We sequenced the genomes of both the purple sulfur gammaproteobacterium Marichromatium gracile HOL-1 and another purple photosynthetic organism, strain H1R, that was originally isolated as an unidentified contaminant. Through genome sequencing, we have now identified organism H1R as a species of Afifella. A whole-genome-based phylogenetic analysis of both species is provided.

Phaeovibrio sulfidiphilus was reported to be a divergent member of the purple photosynthetic bacteria with limited ability to metabolize organic compounds. Whole-genome-based analysis shows that it is indeed only distantly related to freshwater species of Rhodospirillaceae Unexpectedly, the genome contains unique gene clusters for potential respiratory nitrate reduction and anaerobic glycerol metabolism.

Citation for published version (APA): Hoff, W. D., Sprenger, W. W., Postma, P. W., Meyer, T. E., Veenhuis, M., Leguijt, T., & Hellingwerf, K. (1994). The Photoactive Yellow Protein from Ectothiorhodospira halophila as Studied with a Highly Specific Polyclonal Antiserum: (Intra)cellular Localization, Regulation of Expression, and Taxonomic Distribution of Cross-Reacting Proteins. Journal of Bacteriology, 176(13), 3920-3927.

For several years following the discovery and characterization of the first PYP, from Halorhodospira halophila , it was thought that this photoactive protein was quite unique, notwithstanding the isolation of two additional examples in rapid succession. Mainly because of genomic and metagenomic analyses, we have now learned that more than 140 PYP genes occur in a wide variety of bacteria and metabolic niches although the protein has not been isolated in most cases. The amino acid sequences and physical properties permit their organization into at least seven groups that are also likely to be functionally distinct. Based upon action spectra and the wavelength of maximum absorbance, it was speculated nearly 20 years ago but never proven that Hr. halophila PYP was involved in phototaxis. Nevertheless, in only one instance has the functional role and interaction partner for a PYP been experimentally proven, in Rs. centenum Ppr. Genetic context is one of several types of evidence indicating that PYP is potentially involved in a number of diverse functional roles. The interaction with other sensors to modulate their activity stands out as the single most prominent role for PYP. In this review, we have attempted to summarize the evidence for the functional roles and interaction partners for some 26 of the 35 named species of PYP, which should be considered the basis for further focused molecular and biochemical research.

The dissociation constants for the binding of Rhodobacter capsulatus cytochrome c2 and its K93P mutant to the cytochrome bc1 complex embedded in a phospholipid bilayer were measured by plasmon waveguide resonance spectroscopy in the presence and absence of the inhibitor stigmatellin. The reduced form of cytochrome c2 strongly binds to reduced cytochrome bc1 (Kd = 0.02 microM) but binds much more weakly to the oxidized form (Kd = 3.1 microM). In contrast, oxidized cytochrome c2 binds to oxidized cytochrome bc1 in a biphasic fashion with Kd values of 0.11 and 0.58 microM. Such a biphasic interaction is consistent with binding to two separate sites or conformations of oxidized cytochrome c2 and/or cytochrome bc1. However, in the presence of stigmatellin, we find that oxidized cytochrome c2 binds to oxidized cytochrome bc1 in a monophasic fashion with high affinity (Kd = 0.06 microM) and reduced cytochrome c2 binds less strongly (Kd = 0.11 microM) but approximately 30-fold more tightly than in the absence of stigmatellin. Structural studies with cytochrome bc1, with and without the inhibitor stigmatellin, have led to the proposal that the Rieske protein is mobile, moving between the cytochrome b and cytochrome c1 components during turnover. In one conformation, the Rieske protein binds near the heme of cytochrome c1, while the cytochrome c2 binding site is also near the cytochrome c1 heme but on the opposite side from the Rieske site, where cytochrome c2 cannot directly interact with Rieske. However, the inhibitor, stigmatellin, freezes the Rieske protein iron-sulfur cluster in a conformation proximal to cytochrome b and distal to cytochrome c1. We conclude from this that the dual conformation of the Rieske protein is primarily responsible for biphasic binding of oxidized cytochrome c2 to cytochrome c1. This optimizes turnover by maximizing binding of the substrate, oxidized cytochrome c2, when the iron-sulfur cluster is proximal to cytochrome b and minimizing binding of the product, reduced cytochrome c2, when it is proximal to cytochrome c1.

We investigated the photocycle of mutants Y98Q and Y98F of the photoactive yellow protein (PYP) from Halorhodospira halophila. Y98 is located in the beta4-beta5 loop and is thought to interact with R52 in the alpha3-alpha4 loop thereby stabilizing this region. Y98 is conserved in all known PYP species, except in Ppr and Ppd where it is replaced by F. We find that replacement of Y98 by F has no significant effect on the photocycle kinetics. However, major changes were observed with the Y98Q mutant. Our results indicate a requirement for an aromatic ring at position 98, especially for recovery and a normal I1/I2 equilibrium. The ring of Y98 could stabilize the beta4-beta5 loop. Alternatively, the Y98 ring could transiently interact with the isomerized chromophore ring, thereby stabilizing the I2 intermediate in the I1/I2 equilibrium. For Y98Q, the decay of the signaling state I2' was slowed by a factor of approximately 40, and the rise of the I2 and I2' intermediates was slowed by a factor of 2-3. Moreover, the I1 intermediate is in a pH-dependent equilibrium with I2/I2' with the ratio of the I1 and I2 populations close to one at pH 7 and 50 mM KCl. From pH 5.5 to 8, the equilibrium shifts toward I1, with a pKa of approximately 6.3. Above pH 8, the populations of I1 and I2/I2' decrease due to an equilibrium between I1 and an additional species I1' which absorbs at approximately 425 nm (pKa approximately 9.8) and which we believe to be an I2-like form with a surface-exposed deprotonated chromophore. The I1/I2/I2' equilibrium was found to be strongly dependent on the KCl concentration, with salt stabilizing the signaling state I2' up to 600 mM KCl. This salt-induced transition to I2' was analyzed and interpreted as ion binding to a specific site. Moreover, from analysis of the amplitude spectra, we conclude that KCl exerts its major effect on the I2 to I2' transition, i.e., the global conformational change leading to the signaling state I2' and the exposure of a hydrophobic surface patch. In wild type and Y98F, the I1/I2 equilibrium is more on the side of I2/I2' as compared to Y98Q but is also salt-dependent at pH 7. The I2 to I2' transition appears to be controlled by an ionic lock, possibly involving the salt bridge between K110 on the beta-scaffold and E12 on the N-terminal cap. Salt binding would break the salt bridge and weaken the interaction between the two domains, facilitating the release of the N-terminal domain from the beta-scaffold in the formation of I2'.

Based largely upon analysis of ribosomal RNA, a third domain of life, called archaea, had been proposed in addition to bacteria and eukaryotes. However, quantitative analysis of 73 whole genomes shows only a two-domain division of life: into eukaryotes and prokaryotes. Thousands of orthologous genes in archaea and bacteria show an essentially unimodal distribution of sequence identities. Thus, whole genome analyses indicate that archaea are a phylum of bacteria rather than a separate domain of life. In contrast, archaeal rRNA and that of hyperthermophilic bacteria differ from the rRNA of mesophilic bacteria. Thus, there is a bimodal distribution of rRNA sequence identities which differ by 12%. This discrepancy in rRNA and gene content based analyses of whole genomes is likely due to a 15% elevated C:G content of the rRNA of archaea and hyperthermophilic bacteria. The elevated C:G content is consistent with stabilization against thermal denaturation caused by additional hydrogen bonding (3 bonds) in C:G pairs compared to A:U pairs (2 bonds). Based upon this premise, there is no reliable way to correct rRNA for such differences in base composition and it is not possible to quantitatively compare hyperthermophiles with mesophiles by the rRNA method. Furthermore, quantitative study of whole genomes shows that the extent of change in both bacterial and archaeal genes, including rRNA, has reached a limit. Thus, direct sequence comparisons work with closely related genomes, but it is not possible to differentiate the most divergent prokaryotic species, which are currently designated as separate phyla. We believe that the differences in characteristics of archaeal species is based primarily upon selection of genes and pathways compatible with the extreme environmental lifestyle, i.e., hyperthermophily.

The absorption and CD spectra of wild-type PYP, apo-PYP, and the mutants, E46Q and M100A, were measured between 250 and 550 nm. At neutral pH, the two very weak absorption bands of wild-type PYP at 307 and 318 nm (epsilon(max) = 600 +/- 100 M(-1) cm(-1) at 318 nm) are associated with quite strong positive CD bands (Deltaepsilon(max) approximately 6.8 M(-1) cm(-1)). Both sets of bands are absent in the apoprotein. On the basis of this evidence, we assign these optical signals to the n --> pi* transition of the oxygen of the carbonyl group of the 4-hydroxycinnamic acid chromophore, which is expected to be electric dipole forbidden but magnetic dipole allowed. The progression of narrow bands at 307 and 318 nm with a shoulder in the CD around 329 nm is due to vibrational fine structure with a frequency of about 1050 +/- 50 cm(-1). This is the carbonyl stretch frequency in the electronically excited state and is well-known from the vibrational structure in the CD spectra of carbonyl compounds. The positive sign of the CD in the near UV is in accordance with the octant rule and the high-resolution X-ray structure, if we assume that the NH group of cysteine 69 to which the carbonyl is hydrogen bonded is the principle perturbant. Similar absorption and CD spectra were observed in the range of 300-340 nm for the mutants E46Q and M100A at neutral pH. Protonation of the trans chromophore by lowering the pH in the dark (without photoisomerization) broadens the 307 and 318 nm CD bands in the mutant E46Q but does not significantly affect their positions or alter their sign. For the long-lived I(2) photointermediate of the mutant M100A with protonated cis chromophore, we observed that the sign of the rotational strength in the 310-320 nm range is negative (i.e., opposite to that in the dark state with trans chromophore). This suggests that the light-induced isomerization of the chromophore, which leads to breaking of the hydrogen bond with the backbone amide of C69, brings the carbonyl into a new protein environment with different asymmetry than in the unbleached protein. The observed change in sign is mainly due to this effect, but a change in chromophore twist may also contribute. Thus, the 318 nm CD signal is a sensitive marker for the environment of the chromophore carbonyl, which samples various environments and configurations during the photocycle.

We show from time-resolved fluorescence intensity and depolarization experiments that the fluorescence of the unique tryptophan W119 of PYP is quenched by energy transfer to the 4-hydroxycinnamoyl chromophore. Whereas the intensity decay has a time constant of 0.18 ns in P, the decay in the absence of the cofactor (apo-PYP) has a single exponential lifetime of 4.8 ns. This difference in lifetime with and without acceptor can be explained quantitatively on the basis of energy transfer and the high-resolution X-ray structure of P, which allows an accurate calculation of the kappa2 factor. Fluorescence depolarization experiments with donor and acceptor indicate that both are immobilized so that kappa2 is constant on the fluorescence time scale. Using background illumination from an LED emitting at 470 nm, we measured the time-resolved fluorescence in a photostationary mixture of P and the intermediates I2 and I2'. The composition of the photostationary mixture depends on pH and changes from mainly I2 at low pH to predominantly I2' at high pH. The I2/I2' equilibrium is pH-dependent with a pKa of approximately 6.3. In I2 the lifetime increases to approximately 0.82 ns. This is not due to a change in distance or to the increase in spectral overlap but is primarily a consequence of a large decrease in kappa2. Kappa2 was calculated from the available X-ray structures and decreases from approximately 2.7 in P to 0.27 in I2. This change in kappa2 is caused by the isomerization of the acceptor, which leads to a reorientation of its transition dipole moment. We have here a rare case of the kappa2 factor dominating the change in energy transfer. The fluorescence decay in the light is pH-dependent. From an SVD analysis of the light/dark difference intensity decay at a number of pH values, we identify three species with associated lifetimes: P (0.18 ns), I2 (0.82 ns), and X (0.04 ns). On the basis of the pH dependence of the amplitudes associated with I2 and X, with a pKa of approximately 6.3, we assign the third species to the signaling state I2'. The absorption spectra of the 0.82 and 0.04 ns species were calculated from the pH dependence of their fluorescence amplitudes and of the photostationary light/dark difference absorption spectra. The lambda(max) values of these spectra (372 and 352 nm) identify the 0.82 and 0.04 ns components with I2 and I2', respectively, and validate the fluorescence data analysis. The mutant E46Q allows a further test of the energy transfer explanation, since lowering the pH in the dark leads to a bleached state with an increased spectral overlap but without the isomerization-induced decrease in kappa2. The measured lifetime of 0.04 ns is in excellent agreement with predictions based on energy transfer and the X-ray structure.

We investigated the kinetics of photoreversal from the I(1) and I(2) intermediates of photoactive yellow protein (PYP) by time-resolved optical absorption spectroscopy with double flash excitation. A first flash, at 430 nm, initiated the photocycle. After a variable time delay, the I(1) intermediate was photoreversed by a second flash, at 500 nm, or a mixture of I(2) and I(2)' intermediates was photoreversed by a second flash, at 355 nm. By varying the delay from 1 micros to 3 s, we were able to selectively excite the intermediates I(1), I(2), and I(2)'. The photoreversal kinetics of I(2) and I(2)' at 21 different delays and two wavelengths (340 and 450 nm) required two exponentials for a global fit with time constants of tau(1) = 57 +/- 5 micros and tau(2) = 380 +/- 40 micros (pH 6, 20 degrees C). These were assigned to photoreversal from sequential I(2) and I(2)' intermediates, respectively. The good agreement of the delay dependence of the two amplitudes, A(1) and A(2), with the time dependence of the I(2) and I(2)' populations provided strong evidence for the sequential model. The persistence of A(1) beyond delay times of 5 ms and its decay, together with A(2) around 500 ms, suggest moreover that I(2) and I(2)' are in thermal equilibrium. The wavelength dependence of the photoreversal kinetics was measured at 26 wavelengths from 510 to 330 nm at the two fixed delays of 1 and 10 ms. These data also required two exponentials for a global fit with tau(1) = 59 +/- 5 micros and tau(2) = 400 +/- 40 micros, in good agreement with the delay results. Photoreversal from I(2)' is slower than from I(2), since, in addition to chromophore protonation, the global conformational change has to be reversed. Our data thus provide a first estimate of about 59 micros for deprotonation and 400 micros for the structural change, which also occurs in the thermal decay of the signaling state but is obscured there since reisomerization is rate-limiting. The first step in photoreversal is rapid cis-trans isomerization of the chromophore, which we could not resolve, but which was detected by the instantaneous increase in absorbance between 330 and 380 nm. In agreement with this observation, the spectrum of the I(2)'(trans) intermediate, derived from the A(2) amplitude spectrum, has a much larger extinction coefficient than the spectrum of the I(2)'(cis) intermediate. With a first flash, at 430 nm, and a second flash, at 500 nm, we observed efficient photoreversal of the I(1) intermediate at a delay of 20 micros when most molecules in the cycle are in I(1). We conclude that each of the three intermediates studied can be reversed by a laser flash. Depending on the progression of the photocycle, reversal becomes slower with the time delay, thus mirroring the individual steps of the forward photocycle.

The purple phototrophic bacterium, Thermochromatium tepidum, contains a gene for a chimeric photoactive yellow protein/bacteriophytochrome/diguanylate cyclase (Ppd). We produced the Tc. tepidum PYP domain (Tt PYP) in Escherichia coli, and found that it has a wavelength maximum at 358 nm due to a Leu46 substitution of the color-tuning Glu46 found in the prototypic Halorhodospira halophila PYP (Hh PYP). However, the 358 nm dark-adapted state is in a pH-dependent equilibrium with a yellow species absorbing at 465 nm (pK(a) = 10.2). Following illumination at 358 nm, photocycle kinetics are characterized at pH 7.0 by a small bleach and red shift to what appears to be a long-lived cis intermediate (comparable to the I(2) intermediate in Hh PYP). The recovery to the dark-adapted state has a lifetime of approximately 4 min, which is approximately 1500 times slower than that for Hh PYP. However, when the Tt PYP is illuminated at pH values above 7.5, the light-induced difference spectrum indicates a pH-dependent equilibrium between the I(2) intermediate and a red-shifted 440 nm intermediate. This equilibrium could be responsible for the sigmoidal pH dependence of the recovery of the dark-adapted state (pK(a) = 8.8). In addition, the light-induced difference spectrum shows that, at pH values above 9.3, there is an apparent bleach near 490 nm superimposed on the 358 and 440 nm changes, which we ascribe to the equilibrium between the protonated and ionized dark-adapted forms. The L46E mutant of Tt PYP has a wavelength maximum at 446 nm, resembling wild-type Hh PYP. The kinetics of recovery of L46E following illumination with white light are slow (lifetime of 15 min at pH 7), but are comparable to those of wild-type Tt PYP. We conclude that Tt PYP is unique among the PYPs studied to date in that it has a photocycle initiated from a dark-adapted state with a protonated chromophore at physiological pH. However, it is kinetically most similar to Rhodocista centenaria PYP (Ppr) despite the very different absorption spectra due to the lack of E46.

Through pattern matching of the cytochrome c heme-binding site (CXXCH) against the genome sequence of Shewanella oneidensis MR-1, we identified 42 possible cytochrome c genes (27 of which should be soluble) out of a total of 4758. However, we found only six soluble cytochromes c in extracts of S. oneidensis grown under several different conditions: (1) a small tetraheme cytochrome c, (2) a tetraheme flavocytochrome c-fumarate reductase, (3) a diheme cytochrome c4, (4) a monoheme cytochrome c5, (5) a monoheme cytochrome c', and (6) a diheme bacterial cytochrome c peroxidase. These cytochromes were identified either through N-terminal or complete amino acid sequence determination combined with mass spectroscopy. All six cytochromes were about 10-fold more abundant when cells were grown at low than at high aeration, whereas the flavocytochrome c-fumarate reductase was specifically induced by anaerobic growth on fumarate. When adjusted for the different heme content, the monoheme cytochrome c5 is as abundant as are the small tetraheme cytochrome and the tetraheme fumarate reductase. Published results on regulation of cytochromes from DNA microarrays and 2D-PAGE differ somewhat from our results, emphasizing the importance of multifaceted analyses in proteomics.

The dissociation constants for the binding of oxidized and reduced wild-type cytochrome c(2) from Rhodobacter capsulatus and the lysine 93 to proline mutant of cytochrome c(2) to photosynthetic reaction centers (Rhodobacter sphaeroides) has been measured to high precision using plasmon-waveguide resonance spectroscopy. For the studies reported, detergent-solubilized photosynthetic reaction center was exchanged into a phosphatidylcholine lipid bilayer to approximate the physiological environment. At physiologically relevant ionic strengths ( approximately 100 mM), we found two binding sites for the reduced wild-type cytochrome (K(D) = 10 and 150 nM), with affinities that decrease with decreasing ionic strength (2-5-fold). These results implicate nonpolar interactions as an important factor in determining the dissociation constants. Taking advantage of the ability of plasmon-waveguide resonance spectroscopy to reslove the contribution of changes in mass and of structural anisotropy to cytochrome binding, we can demonstrate very different properties for the two binding sites. In contrast, the oxidized wild-type cytochrome only binds to a single site with a K(D) of 10 nM at high ionic strength, and this site has properties similar to the low-affinity site for binding the reduced cytochrome. The binding of oxidized cytochrome c(2) has a strong ionic strength response, with the affinity decreasing approximately 30-fold in going from high to low ionic strength. The K93P mutant binds to a single site in both redox states, which is similar, in terms of mass and structural anisotropy, to the oxidized wild-type site, with the affinity of the mutant oxidized state being approximately 30-fold weaker than that of the oxidized wild-type cytochrome at high ionic strength. Thus, reduced wild-type cytochrome can bind to both the high- and low-affinity sites, while the oxidized wild-type cytochrome and both redox states of the mutant cytochrome can only bind to the low-affinity site, possibly the consequence of the more stable structure of reduced wild-type cytochrome. In aggregate, the results are consistent with a model in which a transient conformational change in the region 88-102 in the cytochrome three-dimensional structure, the so-called hinge region, drives the dissociation of the oxidized cytochrome from the reaction center-cytochrome complex, facilitating turnover.

A gene for photoactive yellow protein (PYP) was previously cloned from Rhodobacter capsulatus (Rc), and we have now found it to be associated with genes for gas vesicle formation in the recently completed genome sequence. However, the PYP had not been characterized as a protein. We have now produced the recombinant RcPYP in Escherichia coli as a glutathione-S-transferase (GST) fusion protein, along with the biosynthetic enzymes, resulting in the formation of holo-RcPYP following cleavage of the GST tag. The absorption spectrum (with characteristic peaks at 435 and 375 nm) and the photocycle kinetics, initiated by a laser flash at 445 nm, are generally similar to those of Rhodobacter sphaeroides (RsPYP) but are significantly different from those of the prototypic PYP from Halorhodospira halophila (HhPYP), which has a single peak at 446 nm and has slower recovery. RcPYP also is photoactive when excited with near-ultraviolet laser light, but the end point is then above the preflash baseline. This suggests that some of the PYP chromophore is present in the cis-protonated conformation in the resting state. The excess 435 nm form in RcPYP, built up from repetitive 365 nm laser flashes, returns to the preflash baseline with an estimated half-life of 2 h, which is markedly slower than that for the same reaction in RsPYP. Met100 has been reported to facilitate cis-trans isomerization in HhPYP, yet both Rc and RsPYPs have Lys and Gly substitutions at positions 99 and 100 (using HhPYP numbering throughout) and have 100-fold faster recovery kinetics than does HhPYP. However, the G100M and K99Q mutations of RcPYP have virtually no effect on kinetics. Apparently, the RcPYP M100 is in a different conformation, as was recently found for the PYP domain of Rhodocista centenaria Ppr. The cumulative results show that the two Rhodobacter PYPs are clearly distinct from the other species of PYP that have been characterized. These properties also suggest a different functional role, that we postulate to be in regulation of gas vesicle genes, which are known to be light-regulated in other species.

All class I c-type cytochromes studied to date undergo a dynamic process in the oxidized state, which results in the transient breaking of the iron-methionine-sulfur bond and sufficient movement to allow the binding of exogenous ligands (imidazole in this work). In the case of Rhodobacter capsulatus cytochrome c(2), the sixth heme ligand Met96 and up to 14 flanking residues (positions 88-100, termed the hinge region), located between two relatively rigid helical regions, may be involved in structural changes leading to a transient high-spin species able to bind ligands. We have examined 14 mutations at 9 positions in the hinge region of Rhodobacter capsulatus cytochrome c(2) and have determined the structure of the G95E mutant. Mutations near the N- and C-terminus of the hinge region do not affect the kinetics of movement but allow us to further define that portion of the hinge that moves away from the heme to the 93-100 region in the amino acid sequence. Mutations at positions 93 and 95 can alter the rate constant for hinge movement (up to 20-fold), presumably as a result of altering the structure of the native cytochrome to favor a more open conformation. The structure of one of these mutants, G95E, suggests that interactions within the hinge region are stabilized while interaction between the hinge and the heme are destabilized. In contrast, mutations at positions 98 and 99 alter imidazole binding kinetics but not the hinge movement. Thus, it appears that these mutations affect the structure of the cytochrome after the hinge region has moved away from the heme, resulting in increased solvent access to the bound imidazole or alter interactions between the protein and the bound imidazole.

High-potential iron-sulfur protein (HiPIP) has recently been shown to function as a soluble mediator in photosynthetic electron transfer between the cytochrome bc 1 complex and the reaction-center bacteriochlorophyll in some species of phototrophic proteobacteria, a role traditionally assigned to cytochrome c 2 . For those species that produce more than one high-potential electron carrier, it is unclear which protein functions in cyclic electron transfer and what characteristics determine reactivity. To establish how widespread the phenomenon of multiple electron donors might be, we have studied the electron transfer protein composition of a number of phototrophic proteobacterial species. Based upon the distribution of electron transfer proteins alone, we found that HiPIP is likely to be the electron carrier of choice in the purple sulfur bacteria in the families Chromatiaceae and Ectothiorhodospiraceae, but the majority of purple nonsulfur bacteria are likely to utilize cytochrome c 2 . We have identified several new species of phototrophic proteobacteria that may use HiPIP as electron donor and a few that may use cytochromes c other than c 2 . We have determined the amino acid sequences of 14 new HiPIPs and have compared their structures. There is a minimum of three sequence categories of HiPIP based upon major insertions and deletions which approximate the three families of phototrophic proteobacteria and each of them can be further subdivided prior to construction of a phylogenetic tree. The comparison of relationships based upon HiPIP and RNA revealed several discrepancies.

The photoactive yellow protein (PYP) is a bacterial photoreceptor which is the structural prototype for the PAS domain superfamily of regulators and receptors. PYP is known to have a unique p-hydroxycinnamic acid chromophore, covalently attached to a cysteine. To date, it has not been shown how holo-PYP is formed in vivo. Two genes, nearby pyp, were postulated to encode the biosynthetic enzymes, but only one was previously isolated and shown to have the requisite activity. By using a dual plasmid system, one expressing the PYP from Halorhodospira halophila and the other expressing a two-gene operon, consisting of tyrosine ammonia lyase and p-hydroxycinnamic acid ligase, we are able to present evidence that a functionally active holo-PYP can be synthesized in Escherichia coli. Plasmids containing only one of the two enzymes failed to produce holoprotein. Thus, the two genes have been shown to be both necessary and sufficient for production of holoprotein, although the activating group remains unknown. This expression system not only holds great potential for mutagenesis studies but also opens new possibilities in the search for (a) signaling partner(s) of the PYP.