INTRODUCTION

Chlamydomonas has two actin genes, one encoding a conventional actin (CrA) with ~90% amino-acid identity with mammalian actin, and the other a highly divergent actin (NAP, novel actin-like protein) having ~64% identity with mammalian actin (Kato et al., 1993; Sugase et al., 1996; Kato-Minoura et al., 1997; Lee et al., 1997; Kato-Minoura et al., 1998). Phylogenetic analysis has shown that NAP has evolved independently from the conventional actin lineage since algal actins first diversified (Kato-Minoura et al., 2003). The presence of these divergent actins in a single cell suggests that they each have specific functions, but their functional difference is not understood. In Chlamydomonas and many other organisms, actin is a subunit of axonemal inner-arm dyneins, i.e., the protein complexes that produce force for ciliary and flagellar movements (Piperno and Luck, 1979; Muto et al., 1994). The actin subunit may be involved in the docking of inner arm dyneins to the correct loci on the outer doublet microtubules (Yanagisawa and Kamiya, 2001). The mutant ida5, which lacks the gene for CrA (Kato et al., 1993; Kato-Minoura et al., 1997), lacks four of the six species of actin-containing inner-arm dyneins and cannot form fertilization tubules but grows at the same rate as the wild type. The inner-arm dyneins present in ida5 contain NAP instead of CrA. Hence NAP can substitute for CrA in most, but not all, actin-based functions.

Intriguingly, NAP is expressed only negligibly in wild-type cells, although it is expressed abundantly in ida5. In ida5 cells transformed with the genomic CrA gene, the level of NAP message is inversely related to the level of expressed actin protein (Ohara et al., 1998). These observations led to the hypothesis that the expression of NAP is negatively regulated by actin. Recently, Hirono et al. (2003) found that a considerable level of NAP is expressed in wild-type cells during the reflagellation process after flagellar amputation. NAP expression ceases with flagellar growth. This observation suggests that NAP may function in flagellation, and the controlled expression of the two genes may be important in this process.

To determine whether balanced expression of the two actins is important for flagellar regeneration, I transformed cells with chimeric genes in which the 5′-untranslated regions (5′-UTRs) were swapped between the NAP and CrA genes. The ability of cells to regrow flagella was impaired greatly when ida5 was transformed with a chimeric actin gene consisting of the 5′-UTR sequence from the NAP gene and coding sequence from the CrA gene. This result supports the view that NAP plays an important role in the early stage of flagellar regeneration.

MATERIALS AND METHODS

Construction of chimeric actin genes and transformation

The CrA genomic clone (pA7) is described in Sugase et al. (1996). A genomic clone of NAP was obtained by screening a C. reinhardtii genomic library (λFIX II, Stratagene, La Jolla, CA) with an NAP 3′-UTR sequence as a probe (Kato-Minoura et al., 1998). A 4.2-kb NotI fragment derived from a positive phage plaque was subcloned into pBluescript (pNAP4.2N). The chimeric actin genes were constructed as shown in Fig. 1. Primers used for chimeric genes construction were 5′crA1.g (TCGAGGGGGCCACATCCC) and 3′NdcrA738.g (ACACGCAGGATTCAAACATATGCC) to amplify the 5′-UTR of pA7; 5′NdNap97.cL (GGCATATGACTTCCGGCCTTCCA-GACACTGATACT) and 3′KS+30 (AGCTTGATATCGAATTCCTG-CAGCCCGGGG) for the coding region of pNAP4.2N; 5′Nap-93.g (CTGGGAGGCTGGGCTGCG) and 3′NcNap81.c (AACAGCGCA-GACCAAACCCATGGCC) for the 5′-UTR of pNAP4.2N; and 5′NccrA50.c (TTCCATGGCTGACGAGGGCGAGGTCTCTGCTCTG) and 3′crA4534.g (TAAAAATGCTACACACGAAATGGTCCGAGC) for the coding region of pA7. Each amplified fragment was subcloned into pBluescript, and the 5′-UTR and coding regions were linked using either the NdeI or NcoI site. These chimeric genes were linearized by digestion with HindIII and then used to transform ida5/arg2, as described in Ohara et al. (1998).

Fig. 1.

(Upper) Southern blot analyses showing the chimeric genes used for transformation. Genomic DNA from the transformants (TF5 and TF10) and the parent strain (Pa; ida5/arg2) was digested with SmaI (left panel) or XhoI (right panel) and probed with cDNA fragments of protein-coding regions from CrA or NAP. Arrows indicate the restriction fragments from the chimeric genes. Closed and open arrowheads indicate the fragments of wild-type genes. In TF5, several fragments (probably byproducts of the transformation) were seen. (Lower) Structure and restriction map of the two Chlamydomonas actin genes and the chimeric genes constructed in the present study. pA7, CrA genomic clone; pNAP4.2N, a NAP genomic clone; pNAPΛpA7 and pA7ΛpNAP, chimeric genes composed of reciprocal portions of the two actin genes. Solid boxes indicate coding exons of pA7, and shaded boxes indicate those of NAP. Open boxes represent 5′- and 3′-UTR, and lines mark introns and flanking DNA. Restriction sites used in the Southern analyses and the translation start (ATG) and stop (TAA or TGA) codons are shown.

RESULTS

Two chimeric genes of NAP and CrA were constructed as illustrated in Fig. 1. In both constructs, the 5′-UTR region of one gene was conjugated to the protein-coding region of the other. Because the 5′-UTR most likely contains regulatory elements for gene expression (Brunke et al., 1984), I expected that swapping of this region between the two genes would disturb their expression. The chimeric genes were transformed into ida5/arg2, and strains TF5 and TF10 are the transformants of pNAPΛpA7 and pA7ΛNAP, respectively. Success of transformation was confirmed by Southern blotting; in both transformants, restriction fragments from the chimeric genes were detected in addition to the fragments from the intrinsic genomic fragments (Fig. 1).

Transformant TF5 had regained the inner-arm dynein heavy chains missing in ida5 (Fig. 2). As shown previously (Ohara et al., 1998), dynein recovery is due to the expression of CrA from the chimeric genes. However, only a small amount of CrA mRNA was transcribed in the TF5 cells, and, unlike in wild-type cells, NAP mRNA also was expressed (Fig. 3). Despite their low mRNA levels, the CrA and NAP proteins were expressed in considerable amounts, as detected with their specific antibodies (Kato-Minoura et al., 1998) (Fig. 4A). To examine the sublocalization of CrA and NAP within the flagella, wild type, ida5, and TF5 flagella were isolated and fractionated (Fig. 4B). In TF5 flagella, CrA was abundant, and only a small quantity of NAP was detected, mainly in the fraction of proteins that remained in the axoneme after ATP extraction.

Fig. 2.

SDS-PAGE patterns of the axonemes of wild type (wt), ida5, and TF5. Only a portion of the gel showing dynein heavy chains is presented. m, membrane contamination; α, β, and γ, outer dynein heavy chains; bands indicated by hatched lines, inner dynein heavy chains; hatched lines with asterisks, heavy chains missing in ida5 (almost overlapping with γ chain).

Fig. 3.

Northern blot analyses showing CrA and NAP mRNA levels in wild type (WT), ida5, and TF5. Total RNA was isolated and probed with a portion of the 3′-UTR of either CrA or NAP. Arrowheads indicate positions of CrA and NAP mRNA (almost the same size).

Fig. 4.

Western blot analysis. (A) Total cell lysates were electrophoresed, and then probed with anti-CrA or anti-NAP antibody. ida5*; ida5/arg2 was used in this experiment. (B) Flagella were isolated and separated into fractions containing 1% NP-40-soluble proteins (M), those released from the axoneme by 10 mM ATP (A), and residual axonemal proteins (R), as described in Pazour et al. (1998). Fractions then were electrophoresed and stained with Coumassie Brilliant Blue (left panel) or probed with anti-CrA, anti-NAP, or anti-p28 antibody (right panels).

Table 1 summarizes the motility properties of the transformants. TF5 recovered swimming velocity to 83% of the wild-type level (about 210% of the ida5 level ). In contrast, TF10 exhibited an even slower swimming velocity than ida5.

Table 1.

Motility of C. reinhardtii clones used in this study.

Closer observation, however, revealed a small population of TF5 that lacked flagella. Because this finding suggested that the transformant might have defects in flagellar growth, I next examined the flagellar regeneration process in the transformants after deflagellation by pH shock (Fig. 5). When wild-type cells are deflagellated, flagella start to grow within several minutes, and almost all cells resume swimming within 40 min; flagella typically recover their original length (~10 μm) within 2 h (Rosenbaum et al., 1969). In TF5, 32% of cells lacked flagella, and the rest of the cells had flagella of various lengths before flagellar amputation (Fig. 5, upper column). After flagellar amputation, only a small population of cells regenerated flagella, and 73% of cells remained non-flagellated even after 60 min. This scenario is in striking contrast with those of wild-type and ida5 cells, >90% of which regained flagella. In contrast, the number of TF10 cells that recovered flagella was comparable to that in ida5, although the transformant also showed some abnormality in the distribution of flagellar lengths.

Fig. 5.

Histograms showing flagellar length recovery during reflagellation. For bars representing >30% of flagellar number, the net values are also indicated. B. D., before deflagellation. Time after deflagellation is noted.

To see whether the reduced flagellation activity of TF5 is due to irregular protein expression as a result of the chimeric gene transformation, the CrA mRNA level was quantified (Fig. 6). When wild-type cells were deflagellated, there was relatively slow upregulation of CrA compared with that of NAP, as previously reported by Hirono et al. (2003). In contrast, in TF5, rapid and robust expression of CrA was observed. As in ida5 and wild-type cells (Hirono et al., 2003), the expression of NAP was conspicuously upregulated in TF5 upon deflagellation.

Fig. 6.

Expression of CrA and NAP after deflagellation. mRNA level was quantified by Northern blotting, and the signal densities of corresponding bands were estimated using NIH image software. Horizontal axes indicate time after deflagellation; vertical axes indicate signal density (arbitrary units).

DISCUSSION

NAP was first discovered in the Chlamydomonas mutant ida5, which lacks the gene for conventional actin. This discovery has caused us to wonder why the NAP gene is preserved in wild-type cells, which express only negligible amounts of NAP under normal conditions. Recently, however, Hirono et al. (2003) found that NAP is expressed shortly after flagellar amputation in wild-type cells. From this observation, these investigators suggested that NAP might somehow be involved in the flagellation process. The findings presented in the current paper support their suggestion.

The most important finding of the current study is that a transformant (TF5) with a chimeric gene combining the 5′-UTR region of NAP with the coding region of CrA, exhibited a notable deficiency in flagellation activity (Fig. 5). The flagellar deficiency seems to be confined to the flagellation process, as flagellated cells recovered dynein composition and motility to the same degree as transformants with wild-type CrA (Fig. 2 and Table 1; Ohara et al., 1998). In contrast, TF10, in which the transforming construct comprises the CrA 5′-UTR region plus the NAP coding region, has normal flagellation activity (Fig. 5), suggesting that the process of transformation with a chimeric gene itself has no influence on flagellation. Rather, the 5′-UTR of NAP that was included in pNAPΛpA7 must have caused a change in the reflagellation activity of TF5. When wild-type cells are deflagellated, the expression of NAP message precedes that of the CrA message; the NAP message is expressed maximally <20 min after flagellar amputation, whereas actin expression reaches a peak 40 min after amputation. In addition, the increase in message expression is much larger for NAP than CrA (Hirono et al., 2003). In the present study, CrA mRNA expression in TF5 occurred very rapidly and robustly after flagellar amputation, as did NAP mRNA from the native gene. Abnormal expression of CrA immediately after flagellar amputation may inhibit proper flagellation, possibly by interfering with the function of NAP, which plays a crucial role during an early phase of flagellar regeneration.

What specific role does NAP play? NAP is present in growing flagella of wild-type cells, but it is gradually lost as flagellar growth is completed (Hirono et al., 2003). During flagellation, several axonemal proteins are transported to the site of assembly by means of intraflagellar transport (IFT), in which moving particles termed “rafts” are observed in the space between the axoneme and membrane (Kozminski et al., 1993; Cole et al., 1998; Scholey, 2003). When flagella are extracted with detergent, rafts and other soluble proteins are released in the “membrane-plus-matrix” fraction. Extraction of the remaining axoneme with ATP releases IFT motors and associated compounds that are bound to the axoneme in an ATP-sensitive manner. In wild-type and TF5 cells, faint bands of NAP were detected in the membrane-plus-matrix fraction (Fig. 4B). This observation suggests that NAP may function in IFT as a minor component of rafts.

Various protein assemblies are transported first as precursors and then converted into mature forms in the axoneme. For example, the radial spoke complex enters growing flagella as a 12S precursor and is converted into a 20S form (Qin et al., 2004). Such conversion occurs also in inner-arm dyneins (Piperno and Mead, 1997). Therefore, actin may well replace NAP in the inner-arm complex during the process of maturation. In TF5, however, the axoneme contained a considerable amount of NAP despite the cells' recovery of dynein composition and motility (Fig. 4B). In this case, abnormal expression of CrA may have interfered with the NAP-actin exchange that would have readily taken place under normal conditions. Alternatively, NAP may participate in the transport or localization of inner arms to specific loci on the outer doublet. Chemical crosslinking experiments have shown that actin is associated indirectly with inner arm heavy chains, through light chains that are bound to the N-terminal half of the heavy chain (Yanagisawa and Kamiya, 2001). Because ATP-independent dynein-micro-tubule association likely occurs at the N-terminal portion of the heavy chain, it seems possible that NAP and actin are somehow involved in the correct localization of inner-arm dyneins. The exact functions of NAP and actin await studies using strains in which the expression of NAP is reduced.