Genes for the peptidoglycan synthesis pathway are essential for chloroplast division in moss

Isolation of Homologues That Are Related to Bacterial Peptidoglycan Synthesis.

To isolate plant homologues that are related to bacterial peptidoglycan synthesis, the full-length EST library of P. patens (15) was searched by using tblastn for the amino acid sequences of peptidoglycan synthetic enzymes from E. coli and cyanobacteria. We found and sequenced the nine genes that are related to peptidoglycan biosynthesis (Fig. 1): MurA, B, C, D, E, and F, and Pbp genes, two genes for d-Ala-d-Ala ligase (Ddl), and dd-carboxypeptidase (Dac). dd-carboxypeptidase is a monofunctional Pbp (16). All of the derived proteins, except for the P. patens MurB (PpMurB) and PpMurC proteins, were predicted to have plastid-targeting sequences by the targetp program (17), whereas the PpMurB protein had a putative mitochondrial-targeting sequence. The predotar program (18) predicted a plastid-target signal for the PpMurC protein. The genome sequence of A. thaliana was determined in 2000 (19). We searched the Arabidopsis genome and found five genes that were related to peptidoglycan biosynthesis: MurE, MraY, MurG, and two Ddl genes. The A. thaliana MurE (AtMurE), AtMraY, and AtMurG genes had putative plastid-targeting signals, which were predicted by the TargetP program. The two AtDdl genes lacked a targeting signal according to the TargetP and Predotar programs. No genes for MraY and MurG in P. patens are known, and although these genes were not found in the full-length cDNA library of P. patens, it is not certain that these genes do not exist.

Characterization of PpMurE Gene.

Because only MurE genes were found with the plastid-targeting signals in both A. thaliana and P. patens, we analyzed the MurE genes. ESTs for MurE genes were identified from plant species such as loblolly pine, lettuce, potato, soybean, cotton, wheat, and rice. Ignoring the putative transit peptides, the similarity between the amino acid sequences of A. thaliana and P. patens MurE proteins was 58% (Fig. 6, which is published as supporting information on the PNAS web site), whereas that between the PpMurE (without the nonidentical transit peptides) and the cyanobacterial MurE was 36%. To detect localization of the PpMurE protein in a cell, we constructed a plasmid for expressing the PpMurE protein fused with a GFP. Polyethylene glycol-mediated transformation with the generated plasmid showed that GFP fluorescence was observed in chloroplasts of P. patens, corroborating the computer prediction (Fig. 2a). This result suggests that the PpMurE protein is located in the stroma of chloroplasts because the PpMurE protein has no transmembrane domains and the bacterial MurE protein exists in the cytosol of E. coli.

In P. patens, a gene-targeting technique has already been established (20). The genomic region of the PpMurE gene was isolated by using genomic PCR. The neomycin phosphotransferase (NPTII) gene, which is driven by the cauliflower mosaic virus (CaMV) 35S promoter and terminated by the polyadenylation signal of the nopaline synthase gene, was inserted into a ClaI site located in the third exon of the cloned PpMurE genomic DNA (Fig. 2c). Polyethylene glycol-mediated transformation of P. patens was carried out with the constructed plasmid. Genomic regions of the PpMurE gene in the transformants were amplified by PCR to verify gene knockout (Fig. 2d). With a primer set located on both sides of the ClaI site, only 2.6-kbp bands corresponding to the size of the MurE genomic region (0.6 kbp) plus the NPTII gene (2.0 kbp) were amplified from the genomic DNA of the transformant lines E1–E3, indicating knockout of the PpMurE gene. When PCR was conducted with primers in the CaMV 35S promoter and in the genomic regions outside of the part used for constructing the PpMurE gene disruption, amplified bands were detected only from the PpMurE knockout transformants. Although these results suggest that the PpMurE gene was disrupted in the E1–E3 transformants, these PCR analyses did not detect additional insertions into the P. patens genome. Therefore, a Southern hybridization analysis with the restriction enzymes, which have no sites in the PpMurE and NPTII genes, was carried out to determine the copy number of the PpMurE gene in the transformant genomes (Fig. 2e). Although the E1 genome had several PpMurE genes, no other insertion of the PpMurE genes was detected in the E2 and E3 genomes. Sizes of the hybridized bands for the E2 and E3 genomes were longer than that of the wild-type because of the size of the NPTII genes, suggesting that the disrupted PpMurE gene was exchanged for the original gene and that there was no additional PpMurE gene. RT-PCR analysis showed that the PpMurE transcripts were not detected in the E1–E3 transformants (Fig. 2f).

Protonema cells in the wild-type plants had ≈50 chloroplasts, whereas three transformant lines, E1–E3, had macrochloroplasts (Fig. 2b). Although development to gametophores from protonemata occurred normally, macrochloroplasts appeared in leaf cells. Moreover, spores were generated despite the low frequency of gametophores. When spores germinated, macrochloroplasts also appeared (Fig. 2b). Time-lapse image analysis revealed that some macrochloroplasts in the transformants divided during cell division, whereas the chloroplasts of wild-type plants did not (Movies 1 and 2, which are published as supporting information on the PNAS web site). Viewed under an electron microscope, no obvious differences in the shape or stacking of thylakoid membranes between the giant chloroplasts of the transformants and those of wild-type plants were observed (Fig. 3). Detailed structure of chloroplasts in the transformants was same to those in the antibiotic-treated cells (data not shown). To carry out transient assay for complementation, we constructed a plasmid with the GFP and PpMurE genes driven by the rice actin promoter. The polyethylene glycol-mediated DNA transfer into protoplasts of the PpMurE knockout transformant was carried out, and regenerated cells were observed under microscopy at 5 days after transformation (Fig. 2g). When transformation was performed with the plasmid including the PpMurE gene, P. patens cells with GFP fluorescence showed normal chloroplast phenotypes. On the other hand, no restoration of phenotypes occurred when protoplasts were transformed with the GFP vector. These results suggest that the PpMurE gene is closely tied to chloroplast morphology, especially to chloroplast division in moss.

Characterization of PpPbp Gene.

The Pbp gene was found in P. patens, and Pbp inhibitors block chloroplast division in moss (11, 14); the Pbp gene was absent in A. thaliana. The predicted amino acid sequence of the PpPbp gene suggests that it has transglycosylase and transpeptidase properties (Fig. 7, which is published as supporting information on the PNAS web site). Computer analysis of the PpPbp amino acid sequence predicted plastid localization. To investigate the subcellular localization of PpPbp, we fused the N-terminal 106 aa to the N terminus of GFP. This PpPbp–GFP fusion protein was clearly targeted to the chloroplasts (Fig. 4a). Endosymbiotic theory suggests that peptidoglycan wall is located between the inner and the outer envelope membranes (see Fig. 5). Therefore, the PpPbp protein may exist in intermembrane space. However, we did not determine the exact location of PpPbp in chloroplasts by the GFP fusion experiment. To observe its location, generation of antibodies against PpPbp is now in progress. The knockout lines for the PpPbp gene were generated in P. patens by the same method used for generation of the PpMurE gene disruption lines (Fig. 4c). We obtained three transformants with macrochloroplasts (Fig. 4b), and Southern analysis showed that these transformants had the PpPbp knockout gene (Fig. 4d). Although the bands in the P1 and P2 transformants were expected because of the restriction site in the NPTII gene, one band in the P3 transformant was larger than expected, possibly because of the rearrangement of genomic DNA during the integration of plasmid DNA. Northern hybridization found no transcripts of the same size as that of the wild-type PpPbp gene transcripts (Fig. 4e). Electron microscopic observations showed no obvious differences in the shape or stacking of thylakoid membranes between the PpPbp knockout transformants and wild-type plants (Fig. 4b). In addition, the PpPbp gene restored giant chloroplast phenotypes of the PpPbp knockout transformant to normal in the GFP-expressing cells (Fig. 4f). These results suggest that the PpPbp gene, as well as the PpMurE gene, functions similarly on chloroplast division in P. patens and that organelle-division mechanisms differ among land plants, even in peptidoglycan-specific ones.

Plastid Evolution.

Plastids of glaucophytes, red algae, and higher plants have evolved as siblings (Fig. 5). In the first stage of endosymbiosis, the protoplastids may have had three membranes with peptidoglycan layer, and then one membrane was lost (9). Several studies have shown that the plastids in glaucophytes, which are sometimes called cyanelles, are surrounded by a peptidoglycan wall (21, 22). The structure of the cyanelle peptidoglycan resembles that of cyanobacteria, and β-lactam antibiotics block cyanelle division (23, 24). These organisms must possess sufficient genes to form the peptidoglycan between the outer and inner plastid envelopes, although no genes tied to the synthetic pathway have been reported. Conversely, red algae have no genes for peptidoglycan synthesis based on the entire genome sequence of the ultrasmall red alga, C. merolae (25), although we searched for homologous genes. Therefore, plants evolved in three distinct ways: glaucophytes retained peptidoglycans, red algae lost them, and green plants retained some peptidoglycan genes. Although the genome of A. thaliana contains five genes that are essential for peptidoglycan synthesis (19), their functions have not been determined. We are now analyzing a T-DNA tagged mutant for the AtMurE gene. The results will demonstrate whether the MurE gene is needed for plastid division in higher plants.

Our results shown in this paper strongly suggests that chloroplasts in moss have wall-like structures. However, under the electron microscope, no obvious structure was detected between the two chloroplast envelopes in moss (Fig. 3). Although cyanelle peptidoglycan is shown as black circle in Fig. 5, no visualized evidence has been shown for its existence by ordinary electron microscopic techniques. Steiner and Löffelhardt (26) detected the peptidoglycan layer of the cyanelle by immunoelectron microscopic observation with antibodies against peptidoglycan from E. coli. Iino and Hashimoto (27) succeeded in showing cyanelle peptidoglycan by adopting silver methenamine staining method. Therefore, because of its thinness, we may not be able to detect peptidoglycan of moss chloroplasts by electron microscopic technique used in this paper, if it exists. If chloroplasts of moss have peptidoglycan wall, its existence must affect maintenance of chloroplasts including protein translocation in addition to morphology and division of chloroplasts. If there is no wall-like structure in moss chloroplasts, the genes found in this study must be related with construction of division site of chloroplasts. Detection with silver methenamine staining or immunodetection may be worthy of further study.

Plant Material and cDNA Clones.

We used Physcomitrella patens (Hedw.) Bruch & Schimp subsp. patens. The full-length cDNA library of P. patens has been described (15), and EST data are available in the PHYSCObase (http://moss.nibb.ac.jp). Sequencing the full-length cDNA for the PpMurC, PpMurD, and PpPbp genes showed that each gene contained two frameshifts in the coding region. Therefore, we used the RT-PCR method with a TaKaRa RNA-PCR kit (Takara Bio, Shiga, Japan) for cloning the cDNA of these genes, and detected cDNA without frameshifts.

Cellular Localization of PpMurE– and PpPbp–GFP Fusion Proteins.

To construct the PpMurE–sGFP fusion gene, the region for the N-terminal 104 aa of the PpMurE gene was amplified with primers 5′-GGGTCGACCATGGCGCTCCAGTGGATCCAGAA-3′ and 5′-CATGCCATGGGTGACACTCTCGCTTCATTCAG-3′ from the full-length EST clone of the PpMurE gene. To construct the PpPbp–sGFP fusion gene, the region for the N-terminal 106 aa of the PpPbp gene was amplified with primers 5′-GGGTCGACCATGGAGTGCATTGTGCTCG-3′ and 5′-CCTCATGAGCTTTACAGTAAAACCCCTGT-3′ from the P. patens genomic DNA. Genomic DNA was isolated from the protonemata of P. patens by using the CTAB method (28). Each amplified DNA was digested with SalI and BspHI to cut restriction sites on the primers and then inserted into the SalI/NcoI-digested 35SΩ-sGFP(S65T) plasmid (29).

Moss Transformation.

The genomic region of the PpMurE gene was amplified from the isolated genomic DNA (10 ng) by PCR, with the primers: 5′-CTCTCCAGTGTCGAAGGTCTAGAC-3′ (MurE-F1) and 5′-ATGCGGGACATTGTACCGAAACTC-3′. The amplified band corresponding to the PpMurE genomic region was cloned into pT7Blue vector (Novagen) with dam⁻ competent cells (SCS110; Stratagene) to remove methylation of the ClaI site. The NPTII gene, with the CaMV 35S promoter and terminator, was generated from the plasmid pTN3 (30) by digestion with EcoRV and inserted into the blunted ClaI site of the cloned PpMurE genomic region. Transformation of P. patens was carried out as described (30). The PpMurE genomic regions in the transformants were amplified by PCR with the MurE primer set: 5′-GGAGACAAGAGTGCAGATGT-3′ and 5′-TTGGACCACATTCTCGGACG-3′ from the genomic DNA of the transformants. Gene disruption was confirmed by PCR with the primer in the external region of the cloned genomic PpMurE gene, 5′-GTTGTTCCAGCTCCAGTATC-3′ and the primer in the CaMV 35S promoter, 5′-GCAATGGAATCCGAGGAGGT-3′. The probe for Southern hybridization was obtained with the PCR DIG probe synthesis kit (Roche Applied Science), the MurE-F1 primer and primer located just before the ClaI site; 5′-CACAGCATCCACTGCTTCGA-3′. There was no EcoRV or SacI restriction site in the MurE genomic region or in the inserted NPTII gene cassette. Total RNA was isolated from the wild-type and transformant protonemata as described (15). RT-PCR was carried out with DNase-treated RNA, downstream primer 5′-ATGCGGGACATTGTACCGAAACTC-3′ and upstream primer 5′-ATGCTGTCGCTAGGCTGTTG-3′. These primers were located on both sides of the ClaI site in the PpMurE genomic region.

The genomic region of the PpPbp gene was amplified by PCR with primers, 5′-TGCACTGTGGTTATGGATTC-3′ and 5′-CGTGCGATATTCAATACTGG-3′. The amplified band was cloned into pT7Blue vector with an XL10 competent cell (Stratagene). Cloned genomic DNA was digested with KpnI and subcloned by self-ligation to remove additional restriction sites. The gene for NPTII with the CaMV 35S promoter and terminator was amplified for the plasmid pTN3 with primers from the BamHI site, npt II-F: 5′-TTCGGATCCTCGAGGTCGACGGTATCGATAAGC-3′ and npt II-R: 5′-TTCGGATCCCGGCCGCTCTAGAACTAGTGGATC-3′. The amplified band was digested with BamHI and inserted into the BamHI site of the cloned PpPbp genomic region. The probe for Southern hybridization was amplified from the P. patens genomic DNA with the PCR DIG probe synthesis kit and primers, 5′-ACTGCCCTAATCGATGGCATTCAT-3′ and 5′-ACTGCCCTAATCGATGGCATTCAT-3′. In the probe used for Northern hybridization, the downstream region of the PpPbp gene was amplified from the cDNA clone by PCR with primers 5′-TTGTCTCTGTCACTTGGAGG-3′ and 5′-GAGCTCGACTAAGATACAAGGTGTGCG-3′ and subcloned into pT7Blue vector. Cloned DNA was digested with BamHI, and the PCR DIG probe was synthesized by in vitro transcription with ThermoT7 RNA polymerase (Toyobo, Osaka, Japan) and DIG-11-UTP.

Microscopic Observations.

Bright-field images of cells were recorded with a charge-coupled device (CCD) camera (Nikon DXM1200 or Zeiss Axiocam) under microscopes (Olympus BX60 or Zwiss Axioskop 2 plus). Movement of chloroplasts was monitored by using a bright field under an inverted microscope (Zwiss Axiovert 200M) equipped with the CCD camera. Under electron microscopy, samples were fixed in 2% glutaraldehyde buffered with sodium cacodylate (pH 7.2) and 1% osmium tetroxide, dehydrated through an ethanol series, and embedded in Spurr’s resin. Thin sections were cut and stained with uranylacetate and lead citrate, and were then observed with a JEM-1200EX electron microscope (JEOL, Tokyo).

Transient Expression of Functional Gene for Complementation.

The PpMurE gene was amplified from the full-length EST clone with the primers: 5′-GGGTCGACCATGGCGCTCCAGTGGATCCAGAAG-3′ and 5′-ATGCGGGACATTGTACCGAAACTC-3′. Blunting and kination of the amplified DNA was carried out with Takara BKL kit (Takara Bio). Then, amplified DNA was inserted into SmaI site between the rice actin promoter and terminator of the rbcS gene of pTFH22.4 vector (31) including the GFP gene, which is driven by CaMV 35S promoter. For amplification of the PpPbp gene, cDNA was synthesized with DNase-treated RNA and oligo(dT) primer. The PpPbp gene was amplified from this cDNA with the primers: 5′-ATGGAGTGCATTGTGCTCGCGTC-3′ and 5′-TTCTGCAGCTAAGAGGGGGTGATGCCAGT-3′. The amplified DNA was cloned into pT7Blue vector (Novagen). The PpPbp gene was excised with the restriction digestions of SalI, EcoRI, and ScaI, blunted with Takara Blunting kit (Takara Bio) and inserted into SmaI site on pTFH22.4. At 5 days after transformation, P. patens cells with fluorescence of GFP were observed under microscopy. Transformants used were the E2 transformant for the PpMurE gene and P3 transformant for the PpPbp gene.