retinal or in binding retinal to BO. It will also be important to determine whether other cellular factors participate in these processes. Such factors may be identified by using the Halobacterium sp. NRC-1 genome sequence (11) and the ura3-based reverse genetics approach as we have demonstrated for brp and blh.