Independent evolution of the prochlorophyte and green plant chlorophyll a/b light-harvesting proteins

Abstract

The prochlorophytes are oxygenic prokaryotes differing from other cyanobacteria by the presence of a light-harvesting system containing both chlorophylls (Chls) a and b and by the absence of phycobilins. We demonstrate here that the Chl a/b binding proteins from all three known prochlorophyte genera are closely related to IsiA, a cyanobacterial Chl a-binding protein induced by iron starvation, and to CP43, a constitutively expressed Chl a antenna protein of photosystem II. The prochlorophyte Chl a/b protein (pcb) genes do not belong to the extended gene family encoding eukaryotic Chl a/b and Chl a/c light-harvesting proteins. Although higher plants and prochlorophytes share common pigment complements, their light-harvesting systems have evolved independently.

Figure 1

Alignment of protein sequences deduced from nucleotide sequences of the pcb genes of the prochlorophytes P. hollandica (pcbA), Prochlorococcus sp. CCMP 1378 (Med strain), and P. didemni, and the isiA genes of the cyanobacteria Synechocystis sp. PCC6803 and Synechococcus sp. PCC7942. Identical residues are in white type on a black background; gray squares are either conservative substitutions (within the groups FILMV, FYW, AG, HNQ, EQ, and DN) or columns where two different amino acids were found in both pcb and isiA proteins. Identity of cloned genes was confirmed by comparison with peptide sequences. Tryptic and N-terminal sequences determined from purified Chl a/b-binding protein P32 of P. hollandica were (2)ATTATPEYG(10), (112)GPEDLXQXDFEFA(124), (126)NFPFEWDDAAQA(137), (286)FSVAPYFVDTIDLPNGA(302), and (341)ALGFDFK(347). N-terminal sequence from the 32-kDa polypeptide of Prochlorococcus sp. (Med) was (1)MQTYGNPDVTYGXXAGN(17). Partial Lys-C digests of the 34-kDa Chl a/b protein of P. didemni gave the peptides (K)E(1)MQTYGNPDVEYGXXAGNSRLA and 118(K)EGPARAPKFDFDXGDGKXLGFI(140). The two additional amino acids consistently detected in the first peptide suggest that the P. didemni protein has an N-terminal extension not found in any other members of the family. The numbers flanking the peptide sequences correspond to residue numbers in the alignment. Double-headed arrows are predicted membrane-spanning helices.

Figure 2

Phylogenetic distance tree (neighbor-joining method; ref. 23) of Pcb, IsiA, and PsbC proteins from prochlorophytes and cyanobacteria. Numbers at the branch points represent the bootstrap values for 100 replicate trees; numbers in parentheses are the bootstrap values for the parsimony analysis that generated a similar tree (not shown). The protein sequences were aligned using clustal v (19); trees were generated using protdist and neighbor (pam matrix) or protpars programs of the phylip package (20). (A) Tree using 308 common residues. All regions with gaps were excluded from the analysis, but the same configuration of branches resulted when gaps were included. (B) IsiA branch from trees that included the IsiA-like sequence from P. didemni, using the 206 common residues.