Effect of Oxygen on Translation and Posttranslational Steps in Expression of Photosynthesis Genes in Rhodobacter capsulatus

Abstract

The formation of the photosynthetic apparatus in Rhodobacter capsulatus is regulated by oxygen tension. Previous studies have shown a regulatory effect of oxygen on the transcription of photosynthesis genes and on the stability of certain mRNA segments. Here we show that oxygen affects puf and puc gene expression posttranslationally and that this regulation depends on the presence of bacteriochlorophyll. Our data suggest that this posttranslational effect of oxygen on puf and puc expression is due to the primary effect of oxygen on bacteriochlorophyll synthesis or assembly of pigment protein complexes. Oxygen does not affect the rates of translation of puf-encoded proteins.

The formation of the photosynthetic apparatus in facultatively photosynthetic bacteria is mainly regulated by oxygen partial pressure in the environment. Rhodobacter capsulatus cells contain low levels of pigment protein complexes under aerobic conditions, and energy is generated by aerobic respiration. A reduction of the oxygen partial pressure strongly induces the formation of pigments and pigment binding proteins, and an intracytoplasmic membrane system develops (for a review, see reference 12). Many investigations have addressed the effect of oxygen on transcription of genes encoding pigment binding proteins and the enzymes that catalyze pigment synthesis. Several trans-acting factors have been identified that activate the transcription of photosynthesis genes under low oxygen tension or function as repressors under high oxygen tension (reviewed in references 3 and 4). The puf operon (Fig. 1A) encodes proteins of light-harvesting complex I (LHI complex) (genes pufB and pufA) and of the reaction center (genes pufL and pufM) and the proteins PufQ, which is involved in regulation of bacteriochlorophyll synthesis (1), and PufX, which most likely assists in organizing the bacterial photosystem for efficient transduction of light energy (15). The non-pigment binding protein of the reaction center is encoded by the puhA gene. The puc operon encodes proteins involved in the formation of the LHII antenna complex.

FIG. 1.

Partial physical maps of puf and puc operons and plasmids used in this study. (A) Operons encoding R. capsulatus photosynthetic apparatus pigment binding proteins. mRNA species detectable by Northern hybridization are indicated by arrows with open heads, and transcriptional starts are indicated by arrows with solid heads. Stabilizing mRNA secondary structures are also indicated (by “pins”) for the puf operon. All mRNAs drawn in Fig. 1 originate from processing of the highly unstable primary puf and puc transcripts, which cannot be detected in Northern blots. (B) Transcriptional aph-puf fusion in pRK415 (19); (C) translational aph-puf-lacZ fusions used in this study (the lacZ gene is not drawn to scale); (D) β-Galactosidase activity expressed from the aph-lacZ fusions listed in panel C and relative increase of expression (fold) after shift from high to low oxygen.

The levels of puf and puc mRNA are determined not only by the rate of transcription but also by the rate of mRNA decay. Individual segments of the polycistronic puf operon mRNA exhibit different stabilities, leading to differential expression of puf-encoded genes (5, 22). The pufBALMX segment decays at a higher rate under high oxygen tension than under low oxygen tension (23). The molecular mechanism responsible for this oxygen effect has not yet been elucidated.

In order to systematically study the effect of oxygen on posttranscriptional steps in gene expression, we have compared the puf and puc mRNA levels and the rate of synthesis of the corresponding proteins.

Correlation of mRNA levels and rates of protein synthesis and incorporation.

If the expression of puf and puc mRNAs were solely regulated on the levels of transcription and mRNA stability, the rates of synthesis of individual puf and puc proteins would strictly correlate with the mRNA amounts. We quantified the levels of puf and puc mRNAs after the transition of R. capsulatus 37b4 (DSM936) grown in minimal malate salt medium (10) from growth under high oxygen (20% [monitored by an Ag-Pt electrode]) to that under low oxygen (1 to 2%) with Northern blots obtained by using a phosphoimager. Cultures (100 ml) were grown in baffled flasks under vigorous agitation to an optical density (at 660 nm) of 0.4 to 0.6. Aliquots (4 to 5 ml) of these cultures were then transferred into a small vessel and further incubated at 32°C. Due to the respiration of the bacteria, the oxygen tension dropped to 1 to 2% within 3 to 5 min. The oxygen tension was then adjusted to 1 to 2% by supplying air and stirring the culture and was monitored throughout the experiment. In order to determine the rates of protein synthesis, the cultures that had been grown with low oxygen for different times after the shift were pulse-labeled with 40 mCi of l-[³⁵S]methionine (Amersham) for 3 min. Membrane fractions were isolated (21), and 20,000 cpm of each sample was separated on sodium dodecyl sulfate gradient-polyacrylamide gels (24). No significant levels of labeled reaction center (RC) and LH proteins could be detected in the cytoplasmic fraction (data not shown). The radioactivity incorporated into puf- and puc-encoded proteins was quantified by using a phosphoimager (Fuji BAS 1000) and TINA software (Raytest). The radioactivity densities of the individual bands were corrected by subtracting the background radioactivity values of representative regions. By this method only those proteins that are synthesized and incorporated into the membrane within 3 min of pulse-labeling are detected.

As displayed in Table 1 there was a strong increase of the puf and puc mRNA levels (6.9- to 9.7-fold, respectively) and of the Puc and Puf proteins (8.4- to 14.2-fold, respectively) that were synthesized and incorporated into the membrane after the drop in oxygen tension. Since the kinetics of mRNA levels and protein synthesis and incorporation rates correlated quite well (not shown), these results did not point to a significant contribution of translational or posttranslational mechanisms to the regulation of puf and puc operon expression. In order to be able to better discriminate between transcriptional and posttranscriptional regulation of puf and puc genes by oxygen, we either blocked mRNA synthesis with rifampin or expressed the puf operon from an oxygen-independent promoter.

TABLE 1.

Relative increases of mRNA levels and rates of synthesis and incorporation of proteins after shift from high- to low-oxygen tension^a

Strain	Relative increase of the indicated mRNA or protein
	RC			LHI			LHII
	2.7-kb pufBALMX mRNA	PufL protein	PufM protein	0.5-kb pufBA mRNA	PufB protein	PufA protein	0.5-kb pucBA mRNA	PucB protein	PucA protein
37b4	9.0 ± 2.2 (30)	8.6 ± 1.9 (30)	8.4 ± 2.3 (30)	6.9 ± 1.6 (60)	14.1 ± 3.4 (60)	12.9 ± 3.9 (60)	9.7 ± 2.0 (60)	12.0 ± 3.1 (60)	14.2 ± 3.2 (60)
Rifampin-treated 37b4	—b	—	1.3 ± 0.2 (5)	—	3.8 ± 1.1 (20)	2.0 ± 0.4 (20)	—	—	1.7 ± 0.5 (5)
ΔRC6(pRK4apuf)	—	—	1.7 ± 0.4 (15)	—	9.0 ± 2.6 (120)	3.6 ± 0.6 (60)	6.1 ± 1.8 (60)	9.0 ± 2.3 (120)	3.2 ± 0.7 (180)
DE335(pRK4apuf)	1.4 ± 0.3 (60)	—	—	1.7 ± 0.4 (60)	—	2.6 ± 0.7 (120)	11.0 ± 2.9 (60)	1.4 ± 0.1 (60)	4.3 ± 0.8 (60)

^aThe data are average values (fold) of at least three independent experiments ± the maximal deviation. The time point (minutes) when maximal increase was observed is indicated in parentheses. 

^bIn some cases no increase (less than 1.3-fold) could be observed. 

Analysis of protein synthesis and incorporation in the presence of rifampin.

To test whether reduction of oxygen leads to an increased incorporation of pigment binding proteins in the membrane in the absence of mRNA synthesis, we added rifampin (150 μg/ml) to the cells when the oxygen tension was reduced from 20% to 1 to 2%. Cells pulse-labeled directly after the shift to low-oxygen tension incorporated 60% of the radiolabeled methionine, whereas only 35% of the methionine was incorporated 60 min after the shift to low oxygen tension. Equal amounts of total proteins were loaded per lane, based on silver staining of the gel. Although the total amount of radiolabeled proteins decreased after the transition of the cells in the presence of rifampin, the amount of certain pigment binding proteins clearly increased for the first 20 min after the oxygen shift (Fig. 2B and C). For the PufM protein we detected a transient 1.3-fold increase of incorporated radioactivity directly after the drop in oxygen and addition of rifampin (Table 1). During further incubation the levels of methionine incorporated in PufM decreased with the same kinetics as the level of the 2.7-kb pufBALMX mRNA (Fig. 2A and C; Table 1), which has a half-life of 8 min. These data suggest that the rate of PufM synthesis and incorporation is almost exclusively determined by the amount of the corresponding mRNA. Both LHI proteins, PufA and PufB, showed an increase in the level of incorporated radioactivity (2.0- and 3.8-fold, respectively), while the level of the 0.5-kb pufBA mRNA decreased with a half-life of 30 min (Fig. 2; Table 1). Primer extension analysis showed that the major 5′ end of the 2.7-kb pufBALMX and that of the 0.5-kb pufBA mRNA disappeared with the kinetics identical to those of the 0.5-kb puf mRNA band on Northern blots (data not shown). The amount of radioactivity incorporated into PucA increased by a factor of 1.7, while no increase for PucB was observed (Fig. 2). The 0.5-kb pucBA mRNA level decreased with a half-life of 28 min during this time. These data suggest that the increase of the puf-encoded LH proteins after a reduction of oxygen is in part due to translational or posttranslational regulation of gene expression.

FIG. 2.

(A) Northern blot analysis of total RNA (8 μg per lane) isolated from wild-type strain 37b4 at different time points after a shift to low oxygen tension with simultaneous addition of rifampin at time point 0. Hybridization was performed with puc- and puf-specific DNA fragments. (B) Autoradiography of the membrane proteins synthesized within 3 min of different time points after shift to low oxygen and simultaneous addition of rifampin at time point 0. (C) Quantification of RC, LHI, and LHII proteins together with their mRNA species (symbolized by circles) from the data shown in panels A and B. The value at time zero (t₀) was set as 1. In panel RC, RC-L (PufL) is symbolized by diamonds and RC-M (PufM) is symbolized by squares. In panels LHI and LHII, the α subunits of the light-harvesting complexes LHI and LHII (PufA and PucA, respectively) are symbolized by diamonds and the β subunits (PufB and PucB, respectively) are symbolized by squares.

Correlation of puf mRNA levels and rates of synthesis of puf-encoded proteins in a strain transcribing the puf operon from an oxygen-independent promoter.

Since a drug like rifampin affects transcription of all mRNAs resulting in the disturbance of many cellular processes, we decided to also study the expression of the wild-type puf genes transcribed from the aph promoter (pRK4apuf) (Fig. 1B), which is known to be unaffected by changes in oxygen (18). Plasmid pRK4apuf was transferred into the mutant strain ΔRC6 (8), which has the puf operon deleted from the chromosome by triparental mating (20). Despite constant puf mRNA levels, we detected significant increased rates of synthesis and incorporation for the PufA and the PufB proteins (Table 1). This increase was, however, smaller than that observed in wild-type cells, confirming the significant influence of oxygen on the transcription of the puf operon. As well, the level of the puc mRNA that is transcribed from the oxygen-controlled promoter on the chromosome as the rates of synthesis of the PucA and PucB proteins showed strong increase (9.0- and 3.2-fold, respectively [Table 1]). The increase in puc-encoded proteins was also considerably smaller than that in wild-type strain 37b4. Since the PufQ protein affects both puf and puc expression (2), this result can be explained by the reduced expression of pufQ under low oxygen tension in strain ΔRC6(pRK4apuf) compared to that in wild-type cells. In summary, these results support the view that oxygen affects puf and puc gene expression on the level of translation or posttranslationally.

Quantification of rates of translation of puf genes by lacZ translational fusions.

In order to discriminate between gene regulation on the translational level and that on the posttranslational level, we constructed translational fusions between different puf genes and the lacZ gene (Fig. 1C) and transferred them into R. capsulatus wild-type strain 37b4. In these constructs we replaced the oxygen-regulated puf promoter by the aph promoter that has been described to be independent of oxygen in R. capsulatus (18). β-Galactosidase levels were determined as described previously (18, 25).

When the lacZ gene was fused to pufL (plasmid pPH6apufL [Fig. 1C]), to pufB (plasmid pPH6apufB), or to pufA (plasmid pPH6apufA), we determined changes of the β-galactosidase levels after reduction of oxygen tension, which were in the range measured for the control, the aph-lacZ fusion on plasmid pPHU264 (18) (Fig. 1C). These data indicate that oxygen does not affect the rates of translation of pufL, pufB, or pufA significantly (Fig. 1D).

As expected from the higher stability of the pufBA mRNA segment compared to that of the pufLMX mRNA segment, the total activity of β-galactosidase in strains carrying the pufB-lacZ or pufA-lacZ fusion (344 to 550 Miller units for pufB, 7,217 to 7,825 Miller units for pufA) was significantly higher than that in strains carrying the pufL-lacZ fusion (12 to 18 Miller units). The different values obtained for the pufB or pufA fusion may indeed reflect differences in the translational rates of the two genes or may be due to differences in the secondary structures of the RNAs transcribed from the two constructs (the secondary structures are artifacts created by the fusion).

Synthesis and incorporation in the membrane of pigment binding proteins in a strain that does not synthesize bacteriochlorophyll.

The results presented above show that the increase of the rates of synthesis and incorporation of the puf- and puc-encoded proteins is due to oxygen regulation’s affecting a posttranslational step of gene expression. One potential target of oxygen control is the incorporation of the proteins into the membrane. During the assembly to photosynthetic complexes, the proteins interact with the bacteriochlorophyll molecules. It is known from many investigations that the rate of bacteriochlorophyll synthesis is also oxygen regulated (3). It is conceivable that the amount of available bacteriochlorophyll determines the rate of incorporation of the pigment binding proteins. In order to test this hypothesis we repeated our in vivo labeling experiments by using strain DE335(pRK4apuf) (for DE335, see reference 27), which does not produce bacteriochlorophyll and does not transcribe the puf mRNA from the chromosome due to the insertion of an Ω cassette but carries plasmid pRK4apuf, allowing puf transcription from the aph promoter. It was shown previously that radioactively labeled pigment binding proteins can be detected in the membrane in the absence of bacteriochlorophyll. They will, however, undergo turnover which can be monitored in chase experiments (9, 21). After the drop in oxygen tension, we found a maximal 1.4-fold increase of the 2.7-kb pufBALMX mRNA in strain DE335(pRK4apuf) and only small increases of the relative amounts of the PufB and PufA proteins synthesized and incorporated into the membrane (Table 1). The 0.5-kb puc mRNA that is transcribed from its own (chromosomal) oxygen-regulated promoter increased by a factor of 11 after the drop in oxygen. The puc-encoded proteins, however, showed maximal increases by factors of 4.3 and 1.4 for PucA and PucB, respectively (Table 1). This finding suggests that the absence of bacteriochlorophyll in strain DE335 allows the incorporation of only a very limited amount of pigment binding proteins into the membrane.

Possible posttranslational steps subjected to oxygen control are (i) protein stability before incorporation, (ii) rate of incorporation into the membrane, and (iii) assembly to photosynthetic complexes and consequent stability of the protein in the membrane. Despite much effort, the sequence of steps in the incorporation of pigment binding proteins and the assembly of pigment protein complexes is not known (reviewed in reference 13). The binding of pigments to LH proteins seems to be an early step that influences assembly. There is some evidence that the PufQ protein may assist in the assembly of bacteriochlorophyll and proteins (16). The expression of some bacteriochlorophyll genes is oxygen regulated (6), and oxygen also affects bacteriochlorophyll synthesis posttranscriptionally (7). Although final proof is lacking, there is some evidence that late steps in this biosynthetic pathway are coupled to the membrane. Our results show that the posttranslational effect of oxygen on the expression of puf and puc genes indeed depends on the presence of bacteriochlorophyll. Blockage of bacteriochlorophyll synthesis at a late step almost abolished the oxygen-dependent increase of the PufB, PufA, and PucB proteins, whereas we could still observe a 4.3-fold increase of PucA. These differential results for the individual LH proteins may reflect differences in the assembly process, like the sequential insertion of the α and β subunits (reviewed in reference 11). An effect of chlorophyll on posttranscriptional steps in chlorophyll apoprotein expression in plants was also described previously (17). In Chlamydomonas rheinhardtii, a role for chlorophyll in the stabilization of certain chlorophyll apoproteins and possibly in the translation of others was demonstrated. A regulatory effect of chlorophyll on the stability of chlorophyll apoproteins was also demonstrated for the plastids of higher plants (26). In vitro data suggest that chlorophyll-dependent accumulation of chlorophyll apoproteins in barley etioplasts is regulated on the level of translation (14).

Our data show that the expression of the puf and puc genes encoding pigment binding proteins in R. capsulatus is affected by oxygen at three levels of gene expression. In addition to showing the well-established regulation at the level of transcription and at the level of mRNA stability, our study suggests an effect of oxygen at the level of incorporation of the proteins into the membrane and assembly to photosynthetic complexes.