Photochemistry and Photobiology VoI. 56,

No. 6 , pp. 929-934, 1992

Printed in Great Britain. All rights reserved

0031-8655/92 $05.00+0.00

Copyright 0 1992 Pergamon Press Ltd

BIOSYNTHETIC INCORPORATION OF M-FLUOROTYROSINE INTO BACTERIORHODOPSIN E. STARR HAZARD III", RAJNI GOVINDJEE?, THOMAS G. EBREY~ and ROSALIE K. CROUCH' 'Departments of Ophthalmology and Pharmacology, Medical University of South Carolina, Charleston, SC 29425-2501, USA and *Department of Physiology and Biophysics, University of Illinois, Urbana, IL 61801, USA (Received 10 January 1992; accepted 4 May 1992) Abstract-Halobacterium halobium, grown in a defined medium where tyrosine had been largely replaced with m-fluorotyrosine, biosynthetically produced purple membrane. Analysis of this membrane by high pressure liquid chromatography of phenylthiocarbamyl derivatized amino acids of membrane acid hydrolysates revealed that up to 50% of the tyrosine was present as the m-fluorotyrosine form. Yields of the purple membrane decreased as the level of incorporation increased. The experimental purple membrane showed a single 19F NMR resonance at -61.983 ppm (relative to trifluoroacetic acid). The bacteriorhodopsin (bR) in the purple membrane was normal as assayed by gel electrophoresis, isoelectric focusing, circular dichroic spectra, and UV-visible spectra. However, the fluorinated tyrosine bacteriorhodopsins at near neutral pH exhibited slightly slower rates of proton uptake and a slower M-state decay with biphasic kinetics reminiscent of alkaline solutions of bR (pH > 9). These results imply that the tyrosines in bacteriorhodopsin may play a role in the photoactivated proton translocation process of this pigment.

The tyrosine ionization state may change during the photocycle and influence its kinetics. Rothschild and The role of tyrosines in the photochemical and procxFworkers'js and Eisenstein and propose ton pumping activities of bacteriorhodopsin (bR)t that, during the photocycle, a tryosine becomes prois controversial. Several different proposals have tonated in going from bR to the K state. Earlier, implicated tyrosines in controlling the absorption several groups'0-'2 suggested that a tyrosine spectra, photochemical behavior and even proton becomes deprotonated in going from bRSMto the pumping. One line of investigation has been studies M intermediate. In contrast, McDermott et al., l3 on the ionization state of one or more of the eleven using NMR, and Ames et a1.,4 using Raman, find tyrosines in bR. Rothschild and co-workers' and no changes in the ionization state of tyrosines during Eisenstein and co-workers2 proposed that some of the transition to the M state. Roepe et aI.14*1Jand the tyrosines exist in an ionized (tyrosinate) form Lin et ~ l . also , ~ proposed models to explain their in light adapted bR at pH 7.0, requiring a very low vibrational spectra which would not lead to a net pK for tyrosine. The presence of a tyrosinate at this change in the tyrosine ionization upon M formation. pH has been denied by Herzfeld et a1,: using an Hanamoto et ~ 1 . 'suggested ~ that the rise of the M NMR technique, and by Mathies and co-worker~~,~ intermediate in the photocycle of bR depends on using resonance Raman. Under alkaline conditions, the ionization state of a tyrosine residue. At low Balashov et 0 1 . ~have used absorbance changes in pH, where tyrosine is protonated, the rate of rise the 240 nm region, where other aromatic amino is much slower than at high pH where tyrosine is acids such as tryptophan have little absorbance, to proposed to be deprotonated. Some uncertainty in show that several tyrosines become ionized as the many of the photochemical measurements may be pH is raised from 7.0 to 11.0. However, the NMR3 due to the contribution of tryptophan to the lightand Raman studies4 do not find any tyrosines ioninduced absorbance changes ca 295 11m.l' ized at high pH (s12). More recently, Rothschild Finally, several workers have considered the et aL7 have suggested that a tyrosine is only partially question of whether tyrosines participate directly or ionized. The question of the ionization state of the even indirectly in the light-driven proton pumping tyrosines of bR at different pH is still under investiprocess. Mogi et a1." have found that the replacegation. ment of tyrosines of bR by phenylalanines does not seem to affect steady state proton pumping rates by more than 50%; in contrast, Soppa et al.19 have *Towhom correspondence should be addressed. ?Abbreviations: bR, bacteriorhodopsin; HPLC, high per- found a tyr57asn mutant that has no photochemical formance liquid chromatography; NMR, nuclear mag- or proton pumping activity, even though some 70% netic resonance; PI, isoelectric point; PTC, phenylthiocarbamyl; SDS, sodium dodecylsulfate; "FA, of the light-adapted pigment is in the normally physiologically active all-trans conformation. Balatrifluoroacetic acid. INTRODUCTION

~~~

929

E. STARR HAZARD I11 et a!.

930

shov et aL6 have proposed that the falling off of the quantum efficiency of proton release at high pH suggested by Liuzo may be due to the ionization of a tyrosine residue and therefore this tyrosine could have a direct role in the proton release pathway. Fluorination of a tyrosine residue alters its pK. Following the early work of Kuryatov et a1.21 and Ovchinnikov et al.,” we have used the biosynthetic incorporation of fluorinated tyrosine into bR to study the effects of this altered tyrosine on the photochemical and proton pumping properties of bR .

cooH~N+-&-H I

H

~

F

OH Figure 1. Structure of m-fluorotyrosine.

MATERIALS AND METHODS

Bacteriorhodopsin culture. Halobacterium halobium cultures were initiated from slants and grown for about one month in the synthetic medium of Gochnauer and Kushner?’ modified as described by Grey and F i t P with no vitamins and the addition of 5.0 mg/L of tryptophan. After inoculation of the experimental culture medium (minus the normal 200 mg/L Tyr), D , L m-fluorotyrosine (Sigma Chemical, St Louis, MO) was added at 12 and 24 h, pH 6.5. Cultures were left to mature for approximately 2 months under constant illumination at 37°C but without constant agitation. Cells were harvested and the purple membrane isolated using the procedure of Oesterhelt and Stoeckenius.” The extinction coefficient used in protein determination was 63000 M - I cm-’. Spectral measurements. Circular dichroism spectra in the UV region were measured with an AVIV60 spectrometer. Ultraviolet-visible spectra were obtained with a Cary 2200 spectrometer. Dark adaptation was for 1 h and light adaptation was obtained by exposing the sample to a microscope light source with light guides (Dyonics) for 15 min. Nuclear magnetic resonance analysis of lyophilized bR in dimethyl sulfoxide was performed on a 400 MHz Varian VXR400 spectrometer tuned to fluorine resonance frequency but without proton-fluorine decoupling. Trifluoroacetic acid (TFA) was used as a reference standard and m-fluorotyrosine was used as a control. Light-induced absorbance changes were measured on a home built single beam kinetic spectrophotometer similar to that described in Dancshazy et a1.26 The sample was thermostatted at 20°C. The source of actinic illumination was the second harmonic of a Quanta Ray DCR-11 Nd:YAG laser (532 nm; 7 ns pulse; 5 ndlpulse; Spectra Physics, Mountain View, CA). Flash-induced proton release and uptake were observed by measuring the absorbance changes of the pH sensitive dye, pyranine, at 458 nm at pH 7.0. Light-induced absorption changes were measured at 458 nm in the absence and presence of pyranine. The two traces were then subtracted to get the absorbance change due to the dye alone. Flash-induced difference absorption specta (where 20% of the sample was photoconverted) were measured with a gated optical multichannel analyzer (Princeton Instruments, Princeton, NJ). The detector was a PAR OMA model NO. IRYS12G; gate-width for the detector was 4 ks. Electrophoresis. Analytical electrophoresis followed the method of Laemmliz7using sodium dodecyl sulfate (SDS)polyacrylamide 12.5% gels with a 4.5% stack and appropriate molecular weight standards. Gels were stained with Coomassie Blue R250 and destained with methanouacetic acid. Isoelectric focusing was performed using 7% polyacrylamide gels with a final concentration of 5% pH 4.0-6.5 ampholytes (Sigma Chemical). Samples were solubilized in triton X-100 for 24 h prior to focusing, mixed with glycerol and focused on the tube gels (15 X 0.5 cm) at 1 mA per tube for 24 h at 4°C. Gel slices (1 cm) were equilibrated in 50 mM KCI for 30 min and the pH rneas-

ured. Bacteriorhodopsin was readily visible on gels, so gels were not routinely stained. When small amounts of bR were examined, a stain consisting of 27% isopropanol, 10% acetic acid, 0.04% Coomassie brilliant blue R-250, 0.5% CuSO, and 0.05% crocein scarlet in water was used. Gels were destained in isopropanol (12%): acetic acid (7%): c u s o , (0.05YOL Amino acid ~ n a l y s kSamples . were lyophilized and then digested in 0.5 mL 6M HCI in vacuo for 24 h, dried and derivatized to the corresponding phenylthiocarbamyl (PTC) compounds. Identification of amino acids was by HPLC (Pico-Tag column, Waters) and comparison to a standard amino acid mixture (Pierce). .

I

RESULTS

Bacteriorhodopsin was successfully grown in medium in which the tyrosine had been replaced by m-fluorotyrosine (Fig. 1).The level of incorporation of m-fluorotyrosine was determined by HPLC profiles of the hydrolyzates. A representative HPLC chromatogram is shown in Fig. 2. Panel A shows a HPLC chromatogram of wild-type bR for comparison. Note the absence of a FTC-amino acid peak after FTC-valine. The lower two chromatograms are from the same hydrolysate of m-fluorotyrosine incorporated bR. Panel B shows the unadulterated bR hydrolysate. Panel C shows that an authentic m-fluorotyrosine standard co-elutes with peak 3. Incorporation ranged from 10 to 50% m-fluorotyrosine (expressed as a percentage of the total tyrosine pool) for different batches. The yields of bR tended to fall off at the higher incorporation levels. The UV-visible absorption characteristics of the fluorinated bRs were indistinguishable from those of native bR. Spectra of the light and dark adapted forms of 48% m-fluorotyrosine incorporated bR are shown in Fig. 3. The circular dichroism spectra of these fluorinated bR were likewise indistinguishable from those of native bR.29 A sample of approximately 5 mg of 28% m-fluorotyrosine incorporated bR gave a single weak ‘9 NMR resonance at -61.984 ppm (relative to the TFA reference standard; recorded transients >75000; signal to noise ca 11). A reference sample of m-fluorotyrosine in dimethylsulfoxide gave a resonance peak at -61.345 ppm relative to the TFA

m-Fluorotyrosine bacteriorhodopsin analogue

I

Ilms M

1

I

931

n

E c

X

e

s;

0

Y

300

400

500

600

700

nm I

I

0

I

5 Time (min)

Figure 4. Flash-induced difference spectra of 48% mfluorotyrosine bR at 1 ms (top panel) and 4 ms (bottom panel) after the actinic flash. Actinic flash, 532 nm; average of 32 flashes, at 2 s intervals; 5 mllpulse; pH 6.5; temperature 20°C.

10

Figure 2. HPLC chromatograms of phenylthiocarbamyl (FTC) derivatives of amino acids from in V ~ C U Oacid hydrolysis of bR. Peak 1 is the FTC-Tyr, 2 is PTC-valine, 3 is PTC-m-fluorotyrosine, and 4 is PTC-methionine. (A) Hydrolysate of wild-type bR; note the absence of peak 3 in this sample. (B) Hydrolysate of bR grown in the presence of m-fluorotyrosine, Peak 3 represents 48% of the total tyrosine in this sample. (C) Injection of a duplicate sample of the bR hydrolysate [e.g. the same as (B)] to which had been added an authentic m-fluorotyrosine standard. The differences in elution time for the quartet (elution time about 5 min) of PTC-amino acids in these chromatograms are evidence of the normal variation in retention times for duplicate [e.g. (B) and (C)] injections.

1

300

500

700

nm

Figure 3. UV-visible absorption spectra of light and dark adapted 48% m-fluorotyrosine incorporated bR. The for this respective visible light and dark adapted,,A sample were 567 nm and 562 nm. The corresponding UV A,, were 275 and 276 nm for the light and dark adapted respectively.

standard. We concluded from this that the incorporation had indeed been successful, corroborating the amino acid analysis. SDS gel electrophoresis of the purple membrane produced by the growth procedures described above shows the pigments to be of a molecular weight of about 26000. Isoelectric focusing gels show a PI of 5.24 for both normal bR and derivatized bR. The PI results correspond well to those of Miercke et a1.3O and Ross et a1.3' except that we were unable to detect any of the minor 5.6 or 6.1 proteins. Some minor banding (