Gonium pectorale has flattened 16-celled colonies and produces isogametes in sexual reproduction. Heterothallic sexuality in G. pectorale with two mating types, plus and minus, was studied by Schreiber (1925) and Stein (1958). Although only one (plus) of the two conjugating isogametes of C. reinhardtii has a tubular mating structure (TMS), both isogametes of G. pectorale extend a TMS toward the other (Nozaki 1984; Nozaki and Itoh 1994). Phylogenetic analyses imply that anisogamous/oogamous species of the colonial Volvocales evolved from an ancestral colonial species that exhibits isogamy as in G. pectorale (Nozaki and Itoh 1994; Nozaki et al. 2000). The genus Gonium is phylogenetically important, as it represents the most basal lineage within the relatively advanced volvocine algae composed of isogamous genera and anisogamous/oogamous members (Nozaki et al. 2000). In C. reinhardtii the mating-type (MT) loci—mating type plus (MT⁺) and mating type minus (MT⁻)—consist of a 200-kb region dimorphic between the two mating types, referred to as the rearranged (R) domain, and include genes involved in mating-type determination and in the function of the TMS, as well as housekeeping genes with alleles present in both MT loci. The dimorphism results in recombinational suppression over a region of ~1 Mb (Ferris and Goodenough 1994; Ferris et al. 2002).

The minus-dominance (MID) gene is one of the minus-specific genes in the MT⁻ locus of C. reinhardtii and was demonstrated to be the master regulator in mating-type minus determination (CrMID: Ferris and Goodenough 1997). CrMID contains an RWP-RK domain, which is a characteristically conserved putative DNA-binding domain observed in plants, oomycetes, and cellular slime molds (Schauser et al. 1999, 2005; Nozaki et al. 2006). Ferris et al. (1997) identified only a single MID ortholog, CiMID, from Chlamydomonas incerta, the closest known relative to C. reinhardtii, although they attempted to find MID orthologs from other Chlamydomonas species, G. pectorale, and Volvox carteri by means of low-stringency DNA gel blot hybridization. They concluded, from comparing divergence of housekeeping genes and MID, that sex-related genes evolve very rapidly (Ferris et al. 1997; P. J. Ferris, unpublished results). Recently, however, the Pleodorina starrii minus-dominance gene ortholog (PlestMID) was obtained by reverse transcribed (RT)–polymerase chain reaction (PCR) from nitrogen-starved, sexually induced males using degenerate primers designed from the RWP-RK domains of C. reinhardtii and C. incerta (Nozaki et al. 2006). This result motivated us to identify additional volvocine MID orthologs and, in turn, the MT loci. The genomic changes accompanying the evolution of sex will be elucidated by a step-by-step analysis of the volvocine MT loci.

The G. pectorale MT locus, by analogy with the C. reinhardtii MT locus, is likely to be a complex, recombinationally suppressed region of some hundreds of kilobases, containing a mix of housekeeping and sex-specific genes, which would include the MID gene. To fully clone and characterize this locus, we adopted a two-pronged strategy: (1) identify the Gonium MID gene following the Pleodorina example and (2) use genetics to confirm that the GpMID gene does indeed map to the sex-determining MT locus and to identify housekeeping genes contained within the Gonium MT locus to provide additional access points for a BAC-clone-based chromosome walk.

Here, as the first step in characterizing the G. pectorale MT locus, we isolated the G. pectorale ortholog of the MID gene (GpMID). GpMID is encoded specifically in minus strains of G. pectorale and strictly linked to certain nuclear genes, suggesting the presence of a MT⁻ locus that is comparable to that of C. reinhardtii. In addition, modes of organellar genome inheritance in G. pectorale were examined and compared with those in C. reinhardtii (Boynton et al. 1987) and in V. carteri (Adams et al. 1990). The evolutionary significance of GpMID and the uniparental inheritance of chloroplast and mitochondrial genomes are discussed in this article.

Experimental organisms and culture and mating methods: The names of the two mating types of G. pectorale (plus and minus) were assigned arbitrarily (Stein 1958), and since we show here that the MT⁻ loci of both G. pectorale and C. reinhardtii carry the MID gene, we have chosen to continue with this usage. Four strains of G. pectorale were used here: Kaneko3 (minus) and Kaneko4 (plus), originating from Okinawa Prefecture, Japan (Yamada et al. 2006), and Mongolia 4 (minus) and Mongolia 1 (plus). The Kaneko strains have been deposited in the Microbial Culture Collection at the National Institute for Environmental Studies (Tsukuba, Japan; Kasai et al. 2004) as NIES-1710 and -1711; the Mongolia strains are deposited in the Culture Centre of Algae and Protozoa (CCAP, Ambleside, Scotland; Gachon et al. 2007) as CCAP32/13 and CCAP32/14. The Mongolia strains were isolated as single cells from incubated petri dishes in which a small amount of dried mud had been rewetted with distilled water, and they were maintained in soil-water medium (Pringsheim 1946). The mud samples were collected by R. A. Lewin (University of California, San Diego) in October 1997. The mud came from two small pools of freshwater adjacent to a large hypersaline lake (Cha-gan-nur, Hao-tong-yin) in Inner Mongolia. Mating types of Mongolia 4 and 1 were determined by crossing with Alaska 1 and 2 strains (Fabry et al. 1998), whereas those of Kaneko3 and -4 were tested against Mongolia 4 and 1.

The cultures of Kaneko3 and -4 were grown in AF-6, VTAC, or standard Volvox media (SVM) (Starr 1969; Kasai et al. 2004) at 20°–25°, with alternating periods of 14 hr light and 10 hr dark at a light intensity of ~30–200 μmol photons · m⁻² · s⁻¹ provided by cool white fluorescent lamps. For crossing, cells of these strains were grown in liquid Tris–acetate–phosphate (TAP) medium (Harris 1989) on a light shelf under constant illumination until growth appeared saturated (typically 4–5 days). The cells were then pelleted and resuspended in nitrogen-free SVM medium (Starr 1969). After sitting overnight, strains of opposite mating type were mixed. Flagellar agglutination could be observed just after mixing. Several hours after, the mixture was transferred to 4% agar SVM plates and placed in the dark for at least 10 days. In successful matings, the orange-colored zygotes were clearly visible using a dissecting microscope. The zygotes were then transferred from 4 to 1.5% agar TAP plates and manipulated by hand to separate individual zygotes; germination was high, but poor progeny survival obviated tetrad analysis. Consequently, a random progeny approach was adopted for the genetic analysis. Matured zygotes on the plates were flooded with liquid SVM, and cells loosened with a glass hockey stick were transferred to a sterile tube and placed in a −20° freezer overnight to kill unmated cells. After thawing, the zygotes were spread on a TAP plate and incubated under light until colonies were visible by the naked eye. Each colony represents the surviving progeny from a single zygote. Colonies were subcloned before further characterization.

Cloning of GpMID: For isolation of the G. pectorale MID gene, 14-day-old cultures of Kaneko3 and -4 in VTAC medium were separately placed in petri dishes with an equal volume of autoclaved MilliQ water. The petri dishes were then incubated at 25° under a 14 hr light/10 hr dark cycle. After 2 days, both cultures were largely unicellular and commenced flagellar agglutination within 1 hr if an aliquot from each was mixed. Such sexually activated, nonmixed cells were used for isolation of gamete RNA. Total RNA was isolated with the RNeasy Midi kit (QIAGEN, Hilden, Germany; protocol for heart, muscle, and skin tissue) after the cells had been homogenized with ceramic beads and a wash brush (Nozaki et al. 1997, 2006). Full-length cDNA synthesis from the total RNA was carried out with the CapFishing full-length cDNA premix kit (Seegene, Seoul, Korea). Nested RT–PCR using this cDNA yielded the partial fragment of GpMID; the primers used in the first PCR were the Seegene kit's 3′-rapid amplification of the cDNA end (3′-RACE) primer and dMT-dF3 (primers designed in this study are listed in Table 1); the primers used in the second PCR were dMT-dF3 and mt-R4 (Nozaki et al. 2006). The PCR reactions were carried out using TAKARA Taq polymerase (TAKARA, Osaka, Japan) using the cycling conditions described previously (Nozaki et al. 1995). To determine the C terminus sequence of GpMID, 3′-RACE was performed with the 3′-RACE primer of the kit and GpMidF1. The N terminus sequence was determined using the 5′-RACE system (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol; the first antisense gene-specific primer was GpMidR3; the second gene-specific primer was GpMidR1. The resulting fragments were TA subcloned using the pGEM T-easy kit (Promega, Madison, WI) and sequenced as described previously (Nozaki et al. 2006).

TABLE 1 Primers newly designed in this study

Genomic PCR and sequencing of GpMID: Genomic DNA from the four parental strains (Kaneko3 and -4; Mongolia 4 and 1) and their F₁ progeny were prepared either by the “miniprep” method described by Miller et al. (1993) or using the DNeasy plant mini kit (QIAGEN). For determining the genomic sequence of GpMID from Kaneko3 and Mongolia 4, PCR amplification was performed with specific primers for GpMID (GPMID5P-1F and GPMID3P-1R). The PCR reaction was done using LA Taq DNA polymerase with GC buffer I (TAKARA) and GeneAmp PCR system 9700 (Applied Biosystems, Foster City, CA). PCR products were sequenced directly or after TA cloning, as described above.

DNA gel blot analysis of GpMID: Southern blot analysis was done following the protocols of Nishii et al. (2003) modified to use the nonradioisotope detection system outlined below. Restriction enzyme digests (6 μg) of genomic DNA were separated by 1.0% agarose gel electrophoresis and transferred onto a Hybond-N+ membrane (GE Healthcare Bio-Sciences, Little Chalfont, UK) (Sambrook and Russell 2001) using a vacuum blotter (Bio-Rad, Hercules, CA). A probe containing the GpMID ORF region was prepared by PCR with the primer pair GPMID-5P1F and GpMidR3 using Kaneko3 genomic DNA as template. The probe was labeled with fluorescein-11-dUTP (Gene Images random prime labeling module, GE Healthcare Bio-Sciences) and hybridized at 68° according to the manufacturer's protocol. The signals were detected with Gene Images CDP-Star detection kit (GE Healthcare Bio-Sciences) and VersaDoc Model 5000 (Bio-Rad). The resulting image was processed with a median filter (diameter: 1 pixel) in ImageJ (National Institutes of Health, Bethesda, MD) to remove random noise produced by long exposure (2 hr).

Comparison of GpMID transcript level: Because the expression of MID genes from C. reinhardtii (Ferris and Goodenough 1997; Lin and Goodenough 2007) and P. starrii (Nozaki et al. 2006) increased in nitrogen-starved cultures, expression of GpMID was examined with or without nitrogen starvation for 18 hr; ~10⁵–10⁶ cells of G. pectorale Kaneko3 were pelleted by centrifugation and resuspended in N-free SVM (SVM modified by omitting urea and replacing CaNO₃ with CaCl₂). Polyadenylated mRNA was extracted from algae using Dynabeads (Invitrogen) as described in the manufacturer's protocol. The GpMID expression was assayed by semiquantitative RT–PCR, using the EF1-like gene as an internal control: GpMID sense, GPMID_int1F; GpMID antisense, GPMID_int4R; EF1-like sense, CV_EF1A1-R2; and EF1-like antisense, GpEF1A-INT3-R. The PCR reaction was done as follows: 95° for 2 min, followed by 28 cycles of 95°/15 sec, 63°/30 sec, and 68°/40 sec using KOD plus DNA polymerase (TOYOBO, Osaka, Japan) and GeneAmp PCR system 9700.

Phylogenetic analysis of GpMID: Both the C. reinhardtii and V. carteri genome databases of the Department of Energy's Joint Genome Institute (JGI) (http://www.jgi.doe.gov/) were screened for RWP-RK domain genes by TBLASTN (NCBI) using the CrMID protein sequence. We identified 15 and 10 predicted RWP-RK domain sequences of C. reinhardtii and V. carteri, respectively. These numbers do not include the MID genes because the strains chosen for sequencing—plus and female, respectively—do not possess the MID gene. The computer-generated gene models associated with these RWP-RK domains were assessed and modified or new models were created when necessary to ensure, for example, that the RWP-RK domain was contained in the models that we ultimately annotated on the JGI websites and to incorporate unpublished cDNA data where available. For C. reinhardtii, the resulting models were named RWP1-RWP14; the 15th gene had previously been named NIT2 (Schnell and Lefebvre 1993). The RWP-RK domain of RWP12 is located at the N terminus; the initial methionine is set at the seventh residue of the multiple alignment. Since it seemed likely that the RWP-RK domain should extend farther into the 5′-region, the DNA sequence neighboring RWP12 was analyzed by GENSCAN (Burge and Karlin 1997) with the “organism” option as Arabidopsis, and a longer protein was used for the phylogenetic analysis. The 10 possible RWP-RK protein sequences from the V. carteri genome were also assessed using GENSCAN. Most of the models from both species have little or no support from expressed sequence tags and so should be viewed as tentative. The 25 gene models above, as well as CrMID, CiMID, PlestMID, and GpMID, were aligned using ClustalX (Thompson et al. 1997) with the default option. The 29 RWP-RK domains of 47 amino acids in length (supplemental Figure 1 at http://www.genetics.org/supplemental/) were subjected to a maximum-likelihood analysis based on the Whelan and Goldman (WAG) model (Whelan and Goldman 2001) in PHYML online (Guindon et al. 2005). The robustness of the result was examined using a bootstrap analysis (Felsenstein 1985) with 500 replications. On the basis of the same alignment data, a maximum parsimony analysis was performed by PAUP*4.0b10 (Swofford 2003) with a bootstrap analysis based on 1000 replications of the general heuristic search using the tree-bisection-reconnection branch-swapping algorithm. A neighbor-joining analysis (Saitou and Nei 1987) based on the Jones–Taylor–Thornton (JTT) model (Jones et al. 1992) was also performed by MEGA 4.0 (Tamura et al. 2007) with a bootstrap analysis based on 500 replications. Phylogenetic trees including subsets of these RWP-RK domains have also been published by Nozaki et al. (2006) and Lin and Goodenough (2007).

Scoring of genetic markers in the F₁ progeny:

Mating phenotype: The mating type of the progeny was determined by mating tests with the two parental strains to determine with which parent zygotes were formed. Each F₁ was co-inoculated separately with both parent strains in soil–water medium tubes and allowed to grow up under constant illumination. If mating was abundant, a zygote pellicle could be seen after 4–5 days, but in any case, after 5–7 days the thick-walled, orange-colored zygotes were readily identifiable by light microscopy even if only a few percent of the cells had mated. A few progeny failed to mate with either parent.

ACT and YPT4: The actin (ACT) gene and small G-protein gene YPT4 (Fabry et al. 1998) were chosen as controls because they are unlinked to MT in Chlamydomonas and because PCR primers had been designed previously by Liss et al. (1997) for amplifying these two genes from Volvocales. On the JGI Chlamydomonas genome website, these two genes are annotated as IDA5 and RABB1, respectively.

The A/9-5′ and A/9-3′ primers were used to amplify intron IX of ACT by PCR from both the Mongolia 1 and 4 strains. The sequences of these two PCR products were used to design two nondegenerate primers, gonact1 and gonact2, that PCR amplify both ACT alleles. That ~380-bp PCR product was double digested with EcoRI (one site in the Mongolia 1 allele, yielding bands of ~180 and 200 bp) and XhoI (one site in the Mongolia 4 allele, yielding bands of ~140 and 240 bp) and analyzed by agarose gel electrophoresis.

The 4/6-5′ and 4/6-3′ primers were used to amplify intron VI of YPT4 by PCR from the Mongolia 1 and 4 strains. The sequences of these two PCR products were used to design two nondegenerate primers, gonypt1 and gonypt2, that PCR amplify both YPT4 alleles. The ~360-bp PCR product was then digested with HinfI and analyzed by agarose gel electrophoresis to visualize the different patterns of bands in the two strains (in Mongolia 1: 260, 70, and 30 bp; in Mongolia 4: 120, 110, 70, and 60 bp).

PR46a: The PR46a gene is within segment 3 of the sexually dimorphic R domain of the C. reinhardtii MT loci (Ferris et al. 2002). Its predicted amino acid sequence was used to design primers for amplifying the PR46a gene from G. pectorale genomic DNA. PCR of Mongolia 4 genomic DNA using primers pr46for and pr46rev yielded an ~200-bp product, which was TA subcloned using the pGEM T-easy kit. The gel-purified insert was radiolabeled with ³²P and hybridized at a reduced stringency (58°) to the PstI-digested genomic DNA Southern blots used to score chloroplast and mitochondria markers (see below). The RFLP visualized in Mongolia 1 is 1.0 kb and in Mongolia 4, 1.8 kb.

LEU1S: Identified in the Chlamydomonas genome sequence, the gene encoding the small subunit of isopropylmalate dehydratase (LEU1S; protein ID 126865 in the C. reinhardtii genome database version 3.0) is found in segment 4 of the sexually dimorphic R domain of the Chlamydomonas MT loci near the position of probe Pr65 in Ferris et al. (2002). The G. pectorale LEU1S gene was cloned by screening a Mongolia 1 genomic EMBL3 library using as probe a 600-bp ClaI/HindIII fragment (roughly corresponding to the coding region) purified from a C. reinhardtii LEU1S cDNA clone (provided by the Kazusa Institute; corresponds to AV388014). The portion of the phage insert containing the LEU1S gene was sequenced. To score progeny for the LEU1S marker, a PCR reaction was performed using primers gonleu13 and gonleu14. The resulting ~1050-bp PCR product was then digested with MfeI and subjected to agarose gel electrophoresis, producing different band patterns in the two alleles (in Mongolia 1: 470 and 580 bp; in Mongolia 4: 360 and 690 bp).

ALB3.1: The ALB3.1 gene, also known in the Chlamydomonas literature as AC29, is ~40 kb centromere distal of the MT locus R domain (Ferris et al. 2002). The G. pectorale ALB3.1 gene was cloned by screening an Alaska2 genomic EMBL3 library using a portion of the C. reinhardtii ALB3.1 cDNA as the probe. The portion of the phage insert corresponding to the ALB3.1 gene was sequenced. The predicted protein product of the G. pectorale gene was clearly more similar to the C. reinhardtii ALB3.1 gene linked to the mating-type locus than to the unlinked ALB3.2 gene (Bellafiore et al. 2002). To score progeny for the ALB3.1 marker, a PCR reaction was performed using primers gac30 and gac31, and the ~350-bp PCR product was then digested with PstI, which cuts only the Mongolia 1 allele (into 240 and 110 bp), and analyzed by agarose gel electrophoresis.

MID: To determine the presence or absence of the GpMID gene in the genome, PCR was carried out with specific primers: Mong4F1 and Mong4R1 for GpMID. As the internal control, ACT gene-specific primers gonact1 and gonact2 were used.

Organellar markers: Chloroplast and mitochondrial genomes from the two parents were scored by taking advantage of RFLPs. Genomic DNA (~500 ng) of the parental and progeny strains was digested with PstI, electrophoresed on a 0.75% agarose gel in TBE buffer, and transferred to nitrocellulose. The blots were hybridized (Church and Gilbert 1984) sequentially with the PR46a probe (see above), the mitochondria probe, and the chloroplast probe, stripping the filters to remove the previous probe before rehybridizing. The probe for mitochondrial DNA was the 1.6-kb EcoRI/HindIII fragment of the C. reinhardtii mitochondrial genome, purified from the P318 plasmid obtained from the Chlamydomonas Center. Hybridization was carried out at a reduced stringency: 60° rather than the usual 65°.

To visualize chloroplast DNA, a section of the G. pectorale psaA gene (AB044242) was PCR amplified using Alaska2 genomic DNA as template with primers gonpsa1 and gonpsa2. The ~850-bp PCR product was used as probe at normal stringency (65°). All probes were radiolabeled with [α-³²P]dCTP by random priming.

Identification and characterization of GpMID: Differential nested RT–PCR using cDNA prepared from “sexually activated cultures” of Kaneko3 (minus) and -4 (plus) displayed a fragment of the expected length for MID orthologs (149 bp) only in Kaneko3. The full-length cDNA sequence indicated that the mRNA is 1.5–1.6 kb in length. The predicted open reading frame encodes 164 amino acids; the hypothetical molecular weight is ~18 kDa; the theoretical pI is 9.25. The hypothetical polypeptide sequence contains a leucine zipper and an RWP-RK domain near the C terminus (Figure 1). A BLASTP search (http://www.ncbi.nih.gov/) with this sequence indicated that it is similar to PlestMID (E = 5e-46), CiMID (E = 4e-28), and CrMID (E = 2e-27). The percentage of identity of GpMID protein with CrMID, CiMID, and PlestMID is 38.4, 38.4, and 53.3, respectively. A BLASTN search (http://www.ncbi.nih.gov/) using an expect threshold of 10 and a word size of 11 with the GpMID cDNA sequence found matches with PlestMID but none with CrMID or CiMID. The coding region is 52.9% GC and the primary transcript is 51.4% GC.

Figure 1.— Alignment of four MID proteins from G. pectorale (GpMID), P. starrii (PlestMID), C. reinhardtii (CrMID), and C. incerta (CiMID). Solid and shaded backgrounds indicate identity in 100% or in 75% of the sequences aligned, respectively. Five amino acids (more ...)

The genomic GpMID gene sequenced in this study (1603 bp) covered nearly the entire cDNA sequence and demonstrated that it contains four introns, three of which are in positions very similar to those of the three introns that are present in the CrMID and CiMID genes, and all four are shared with PlestMID (Figure 2).

Figure 2.— Exon–intron structure of MID genes from G. pectorale (GpMID), P. starrii (PlestMID), C. reinhardtii (CrMID), and C. incerta (CiMID). Corresponding parts of the coding regions (CDS) are interconnected by dotted lines. The line displayed above (more ...)

A genomic DNA gel blot analysis of restriction-digested plus and minus DNA probed with a PCR fragment of GpMID demonstrated a minus-specific single band (Figure 3), indicating that GpMID is a single-copy gene in minus (Kaneko3). The lack of hybridizing signal in the plus strain (Kaneko4) suggests genomic asymmetry between mating types. A similar result was observed with Mongolia 1 and 4 strains; the signal was detected not from Mongolia 1 (plus) but from Mongolia 4 (minus; data not shown). This conclusion was confirmed when specific GpMID primers were used to perform PCR with DNA samples from plus and minus F₁ progeny derived from Kaneko3 × -4, consistent with their mating phenotypes (data not shown). These results clearly suggest that the mating type minus of G. pectorale correlates with the presence of GpMID.

Figure 3.— GpMID is present only in MT−. Genomic DNA isolated from a minus strain, Kaneko3 (labeled “−” below the gel) and a plus strain, Kaneko4 (labeled “+”), was digested with either SacII or PstI and hybridized (more ...)

Semiquantitative RT–PCR with the GpMID cDNA-specific primers showed that GpMID transcription was upregulated in nitrogen-free culture (Figure 4). The mating reaction commenced immediately after mixing the two complementary mating types when the cells were nitrogen starved for 18 hr, whereas no mating reaction occurred within 24 hr after mixing of cells cultured in SVM. These observations suggest that expression of GpMID increased in minus gametes induced by nitrogen starvation.

Figure 4.— Semiquantitative RT–PCR analysis of GpMID. The Kaneko3 strain was cultured either in nitrogen-free SVM (N-free SVM) for 18 hr or in ordinary SVM. Poly(A)+ RNA from these cultures was reverse transcribed and PCR amplified using primers (more ...)

Phylogenetic analysis of genes containing the conserved RWP-RK domain: To confirm that GpMID is orthologous to the MID genes, a phylogenetic analysis was carried out using the alignment of four MID proteins and 25 other RWP-RK domains of C. reinhardtii and V. carteri. As shown in Figure 5, the four MID proteins formed a robust monophyletic group (with 74–91% bootstrap values). Within the MID clade, GpMID and PlestMID are sister to each other with 81–99% bootstrap values. Similar results were reported by Lin and Goodenough (2007) using a smaller set of genes.

Figure 5.— Maximum-likelihood (ML) tree (based on WAG model) of four MID proteins and 25 RWP-RK domains from C. reinhardtii (Cr) and V. carteri (Vc) genome databases. Branch lengths are proportional to the estimated amino acid substitutions, which are indicated (more ...)

Nuclear and organellar genetics: The bulk of the genetic mapping data was generated using 78 progeny of a single cross between Mongolia 1 and 4 (Tables 2 and 3; Figures 6–8). Another 20 progeny were eliminated from further analysis after being scored with only the ACT marker because they contained both parental alleles and either were diploid/aneuploid or were not properly subcloned. As shown in Table 2, recovery of the alleles for some of the markers deviates from 1:1 in several cases, especially YPT4.

TABLE 2 Ratio of parental alleles in the progeny for each locus

TABLE 3 Fraction of progeny recombinant for each pair of markers

Figure 6.— Scoring the nuclear markers in the F1 progeny. A sampling of the DNA gel blot and PCR–RFLP analyses on G. pectorale progeny strains for the six nuclear markers. Diploid strains are marked with asterisks. Parental strains were Mongolia 1 (M1) and (more ...)

Figure 7.— A complete representation of which alleles were inherited in the set of G. pectorale F1 progeny examined. Shaded boxes indicate inheritance of the phenotype (mating type: MT) or genotype (chlp, chloroplast; mito, mitochondria) of the plus parent [M1 (more ...)

Figure 8.— Uniparental inheritance of organellar genomes in progeny of G. pectorale crosses. (A) Genomic DNA of the two parent strains (Mongolia 1 and 4, M1 and M4, respectively) and 11 F1 progeny were digested with PstI, and the resulting Southern blot was hybridized (more ...)

Six nuclear genes were scored by taking advantage of sequence polymorphisms between the parental strains. Since this is the first G. pectorale genetic analysis and G. pectorale is not routinely in laboratory use, it was important to ensure that the progeny were indeed products of meiosis and not, for example, unmated gametes that had somehow survived freezing. Hence, two genes (ACT and YPT4) that are not expected to be MT linked (Ferris et al. 2002) were chosen as controls to confirm independent segregation of unlinked markers. The remaining four markers are all linked to the MT locus in C. reinhardtii (Ferris and Goodenough 1994, Ferris et al. 2002): three (LEU1S, PR46a, and MID) are in the dimorphic R domain of the MT locus, and one (ALB3.1) is located just centromere distal of the R domain. The mating phenotype was also scored to be either plus or minus.

Figure 6 shows a representative DNA–DNA gel blot and AFLP analyses used to score the nuclear markers in the F₁ progeny, and Figure 7 shows the genotype of the entire set of progeny. Table 3 shows, for each pair of markers, the proportion of progeny that are recombinant. As anticipated, the ACT and YPT4 genes are not linked to each other or to any other marker. However, the LEU1S gene was strictly linked to mating type. The PR46a and ALB3.1 genes are loosely linked to each other, recombining in only 12 of 77 progeny, a map distance of 16 cM (Table 3), but are not linked to mating type. The presence of GpMID is strictly linked to having a minus mating phenotype; no GpMID was found in plus progeny.

Of the 20 progeny strains that showed both alleles of the ACT marker, 3 were resubcloned and proved to have both alleles for all nuclear markers, as well as the GpMID marker, and hence are likely diploid (Figure 6). All three of them mated with the plus parent. Therefore, the dominant mating type of G. pectorale is minus.

The direction of uniparental inheritance was tested in the F₁ progeny as well. Radiolabeled probes to the chloroplast and mitochondrial DNA were hybridized to Southern blots of PstI-digested DNA from parents and progeny. The 11-kb chloroplast DNA fragment from the plus parent was inherited in 66 progeny (UP+), while only 5 progeny inherited the 2.8-kb fragment of the minus parent (Figure 8A). Seven progeny showed a band of 20 kb not present in either parent, either in conjunction with the 11-kb band (5 progeny, one of which is shown in Figure 8A) or as the sole band (2 progeny). Perhaps this represents a polymorphism that appeared in a subpopulation of the plus parent strain after the DNA preparation was made. We chose to score these 7 progeny as having inherited from the plus parent.

Seventy progeny inherited exclusively the 7-kb mitochondrial DNA fragment from the minus parent (UP−), while the 10-kb fragment from the plus parent was inherited in only a single progeny (Figure 8B, top). Seven progeny were biparental, but the Mongolia 4 (minus)-type band predominated (e.g., the fourth progeny in Figure 8B, top).

As a control for the possibility that a replication advantage of the mitochondrial DNA from the Mongolia 4 parent might be mistaken for uniparental inheritance, the single progeny (marked as P2 in both panels of Figure 8B) that was minus with a Mongolia 1 (plus)-type mitochondrial genome was crossed with another F₁ progeny (lane P1 in the bottom of Figure 8B) that was plus and had inherited the Mongolia 4 mitochondrial DNA. All 12 progeny inherited the mitochondrial DNA of the minus parent (P2), which in this cross was the Mongolia 1 allele. The 12 F₂ progeny were also tested for segregation of several nuclear markers to ensure that they were indeed meiotic progeny (not shown).

GpMID is an ortholog of CrMID: The structure of GpMID is essentially consistent with that of the other MID orthologs, CrMID, CiMID, and PlestMID (Ferris and Goodenough 1997; Ferris et al. 1997; Nozaki et al. 2006). They have the homologous RWP-RK domain, the putative bZIP DNA-binding region containing the regularly repeated hydrophobic amino acids (Figure 1). Moreover, the present phylogenetic analysis clearly demonstrates that all four MID genes form a robust monophyletic group relative to other RWP-RK domain-containing genes in C. reinhardtii and V. carteri (Figure 5). Although the GpMID protein sequence shows similarity to the other three MID orthologs, the present analysis using the nucleotide sequence (BLASTN of NCBI) could not find significant matches between GpMID and the two MID orthologs of Chlamydomonas, consistent with the fact that Ferris et al. (1997) could not detect any MID signal in G. pectorale by the low-stringency DNA gel blot analysis. As with CrMID and PlestMID (Ferris and Goodenough 1997; Nozaki et al. 2006), GpMID is a mating-type-specific gene that is expressed in nitrogen-starved cultures of minus/male strains. In both DNA gel blot analysis and genomic PCR, GpMID was detected in minus strains, but not in plus (Figures 3 and 7). An increase in expression of GpMID was detected by RT–PCR in nitrogen-starved cultures of minus (Figure 4). PlestMID, the MID ortholog from P. starrii, was shown to be localized in the sperm (male gamete) nucleus by immunofluorescent microscopy (Nozaki et al. 2006). Therefore, GpMID is supposed to be functional in the nucleus of the G. pectorale minus gamete.

Diploid progeny from Mongolia 1 (plus) and Mongolia 4 (minus) exhibited a minus phenotype. C. reinhardtii diploids also exhibit this minus dominance (Ebersold 1967), which is the result of the presence of the MID gene (Ferris and Goodenough 1997). The GpMID gene presumably makes minus the dominant mating type in G. pectorale. Transformation of GpMID DNA into a G. pectorale plus strain would confirm this point but has not been performed because there is no stable transformation method in G. pectorale. Recently, however, the highly efficient transgenic method in C. reinhardtii has been applied to the colonial volvocalean algae (Sizova et al. 2001; Jakobiak et al. 2004; Hallmann and Wodniok 2006). This transgenic approach may be available in G. pectorale to examine the GpMID function in the plus strain.

Regulation of the TMS formation might have changed in the course of evolution: In C. reinhardtii, gametes of only one of the two mating types, plus, bear TMS, but mid-1, the CrMID-defective mutant of the minus strain, forms a TMS, suggesting that CrMID directly or indirectly suppresses the formation of TMS (Goodenough et al. 1982; Ferris and Goodenough 1997; Lin and Goodenough 2007). However, gametes of both mating types of G. pectorale form TMS (Nozaki 1984), whereas GpMID is present only in minus genomes (Figure 3). Therefore, G. pectorale may be different from C. reinhardtii in that its MID gene (or a downstream MID responsive gene) may not suppress formation of TMS. Because G. pectorale and the anisogamous/oogamous alga P. starrii (Nozaki et al. 2006) have MID orthologs that are mating-type-specific as in C. reinhardtii (Ferris and Goodenough 1997), the mating-type-specific feature of MID orthologs seems to be conservative within the colonial Volvocales. In addition, almost all of the isogamous colonial Volvocales (Gonium, Astrephomene, Pandorina, Volvulina, and Yamagishiella) have TMS in each of the two conjugating gametes, and this type of TMS might have evolved from TMS as found in C. reinhardtii (Nozaki and Itoh 1994; Nozaki et al. 2000). Therefore, the loss of suppression of TMS formation as suggested in the GpMID contol pathway might have evolved in the common ancestor of these members of the colonial Volvocales, possibly at an early evolutionary stage (a four- to eight-celled colonial stage) within the colonial Volvocales (Nozaki and Itoh 1994; Nozaki et al. 2000).

Nuclear genetics suggests dynamic reorganization in the MT locus: On the basis of this genetic analysis of G. pectorale F₁ progeny, the LEU1S gene and presence/absence of GpMID are strictly linked to mating phenotypes (Table 3; Figure 7). In contrast, the other four nuclear genes (PR46a, ACT, YPT4, and ALB3.1) are genetically independent of plus or minus in G. pectorale, although ALB3.1 and PR46a are located 16 cM apart. We suggest that GpMID and LEU1S are likely harbored in a chromosomal region of recombinational suppression that is comparable to the C. reinhardtii MT locus, although this linkage could also be explained if the two genes are close together—they are only 50 kb apart in Chlamydomonas. In C. reinhardtii, two genes (ACT and YPT4) are also not MT linked (Ferris et al. 2002) whereas all four remaining markers are linked to MT in C. reinhardtii (Ferris and Goodenough 1994; Ferris et al. 2002): one centromere distal of the R domain (ALB3.1) and the remaining three (LEU1S, PR46a, and CrMID) in the R domain. Therefore, gene rearrangement within and around the MT locus must have occurred during the evolution from the common ancestor of C. reinhardtii and G. pectorale. Whether the PR46a and ALB3.1 genes have relocated to another chromosome or to a more distant location on the same chromosome is not yet determined.

One of the unusual features of the MID genes is the relatively low GC content (GpMID: 52.9%; PlestMID: 49.1%; CrMID: 50.5%; CiMID: 49.4%). GC content of coding regions is high in C. reinhardtii (68%; Merchant et al. 2007) and in C. incerta (64.7%; Popescu et al. 2006) and, although based only on the two gene sequences (ALB3.1 and LEU1S) reported here (68.5%), G. pectorale may be similar. The GC content of MID is low in the introns and UTRs as well. In Ferris et al. (2002), it was suggested that the low GC content of both MID and FUS1 (which exists only in the MT⁺ locus; Ferris et al. 1996) could result from being restricted to only one MT locus over evolutionary time periods. One possible mechanism consistent with this idea is “biased gene conversion toward GC” (BGC_GC; reviewed in Marais 2003). In the BGC_GC scheme, when recombination or gene conversion occurs, repair mechanisms tend to replace the mismatches in the resulting DNA heteroduplexes toward G or C; recombination rate correlates with the GC content in yeast, invertebrate, and mammalian genomes. The MID gene never undergoes recombination since it is hemizygous in the diploid stage and would escape biased conversion, accounting for the lower GC content over both coding and noncoding sequences. Since MID has been conserved throughout the Volvocales (being present also in the V. carteri male mating type; P. J. Ferris, T. Hamaji, I. Nishii and H. Nozaki, unpublished results), it would have been subject to BGC_GC for at least 50 MY (Kirk 2005). The lower GC content of Y-specific genes in humans has been postulated to be the result of BGC_GC as well (Galtier et al. 2001).

Biased gene conversion can also explain the similarly low GC content (47.7%) of FUS1; however, no orthologs of FUS1 have been characterized, so there is no information on how long FUS1 has been restricted to only one mating type. Three additional genes in the C. reinhardtii MT loci—MTA1, EZY2, and MTD1—are sexually dimorphic but have coding-region GC contents typical for C. reinhardtii (65, 68, and 68%, respectively). The MT+ locus-specific MTA1 appears to have been created within a recent duplication from another chromosome (Ferris et al. 2002). Hence, MTA1 likely exists only in C. reinhardtii, and there has been only limited time for BGC_GC to affect this sequence. The EZY2 gene is a tandemly repeated gene in the MT⁺ locus, but this is also likely a recent event, with a single well-conserved (albeit pseudogene) copy of EZY2 still present in MT⁻ (Ferris et al. 2002). In addition, BGC_GC can occur via the mechanisms responsible for concerted evolution in tandem arrays (Galtier et al. 2001). No orthologs of the EZY2 gene have been identified. Finally, MTD1 is limited to the MT⁻ locus (Ferris et al. 2002; Lin and Goodenough 2007). Although its coding region has a typical GC content, its noncoding sequences (introns plus UTRs) have a GC content of only 54% (P. J. Ferris and J. Umen, unpublished results), while the G. pectorale ortholog is 55.5% GC coding, and 50.2% GC noncoding (T. Hamaji, P. J. Ferris, I. Nishii and H. Nozaki, unpublished results). The amino acid composition of the C. reinhardtii MTD1 is biased toward GC-rich codons, suggesting that other selection pressures on the GC content of the coding region may obscure the effect of BGC_GC.

Organellar inheritance between isogamy and anisogamy/oogamy might have changed: In C. reinhardtii, the uniparental inheritance of organellar genomes differs in plastids and mitochondria (Boynton et al. 1987). The plastid genome in F₁ progeny is usually transmitted from the plus parent whereas the mitochondrial DNA of the progeny is from the parental minus, which has CrMID. The mode of uniparental inheritance of plastid and mitochondrial genomes in G. pectorale is UP+ and UP−, respectively (Table 3), the same as that of C. reinhardtii. However, both the chloroplast and mitochondrial genomes are transmitted from the female parent in the oogamous V. carteri (Adams et al. 1990). Both V. carteri and its close relative P. starrii have the MID ortholog (PlestMID) in the male strains (Nozaki et al. 2006 and P. J. Ferris, T. Hamaji, I. Nishii and H. Nozaki, unpublished results), suggesting that the male in the anisogamous/oogamous volvocaleans evolved from the MID-containing, minus mating type of C. reinhardtii (Nozaki et al. 2006). Therefore, the uniparental inheritance of the mitochondrial genome has changed at some point during the evolution of the colonial Volvocales from uniparentally from the MID-containing (minus) parent to uniparentally from the female, which does not carry MID.

Conclusion: This characterization of the minus-specific gene GpMID and analyses of F₁ progeny suggested that GpMID and LEU1S are harbored in a chromosomal region under recombinational suppression that is comparable to the C. reinhardtii MT locus, which consists of a ~1-Mb region of recombinational suppression and harbors several mating-type-specific genes (Ferris and Goodenough 1994; Ferris et al. 2002); the linkage of the two genes could also be explained by their close proximity (e.g., within 50 kb in Chlamydomonas). Nozaki et al. (2006) demonstrated that PlestMID and its pseudogene PsPlestMID in the anisogamous colonial volvocalean P. starrii are also mating type (male) specific and have neutral GC contents that may represent suppression of recombination in the regions of the two genes. Therefore, the presence of a MT locus under recombinational suppression may be conserved within the isogamous and anisogamous/oogamous members of the colonial Volvocales. However, these genetic analyses of G. pectorale demonstrated that chromosomal rearrangement including the MT locus must have occurred during the evolution from the common ancestor of C. reinhardtii and G. pectorale. Genomic analyses of MT loci using BAC libraries from colonial Volvocales including G. pectorale and P. starrii will resolve important gene changes in the MT loci during the evolution of TMS and the origin of female and male within this model lineage of “sex evolution.”

We thank Linda Small for her technical support and Ursula Goodenough and Jim Umen for their encouragement. The C. reinhardtii sequence data were produced by the U. S. Department of Energy Joint Genome Institute (http://www.jgi.doe.gov/). The V. carteri genome sequencing work was performed by the Joint Genome Institute under the auspices of the U. S. Department of Energy's Office of Science, Biological and Environmental Research Program and the University of California, Lawrence Livermore National Laboratory, under contract no. W-7405-ENG-48; Lawrence Berkeley National Laboratory under contract no. DE-AC03-76SF00098; and Los Alamos National Laboratory under contract no. W-7405-ENG-36; and was provided for use in this publication only. This work was supported by a Grant-in-Aid for Creative Scientific Research (no. 16GS0304 to H.N.); by a Grant-in-Aid for Scientific Research (no. 17370087 to H.N.) from the Ministry of Education, Culture, Sports, Science and Technology, Japan; by the Japan Society for the Promotion of Science (nos. S05750 and L06701 to P.J.F.); and by the National Science Foundation (no. 9904667).