UAG is a sense codon in several chlorophycean mitochondria

( Springer-Verlag 1996

Curr Genet (1996) 30: 29—33

OR I G I N A L P AP E R

Yasuko Hayashi-Ishimaru · Takeshi Ohama Yoshimi Kawatsu · Keiko Nakamura · Syozo Osawa

UAG is a sense codon in several chlorophycean mitochondria

Received: 23 October/6 December 1995

Abstract The mitochondrial genetic code of those land plants and green algae that have been examined does not deviate from the universal one. A red alga, Chondrus crispus, is the sole reported example throughout the algae that uses a deviant (non-universal) mitochondrial genetic code (UGA"Trp). We have analyzed 366-bp DNA sequences of the gene for mitochondrial cytochrome oxidase subunit I (COXI) from ten chlorophyceaen algae, and detected 3—8 in-frame UAG codons in the sequences of five species. Comparisons of these sequences with those of other algae and land plants have shown that most of the UAG sites in Hydrodictyon reticulatum, Pediastrum boryanum and ¹etraedron bitridens correspond to alanine, and those of Coelastrum microporum and Scenedesmus quadricauda to leucine. The three species in which UAG probably codes for alanine are characterized by zoospore formation in asexual reproduction and form a clade in the COXI phylogenetic tree. The two species in which UAG codes for leucine are known to form daughter coenobia and pair in the tree. This is the first report on a deviant mitochondrial genetic code in green algae. Mutational change(s) in the release factor corresponding to UAG would be involved in these code changes. No genetic code deviation has been found in five other species examined. Key words Deviant genetic code · COXI · Chlorophyceaen mitochondria · Release factor

Y. Hayashi-Ishimaru · T. Ohama ( ) · K. Nakamura · S. Osawa Biohistory Research Hall, 1-1 Murasaki-cho, Takatsuki, Osaka 569, Japan Y. Kawatsu Division of Natural Science, Osaka Kyoiku University, Kasiwara, Osaka 582, Japan Communicated by M. Yamamoto

Introduction The classification of algae has been based mainly on a combination of modes of cell division and reproduction as well as ultrastructural and biochemical properties. In some cases, comparisons of DNA or RNA sequences have been considered. We have been constructing phylogenetic trees of algae based on the mitochondrial (mt) gene for COXI both because of its slow evolutionary rate among mt genes and the sequencing data available for this gene in several algae and land plants (Ku¨ck and Neuhaus 1986; Kadowaki et al. 1989; Oda et al. 1992; Wolff et al. 1994). In some instances, the phylogenetic relationships from DNA sequence comparisons have been evaluated by molecular genetic characteristics, such as the position of introns (Kuhsel et al. 1990; Manhart and Palme 1990; Ohta et al. 1993), RNA editing (Hiesel et al. 1994), gene organization (Kuhsel et al. 1990; Oda et al. 1992), and deviant genetic code (Ohama et al. 1993; Boyen et al. 1994). The deviant codon, if it exists, should be useful, because it is vertically inherited and is not transferred laterally to other organisms. The mitochondria of land plants and green algae have been believed to use the universal genetic code (Ku¨ck and Neuhaus 1986; Boer and Gray 1988; Oda et al. 1992; Wolff et al. 1994; for a review see Osawa 1995), because no examples of deviant codes have been reported in them. In the present study we have found two types of deviant codes in green algae. The distribution of these codes within the algae is consistent with the COXI phylogenetic tree.

Materials and methods Materials. The list of green algae used in this study is shown in Table 1. Most of them were purchased from the Culture Collection of Algae and Protozoa (CCAP, England) and the National Institute of Environmental Study (NIES, Japan). Hydrodictyon reticulatum and »olvox sp. were kindly supplied by Dr. Kyoko Hatano (Kyoto

30 Table 1 List of green algae in the COXI tree and their characteristics Species Order Volvocales Chlamydomonas reinhardtii »olvox sp. Order Chlorococcales Actinastrum hantzschii (NIES-415) Chlorella reisiglii (CCAP-211/59) Chlorella vulgaris (NIES-227) Kirchneriella lunaris (CCAP-243/1) Prototheca wickerhamii Coelastrum microporum (CCAP-217/1A) Scenedesmus quadricauda (NIES-96) Hydrodictyon reticulatum Pediastrum boryanum (NIES-209) ¹etraedron bitridens (CCAP-282/1)

Colony formation

Motility of vegetative cell

Mode of asexual reproduction

UAG codon assignment

COXI GC%

Data source

! #

# #

Bipartition Daughter colony

Stop !*

44% 42%

U03843 D63661**

# ! ! # ! # # # # !

! ! ! ! ! ! ! ! ! !

Autospore Autospore Autospore Autospore Autospore Daughter coenbium Daughter coenbium Zoospore Zoospore Zoospore

— — — — Unassigned Leu Leu Ala Ala Ala

37% 43% 36% 39% 34% 44% 41% 41% 41% 42%

D63660** D63652** D63763** D63653** U02970 D63656** D63658** D63654** D63659** D63657**

*!"not examined; **"this study University, Japan). They were cultured in the recommended medium and stored at !120 °C until use (Table 1). Preparation of DNA for PCR. The frozen algal cells (0.1—0.5 g) were broken mechanically by a micro-dismembrator (B. Brown, Germany) in a Teflon capsule with a tungsten ball (diameter 8 mm) for 1 min at maximum power. The pulverized material was incubated in 1 ml of TE buffer [10 mM Tris · HCl (pH"7.4), 0.1 mM EDTA], containing 200 lg/ml of proteinase K (Boehringer) and 0.6% SDS, for 12 h at 50 °C. Nucleic acids were extracted once with phenol, precipitated with ethanol, and dissolved in 0.2 ml of TE buffer. Amplification of a part of the COXI gene by PCR. Using a site in the COXI gene where a series of amino acids is conserved, we designed a pair of primers (5@-¹C¹AGAAC¹AG¹GGA¹CTTYTTYGGNCAYCCNGARGTNTA-3@, and 5@-¹G¹AAAACGACGGCCAG¹GCNACNACRTARTANGTRTCRTG-3@, R: A or G, Y: C or T, N: T, C, A or G) for PCR. To facilitate the direct sequencing of the PCR product, each primer was constructed to have an annealing sequence (shown by italic letters) for the sequence-primer at the 5@-end. The conditions for PCR were as follows: one cycle of PCR contains three steps; 94 °C for 1 min, 39 °C for 2 min and 72 °C for 2 min. The cycle was repeated 30 times using DNA Thermal Cycler Model 480 (Perkin Elmer). Direct sequencing of PCR product. After the PCR reaction, the amplified DNA fragment was purified through a 1.0—1.5% agarose gel and extracted from the gel slice by centrifugation in a disposable capsule having a membrane filter inside (Suprec-1, Takara Co., Japan), according to the supplier’s manual. Sequencing reaction. The purified PCR product was sequenced by the cycle sequencing method using an automatic sequencer (Applied Biosystems, Model 373A). Preparation of total RNA and cDNA synthesis. Scenedesmus quadricauda or H. reticulatum cells (0.2 g) were broken mechanically as described above. Nucleic acids were dissolved in 4 M guanidinium thiocyanate. The pellet obtained after centrifugation was dissolved in TMG buffer (TE buffer with 5 mM of magnesium) containing 6 units of DNase (Stratagene), and incubated at 37 °C for 30 min. Twelve Micrograms of crude RNA were obtained. The first cDNA strand was synthesized using this RNA (5 lg) as a template by the addition of avian myeloblastosis virus reverse transcriptase

(Pharmacia) with the first primer DNA (5@-GTATTAATTTTACCATAGCTTGGT-3@ for H. reticulatum; 5@-GTATAGATTTAGCCTGCATTTGGT-3@ for S. quadricauda) having the sequence complementary to that of COXI. By the addition of the second primer (5@-ATCAACACCTGCGTTTGCTAGTA-3@ for H. reticulatum, 5@GTCTACTCCTGCATTAGCAAGTA-3@ for S. quadricauda) and ¹aq DNA polymerase, the first cDNA strand was converted to double-stranded DNA and amplified under the same PCR conditions as described above. A control experiment was carried out without addition of the reverse transcriptase to confirm that no PCR products were formed. Construction of the COXI phylogenetic tree. The COXI phylogenetic tree was constructed by the neighbor-joining method (Saitou and Nei 1987) based on the evolutionary distance matrix calculated by Kimura’s two-parameter model (Kimura 1980) using the software package MEGA (Kumar et al. 1993). The sequences were mainly aligned manually with the aid of the computer software package Clustal V (Higgins et al. 1992). Bootstrap (Felsenstein 1985) re-samplings (1000 times) were carried out to evaluate the branching pattern.

Results We determined a 366-bp region of the COXI gene for the ten species of green algae listed in Table 1. Neither insertions nor deletions were required for multiple alignment. In the sequenced region for Coelastrum microporum, H. reticulatum, Pediastrum boryanum, S. quadricauda and ¹etraedron bitridens, 3, 6, 5, 4 and 8 in-frame TAG codons were detected, respectively, whereas the other species had no TAG. Most of the TAG sites in H. reticulatum, P. boryanum and ¹. bitridens corresponded to alanine in other green algae and land plants (Fig. 1). In C. microporum and S. quadricauda, the commonest amino acid for the TAG sites was leucine. A part of the COXI cDNA of H. reticulatum and S. quadricauda was amplified by PCR, and sequenced directly. All TAGs were confirmed to be in the cDNA,

31

Fig. 1 Multiple alignment of predicted amino-acid sequences of a part of the COXI gene based on a 366-bp DNA sequence. The sequences of the species in parentheses were from the data base. UAG sites are shown by ‘ * ’. A serine site coded by a UCN codon is shown by ‘s’ and that coded by AGY as ‘S’. The TAG sites detected in the sequence of H. reticulatum, P. boryanum and ¹. bitridens mostly correspond to alanine (A) sites of other species, while those in S. quadricauda and C. microporum to leucine (¸). Chlorella r."Chlorella reisiglii, Chlorella l."Chlorella vulgaris

Fig. 2 COXI phylogenetic tree based on the neighbor-joining method at the DNA level. The predicted amino-acid assignment of UAG is shown in parentheses. The dotted line indicates the lineage in which the deviant codons are used. Bootstrap percentage values are shown at the branching point

indicating that these sites were not edited in mRNA. The results suggest that UAG is used as a sense codon most likely for alanine in H. reticulatum, P. boryanum and ¹. bitridens and for leucine in C. microporum and S. quadricauda. The TAG alanine codon was commonly used in the above three species and the frequency reached, on average, 43% of all alanine codons (6 UAG codons out of 14 alanine sites [6/14 (43%)], 5/15 (33%) and 8/15 (53%) in H. reticulatum, P. boryanum and ¹. bitridens, respectively). Four TAG leucine codons out of 14 leucine sites [4/14 (29%)] and 3/16 (19%), were detected in S. quadricauda and C. microporum, respectively. The COXI phylogenetic tree was constructed using the sequences determined in this study together with those reported for Chlamydomonas reinhardtii (Boer et al. 1985) and Prototheca wickerhamii (Wolff et al. 1994) (Class Chlorophyceae), taking the Euglena gracilis COXI sequence (DDBJ nucleotide sequence database, accession no. D63655) as an outgroup. The tree is composed of three clades (Fig. 2). The first included Chlorella reisiglii, Chlorella vulgaris, P. wickerhamii and Actinastrum hantzschii. Except for A. hantzschii, these green algae have common characteristics, i.e. they have no flagellum in the vegetative cells and do not form colonies. All of them reproduce through autospores in asexual reproduction. The second clade included »olvox sp. and C. reinhardtii. Their vegetative cells are motile with flagella. »olvox forms colonies, whereas Chlamydomonas does not. The third clade is composed of C. microporum, H. reticulatum, Kirchneriella lunaris, P. boryanum, S. quadricauda and ¹. bitridens. Except for ¹. bitridens, all of them form colonies, are non-motile in the vegetative cells, and reproduce through zoospores. The UAG codon is used as a sense codon (either for alanine or leucine) in mt

32

of the species belonging to the third clade, except in the case of K. lunaris which is the outgroup for this clade. The positions of the species in the above three clades in the phylogenetic tree of nuclear ribosomal RNA gene were consistent with the COXI tree. Close relationships between H. reticulatum and S. quadricauda (Wilcox et al. 1992, 1993) and C. reinhardtii and »olvox sp. (Rausch et al. 1989), and a remote relationship between the two Chlorella species (Huss and Sogin 1990), were observed in both the rRNA and the COXI trees.

Discussion On average, 43% of alanine and 23% of leucine are assumed to be coded by a UAG codon in the mt of three and two species of green algae, respectively. Considering the high frequency of UAG usage for alanine or leucine in these species, green algae in which no in-frame UAG codons were detected would either use UAG as a stop codon or would not use it at all (unassigned codon). The COXI phylogenetic tree suggests that the code changes occurred in the third clade (see Fig. 2) after branching off from the K. lunaris line. First, UAG would have become an unassigned codon presumably as a result of gradual accumulation of mutations on a release factor so that UAG can not be recognized by it. The motive force for the relaxation of the functional constraints that led to an accumulation of otherwise deleterious mutations on the release factor could be the disappearance of UAG by AT-biased mutation pressure, although the AT-content of the sequenced COXI region carrying the deviant codons is not significantly high (on average GC"42%) as compared with the mt of other green algae that use the universal genetic code (on average GC"39%) (Table 1). Note, however, that the GC content of the genome is changeable during evolution (Osawa 1995). Another possibility would be that UAG-suppressor tRNA plays a role in the code changes. In this case the reading of the UAG codon is ambiguous, i.e. UAG does not disappear completely. In lower eukaryotic mitochondria, specific stop codons are frequently not used at all, e.g. the disappearance of UAG and UGA in mitochondria of the chlorophyte alga P. wickerhamii (Wolff et al. 1994) and of the UAG stop codon in Hansenula wingei (Sekito et al. 1995). These facts indicate that a specific stop codon(s) can disappear from mitochondrial genomes. Thus, a UAGsuppressor would not be responsible for the code changes in our cases. A possible scenario for the codon changes would involve the UAG codon becoming unassigned, as a result of mutations in the corresponding release factor as discussed above, in the common ancestry of the

S. quadricauda/C. microporum lineage and the H. reticulatum/P. boryanum/¹. bitridens lineage. Following this, two independent code changes would have taken place, i.e. nonsense to leucine in the ancestor of S. quadricauda/C. microporum, and nonsense to alanine in the ancestor of H. reticulatum/P. boryanum/¹. bitridens. A UUG leucine codon can change to UAG by a one-step mutation. The tRNA decoding UAG codon as leucine can be created by a single mutation of tRNAL%6 to tRNAL%6 . UAG would then have been UUA UAA captured by the new tRNAL%6 . The creation of a tRNA UUA that decodes a UAG codon as alanine requires two or more point mutations starting from tRNAA-! and UGC resulting in tRNAA-! . RNA editing of the anticodon UUA UGC in tRNAA-! to UUA might be responsible for this, as in the case of marsupial liver mitochondria in which the tRNAA41 gene has a GCC anticodon sequence and is edited to GUC in the matured form (Janke and Pa¨a¨bo 1993). There are other possibilities, but we reserve further discussion until information on the responsible tRNA becomes available. The change of the GCN alanine codon to a new UAG alanine codon is not possible by one point mutation, and must pass through an intermediate amino acid. The most likely process would be that GCN alanine was converted by a single mutation to UCN serine temporarily, and then to UAG alanine. Actually, positions 19 and 34 are occupied by either alanine or serine among different organisms (Fig. 1). In addition to these sites, the occasional appearance of an alanine or serine codon in the homologous site was observed in the multiple predicted amino-acid alignment of the COXI gene (Fig. 1). Here conserved serine sites exist, some coded by UCN (shown by ‘s’ in Fig. 1) and others by AGY (shown by ‘S’ in Fig. 1). The one-step conversion of UCN serine to AGY serine, or vice versa, is not possible without passing through another amino-acid codon. This suggests that, even at the well-conserved serine sites, conversion of UCN%AGY took place via an intermediate amino-acid codon without serious negative effects. As shown in Fig. 2, all the green algae which use UAG as a sense codon in the mt genetic code are clustered together; the green algae that use UAG for alanine (H. reticulatum, P. boryanum and ¹. bitridens) form a subclade, those for leucine (C. microporum and S. quadricauda) form another subclade, and K. lunaris, in which UAG is not a sense codon, is the outgroup of the two subclades. All other green algae that form clades distinct from the above algae do not use UAG as a sense codon. The results suggest that a deviant genetic code may be taken as a good marker to check the COXI phylogenetic tree, especially when the bootstrap value of a node is not significantly high. Acknowledgments We thank Ms. Hideko Tanaka for her technical assistance, and Dr. Kyoko Hatano (Kyoto University, Japan) for the generous gift of the algae used in this study.

33

References Boer PH, Gray MW (1988) Transfer RNA genes and the genetic code in Chlamydomonas reinhardtii mitochondria. Curr Genet 14:583—590 Boer PH, Bonen L, Lee RW, Gray MW (1985) Genes for respiratory chain proteins and ribosomal RNAs are present on a 16-kilobase-pair DNA species from Chlamydomonas reinhardtii mitochondria. Proc Natl Acad Sci USA 82:3340—3344 Boyen C, Leblanc C, Bonnard G, Grienenberger J-M, Kloareg B (1994) Nucleotide sequence of the COX3 gene from Chondrus crispus: evidence that UGA encodes tryptophan and evolutionary implications. Nucleic Acids Res 22:1400—1403 Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783—791 Hiesel R, Combettes B, Brennicke A (1994) Evidence for RNA editing in mitochondria of all major groups of land plants except the Bryophyta. Proc Natl Acad Sci USA 91:629—633 Higgins DG, Bleasby AJ, Fuchs R (1992) Clustal V: improved software for multiple sequence alignment. Comput Appl Biosci 8:189—191 Huss VAR, Sogin ML (1990) Phylogenetic position of some Chlorella species within the chlorococcales based upon complete smallsubunit ribosomal RNA sequences. J Mol Evol 31:432—442 Janke A, Pa¨a¨bo S (1993) Editing of a tRNA anticodon in marsupial mitochondria changes its codon recognition. Nucleic Acids Res 21:1523—1525 Kadowaki K, Suzuki T, Kazama S, Oh-fuchi T, Sakamoto W (1989) Nucleotide sequence of the cytochrome oxidase subunit I gene from rice mitochondria. Nucleic Acids Res 17:7519 Kimura M (1980) A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16:111—120 Ku¨ck U, Neuhaus H (1986) Universal genetic code evidenced in mitochondria of Chlamydomonas reinhardii. Appl Microbiol Biotechnol 23:462—469 Kuhsel MG, Strickland R, Palmer JD (1990) An ancient Group I intron shared by eubacteria and chloroplasts. Science 250: 1570—1573 Kumar S, Tamura K, Nei M (1993) Molecular evolutionary genetics analysis, version 1.01. The Pennsylvania State University, University Park, Pennsylvania

Manhart JR, Palme JD (1990) The gain of two chloroplast tRNA introns marks the green algal ancestors of land plants. Nature 345:268—270 Oda K, Yamato K, Ohta E, Nakamura Y, Takemura M, Nozato N, Akashi K, Kanegae T, Ogura Y, Kohchi T, Ohyama K (1992) Gene organization deduced from the complete sequence of liverwort Marchantia polymorpha mitochondrial DNA. J Mol Biol 223:1—7 Ohama T, Suzuki T, Mori M, Osawa S, Ueda T, Watanabe K, Nakase T (1993) Non-universal decoding of the leucine codon CUG in several Candida species. Nucleic Acids Res 21:4039—4045 Ohta E, Oda K, Yamato K, Nakamura Y, Takemura M, Nozato N, Akashi K, Ohyama K, Michel F (1993) Group I introns in the liverwort mitochondrial genome:the gene coding for subunit 1 of cytochrome oxidase shares five intron positions with its fungal counterparts. Nucleic Acids Res 21: 1297—1305 Osawa S (1995) The evolving genetic code. In: Evolution of the genetic code. Oxford University Press, Oxford, pp 103—113 Rausch H, Larsen N, Schmitt R (1989) Phylogenetic relationships of the green alga »olvox carteri deduced from small-subunit ribosomal RNA comparisons. J Mol Evol 29:255—265 Saitou N, Nei M (1987) The neighbor-joining method:A new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406—425 Sekito T, Okamoto K, Kitano H, Yoshida K (1995) The complete mitochondrial DNA sequence of Hansenula wingei reveals new characteristics of yeast mitochondria. Curr Genet 28:39—53 Wilcox LW, Lewis LA, Fuerst PA, Floyd GL (1992) Assessing the relationships of autosporic and zoosporic chlorococcalean green algae with 18S rDNA sequence data. J Phycol 28: 381—386 Wilcox LW Fuerst PA, Floyd GL (1993) Phylogenetic relationships of four charophycean green algae inferred from complete nuclear-encoded small-subunit rRNA gene sequences. Am J Bot 80:1028—1033 Wolff G, Plante I, Lang BF, Ku¨ck U, Burger G (1994) Complete sequence of the mitochondrial DNA of the chlorophyte alga Prototheca wickerhamii. Gene content and genome organization. J Mol Biol 237:75—86

.