Biogenesis of Phycobiliproteins

All phycobiliproteins contain a conserved, post-translational modification on asparagine 72 of their β-subunits. Methylation of this Asn to produce γ-N-methylasparagine has been shown to increase energy transfer efficiency within the phycobilisome and to prevent photoinhibition. We report here the biochemical characterization of the product of sll0487, which we have named cpcM, from the cyanobacterium Synechocystis sp. PCC 6803. Recombinant apo-phycocyanin and apo-allophycocyanin subunits were used as the substrates for assays with [methyl-3H]S-adenosylmethionine and recombinant CpcM. CpcM methylated the β-subunits of phycobiliproteins (CpcB, ApcB, and ApcF) and did not methylate the corresponding α-subunits (CpcA, ApcA, and ApcD), although they are similar in primary and tertiary structure. CpcM preferentially methylated its CpcB substrate after chromophorylation had occurred at Cys82. CpcM exhibited lower activity on trimeric phycocyanin after complete chromophorylation and oligomerization had occurred. Based upon these in vitro studies, we conclude that this post-translational modification probably occurs after chromophorylation but before trimer assembly in vivo.

Post-translational modification of a protein is an important way to alter its activity or conformation. Phycobiliproteins (PBP), 4 which form the light-harvesting antenna in many pho-tosynthetic organisms, such as cyanobacteria, red algae, and cryptomonads, (1) contain a unique post-translationally modified residue, ␥-N-methylasparagine, on ␤-subunits (2,3). This modification of PBP contributes to the efficient, radiationless transfer of energy to photosystem (PS) I and PS II through the light-harvesting complex called the phycobilisome (PBS) (4). PBP contain other post-translational modifications, which are performed by lyases that attach one or more bilin chromophores to Cys residues on the ␣and ␤-subunits via thioether linkages (5,6).
The ␥-N-methylasparagine residue, found in nearly all PBP isolated from cyanobacteria, red algae, and cryptomonads (4,7,8,10), is located at the ␤-72-position (numbering according to CpcB of Synechococcus sp. PCC 7002; see supplemental Fig. 1), and no such modification occurs on the homologous ␣-position (2)(3)(4). Radiotracer experiments have shown that the methyl group of ␥-N-methylasparagine is derived from S-adenosylmethionine (SAM) (3). The SAM-dependent methyltransferases transfer a methyl group from the sulfonium atom of SAM to a variety of nucleophiles, including oxygen, sulfur, and carbon atoms, and in the case of ␥-N-methylasparagine, a nitrogen atom, on proteins, nucleic acids, carbohydrates, lipids, and small molecules in all organisms (11,12).
Methylation of some amino acid residues is reversible, and in these instances, the function of the modification is to modulate the conformation/activity of an enzyme and for signaling (12). For example, in Escherichia coli and Salmonella enterica, adaptation during bacterial chemotaxis is in part mediated by reversible covalent modifications of transmembrane chemoreceptors, also referred to as methyl-accepting chemotaxis proteins, in which the side-chain carboxyl group of specific glutamate residues within the cytoplasmic domains is methylated by methyltransferase CheR and demethylated by methylesterase CheB (13). Methylation of amino acid residues can also be a way to modulate the function of a protein permanently, and the methylation of PBP ␤-subunits falls into this category. Similarly, the translation release factors RF-1 and RF-2 of E. coli contain a conserved glutamine residue that is methylated by HemK/PrmC (to produce ␦-N 5 -methylglutamine) in order to increase the efficiency of protein synthesis termination by ribosomes (14).
The side chain of ␥-N-methyl-Asn 72 of PC is located in close proximity to the phycocyanobilin (PCB) chromophore at CpcB-Cys 82 (4,15), which serves as the terminal energy acceptor in phycocyanin (PC) (16). This suggested that methylation of Asn 72 might affect the spectroscopic properties of the PCB chromophore at CpcB-Cys 82 . Swanson and Glazer (4) isolated two mutants of the cyanobacterium Synechococcus sp. PCC 7942, pcm-1 and pcm-2, after incubation with the mutagen N-ethyl-NЈ-nitro-N-nitrosoguanidine, which produced PC and allophycocyanin (AP) that lacked methylation at CpcB-Asn 73 and ApcB-Asn 71 (homologous to CpcB-Asn 72 in Synechococcus sp. PCC 7002; see supplemental Fig. 1), respectively, and were shown to lack the methylase activity in extracts (4). Further analyses of the mutants showed that PBP methylation contributes significantly to the efficiency of directional energy transfer to the terminal energy acceptors in the PBS. Using PC mutants of the cyanobacterium Synechococcus sp. PCC 7002, which contained either Asp or Gln in place of the Asn 72 of CpcB, Thomas et al. (17) demonstrated that these substitutions affected both the ground to excited state transition and the excited state characteristics of the CpcB-Cys 82 PCB chromophore. They concluded that ␥-N-methyl-Asn 72 plays an important role in establishing the local environment surrounding the CpcB-Cys 82 chromophore and thereby minimized nonradiative energy loss (17).
We recently reported the identification of the gene, denoted cpcM, which encodes the methyltransferase that modifies the Asn 72 position of PBP ␤-subunits in two cyanobacteria (18). We report here the in vitro characterization of recombinant CpcM from the cyanobacterium Synechocystis sp. PCC 6803. CpcM methylated CpcB, ApcB, and ApcF but not PBP ␣-subunits (CpcA, ApcA, and ApcD). CpcM was more active with CpcB carrying the PCB chromophore at Cys 82 than with apo-CpcB. Last, although CpcM could slowly methylate PC trimers isolated from a cpcM mutant (18), this reaction was much slower than that with apo-PC.

EXPERIMENTAL PROCEDURES
Plasmid Construction-Primers used to amplify the Synechocystis sp. PCC 6803 apcAB genes were 6803 apcA5 (5Ј-CACC-ATGAGTATCGTCACGAAATCAATCTG-3Ј) and 6803 apcB3 (5Ј-GGAATTCAGGACTAGCTCAAGCCAGA-3Ј). Pfx polymerase (Invitrogen) was used in the PCRs using genomic DNA following the manufacturer's instructions. The 1.1-kb amplicon encoding apcAB was cloned into pET100 using the Invitrogen Champion TM pET Directional TOPO expression kit, following the manufacturer's directions. The apcAB genes were cloned into this expression vector in such a way that ApcA was produced with a His 6 tag at its N terminus (HT-ApcA). The calculated molecular mass of HT-ApcA is 21,539 Da, and that of ApcB is 17,217 Da. The molar extinction coefficient at 280 nm for HT-ApcA/ApcB was estimated to be 19,440 M Ϫ1 cm Ϫ1 , and this value was used to calculate concentrations of these proteins.
The cpcBA genes were amplified by PCR from Synechocystis sp. PCC6803 chromosomal DNA using two oligonucleo-tides (cpcB.1, 5Ј-GGAGATTAATCATATGTTCGACGTA-TTCAC-3Ј; cpcA.3, 5Ј-CCCAAGCTTCCAGGCCAGCTA-GAAT-3Ј). The 1150-bp PCR product was digested with NdeI and HindIII and cloned into the NdeI and HindIII sites of pBS150v (19), a derivative of a expression vector that confers resistance to spectinomycin (19,20). Both cpcB and cpcA are expressed from the p trc promoter and can be induced by the addition of isopropyl 1-thio-␤-D-galactopyranoside, but only the CpcB protein has an N-terminal His 6 tag. Apo-CpcA is copurified through its association with apo-HT-CpcB (19). The calculated molecular masses of the proteins are 17,588 Da for CpcA and 21,005 Da for HT-CpcB; the estimated molar extinction coefficient at 280 nm used to determine the concentration of HT-CpcB/CpcA was 29,330 M Ϫ1 cm Ϫ1 .
The apcF gene was amplified by PCR from wild-type Synechococcus sp. PCC 7002 chromosomal DNA using the primers 7002ApcF5 (5Ј-CACCATGCGGGACGCTGTTAC-AAGTC-3Ј) and 7002ApcF3 (5Ј-TCAGATATCTTAGAGA-TCCACTTCGCTCAGTTC-3Ј). After PCR purification, the resulting 0.6-kb product was cloned into the pET100 vector using the Champion TM pET Directional TOPO expression kit (Invitrogen) by following the manufacturer's guidelines. Once the construct was verified, the pET100::apcF plasmid was used to construct another clone. The apcA gene from Synechococcus sp. PCC 7002 was amplified by PCR using the primers apcA5Sac (5Ј-ATTTGAGCTCATATCTAGAGGAGGAATC-3Ј) and apcA3Sac (5Ј-AACGGAGCTCCAGAGGGTAAAAA-CCAC-3Ј). The resulting 0.6-kb product was digested with SacI (restriction site underlined) and cloned into SacI-digested pET100::apcF, such that apcA was downstream of and would be cotranscribed with apcF. Restriction mapping and DNA sequencing were performed to verify this construction. ApcA (17,297 Da) co-purifies with HT-ApcF (22,824 Da). The estimated molar extinction coefficient at 280 nm used to calculate the concentration of HT-ApcF/ApcA was 22,000 M Ϫ1 cm Ϫ1 .
The apcD gene was amplified by PCR from wild-type Synechococcus sp. PCC 7002 chromosomal DNA using the primers 7002 apcD5 (5Ј-CACCATGAGCGTCGTTAGTCAAGTT-ATC-3Ј) and 7002 ApcD3 (5Ј-AACGATATCCTAGGACAT-TGCTTGGGTAATGAAG-3Ј). After PCR amplification, the resulting 0.6-kb product was cloned into the pET100 vector using the Champion TM pET Directional TOPO expression kit (Invitrogen) by following manufacturer guidelines. Once the construct was verified, the pET100::apcD plasmid was used for a second construction. The apcB gene from Synechococcus sp. PCC 7002 was amplified by PCR using the primers 7002apcB5Sac (5Ј-TTAAGAGCTCCTAAGAGAAGGAGTT-ATAACAATGCAAGA-3Ј) and 7002apcB3Sac (5Ј-GATAGA-GCTCGTAAAAGCTTATAAAGAGCTAGACAG-3Ј). The resulting 0.6-kb product was digested with SacI (restriction site underlined) and cloned into SacI-digested pET100::apcD, such that apcB was downstream of and would be cotranscribed with apcD. Restriction mapping and DNA sequencing were performed to verify this construction. ApcB (17,220 Da) co-purifies with HT-ApcD (21,826 Da). The estimated molar extinction coefficient at 280 nm used to calculate the concentration of HT-ApcD/ApcB was 33,050 M Ϫ1 cm Ϫ1 .
Strains and Culture Conditions-The Synechocystis sp. PCC 6803 holo-HT-CpcA protein was produced in E. coli cells as described (22). For protein overproduction in E. coli, expression plasmids were transformed separately into E. coli strain BL21 (DE3) (Invitrogen). All E. coli strains were cultured in Luria-Bertani medium with appropriate antibiotics (ampicillin or spectinomycin, 100 g ml Ϫ1 ). Protein overproduction was induced with the addition of 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside. After further incubation for 4 h, cells were harvested by centrifugation and stored at Ϫ20°C until required.
Purification of Synechococcus sp. PCC 7002 PC and Recombinant Apo-PC-PC lacking methylation at Asn 72 was purified from the cpcM mutant of Synechococcus sp. PCC 7002 (18) cells using ammonium sulfate fractionation and DEAE chromatography as described (23). This protein was used as a substrate in methyltransferase assays as described below. Recombinant Synechococcus sp. PCC 7002 apo-PC (CpcB/CpcA) was purified from E. coli as described (24). The calculated molecular masses for apo-CpcB and apo-CpcA were 18,335 and 17,621 Da, respectively. The molar extinction coefficient at 280 nm was estimated to be 26,770 M Ϫ1 cm Ϫ1 .
Purification of His 6 -tagged Proteins-Metal affinity chromatography was used to purify recombinant HT-ApcA/ApcB, HT-CpcB/CpcA, and HT-CpcA. Cell pellets were thawed and resuspended in 20 ml of cold buffer 0 (20 mM Tris-HCl, pH 8.0, 50 mM NaCl, 50 mM KCl). Cells were lysed by three passages through a French pressure cell at 138 megapascals. After centrifugation of the cell extract at 8,000 ϫ g for 30 min, the supernatant was applied to a column containing 5 ml of nickel-nitrilotriacetic acid resin (Qiagen) as described (19). Purified His 6 -tagged proteins were dialyzed against buffer 0 with 1 mM 2-mercaptoethanol.
Purification of GST Fusion Proteins-GST-CpcM (calculated molecular mass 72,825 Da and ⑀ 280 nm ϭ 106,360 M Ϫ1 cm Ϫ1 ) and the GST (calculated molecular mass 26,989 Da and ⑀ 280 nm ϭ 41,160 M Ϫ1 cm Ϫ1 ; produced from pGEX-2T with no insert) were both purified using glutathione affinity chromatography. Cell pellets were thawed, resuspended in 20 ml of cold buffer 0, and lysed by three passages through a chilled French pressure cell at 138 megapascals. After centrifugation of the cell lysate at 17,000 ϫ g for 30 min, the supernatant was applied to a glutathione-agarose column (5 ml; Sigma) that had previously been washed with 10 ml of equilibration buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl). After washing with 10 ml of equilibration buffer, the recombinant GST-CpcM was eluted with 50 mM Tris-HCl, pH 8.0, 5 mM glutathione. Protein solutions were dialyzed against 50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM 2-mercaptoethanol.
Methyltransferase Assay-Methyltransferase activity was assayed using the protocol described by Swanson and Glazer (4). Each reaction mixture contained 14.5 M substrate (HT-ApcA/ApcB, HT-CpcB/CpcA, CpcB/CpcA, or demethyl-PC from the cpcM mutant), 1.25 M GST-CpcM (or 1.25 M GST for controls), 50 mM Tris⅐HCl (pH 8.0), 10 mM EDTA, 10 mM dithiothreitol, 1.6 M S-adenosylmethionine (Sigma), and 0.4 M [methyl-3 H]S-adenosylmethionine (Amersham Biosciences; 15.0 Ci/mmol) in reaction volumes ranging from 50 to 200 l. Reaction mixtures were incubated at 30°C and terminated by the addition of an equal volume of cold 10% (v/v) trichloroacetic acid. After overnight incubation at 4°C, protein precipitates were recovered and analyzed by SDS-PAGE as described below or by scintillation counting. For quantitation of methyltransferase activity using a scintillation counter, precipitates were collected on 25-mm nitrocellulose filters (Fisher) and washed with 30 ml of 5% (w/v) trichloroacetic acid, followed by 20 ml of 95% (v/v) ethanol. Filters were dried at 37°C for 15 min and then placed in a scintillation vial with 4 ml of scintillation fluid and assayed by scintillation counting to determine the 3 H counts/min. Time course assays were also performed in methyltransferase assays (80-l reaction volume) with apo-HT-CpcA, holo-HT-CpcA, Synechococcus sp. PCC 7002 apo-CpcA/CpcB, or CpcA/CpcB-Cys 82 -PCB (generated from in vitro reaction with CpcS I /CpcU (25). Assays with 14.5 M 7002 apo-CpcA/CpcB and demethyl-PC trimers from a cpcM mutant (18) were performed in a total volume of 85 l and were terminated by precipitation with 5% (w/v) trichloroacetic acid prior to analysis by SDS-PAGE.
Size Exclusion HPLC-Size exclusion chromatography was performed on a Bio-Sil SEC250 column (300 ϫ 7.8 mm; Bio-Rad), equipped with a guard column of the same material (80 ϫ 7.8 mm) using a Waters HPLC pump and detector as described (25).
Polyacrylamide Gel Electrophoresis and Fluorographic Detection-Assays for analysis by SDS-PAGE were performed in 50-l volumes. Proteins were precipitated by incubation with 5% (w/v) trichloroacetic acid at 4°C and resuspended in 20 l of 50 mM Tris⅐HCl (pH 8.0) and an equal volume of 2ϫ SDS loading buffer (26). Proteins from different assays were resolved by SDS-PAGE using a 15% (w/v) acrylamide gel as described in Ref. 25. For fluorographic detection, gels were first soaked in a fixing solution (isopropyl alcohol/water/acetic acid in the ratio 25:65:10, v/v/v) and then in a fluorographic enhancer solution (Amplify; Amersham Biosciences) for 30 min each. Gels were transferred to Whatman 3M paper, dried under vacuum, and exposed to film (BioMax MS film; Eastman Kodak Co.). Visualization of proteins by PCB fluorescence under UV light was performed after soaking gels in 100 mM ZnSO 4 for 2 min (27).

Purification and Methyltransferase Activity of Recombinant
CpcM-As shown in our recent study (18), the cpcM gene is present in all PBP-containing cyanobacteria whose genomes have been sequenced. Mutants lacking CpcM, which is predicted to be a SAM-dependent methyltransferase, no longer synthesize methylated PBP and are sensitive to high light intensity. For overproduction of CpcM, we cloned the cpcM gene Asn Methyltransferase for Phycobiliprotein ␤-Subunits from Synechocystis sp. PCC 6803 into an expression vector to produce N-terminal fusion to GST. Supplemental Fig. 2 shows SDS-PAGE analyses of some of the partly purified recombinant proteins employed in the studies reported here.
To determine if recombinant GST-CpcM had PBP methyltransferase activity, GST-CpcM was initially assayed with recombinant apo-HT-AP and apo-HT-PC (Fig. 1). With [methyl-3 H]SAM as the methyl donor, 3 H was incorporated into trichloroacetic acid-precipitable material in the presence of GST-CpcM but not in control reactions containing GST alone (Fig.  1). Under the assay conditions employed, methylation of both apo-HT-AP and apo-HT-PC was evident after 1 h and was apparently complete after 2 h. Scintillation counting of a sample incubated for 2 h revealed that PC contained 23 times more 3 H than the control containing no CpcM (GST only). Similarly, apo-HT-AP contained ϳ34 times more 3 H than the control incubated with GST for 2 h (Fig. 1). Although the data in Fig. 1 indicated that methylation increased for at least 2 h, 1-h incubations were employed for most other assays, because sufficient 3 H was incorporated within this time to allow facile detection of the reaction products.
When one aligns the sequences of Synechocystis sp. PCC 6803 CpcB, CpcA, ApcA, and ApcB and Synechococcus sp. PCC 7002 ApcD and ApcF, all six proteins contain the conserved Asn at position 72 (Asn 72 ; numbering according to CpcB of Synechococcus sp. PCC 7002) (see supplemental Fig. 1). In order to verify that the recombinant CpcM was displaying the appropriate substrate specificity for ␤-subunits, the products of methylation assays were analyzed by SDS-PAGE and detected by fluorography (Fig. 2B). The migration of the substrates used in the reactions on SDS-PAGE are shown in Fig. 2A. The results showed that 3 H was incorporated into recombinant apo-ApcB (Fig. 2B, lane 1) and apo-HT-CpcB (Fig. 2B, lane 3) only when GST-CpcM was included in the assay mixture. Moreover, GST-CpcM did not methylate CpcA or HT-ApcA detectably. Thus, we concluded that recombinant GST-CpcM had the expected specificity for ␤-subunits of AP and PC.
CpcM Methylates ApcF but Not ApcD-Each PBS core substructure contains two copies of ApcF, a variant AP ␤-subunit, which is important for energy transfer from PBS to PS I and PS II (28 -30), and two copies of ApcD, the variant ␣-AP-subunit (also known as ␣ AP-B ), which also plays a role in energy transfer from PBS to PS I and PS II (28,29,31,32). Like PC and AP, ApcF also contains a ␥-N-methyl-Asn residue (10,18). Recombinant Synechococcus sp. PCC 7002 HT-ApcF was overproduced with ApcA, and recombinant Synechococcus sp. PCC 7002 HT-ApcD was overproduced with ApcB in E. coli cells and copurified for use in CpcM assays (Fig. 3, A and C, respectively). The 1-h assays were performed using purified apo-HT-ApcF/ApcA or apo-HT-ApcD/ApcB as the substrate to determine if CpcM methylates ApcF and ApcD. Trichloroacetic acid precipitates were analyzed by SDS-PAGE, and the fluorograms of the gel are shown in Fig. 3, B and D, respectively. Only one tritiated polypeptide appeared on each fluorogram. In Fig. 3B (lane 1), the size of the tritiated protein matched that of the HT-ApcF protein; the copurified ApcA again showed no tritium incorporation, in agreement with the results in Fig. 2B and the expected specificity of PBP Asn methyltransferase in vivo. In Fig. 3D (lane 1), the size of the tritiated polypeptide matched the ApcB protein (similar to results seen in Fig. 2B); HT-ApcD showed no tritium incorporation, as expected. These results show that CpcM methylates not only the PC and AP ␤-subunits but also the AP-␤-like polypeptide, ApcF, found in PBS cores. CpcM Substrate Preference: Apo-PC or Demethyl-PC?-To determine when methylation of the ␤-Asn 72 residue occurs during the post-translational maturation of PC, assays were performed to compare the ability of GST-CpcM to methylate apo-PC and demethyl-PC substrates. If CpcM methylates one substrate faster than another, then one might infer whether Asn 72 methylation occurs before or after chromophore attachment to the PBP. First, CpcA was used in assays to verify that neither apo-CpcA nor holo-CpcA alone was a substrate for GST-CpcM. Neither apo-CpcA nor holo-HT-CpcA (22) from Synechocystis sp. PCC 6803 was detectably methylated by GST-CpcM after a 1-h incubation. The apo-HT-CpcA was purified from E. coli cells containing pBS414v (encoding the CpcA protein and the ␣-PC lyase CpcE/CpcF) (Fig. 4A, lane 1). Holo-HT-CpcA was purified from E. coli cells containing pBS414v and pAT101. The latter plasmid carries the ho1 and pcyA genes, which encode the enzymes required to synthesize PCB from heme (22) (Fig. 4B, lane 1). This experiment confirmed that GST-CpcM does not methylate the chromophorylated ␣-PC subunit. Absorption spectroscopy (data not shown) and zincenhanced fluorescence of the covalently bound PCB chro-mophore (see supplemental Fig. 3) verified that the HT-CpcA actually contained PCB. After a 1-h incubation with GST-CpcM, the proteins were precipitated and analyzed by SDS-PAGE, and fluorography of the gel was performed. Consistent with the expected specificity for CpcM for ␤-subunits, no 3 H incorporation into apo-CpcA or holo-␣-PC (CpcA-Cys 84 -PCB) subunit was detected (Fig. 4D, lanes 3 and 4).
GST-CpcM activity was also tested to compare its activity with two other substrates: apo-CpcB and CpcB carrying a PCB chromophore at Cys 82 (CpcB-Cys 82 -PCB). Fig. 4D shows the results for assays containing either Synechococcus sp. PCC 7002 apo-CpcA/CpcB or apo-CpcA/CpcB-Cys 82 -PCB as substrates. The apo-CpcA/CpcB-Cys 82 -PCB had been chromophorylated at Cys 82 on CpcB by CpcS I /CpcU lyase, as described (25). Both of these substrates are heterodimeric ␣␤ protomers, but one carries no PCB chromophore and the other has a single PCB attached to the Cys 82 residue of CpcB (see Figs. 4 and 5 in Ref. 25). SDS-PAGE analyses of the proteins employed in these assays are also shown in Fig. 4, A-C. After a 1-h incubation with GST-CpcM, both the apo-CpcB and CpcB-Cys 82 -PCB subunits were radiolabeled, and the two forms of CpcB appeared to be equally methylated (Fig. 4D, lanes 1 and 2). This experiment verifies that the recombinant GST-CpcM enzyme from Synechocystis sp. PCC 6803 recognized the Synechococcus sp. PCC 7002 PC CpcB subunit as a substrate (and not CpcA) and that the presence of a PCB chromophore at Cys 82 did not prevent GST-CpcM from methylating Asn 72 .
Because both substrates appeared to be equally methylated after a 1-h incubation with the methyltransferase, methylation assays were performed for shorter time periods with Synechococcus sp. PCC 7002 apo-CpcA/CpcB and Synechococcus sp.  PCC 7002 CpcA/CpcB-Cys 82 -PCB as substrates. The relative amounts of the substrate proteins used in these assays are shown in the SDS-PAGE analysis of Fig. 4C. Assays were acidquenched after various time increments (Fig. 5). The first protein to appear on the fluorogram, CpcB-Cys 82 -PCB, was visible after a 1.0-min incubation with GST-CpcM (Fig. 5, lane 2). Although it is clear that both apo-CpcB and CpcB-Cys 82 -PCB are substrates for GST-CpcM, the latter protein was more rapidly labeled. The CpcB-Cys 82 -PCB subunit was obviously not fully chromophorylated, because it lacked the PCB at Cys 153 (24,25,33). However, GST-CpcM clearly showed greater activity with the CpcB substrate that carried a PCB chromophore at Cys 82 . We tentatively concluded from these assays that methylation might occur after chromophorylation of the apo-PC subunits.
To investigate whether methylation occurs before or after trimerization of PC monomers, assays were performed using Synechococcus sp. PCC 7002 recombinant apo-CpcA/CpcB (apo-PC) and demethyl-PC purified from a Synechococcus sp. PCC 7002 cpcM mutant (18). Size exclusion chromatography was used to verify that the demethyl-PC purified from the cpcM mutant was in the trimeric (␣␤) 3 form (supplemental Fig. 4). The mass of the PC complexes was determined to be 112,000 Da; this value is very close to the calculated value of 113,208 Da, the mass for (␣␤) 3 PC trimers (holo-CpcA and demethyl-PC have masses of 18,210 and 19,512 Da, respectively) (18). Therefore, we were certain that the unmethylated PC isolated from the cpcM mutant was in the trimeric form. SDS-PAGE was performed to verify that approximately equal concentrations of these two substrate proteins were used (Fig. 6A). Individual assays were set up and quenched with acid at various time intervals (from 1 to 17.5 min). The tritiated proteins were separated by SDS-PAGE and visualized by fluorography (Fig. 6B). Methylation of apo-CpcB was observed within 5 min (lane 5), when the substrate was recombinant apo-CpcA/CpcB. When the substrate was trimeric demethyl-PC from the cpcM mutant, a similar level of methylation took more than 15 min (lane 14). Longer incubations of 1 and 2 h confirmed that CpcM methylated trimeric demethyl-PC (Fig. 6C). Therefore, recombinant GST-CpcM methylates apo-CpcB in ␣␤ monomers faster than chromophorylated CpcB in (␣␤) 3 trimers. Based upon these in vitro studies with purified proteins, we concluded that methylation of Asn 72 takes place after chromophore addition but before trimer assembly during PC maturation.

DISCUSSION
The presence of the unusual post-translational modification, ␥-N-methylasparagine, at the Asn 72 residue of PBP ␤-subunits in most cyanobacteria, red algae, and cryptomonads has been well documented (2,3,7,8). Klotz and Glazer (3) examined a range of cyanobacterial PBP for the release of methylamine after acid hydrolysis and found that PBP ␤-subunits were only rarely unmethylated. Exceptions that were noted were Anabaena variabilis phycoerythrocyanin, Gloeobacter violaceus PC, and Synechococcus sp. WH8103 R-phycoerythrin II. Other PBP from these same organisms did have the modification, however. When the sequence around the ␤-72-position was examined for each of these nonmethylated ␤-subunits, in most cases, the amino acid at the equivalent position was not Asn. Although Apt et al. (34) noted that a few of the 100 PBP sequences in the data base (in 1995) were missing the Asn residue at the position equivalent to ␤-72, this was a very rare occurrence. Moreover, several of the sequences in the data base were obtained by protein sequencing, a method that would result in the misidentification of this modified ␤-Asn 72 residue. For example, using amino acid sequencing of peptides, ␤-72  was identified as Asp in ApcB and Ser in ApcF from Mastidocladus laminosus (35); later, the same group performed the methylamine test to show that both ApcF and ApcB had ␥-Nmethylasparagine at this position (10). Therefore, the majority of PBP appear to have this modification.
Researchers have examined the consequences of this methylation upon the structure and function of PBP and of PBS (4). Several studies have shown that this site-specific methylation of PBP ␤-subunits contributes significantly to the efficiency of energy transfer in PBP by minimizing nonradiative energy losses (4,36). This prevention of nonradiative energy loss occurs by "fine tuning" the spectroscopic properties of the neighboring ␤-Cys 82 chromophore (17). This protection of quantum energy yield presumably provides a selective advantage to cells growing in aquatic niches that are light-limiting (4). More recent characterization of cyanobacterial cpcM mutants revealed a second phenotype associated with this methylation (18). A Synechocystis sp. PCC 6803 mutant lacking the ability to modify Asn 72 of PBP was severely photoinhibited and was unable to grow at even moderately high light intensity. The data presented here show that recombinant GST-CpcM from Synechocystis sp. PCC 6803 modifies the Asn 72 residues of CpcB, ApcB, and ApcF and that GST-CpcM does not modify the Asn residues of ApcA, CpcA, or ApcD. We have verified that tryptic peptides containing Asn 72 derived from PC isolated from the cpcM mutants of two cyanobacteria lack methylation by mass spectrometry (18).
Given the extremely similar primary and tertiary structures of PBP, one might question the origin of such strong specificity for ␤-subunits (see supplemental Fig. 1) (37,38). Comparing the sequences surrounding Asn 72 in various PBP subunits, there is strong conservation of a short sequence motif among the ␤-subunits that is not present in the ␣-subunits. This signature sequence is PGGN(M/A)YT (2) (see supplemental Fig.  1), and it is possible that CpcM can specifically and locally recognize this motif. Alternatively, the enzyme may recognize structural determinants that are distributed across a larger region of the ␤-subunits.
Although there are rather slight structural differences between the ␣and ␤-subunits of PBP, only the ␤-subunit carries the terminal energy "acceptor" ("fluorescing") bilin chromophore at Cys 82 . The methyl group on Asn 72 may help to restrict the conformational flexibility of the Cys 82 chromophore of the ␤-subunit. By helping to lock the chromophore in a rigid and extended conformation, processes such as excited state proton transfer reactions, intersystem crossing, or photoisomerization are decreased (17). It is also possible that the methyl group restricts the approach of oxygen to this chromophore, since the Asn 72 side chain lies very close to the surface of the protein. The modified residue is close to the propionic acid side chain on ring B and probably helps shield the chromophore from solvent (15).
Another way that this post-translational modification affects Asn 72 within the ␤-subunits is by adding bulk to the side chain amide group; this eliminates the possible participation of one of the two amide hydrogen atoms in hydrogen-bonding schemes. Studies on synthetic peptides have suggested that methylation can slow the spontaneous deamidation of the side chain of Asn residues by 45-fold (39). This might be very important if Asn 72 is supposed to hold the ␤-Cys 81/82 chromophore rigid and in place for efficient light harvesting and energy transfer. If the chromophore were to become more flexible, nonradiative relaxation of the excited state might significantly decrease the photosynthetic quantum efficiency of the organism.
A trimeric PC fraction absorbing at 612 nm, which copurified with AP from Thermosynechococcus vulcanus, was purified, and its crystal structure was determined (40). This form of PC was unmethylated at Asn 72 . The authors hypothesized that this form of PC was a connector between rods and the core, but this assertion was based only on the fact that this form of PC copurifies with some AP, and stoichiometric competence for this assertion was not demonstrated. If this type of "connector" were to be required to attach rods onto cores, then the extensive electron microscopic analyses performed on isolated PBS would probably have revealed this extra rod segment, and demethyl-PC should have accounted for a major fraction of the total PC (41)(42)(43)(44)(45)(46)(47)(48)(49). We have also observed small amounts (ϳ1-3%) of unmethylated PC subunits in Synechococcus sp. PCC 7002 and Synechocystis sp. PCC 6803 by mass spectrometry of tryptic peptides, 5 but the levels of these proteins were far too small to function in rod attachment to phycobilisome cores. Therefore, the reported copurification of unmethylated trimeric PC with AP is in our opinion merely a coincidence that does not reveal any functional role.
Methyltransferases are known to modify several other amino acids post-translationally, including Arg, Glu, and Gln. Thomas et al. (17) showed that when the Asn 72 residue was replaced with a Gln residue in CpcB, methylation was still detected. Their experiments suggest that CpcM might recognize either an Asn or Gln residue after the two glycine residues in the signature sequence surrounding the Asn 72 residue of the PC ␤-subunit (17). The Asn residue methylated in PBP is located in a sequence context similar to that of the Gln residues in the RF1 and RF2 release factors that are methylated by the HemK/PrmC methyltransferase (11). As is the case for all substrates for CpcM, a pair of Gly residues precedes the methylated Gln residue in both of these proteins. Klotz and Glazer (3) found that this modification occurred prior to PBS assembly, because they found ␥-N-methylasparagine in free hexamers of PE in a mutant that lacked the linker protein required for PE attachment to PBS rods. We further probed the timing of this post-translational methylation during PBP biogenesis by performing assays on PBP with and without PCB chromophores. Methylation occurred more rapidly when a PCB chromophore was bound to Cys 82 than when no chromophores were bound to CpcA or CpcB, although the methylation rate was only about 3-fold faster. On the other hand, fully chromophorylated, trimeric (␣␤) 3 PC isolated from the cpcM mutant of Synechococcus sp. PCC 7002 was a much poorer substrate for GST-CpcM than monomeric apo-CpcA/CpcB (␣␤). One explanation might be that CpcM methylates the ␤-subunits early in the PBP maturation, either before or just after bilins are added. CpcM showed less activity once the PBP had been assembled into (␣␤) 3 trimers, perhaps because a portion of the substrate recognition determinants are masked and/or because of steric interference. In examining the structure and position of the Asn 72 residue within PC trimers, Duerring et al. (15) noted that the Asn 72 residue was readily accessible on the exterior of the trimer in the flexible loop region between helices B and E, and they speculated that a methyltransferase might recognize this oligomer as a substrate. Although the two Gly residues that occur immediately before Asn 72 may be important for the specificity of CpcM for ␤-subunits, these two amino acids are also present within this flexible loop region and should be accessible in PBP trimers. Therefore, it is likely that CpcM recognizes and binds to a much larger region of the ␤-subunit, and thus the activity level is lower with trimeric or hexameric PC. Further studies with GST-CpcM using other possible PC maturation intermediates generated with the lyases CpcE/CpcF, CpcS I /CpcU, and CpcT might allow a complete assessment of the influence of each PCB chromophore on the methylation reaction. The conclusion from our research is that chromophorylation precedes methylation, which is followed by oligomerization, linker protein insertion, and the assembly of the entire PBS.
In conclusion, with the identification (18) and characterization (this study) of CpcM, the Asn 72 PBP methyltransferase, all of the enzymes required for the post-translational modification of PBP in Synechococcus sp. PCC 7002 have now been identified: CpcE/CpcF phycocyanobilin lyase for the PCB addition at Cys 82 of CpcA (5,9,50); CpcT, the phycocyanobilin lyase for PCB addition at Cys 153 of CpcB (24); and CpcS I /CpcU, the phycocyanobilin lyase for PCB addition at Cys 82 on CpcB and Cys 81 of the ␣and ␤-subunits of AP (25,33). In addition, our results suggest for the first time the order of these post-translational modifications during PBP biogenesis.