Photokinetic, Biochemical and Structural Features of Chimeric Photoactive Yellow Protein Constructs
Article first published online: 9 OCT 2012
© 2012 Wiley Periodicals, Inc. Photochemistry and Photobiology © 2012 The American Society of Photobiology
Photochemistry and Photobiology
Volume 89, Issue 2, pages 349–360, March/April 2013
How to Cite
Kyndt, J. A., Meyer, T. E., Olson, K. T., Van Beeumen, J. and Cusanovich, M. A. (2013), Photokinetic, Biochemical and Structural Features of Chimeric Photoactive Yellow Protein Constructs. Photochemistry and Photobiology, 89: 349–360. doi: 10.1111/j.1751-1097.2012.01235.x
- Issue published online: 1 MAR 2013
- Article first published online: 9 OCT 2012
- Accepted manuscript online: 7 SEP 2012 06:16AM EST
- Manuscript Accepted: 29 AUG 2012
- Manuscript Received: 4 JUN 2012
Of the 10 photoactive yellow protein (PYPs) that have been characterized, the two from Rhodobacter species are the only ones that have an additional intermediate spectral form in the resting state (λmax = 375 nm), compared to the prototypical Halorhodospira halophila PYP. We have constructed three chimeric PYP proteins by replacing the first 21 residues from the N-terminus (Hyb1PYP), 10 from the β4–β5 loop (Hyb2PYP) and both (Hyb3PYP) in Hhal PYP with those from Rb. capsulatus PYP. The N-terminal chimera behaves both spectrally and kinetically like Hhal PYP, indicating that the Rcaps N-terminus folds against the core of Hhal PYP. A small fraction shows dimerization and slower recovery, possibly due to interaction at the N-termini. The loop chimera has a small amount of the intermediate spectral form and a photocycle that is 20 000 times slower than Hhal PYP. The third chimera, with both regions exchanged, resembles Rcaps PYP with a significant amount of intermediate spectral form (λmax = 380 nm), but has even slower kinetics. The effects are not strictly additive in the double chimera, suggesting that what perturbs one site, affects the other as well. These chimeras suggest that the intermediate spectral form has its origins in overall protein stability and solvent exposure.
PAS signaling domains (acronym formed of the names of the first three proteins recognized as sharing this sensor motif, periodic clock protein of Drosophila, aryl hydrocarbon receptor nuclear translocator of vertebrates, single-minded protein of Drosophila) occur in all forms of life. Although these domains respond to a diverse range of compounds and stimuli, they have been postulated to utilize a common mechanism of intramolecular signaling [1-3]. The PAS fold is composed of a central, twisted, six stranded antiparallel β-sheet, which, in various species, is additionally flanked on both sides by loops and helices . There are hydrophobic cores on each side of the central β-sheet. The chromophore, or ligand, bound by the PAS domain is generally found in the major hydrophobic groove, while, as far as known, the protein–protein interaction with a signaling partner occurs on the other side of the central β sheet. It is the presumed exposure of the hydrophobic face by movement of the N-terminal or C-terminal regions away from the central β-sheet that provides a binding site for interaction partners. Photoactive yellow protein (PYP) is considered to be the structural prototype of PAS domains  and has several advantages over others as a study system, e.g. its high stability and ease of production, the facile initiation of intramolecular signaling by light and the possibility to monitor chromophore-related conformational changes by UV–VIS spectroscopy. These have made PYP an elaborately studied model system for light-to-signal conversion in proteins, and intramolecular signaling in PAS domains in general. Although most studies have been performed on the prototypic Hhal PYP, it has become obvious in more recent years that there are many variations on the PYP theme, not only in terms of different photocycle kinetics but also in terms of spectrally different ground states and structural differences in certain defined regions of the protein.
To date, the PYP protein has only been isolated from three organisms, which happen to be halophilic: Hr. halophila [5, 6], Rhodothalassium salexigens  and Halochromatium salexigens , whereas recombinant PYP has been produced in E. coli from pyp genes of Rhodocista centenaria [9-11], Thermochromatium tepidum , Rhodobacter capsulatus , Rhodobacter sphaeroides [14-16], Idiomarina loihiensis  and Salinibacter ruber [17, 18]. In addition, pyp genes have been found, but the proteins not yet characterized, during genome sequencing of Stigmatella aurantiaca, Rhodopseudomonas palustris and Burkholderia phytofirmans . Both Halorhodospira and Stigmatella contain two copies of pyp [17, 19]. Thus, PYP genes are found in alpha-, beta-, gamma-, and delta-proteobacteria, and also in Bacteroides. It remains clear that PYP is not restricted to photosynthetic organisms, although it may be assumed that the PYP-containing bacteria experience light at some point in their lifecycle.
The diversity of PYP is also reflected in the presumed function, ranging from regulation of phototaxis and twitching motility to cell buoyancy, polyketide synthase and cyst formation, biofilm formation, and bacteriophytochrome regulation, depending on the source [9-11, 17, 18, 20]. However, in each instance, the cellular function still needs to be determined unambiguously. In at least four examples, PYP is found as a fusion protein with its direct interaction partner, which is a bacteriophytochrome domain and either a histidine kinase (Ppr) or a diguanylate cyclase plus phosphodiesterase (Ppd; [9, 11, 12, 21]).
There is also a significant variation in the photocycle kinetics among PYPs. The photocycle of Hhal PYP can be initiated by blue light near the absorbance maximum of 446 nm, which results in trans to cis isomerization of the chromophore, p-hydroxycinnamic acid [1, 22, 23]. The photocycle includes a number of spectrally detectable intermediates. The red-shifted intermediate, I1 (λmax = 465 nm), decays in ca 500 μs to the blue-shifted I2 (λmax = 360 nm). In about 1.5 ms, a major conformational change of the protein exposes a hydrophobic site to solvent [24-28], while forming the I2′ intermediate (λmax = 340 nm). The I1, I2 and I2′ intermediates are in equilibrium, which is pH dependent with I2 favored below a pKa of 6.4 and I2′ above [29-31]. The pKa of 6.4 is thought to be due to E46 in the bleached state. The return to the dark-adapted state occurs with a lifetime of about 160 ms. Since I2′ is the longest-lived intermediate and undergoes a major conformational change, it is generally believed to be the signaling state that interacts with other proteins. Changes in diffusion constant  and H-D exchange, detected using NMR  or monitored using mass spectroscopy , indicates that the N-terminal domain becomes detached from the remainder of the protein during formation of I2′. Recently, a variety of spectroscopic techniques were combined to produce a structural model for the I2′ state . The photocycles of the homologous PYPs are not as well characterized, but where studied, they appear to form comparable intermediates.
The recovery (I2′ to dark state transition) is an order of magnitude slower in the case of Rs. centenaria (Rcen) PYP [9, 10] and Tc. tepidum (Ttep) PYP . A structural explanation for the slower recovery was found in a different orientation of the β4–β5 loop . This loop contains residue M100, which has an effect on reisomerization of the chromophore during recovery [37, 38]. The different conformation of this loop in Rcen PYP positions M100 in such a way that it can no longer efficiently influence reisomerization . The same is thought to be true for Ttep PYP. In addition, we have shown that other substitutions in the same loop region (e.g. Y98Q) can result in structural changes that cause a similar effect in Hhal PYP . On the other hand, the photocycle is about 100-fold faster in Rb. capsulatus and Rb. sphaeroides PYP than in the prototypical Hhal PYP [13, 16]. Interestingly, both Rcaps and Rsph PYP have an M100G substitution in their sequences, which would seem to demand an order of magnitude slower recovery. It was suggested that the β4–β5 loop could possibly adopt another conformation, different from both Hhal and Rcen PYP, and that the reisomerization could be promoted by another residue in the loop . However, two single point mutations to replace residues in the β4–β5 loop with the corresponding residues of Hhal PYP (G100M and K99Q) did not significantly change the kinetics . Rcaps PYP was characterized as a rather unstable protein and initial attempts to obtain crystals for structure determination failed. In addition, it was found that the dark state of Rcaps PYP is composed of three spectrally distinct forms, which may be one reason for impaired crystallization. The absorption spectrum shows an absorbance maximum at 375 nm, besides the 435 nm maximum, with a 1:1 ratio. This 375 nm form is similar to the so-called intermediate spectral form, which was first described in the Y42F Hhal PYP mutant , but was later shown to occur to a small extent in other mutants as well, and has a variable absorption maximum of 370–400 nm, depending on solvent and temperature . Studies of Hhal PYP mutant Y42F showed that this intermediate form consists of a dark-state like structure with a protonated chromophore , which is in rapid equilibrium with the 457 nm yellow form . The two forms share the same photocycle and merge paths early on, presumably through excited state proton transfer. It thus appears that the intermediate spectral form has a protonated trans chromophore.
Additional spectral and photokinetic evidence obtained with Rcaps PYP suggested that the 375 nm form is actually composed of the intermediate spectral form plus an additional form that has an absorption maximum closer to 350 nm and has the chromophore in the cis conformation . It was shown that Rcaps PYP can be photoactivated with both blue (435 nm) and UV light (365 or 386 nm), however, the recovery at 435 nm after UV illumination goes above the preflash baseline. The excess 435 nm form returns to the preflash baseline with a lifetime of several hours, indicating that the 350 nm form is like the photoproduct of the 435 nm form and that there exists a 350–435 nm form dark equilibrium . A similar phenomenon was also observed with Rb. sphaeroides PYP . A working hypothesis is that the Rhodobacter PYPs exist in equilibrium among ionized trans (yellow form), protonated trans (intermediate spectral form) and cis protonated (350 nm) form in the ground state.
Another region of PYP that has been shown to be important for maintaining normal photokinetics, besides the M100-containing β4–β5 loop, is the N-terminal cap (residues 1–25). Deletion of 6, 15 or 23 residues in Hhal PYP results in a progressive 100-to ca 5000-fold slower recovery rate , whereas a 1–25 deletion mutant results in 100-fold slower recovery . The importance of the N-terminus in signaling state formation was recently further explored in a study that provided evidence for the existence of an “ionic lock” between E12 (on the N-terminal cap) and K110 (on the central β sheet) in the ground state, which is broken during I2′ formation, and reformed during recovery . It was the observation of a large salt effect on the formation and decay of I2′, and the lack of this effect in N-terminal deletion mutants, which led to the discovery of this ionic lock . Furthermore, mutagenesis of E12 and especially K110 led to elimination of the salt effect. Although E12 is substituted by a shorter Asp residue in both Rcaps and Rsph PYP, their electrostatic properties and the interaction partner K110 are conserved. Since no structural data are available, it is not known if these residues are also involved in such an ionic lock interaction in Rhodobacter. In addition, Harigai et al.  reported the interaction of Phe6 with the methylene groups of Lys123 near the C-terminus of the protein which also underscores the importance of the N-terminal domain for maintaining photocycle kinetics.
The unusually rapid photocycle of the Rcaps PYP would have profound effects on function, as it determines how long the protein remains in its signaling state. The cis form and intermediate spectral form would also have an effect on function as they increase the wavelength range for a signal response and/or allow the response to be modulated by light quality. As indicated above, previous studies have shown that residues within the β4–β5 loop can have a marked influence on the photocycle in Hhal PYP, but when we made point mutants of these residues in Rcaps PYP, it resulted in very small effects. Moreover, shortening the N-terminus of Hhal PYP by increments resulted in progressively slower recovery. We reasoned that hybrids of Hhal and Rcaps PYP might indicate on which regions to focus our future research. If a special orientation of the Rcaps loop were responsible, it might result in a faster photocycle in the hybrid PYP. If on the other hand, a stronger association of the Rcaps N-terminus with the PAS core were responsible for faster recovery, that should also show up in the hybrid, provided that it folded correctly. It is possible that the Rcaps N-terminus may not fold against the Hhal core, but if so, recovery of the hybrid should be very slow like the truncated proteins. The intermediate spectral form does not necessarily have to be due to the same cause or a single cause, but may be a result of greater solvent exposure of the chromophore, perhaps due to substitution by smaller amino acid residues  or due to lower overall protein stability. Thus, it could be magnified in the double hybrid. A previous attempt at producing hybrids was reported , but it combined so many mutations that conclusions were severely limited. Our approach differs in that we attempted to swap regions of two functional proteins without mutating any specific residues.
Materials and Methods
A structural model of Rcaps PYP was created using SWISSMODEL (http://swissmodel.expasy.org//SWISS-MODEL.html; [50-52]). A single template file was used (2PHY.pdb), which is the structure file for the dark state of Hhal PYP. Sequence identity between Hhal and Rcaps PYP is 44%. The model was generated with ProModII and energy minimization of the model was performed with Gromos96, resulting in a backbone rmsd of 0.22 Å. Hyb3PYP was modeled in the same way to the Hhal PYP structure.
Construction of PYP chimeras
The gene for Hyb1PYP (N-terminal mutant) was designed using polymerase chain reaction (PCR) with pET15b (WTHhalPYP; ) as template and the following primers: HYB1pypF (5′-ATGGAAATCATTCCGTTCGGGACGAACGACATCGACAACATCCTGGCGCGCGAGCCCGCGCAGCTCGACGGCCTGGC-3′) and HhalpypR (5′-CTAGACGCGCTTGACGAAGAC-3′). The Rcaps PYP sequence is underlined. This generated a PCR product of 370 bp that contains the first 60 bp of Rcaps PYP followed by the last 310 bp of Hhal PYP. A second round of PCR was performed with primers designed with restriction sites for NdeI and BamHI on the 5′ and 3′ ends, respectively (NdeIcapsPYPF [GGAATTCCATATGGAAATCATTCCGTTCGG] and BamHIHhalPYPR: [CGCGGATCCCTAGACGCGCTTGACGAAGAC]). Restriction enzyme recognition sites are underlined. The resulting PCR fragment was purified using gel electrophoresis and digested with NdeI and BamHI, after which it was ligated into the predigested pET20b. This resulted in pET20b (Hyb1PYP). The gene sequence was confirmed using DNA sequencing.
The gene sequence for Hyb2PYP (β4–β5 loop mutant) was synthesized by Entelechon (Regensburg, Germany, www.entelechon.com). The gene has a 30 bp region (between 283 and 312 bp, Hhal numbering) replaced with the corresponding region from Rcaps PYP (from 280 to 309 bp, Rcaps numbering). The gene was provided as a pCR-BluntII-TOPO (Hyb2PYP) construct. NdeI and BamHI restriction sites were engineered on the 5′ and 3′ ends during synthesis. The Hyb2PYP gene was digested from pCR-BluntII-TOPO (Hyb2PYP) with NdeI and BamHI and recloned into a pET20b vector, which was predigested with the same enzymes. This resulted in the pET20b(HYB2PYP) construct.
The Hyb3PYP gene (N-terminal/β4–β5 loop double mutant) was designed to be constructed by PCR in a similar way as Hyb1PYP; however, pCR-BluntII-TOPO (Hyb2PYP) was used as the template.
The P102E mutant of Hhal PYP was constructed with the Quick Change Site-Directed Mutagenesis kit (Stratagene) with changes to the manufacturers' protocol as described . After sequencing to confirm the mutation, the gene was expressed using pET20b as described for the chimeric proteins.
Protein production and purification
The three PYP chimeric proteins (Hyb1PYP, Hyb2PYP and Hyb3PYP) were produced as holo-proteins in E. coli BL21(DE3) by coproduction of the biosynthetic proteins. Details of the expression were as described [13, 43]. After harvesting the cells using centrifugation, they were resuspended in 50 mm Tris-HCl (pH 7.5) and cell lysis was performed through sonication for Hyb2PYP and automated French press (Ribi) for Hyb1PYP and Hyb3PYP. An ultracentrifugation step (40 000× g for 1 h) removed all cell debris and the clear yellow-colored supernatants were applied to a 1 mL Q-sepharose FF column (Amersham Biosciences). This column was washed with 0.7 m NaCl and the wash fraction pooled with the flow-through. The sample was dialyzed and applied to a 20 mL Q-sepharose FF column. Elution was performed with 50 mm Tris-HCl (pH 7.5) with increasing amounts of NaCl (steps of 50, 100, 250, 350 and 500 mm NaCl). The yellow fractions started eluting from the column with 250 mm NaCl, except for Hyb1PYP, which eluted in two fractions with about 10 and 250 mm NaCl (labeled a and b). Yellow fractions were pooled and concentrated on an Ultrafree-4 centrifugal filter (5 kDa MW cutoff; Millipore, Bedford, MA), after which Hyb1PYP and Hyb2PYP samples were applied to a Superdex 75 Hiload column (GE Healthcare). Purification was performed on an AKTA Explorer fast protein liquid chromatography system with 50 mm Tris-HCl (pH 7.2) and 200 mm NaCl as running buffer. Hyb3PYP was purified on a Sephadex G50 column by gravity, with the same buffer. Yellow samples were analyzed by SDS-PAGE and UV–Vis absorbance. The pooled samples were dialyzed on an Ultrafree-4 centrifugal filter and applied to a MonoQ anion exchange column (Tricorn MonoQ 5/50 GL; GE Healthcare) on the AKTA Explorer system. Elution was performed with a linear gradient of 0–1 m NaCl in 50 mm Tris-HCl (pH 7.5). After this step, the yield of PYP varied between 10 and 30 mg L−1 of culture.
All PYP protein samples were analyzed with mass spectrometry, to confirm the substitutions and analyze the purity. A mass of 13912.9 kDa was obtained for Hyb2PYP and of 13834.2 kDa for Hyb3PYP, which is close to the theoretical mass of PYP plus chromophore (13915.4 kDa for Hyb2PYP and 13835.4 kDa for Hyb3PYP). The measured masses of Hyb1PYPa and Hyb1PYPb under denaturing conditions were 13938.9 and 13939.9 kDa, which are close to the theoretical mass of 13940.5 kDa. The sole contaminating protein in all samples was identified as apo-PYP by MS/MS, which although not quantitative appeared to be about half the total. This is significantly higher than in wild-type Hhal PYP preparations (which generate ca 10% apo-PYP). However, the low yields are likely due to chimeric mutations as wild-type Rcaps PYP has been observed to generate 90% apo-PYP under the same conditions . In addition, we observed lower yields of holo-PYP when more regions were exchanged in the chimeras.
Absorption spectra were obtained with either a UVIKON 943 (NorthStar Scientific, Leeds, UK) double beam spectrophotometer or a Cary 300 (Varian, Palo Alto, CA, USA) spectrophotometer. All studies were carried out in a universal buffer (5 mm CAPS, CHES, Bicine, HEPES, MES, sodium citrate, 2.5 mm PIPES), unless otherwise indicated. In the steady-state recovery experiments, the protein was first bleached by 1 min exposure to a tungsten lamp (60 W) and the subsequent recovery was measured in the dark. A 410 nm cut-off filter was used to excite only the 445 nm form. This eliminated possible photoreversal from UV absorbing photocycle intermediates. Absorbance changes were measured over a 40–300 min time range, and the kinetic data were fit using SIGMAPLOT 8.0 (Systat Software Inc., San Jose, CA, USA).
Time resolved spectroscopy
The laser flash photolysis apparatus and data analysis protocol were as described previously [6, 13].
Structural homology model
There were two reasons to model Rcaps PYP. The first was to suggest possible structural reasons for the differences in properties, and the second was to help design constructs to test the resulting proposals. Rcaps PYP was modeled into the Hhal PYP dark state structure (2PHY) using SWISSMODEL. The predicted model is very similar overall to the Hhal structure and all the structural elements are maintained (Fig. 1A). The structures can be overlaid with an rmsd of 0.22 Å for all Cα atoms.
A more detailed examination shows that the two regions show a slightly higher degree of variation than the rest. Residues 15–19, which are part of the N-terminal cap region, have adopted a different conformation, with the largest change being that residues Lys17 and Met18 are replaced by a single Arg residue, thereby shortening the α1–α2 loop and orienting Arg 17 outward to the solvent (Fig. 1B). Note that Rcaps PYP is one residue shorter than Hhal PYP (124 versus 125 aa). In addition, the replacement of Asp20 by a Pro residue in Rcaps PYP is expected to induce additional changes in this region. Although the first 14 residues of the model overlay reasonably well with the Hhal structure, the changes in the α1–α2 loop will undoubtedly also exert an effect on the position of the first α helix in the native Rcaps PYP structure. This is even more plausible if one takes into consideration that the N-terminal cap region has the highest temperature factors in all known PYP structures, consistent with a high degree of flexibility. In addition, NMR spectroscopy and deuterium exchange MS analysis have also yielded evidence for the flexibility of the N-terminal region [33, 34]. To study the effect of these differences in the N-terminal region in our model, we chose to replace all of the first 21 Hhal PYP residues with the first 20 residues of Rcaps PYP in the chimera design.
A second difference between the model and the Hhal structure can be found in the β4–β5 loop region (Fig. 1C). Although the Cα chain of the loop conforms to the Hhal structure, several amino acid replacements may lead to significant differences in structure and/or kinetics. The loop clearly lacks M100, which is thought to promote reisomerization of the chromophore to the dark state, yet it is known that Rcaps PYP kinetics of recovery are 100 times faster without this residue. Previous point mutations of G99 and Q98 to the corresponding residues in Hhal (M100 and K99) did not change the PYP properties significantly . The β4–β5 loop in the Rcaps model also lacks a proline residue (P102, replaced by E101 in Rcaps PYP). The lack of this proline is expected to change the conformation, or at least the mobility, of this loop. We therefore constructed the P102E mutant of Hhal PYP; however, that did not change the spectrum, but it did slow the recovery about six-fold (k = 0.94 s−1 versus ca 6.0 s−1 for wild-type HhalPYP; data not shown). This indicates that there is a change in orientation of the β4–β5 loop that repositions M100 to where it is less effective. The model also shows that three residues of the β4–β5 loop are positioned in close enough proximity to form H-bonds and electrostatic interactions that would stabilize the conformation of the loop (N96-E101-K94; see Fig. 1C). These interactions are not possible in Hhal PYP. If N96 is involved in H-bonding with E101, it would imply that the chromophore–protein interaction between the corresponding Asp residue in Hhal PYP (D97) and the Cβ of Cys 69 is lost in Rcaps, which, as a consequence, may have an impact on the chromophore and possibly render it more solvent accessible. As all these are merely predictions made on a static model, and the model does not give any leads toward compensatory substitutions in other parts of the molecule, we decided to replace the entire loop (residues 95–104) with the corresponding loop from Rcaps PYP (residues 94–103) in our chimera. It is known that the entire loop can adopt variable conformations depending on the sequence [36, 39], with a large impact on the photocycle recovery kinetics, which further justifies exchanging the entire loop.
On the basis of the homology model, we constructed three PYP hybrids that had the N-terminal region exchanged (Hyb1PYP), the β4–β5 loop exchanged (Hyb2PYP) and both the N-terminal region as well as the β4–β5 loop exchanged (Hyb3PYP). The chimeric regions are on opposite sides of the beta sheet core structure and are indicated in the modeled Hyb3PYP structure in Fig. 1A.
Hyb1PYP chimera (N-terminal mutant)
|PYP (vis abs. maximum; nm)||Recovery rate||Corresponding lifetimes|
|WT Hhal PYP (445)||6.0 s−1||160 ms|
|WT Rcaps PYP (435)||820 s−1||1.2 ms|
|Hyb1PYPa (445)||5.9 s−1||170 ms|
|Hyb1PYPb (452)|| |
0.31 min−1 (41%)
0.10 min−1 (59%)
|Hyb2PYP (445)||0.018 min−1||57 min|
|Hyb3PYP (444)||0.012 min−1||86 min|
During the first purification step (an anion exchange column) of Hyb1PYP, the yellow fraction was separated into two parts. A major fraction (>90%) eluted from the column with minimal salt (5–10 mm), whereas a second, minor form (<10%), eluted at about 200 mm NaCl. We labeled the fractions as Hyb1PYPa (major fraction) and Hyb1PYPb (minor fraction). The two forms have slightly different absorbance maxima (Fig. 2). The major form shows a typical HhalPYP 445 nm absorbance peak, whereas the maximum of the second form is at 452 nm. On size exclusion chromatography, the 452 nm form eluted with higher MW (possibly as a dimer), whereas the 445 nm form eluted as a monomer. After purification, a mass spectrometric analysis under denaturing conditions showed that both the major and the minor fractions had the expected mass of holo-Hyb1PYP. The two fractions were analyzed separately for their biochemical and kinetic properties.
When Hyb1PYPa was illuminated by a 445 nm laser flash, we found the kinetics of recovery at 445 nm and decay at 360 nm to be very similar to wild-type Hhal PYP (see Fig. 3A and Table 1). The sample was unaffected by room light illumination. The pH dependence is bell-shaped with a maximum recovery rate at pH 7.3 and two pKa's: 5.6 and 9.0 (Fig. 4A). This is similar to Hhal PYP (pKa's of 6.4 and 9.4). As Hyb1PYP has part of the ionic lock region exchanged, we examined the salt dependence of the recovery kinetics. It is known that chaotropic and kosmotropic salts have a distinct effect on the PYP recovery and can be an indication for the presence of the ionic lock . Typically, an initial decrease in recovery rate is observed (up to ca 600 mm salt) due to the breaking of the ionic lock and the shift of the I2/I2′equilibrium to I2′, followed by a reversal of these effects at higher ionic strength . Ammonium sulfate, a known kosmotrope, causes a significant increase in the recovery rate of Hyb1PYP as shown in Fig. 5A.
There is an initial small decrease (up to 0.1 m) followed by a three-fold increase up to 2 m. This indicates that, similar to Hhal PYP, I2 is favored in the I2/I2′ equilibrium under high kosmotropic salt concentrations. To further investigate this, we studied the spectral differences that occur upon illumination. Difference spectra of Hyb1PYPa (taken at 8 and 120 ms, reflecting formation of I2 λmax ca 360 and I2′ λmax ca 340) without salt and in 2 m ammonium sulfate show that the maximum of the photoproduct without salt is close to 345 nm, with a shoulder around 360 nm, whereas in the case of 2 m ammonium sulfate, the absorbance is broader and the maximum is closer to 360 nm (Fig. 5C). This indicates that there is a change in the I2/I2′ ratio under different salt conditions, favoring I2 at high kosmotropic salt concentrations, similar to Hhal PYP. The behavior toward NaCl was also examined. There is a two-fold decrease in recovery rate for Hyb1PYPa when the NaCl concentration is increased up to 2 m (Fig. 5B). The initial decrease in recovery rate upon increasing salt concentration is similar to the Hhal PYP behavior and is likely an indication of the presence of an ionic lock [46, 47]. However, unlike Hhal PYP in the presence of KCl, there is no increase of recovery rate above 600 mm. The high salt effect in Hhal PYP was interpreted as favoring I2 in the I2/I2′ equilibrium (after I2′ was favored up to ca 600 mm). With Hyb1PYPa on the contrary, the decrease appears to be biphasic, with a fast decrease up to 1M NaCl, which is followed by a further slow decrease. This is similar to Hhal behavior with chaotropic salts where, at low ionic strength, the I2/I2′ equilibrium is shifted toward I2′ and shifted even further at higher concentrations . The initial decrease in recovery rate is observed both with NaCl and (NH4)2SO4. Note that the scale of the y-axes in Figs. 5A and B are different, however, the magnitude of the initial decrease is largely the same. At higher concentrations the effects of the two salts are opposite, as one is to expect, as NaCl is more chaotropic than (NH4)2SO4. It has previously been shown that at low salt, the effect on recovery is due to ionic strength, however, at high salt, specific anion effects occur according to the strength of these effects in the Hofmeister series . Thus, at least at low salt concentration, there is little difference between Hyb1PYPa and wild-type Hhal PYP.
The decay rate of the photoproduct is the same at 360 as at 340 nm, in either salt condition (data not shown), indicating that both intermediates return to the ground state with the same rate, which is as expected because they are thought to be in rapid equilibrium. In both cases, the decay data were best fit with a monoexponential curve and the rate was virtually the same as the recovery rate measured at 445 nm under the same reaction conditions.
Hyb1PYPb, the dimer form, was more readily bleached by both blue and white light than Hyb1PYPa, and showed biphasic recovery kinetics in which the fastest phase was about 1000-fold, and the slower phase about 3000-fold slower than in wild-type Hhal PYP (Fig. 3B and Table 1). In addition, the absorption maximum of Hyb1PYPb is red-shifted toward 452 nm and a small shoulder with a broad maximum around 380 nm is present (Fig. 2). The light minus dark difference spectrum of Hyb1PYPb shows the formation of a 358 nm form at the expense of the 454 nm difference maximum (Fig. 6). Due to the shifted dark state maxima, it is uncertain whether the 358 nm absorbance is due to a shifted I2/I2′ equilibrium toward I2 or whether both absorbance maxima are shifted. Interestingly, the bleach of Hyb1PYPb is incomplete and there is a residual form with an absorption maximum at 447 nm, with ca 25% amplitude, after both blue light and white light illumination. When Hyb1PYPa and Hyb1PYPb were analyzed on native PAGE, it was seen that the former migrates as a single band with the expected molecular weight, whereas Hyb1PYPb fractions show both monomer and dimer bands, with an estimated ratio of 1–4, indicating that the residual absorbance after bleaching is likely due to the presence of monomeric Hyb1PYPa in the sample. Increasing the salt concentration to 0.5 m NaCl did not seem to change the dimer-to-monomer ratio of Hyb1PYPb (data not shown). These results suggest that the N-terminal chimera can adopt several conformations, only one of which (Hyb1PYPa) is similar to HhalPYP. The other conformations presumably do not completely cover the hydrophobic interface between the loop and the core beta sheet, but allow dimerization. Whether the dimer dissociates during the photocycle and can adopt the monomeric Hyb1PYPa conformation remains unclear.
Hyb2PYP chimera (β4–β5 loop mutant)
The dark-adapted Hyb2PYP spectrum is very similar to the Hhal PYP spectrum; however, there is an additional small spectral shoulder with a broad maximum around 365 nm (Fig. 2). Note that Rcaps PYP has a significant spectral peak at 375 nm, but the ratio of 375–435 nm is close to 1, whereas in Hyb2PYP the ratio of 365–445 nm is merely 0.3. Therefore, the β4–β5 loop is clearly not solely responsible for the spectral shoulder or for the 11 nm blueshift in Rcaps PYP. In addition to the spectral difference with Hhal PYP, it became clear early in the purification that the 445 nm absorbance of Hyb2PYP could be readily bleached by room light, which is not the case for Hhal PYP.
Figure 6 gives a difference spectrum of Hyb2PYP after illumination for 1 min with blue light (BL), which shows that the 445 nm form is easily bleached (ca 90%) and converted into a 355 nm absorbing form. This difference spectrum is similar to that of the previously characterized PYPs and indicates that the same intermediates (I2/I2′) are formed (compare to the Hhal M100A PYP difference spectra in Fig. 6). It has been shown that I2 and I2′ are spectrally distinct, with the I2 maximum being around 370 nm and the I2′ maximum around 350 nm . It appears from the Hyb2PYP difference spectra that both forms could be present (based upon broad features in Fig. 6). The kinetics of recovery, however, are significantly different from WT Hhal PYP (Fig. 3C and Table 1).
Recovery monitored at 445 nm has a lifetime close to 1 h (k = 0.018 min−1), which is ca 20 000-fold longer than for wild-type Hhal PYP at the same pH and ionic strength, but more similar to that of the HhalPYP M100A mutant (k = 0.081 min−1; 37). In addition, the pH dependence of the recovery kinetics for Hyb2PYP changes from a bell-shaped curve, with a maximum around pH 8 in both Hhal PYP and Rcaps PYP [13, 54], to a sigmoidal pH dependence that can be fit with a single pKa of ca 10 (Fig. 4).
Hyb3PYP chimera (double mutant)
The absorption spectrum of the purified double chimera, Hyb3PYP, can be found in Fig. 2. This chimera has a significant maximum around 380 nm, besides the usual 444 nm absorption peak. Of all the chimeras, this spectrum is most similar to the Rcaps PYP spectrum, although Rcaps PYP has an even more pronounced 375 nm peak. The ratio of 380–444 nm in Hyb3PYP is only about 0.6, whereas the 375–435 nm ratio in Rcaps PYP is close to 1. The difference spectrum of Hyb3PYP is included in Fig. 6. It can be seen that, with blue light illumination, there is a bleach of mainly the 444 nm form, although there is still some minor bleach of the 380 nm form. Due to the bleach of the latter form, the isosbestic point in the difference spectrum changes from 381 nm for Hhal M100A PYP to 379 nm for Hyb2PYP and to 371 nm for Hyb3PYP.
Although the spectrum resembles Rcaps PYP, the kinetics of recovery at 444 nm of Hyb3PYP are more similar to Hyb2PYP, with a lifetime of about 86 min at pH 7.2 (Fig. 3 and Table 1). White light and blue light gave essentially the same recovery kinetics (lifetimes of 88 and 86 min, respectively). The pH dependence of the 444 nm recovery kinetics was sigmoidal, as for the loop mutant, and not bell-shaped (Fig. 4B). Although obtaining datapoints above pH 9.5 was impaired by instability of the protein, it appears that the pH dependence of Hyb3PYP is very similar to that of Hyb2PYP.
To investigate the nature of the 380 nm shoulder in the double mutant, we determined its temperature and salt dependence. Rcaps PYP studies and previous studies on PYPs that have the intermediate spectral form (between 365 and 400 nm), such as the Y42F Hhal PYP mutant, Rb. sphaeroides PYP and Sb. ruber PYP, have demonstrated that temperature, pH and salt have a distinct effect on the 375–435 nm ratio [13, 15, 16, 18, 41]. In Rcaps PYP and Y42F Hhal PYP, the intermediate form is favored in the presence of chaotropes such as ammonium chloride, whereas the yellow form is predominant in higher concentrations of kosmotropes like ammonium sulfate [13, 55]. Figure 7A shows that an increase in NaCl significantly favors the 380 nm form of Hyb3PYP. At 2 m NaCl, about 75% of the 444 nm form is converted into the 380 nm form. The same is true when the titration is done with NH4Cl (Fig. 7B).
Concomitant with the conversion, there is a slight blueshift to 372 nm. With increasing kosmotrope concentration, e.g. ammonium sulfate, there is initially a decrease in 444 nm absorbance, followed by an increase with concentrations of (NH4)2SO4 above 0.5 m (Fig. 7C). Hyb3PYP behaves similarly to Rcaps PYP and Y42F Hhal PYP where chaotropic and kosmotropic salts have opposing effects on the different forms, which is opposite from Sal PYP that is more stable at high concentrations of NaCl . However, Hyb3PYP is unique in its behavior toward kosmotropes because, at lower concentrations (<0.5 m), the 444 nm form is disfavored, before the absorbance increases at higher concentrations (see insert in Fig. 7C), similar to Rcaps PYP. The disagreement that can be observed in Fig. 7C between the 380 and 444 nm changes is likely due to the presence of multiple conformations underlying the 380 nm peak, similar to what was described with Rcaps PYP .
The temperature dependence of the 444 and 380 nm forms of Hyb3PYP is given in Fig. 8. With increasing temperature, there is a decrease of the 444 nm form, which at lower temperature (up to ca 25°C) is not isosbestic. At higher temperature, however, the decrease of the 444 nm form is clearly isosbestic, with the ultimate formation of a 355 nm form up to 65°C, after which precipitation occurred. Hhal PYP spectra are stable up to ca 80°C (melting temperature is 87°C), prior to denaturation and blueshift . This indicates that Hyb3PYP is clearly destabilized compared to HhalPYP. When the absorbance changes are plotted against temperature, the data cannot be fit to a simple sigmoidal curve, as expected on the basis of Hhal PYP data, but appears to be more complex.
The fact that increasing NaCl concentrations favor the 380 nm form, allowed us to study its photokinetics. A difference spectrum of Hyb3PYP in 2 m NaCl shows that blue light illumination not only bleaches the residual 444 nm form but also bleaches the 380 nm form. Further treatment with white light does not cause any additional bleaching of the 380 nm form (Fig. 6B). This indicates that the bleach we observe for the 380 nm form is due to an adjustment of the 380/444 nm equilibrium after photobleaching of the 444 nm form, as the 380 nm form does not absorb blue light (>410 nm). We know from the salt dependence that the 380 and 444 nm forms are in equilibrium, but further evidence for such a fast dark equilibrium will have to come from more detailed fast bleach and recovery kinetics studies, which could not be carried out with our laser set-up (ms time frame). Recent studies with Y42F Hhal PYP have shown that an excited state proton transfer equilibrium forms the basis for such a fast equilibrium between the yellow and the intermediate spectral form in that mutant .
The salt effect on recovery kinetics was also examined. As it can be seen in Fig. 9, there is no significant effect on these kinetics in the presence of NaCl, even up to 2 m concentration. The high salt effect, characterized in Hhal PYP by an increased recovery rate above 600 mm NaCl, appears to be eliminated in Hyb3PYP. Performing the same experiment with the more kosmotropic (NH4)2SO4 showed that there is an increase in the recovery rate at higher salt concentrations, indicating that there is still a kosmotropic effect of salt in Hyb3PYP. However, in this case there was no initial decrease at lower salt concentration observed as there is in Hhal PYP (Fig. 9).
On the basis of the structural homology model we created for Rcaps PYP, we selected two regions that have significant differences among PYPs, the N-terminal 21 amino acids and the 10 amino acid β4–β5 loop, and engineered three new chimeric PYPs that have the core Hhal PYP structure, but contain either the N-term region (Hyb1PYP), the β4–β5 loop (Hyb2PYP) or both (Hyb3PYP) from Rcaps PYP (see Fig. 1A). By creating these chimeras, we exchanged a total of 25% of Hhal PYP with Rcaps PYP. Both Hyb2PYP and Hyb3PYP have a spectral shoulder (at 365–380 nm), besides the main 445 nm absorbance peak. This makes these chimeras spectrally similar to Rcaps PYP; however, the ratio of the spectral shoulder to the main peak is still significantly smaller than in Rcaps PYP (ratios 0.3 and 0.6 versus 1). It has been shown that this “intermediate spectral form” can be observed in all Hhal PYP mutants to varying degrees, except for double mutants with E46Q . Studies with Hhal PYP mutant Y42F have recently shown that the intermediate form results from an intramolecular protonation of the chromophore by E46 . In the yellow form of Y42F (457 nm) the chromophore is deprotonated and E46 protonated, as in wild-type Hhal PYP. Overall, the spectral shoulder, that we observed with Hyb2PYP and Hyb3PYP, behaves very similarly to the Hhal intermediate spectral form in response to changes in temperature and ionic strength. The fact that the magnitude of the intermediate spectral form is not the same in the two chimeras, but is larger in the double mutant, suggests that there is increased solvent exposure as more regions are exchanged. Such increased solvent exposure could be further tested with e.g. hydrophobic dye binding, circular dichroism or NMR studies. We also observed lower production of holo-protein as more protein was exchanged, which indirectly suggests decreased stability. It is possible to affect solvent exposure without substantially affecting stability (as in mutant Y42F), however, instability generally leads to increased solvent exposure before complete denaturation.
The structural model provides a clue for the occurrence of this spectral shoulder in these mutants and ultimately also in Rcaps PYP. The chromophore is normally shielded from solvent by the β4–β5 loop in Hhal PYP, which creates a local environment around the chromophore that causes a significant shift in the pKa's of the chromophore and of E46, allowing this unusual electrostatic interaction [22, 56]. Basically, a charge located within a low protein dielectric environment is intrinsically unstable, but more likely to reside on an aromatic ring where it can be delocalized and further stabilized through hydrogen bonding. Thus, E46 is protonated while the aromatic chromophore is ionized and H-bonded to E46 and Y42. The Rcaps model shows that N96 is involved in H-bonding with E101, which are two of three residues in the β4–β5 loop that are positioned in close enough proximity to form H-bonds or electrostatic interactions (N96-E101-K94; Fig. 1C). The corresponding residue in Hhal PYP (D97) is normally not involved in such a network, but forms an interaction with the Cβ of the chromophore-binding Cys 69. This interaction is lost in the Rcaps model which, as a consequence, renders the chromophore-binding pocket more solvent accessible. This will certainly change the polar environment of the chromophore surrounding, and potentially change the H-bonding network around the chromophore. Note that, in solvent, glutamate has a pKa of around 4, and is therefore deprotonated when exposed to solvent, whereas the pKa for the free chromophore is around 9, which results in protonation in solvent at pH 7. Recent studies have shown that there is a fast protonation equilibrium between the yellow form and the intermediate spectral form in Y42F . The same is likely true for these chimeric mutants and rendering the binding pocket more solvent accessible will shift the pKa values of glutamate 46 and the chromophore more toward the free values. This will shift the proton more toward the chromophore, thus yielding more of the intermediate form.
In addition, Rcaps PYP does not have T50 (replaced by alanine), which normally stabilizes the H-bonding network with the chromophore. This may further contribute to the more extensive 375 nm form in Rcaps PYP, as compared to the chimeras, which still have the Hhal T50 residue. Hyb1PYP does not have the β4–β5 loop exchanged and does not show the intermediate spectral form, which is consistent with this interpretation. Haker et al.  created a model of Rb. sphaeroides PYP in which they also noted increased exposure of the chromophore due to T50A and M100G substitutions. Moreover, we reported a G100M mutant of Rcaps PYP, which significantly reduced the magnitude of the intermediate spectral form . We conclude that the β4–β5 loop is, at least partially, responsible for the presence of the intermediate spectral form in Hyb2PYP and Hyb3PYP by increasing solvent exposure of the chromophore.
PYP chimeras were also prepared by Shirai et al. , one of which contained nine substitutions between positions 87 and 125, including four substitutions in the β4–β5 loop that would increase solvent exposure of the chromophore. This chimera has a prominent intermediate spectral form apparent, which is more similar to wild-type Rcaps PYP and Hhal PYP mutant Y42F than our chimeras and likely contains additional residues that have an effect on the absorption spectrum. One such residue is Y118, which is substituted by Phe in the Shirai et al. chimera, but with Cys in the Rhodobacter PYPs. We made mutant Y118F in HhalPYP, but it has wild-type spectra.
Although there may be spectral resemblances between our chimeric PYPs and Rcaps PYP, the recovery kinetics do not resemble it. The exchange of the β4–β5 loop significantly slowed down the recovery after illumination (about 20 000-fold slower). This can mainly be explained by the removal of residue M100 (Gly in the Rcaps loop), which has a large effect on reisomerization. Mutation of M100 to various residues slows down the recovery by a maximum of 1000-fold [37, 38]. Concomitant with the slower recovery is the change from a bell-shaped to a sigmoidal pH dependence, which is also linked to the removal of M100. Hyb1PYPa does not have the Rcaps β4–β5 loop and has normal Hhal PYP recovery kinetics (except for the Hyb1PYPb fraction that dimerizes, see below). We thus conclude that the loop is not responsible for the faster recovery kinetics in Rcaps PYP and in fact behaves as expected from the absence of M100 when placed within the Hhal PYP framework. However, the fact that replacing the entire loop has slowed down the recovery even more than a single M100 mutation implies that there must be additional factors at play that account for the larger decrease in recovery rate.
There are two consequences of the replacement of the N-terminal cap, one being a change in kinetics of recovery and the other in oligomerization state in one form of the hybrid. The dominant form is monomeric and has more or less normal behavior, but the minor dimeric form is significantly slower. The N-terminal domain is thought to undergo significant movement (and even partial unfolding) during the Hhal photocycle at the level of I2′ formation [26, 32, 34]. One important feature in this light-induced N-terminal movement is the existence of an “ionic lock” between E12 (on the N-terminal cap) and K110 (on the central beta sheet) in the ground state, which is broken during I2′ formation and reformed during recovery . This was demonstrated by the loss of the ionic strength effect in mutant K110A . There is a large salt effect on the decay and reformation of this ionic lock , which is mainly expressed as a decrease in recovery rate up to ca 600 mm NaCl. Determining the recovery rate under different salt conditions therefore allows for an indirect measurement of the behavior of the N-terminal cap region. Although E12 is substituted by a shorter aspartic acid side chain in both Rhodobacter PYPs, their electrostatic properties and the interaction partner, K110, are conserved. The structural model for Rcaps PYP positions D12 slightly more distant from K110 (3.63 Å versus 2.91 Å), although small local rearrangements might bring them close enough for ionic interaction in the actual structure. When studied more closely, the model shows that, although the first 14 residues of the model overlay reasonably well with the Hhal structure, there are changes in the α1–α2 connecting loop that will undoubtedly exert an effect on the position of the first α helix, which contains D12 in the native Rcaps PYP structure.
Hyb1PYP and Hyb3PYP both have the N-terminal region exchanged; however, the salt dependence of their recovery appears to be quite different. Although Hyb1PYPa shows a decrease in recovery rate in NaCl up to 500–600 mm, which indicates the presence of an ionic lock similar to Hhal PYP, there seems to be no indication for the presence of such an ionic lock in Hyb3PYP. The latter has the β4–β5 loop exchanged, in addition to the N-terminal region. The fact that this loop is on the opposite side of the protein (Fig. 1A) and still exerts an effect on the N-terminal domain implies that the exchange of both regions introduces an overall destabilization which is consistent with the increased temperature denaturation and a concomitant increase in the intermediate form. As I2′ is typically more unfolded, this overall destabilization could also be related to the shifted I2/I2′ equilibrium and the presence of mainly I2′ in the difference spectrum of the double chimera Hyb3PYP. The single chimera, Hyb1PYP, behaves like wild-type Hhal PYP when monomeric, both in its absorption spectrum and in kinetics of recovery. However, the small fraction that forms a dimer is red-shifted by 6 nm and recovers very slowly. We conclude that the monomer folds more or less similarly to Hhal PYP, but the Hyb1PYP dimer is likely to be folded differently. More precisely, it may dimerize via hydrophobic residues in a small fraction of the N-terminus that is not folded against the core. Interaction via the exposed core itself is less likely because truncated PYP is not known to dimerize. In this case, the bleached protein may simply form a stable complex with another subunit, which slows down the recovery. As this dimer form is a minor fraction, it is possible that this is also formed with Hyb3PYP, which also has the N-terminus exchanged, but may be lost because this chimera is already intrinsically more unstable.
We conclude from these studies that neither the N-terminus nor the β4–β5 loop alone result in rapid recovery of Rhodobacter PYPs. However, they possibly contribute to this effect by lowering the overall stability of PYP. The β4–β5 loop seems to be at least partly responsible for the intermediate spectral form caused by increased solvent access to the chromophore, whereas, when combined with an exchange of the N-terminal region, it causes even further unfolding of the protein. We reason that the Rhodobacter PYPs may undergo a different photocycle from the prototypical Hhal PYP, whereby the protein in the dark state is already in a partially unfolded state, the ionic lock does not need to be broken to initiate the conformational changes and the photocycle may be more or less confined to the movement of the chromophore. Such a destabilized dark state might have a more I2′-like protein conformation, which should allow for a faster photocycle, given that there will be less need for protein backbone rearrangement. To date, nothing is known about the mechanism of Rcaps PYP recovery and as it is 100-fold faster, with no M100 (or obvious equivalent residue) to enhance reisomerization, the rate limiting step for recovery is likely significantly different than in Hhal PYP. Kinetics on a faster time scale combined with directed mutagenesis will be necessary to determine the rate limiting step in Rcaps PYP recovery.
- 241989) Photoactive yellow protein from the purple phototrophic bacterium, Ectothiorhodospira halophila. Quantum yield of photobleaching and effects of temperature, alcohols, glycerol, and sucrose on kinetics of photobleaching and recovery. Biophys. J. 56, 559–564., , and (