Structure and Expression of Fungal Calmodulin Gene *

I report on the isolation, structural analysis, and in vivo expression patterns of a fungal calmodulin gene. The gene is intronless and encodes a protein of 148 amino acid residues that is 92% homologous with vertebrate calmodulins. Through S1 nuclease transcript mapping, it was determined that the cloned gene (a ) is transcribed in vivo, (b) has a 5”untranslated region of about 400 nucleotides, and (c) has a 3”untranslated end of about 300 nucleotides. Southern blot hybridization analysis of the genomic DNA and the cloned gene provide evidence for the existence of only one type of calmodulin gene in the organism. The amino acid sequence deduced from the DNA sequence shows that Achlya klebsiana calmodulin has amino acid substitutions that are a mix of those seen in calmodulins from invertebrates such as Drosophila and trypanosome when compared to mammalian calmodulins. Not surprisingly, it has less resemblance to calmodulins from Saccharomyces and Dictyostelium.

I report on the isolation, structural analysis, and in vivo expression patterns of a fungal calmodulin gene. The gene is intronless and encodes a protein of 148 amino acid residues that is 92% homologous with vertebrate calmodulins. Through S1 nuclease transcript mapping, it was determined that the cloned gene ( a ) is transcribed in vivo, (b) has a 5"untranslated region of about 400 nucleotides, and (c) has a 3"untranslated end of about 300 nucleotides. Southern blot hybridization analysis of the genomic DNA and the cloned gene provide evidence for the existence of only one type of calmodulin gene in the organism. The amino acid sequence deduced from the DNA sequence shows that Achlya klebsiana calmodulin has amino acid substitutions that are a mix of those seen in calmodulins from invertebrates such as Drosophila and trypanosome when compared to mammalian calmodulins. Not surprisingly, it has less resemblance to calmodulins from Saccharomyces and Dictyostelium.
Calcium mediates numerous cellular responses and plays a pivotal role in the regulation of cellular homeostasis (1). These diverse cellular activities of Ca2+ are mediated by cytoplasmic Ca2+ receptors of which calmodulin is the major one in nonmuscle cells ( 2 ) . Calmodulin is a highly conserved acidic monomeric polypeptide (Mr - 16,700) that is present in eukaryotic cells as diverse as mammals (3)(4)(5)(6), invertebrates (7)(8)(9), green plants (10,11), and fungi (12,13). The calmodulin gene has been studied primarily in the cDNA form in human (14), chicken (15), electric eel (16), rat (17), toad (18), trypanosome (19), and slime mold (20). The genomic form of the gene has been well analyzed in chicken (15), Drosophila (21), rat (22), and trypanosome (19), Chicken and rat calmodulin genes have 5 introns in the coding region whereas the Drosophila gene has 3 introns. The three tandemly repeated calmodulin genes of trypanosome are identical and intronless; they are characterized by a 35-nucleotide 5' end leader that is not contiguous with the genes (19). Based entirely on restriction endonuclease DNA fragmentation and hybridization analyses, it was concluded that Xenopus (18) and Dictyostelium (22) calmodulin genes have introns. Thus several eukaryotic calmodulin genes studied so far have introns and the absence of intron in the Achlya gene may be unusual.
We have shown in a series of studies that Achlya klebsiana * This work was supported in part by a grant from the Natural Science and Engineering Research Council of Canada. The costs of publication of this article were defrayed in part by the payment of page charges, This article must therefore be hereby marked "advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence($ reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s1 J05116. has a strong physiological dependence on calcium. Growth (23), energy-linked transport of amino acids (24), nucleosides (25), and sugars (26), and the process of sporulation (27) were all shown to be dependent on the availability of Ca2+. Ca2+, therefore, must play a pivotal role in cellular homeostasis of Achlya. In support of the notion that Ca2+ is important for sporulation, Suryanarayana et al. (28) isolated calmodulin from Achlya ambisexwlk and showed that its induction was associated with cell wall lysis during sporulation (29). I show here that A. klebsiana has a functional intronless calmodulin gene which is induced to produce an abundant quantity of calmodulin transcripts when it is undergoing sporulation.

MATERIALS AND METHODS
Organism+" coenocytic freshwater mold, A. klebsiana, was used as the source of the calmodulin gene studied.
Induction of Sporulation-Hyphal cells derived from sporangiospores germinated and grown vegetatively for 20 h at 24 "C in defined medium, were harvested and resuspended in sporulation induction medium as described (30).
Zsolation of High Molecular Weight DNA-About 20 g (fresh weight) of suction-dried mycelia was frozen with liquid nitrogen and ground to a fine powder with mortar and pestle. The powder was suspended in 50 ml of buffer of composition 0.1 M NaCl, 0.1 M Na2EDTA, 50 mM Tris-C1, pH 8, and 1% Sarkosyl. Proteinase K was added to a final concentration of 50 Fg/ml and incubated for 1 h at 37 "C. An equal volume (70 ml) of buffer (0.1 M NaC1, 0.1 M Na2EDTA, 50 mM Tris-C1, pH 8)-saturated "phenol" (redistilled phenol mixed with 0.1% 8-hydroxyquinoline and 0.5% 2-mercaptoethanol) was added. The solution was gently mixed by hand for 10 min. The mixture was centrifuged at 10,000 X g for 10 min at room temperature to separate the phases. The upper aqueous phase was recovered and phenol extraction of proteins repeated three times or until the interface was free of precipitate. The aqueous fraction was extracted twice with an equal volume of CHCI,. One-tenth volume of 3 M NaOAc, pH 5.2, was added to the aqueous phase and nucleic acids precipitated by layering 2.2 volumes of chilled absolute ethanol over the nucleic acid solution and DNA recovered by spooling with a clean glass rod. The DNA was redissolved in TE (10 mM Tris-C1, pH 7.5, 1 mM EDTA) buffer and respooled. This process was repeated three times and the final spooled DNA was air-dried for 10 min before dissolving in a small volume of TE buffer to a concentration of 0.5- Construction of Genomic Library and Screening for Calmodulin Gene-DNA isolated as described above was consistently greater than 80 kb.' The DNA was partially digested with restriction endonuclease MboI and size-fractioned in sucrose density gradients. DNA fragments of 15-23 kb were selected and ligated into the BamHI site of bacteriophage EMBL3. Ligated products were packaged in vitro and propagated in Escherichia coli P2.392. Ligation and packaging were done according to Stratagene, San Diego, CA and propagated in E. coli P2.392. The library was screened with 32P-labeled electric eel calmodulin cDNA (pCM1116) provided by John Putkey and Anthony Means of Baylor College, Houston, TX by the method of Benton and Davis (31).
Preparation of cDNA-Poly(A)+ RNA was isolated from vegetatively growing and sporulation induced Achlya cells (32) and converted to double stranded cDNA by fast protein liquid chromatogra-The abbreviation used is: kb, kilobase(s). Genomic Southern Blot Hybridization-High molecular weight DNA was digested to completion with a variety of restriction endonucleases, and the DNA fragments were separated in 0.7% agarose gel slabs by electrophoresis and then transferred to Hybond (Amersham Corp.) nylon sheets by the alkaline method of Reed and Mann (33). Hybridization was carried out using, as probes, the 2.0-and 1.7kb DNA fragments representing the 5' and 3' ends, respectively, of Achlya calmodulin gene. Achlya DNA probes were labeled with 32P by the random primer method (34).
RNA Blot Hybridization-Poly(A)+ RNA was isolated as described (32) and electrophoresed in 1.2% agarose denaturing gel containing 6% formaldehyde and transferred to Hybond. RNA on Hybond was probed with a 3.8-kb Sal1 DNA fragment containing the full Achlya calmodulin gene. The probe was 32P-labeled by nick translation (35).
SI Nuclease Mapping-The S1 nuclease transcript mapping procedure of Berk and Sharp (36) was used to evaluate whether the calmodulin gene isolated was transcriptionally active in vivo and to determine the approximate initiation and termination sites of transcription of the gene. Total RNA was isolated from calmodulininduced cells and used to hybridize against the 2.0-and 1.7-kb ClaI DNA fragments representing the 5' and 3' ends of the fungal calmodulin gene. These two DNA fragments contain several hundred nucleotides of untranslated borders of the gene. A single CluI site situated about 50 nucleotides from the 3' end of the coding sequence of the gene was used as the reference point in estimating the transcription initiation and termination sites. S1 nuclease-insensitive DNA-RNA hybrids as well as standards were analyzed in nondenaturing 1.2% agarose gels according to Maniatis et al. (37).
Preparation of Phagemid MIS/Calmodulin DNA Recombinants-Phagemids M13+ and M13-(Stratagene, San Diego, CA) were linearized with ClaI and ligated independently to 2.0-and 1.7-kb ClaI fragments recovered from a recombinant clone isolated from the genome, and known to contain the fungal calmodulin gene. The phagemid recombinants were used to transform E. coli JM107.
Isolation of DNAs from Gels-DNA in agarose gels were recovered by the "Geneclean" technique (Bio 101 Lab., La Jolla, CA).
Exonuclease IZZ and Mung Bean Nuclease Deletions-Phagemid recombinants containing the 2.0-and 1.7-kb DNA fragments were linearized with ApaI to yield 3' extended ends; this was followed by Sal1 digestion which created a 5' extension at one of the two ApaI ends in both phagemids. In case of the 2.0-kb-containing M13 recombinant, a 0.2-kb fragment was removed from the insert by Sa21 resulting in a 1.8-kb/M13 recombinant phagemid. The 1.8-kb/M13 and the 1.7-kb/M13 phagemids were foreshortened unidirectionally by the combined action of exonuclease 111 and mung bean nuclease by the method of Henikoff (38). The deleted phagemids were religated and used to transform E. coli JM107. Phagemids from transformants were screened and nests of deletion recombinants covering the length of the inserts selected. The phagemids were purified by CsCl (39) and used for DNA sequencing.
DNA Sequencing-DNA sequencing was carried out by the dideoxy chain termination method (40). Prior to sequencing, the phagemid was denatured by treatment with alkali as described by Chan and Seeburg (41) and the primer annealed according to Korneluk et al. (42) Autoradiography-DNA sequencing gels were exposed, after drying, to Kodak X-Omat AR film for 1-3 days. Other autoradiograms were prepared with Kodak X-Omat RP films.

RESULTS
Isolation of Fungal Calmodulin Gene-About 50,000 recombinant bacteriophages constructed as described under "Materials and Methods" and representing more than five genome equivalents of the freshwater mold A. klebsiana were screened with 32P-labeled electric eel calmodulin cDNA as probe. Two positive clones were isolated. Both clones had DNA inserts of about 15 kb and identical restriction endonuclease digestion patterns (not shown). Consequently, one was selected for further study.
Restriction Endonuclease Map of Calmodulin Gene-DNA from the recombinant bacteriophage harboring the calmodulin gene was digested with several restriction endonucleases including BamHI, ClaI, DraI, HindIII, HpaI, and SalI, the fragments electrophoresed, Southern blotted, and probed with 32P-labeled electric eel calmodulin cDNA. Although a single strongly hybridizing band of 3.8 kb was detected in the SalI digest products, two positively hybridizing bands of 2 and 1.7 kb observed in the ChI digest products were selected for subcloning and structural analysis of the gene. The rationale was that the two ClaI DNA fragments probably represented the 5' and 3' ends of the gene, as proved to be the case.
A detailed restriction endonuclease map of the cloned 15kb genomic fragment containing the calmodulin gene was obtained using the 2-and 1.7-kb subfragments as probes ( Fig.  1, lower region). Location and orientation of the coding region in the 2-and 1.7-kb fragments was achieved by using as probes the 0.19-kb PstI DNA fragment from electric eel cDNA which represents the 5' end of the gene and the 0.75-kb PstI DNA fragment which represents the 3' end of the gene (16) (Fig. 1, upper region).
Evidence for Existence of Single Calmodulin Gene-Total genomic DNA and 15-kb DNA fragment containing the calmodulin gene were doubly digested with a combination of ClaI and either ApaI, DraI, EcoRV, or SmaI restriction endonucleases. Southern blots of these digests were probed with 2-kb ClaI DNA fragment (representing the 5' end of the gene) and the 1.7-kb ChI DNA fragment (representing the 3' end of the gene). The results (Fig. 2) show that the genomic DNA and the cloned 15-kb DNA have the same hybridization patterns. Two minor hybridizing bands appearing in the ChIIApaI digest of the cloned DNA probed with the 5' end of the gene are artifacts because no ApaI site exists in the 2-or 1.7-kb fragment; a feature that was exploited in preparing progressive unidirectional deletions for sequencing of the gene (see  later). Because of the identical nature of the genomic and cloned DNA hybridization patterns, it is likely that only one form of the calmodulin gene isolated exists in the fungus.
Nucleotide Sequence of Achlya Calmodulin Gene-All of the 2-kb and a portion of the 1.7-kb ClaI DNA fragments containing the fungal calmodulin gene were sequenced (following exonuclease III/mung bean nuclease unidirectional deletions of the linearized plasmids as described under "Materials and Methods") by the dideoxynucleotide chain termination method (40). The strategy used to procure an overlapping nest of deletions is schematically presented in Fig. 3. DNA sequences obtained were assembled and analyzed by the Beckman Microgenie program, and a portion of it displaying the coding sequence of the gene and its 5'-and 3'-noncoding borders is presented in Fig. 4. The amino acid sequence of the  into 2-kb ( a ) and 1.7-kb ( b ) DNA fragments that were ligated and cloned separately in phagemid M13+. Clone A, in both cases, represents undeleted fragments. Clone €3 in a is SaLI truncation of the 2-kb DNA insert. All other clones are exonuclease III/mung bean nuclease deletions. Arrow signifies region of each clone that was sequenced from either the T3 (0) or T7 (0) promoters. The rectangular boxes specify the coding regions of the gene.
fungal calmodulin gene was deduced from the DNA sequence and the information is incorporated into Fig. 4.
The assumption is made that like other calmodulins, the first amino acid of the mature protein succeeds the initiation codon AUG. This means that the protein is 148 amino acid residues long with alanine at the amino end and lysine at the carboxyl end. The protein sequence obtained is 92% homologous to electric eel calmodulin. A comparison of the amino acid sequence of calmodulins from human, electric eel, trypanosome and Achlya (Fig. 5) shows that compared to human, electric eel calmodulin has a single amino acid difference. Achlyu calmodulin has 11 amino acid replacements and trypanosome has 11 also. Assuming that Achlyu calmodulin has a tertiary structure that is similar to those of other eukaryotic calmodulins ( l l ) , four structural Ca2+ binding domains are recognizable (Fig. 6). An interesting feature is that single EcoRI and ClaI sites are present a t identical positions of the coding sequences (codons 67-68 and 130-131, respectively) of Achlya and Drosophila (21) calmodulin genes. This feature is not observed in the calmodulins of chicken (15), Dictyostelium (20), electric eel (16), and rat (17), which reflects, to some extent, differences in the nature of codons used by these organisms.
Within the 5' and 3' flanks of the coding region presented are several TATA box homologies and many CAAT-like (CAAT and CAAAT) consensus sequences at the 5' end and two AAATAA-like homologies at the 3' end. Whether these sequences serve as transcription signals for the gene identified cannot be affirmed with the existing data.
Accumulation of Calmodulin mRNA during Development- Suryanarayana and Thomas (29) have shown that during sporulation in A. ambisexualis, the cells accumulated high levels of calmodulin which were localized by immunofluorescence to sporangial discharge papillae and the spores themselves. Achlyu cells induced to develop sporangia and sporulate were analyzed for calmodulin mRNA content during the induction phase (starvation). Poly(A)+ RNA was isolated from starving cells a t different times, electrophoresed, transferred to nylon filter and probed with a 3.8-kb Sal11 cleaved DNA fragment containing the full length of the calmodulin gene and hundreds of nucleotides of its 5' and 3' end flanks. The results (Fig. 7a) show that there was a single poly(A)+ RNA hybridizing band whose intensity increased with time. The most dramatic increase occurred within the first 2 h. When  (Fig. 7b). This implies that enhanced transcription of the calmodulin gene may be linked to asexual differentiation in this organism. Isolated Calmodulin Gene Is Transcribed-Although the cloned calmodulin gene hybridized to blotted poly(A)+ RNA, it was necessary to show that it represents a gene that is transcribed in uiuo. Indirect evidence for this was obtained by S1 nuclease transcript mapping. Total cell RNA was hybridized to (a) the 2-kb ClaI DNA fragment labeled at the 5' end and represented the 5' end of the gene and ( b ) the 1.7-kb ClaI DNA fragment labeled at the 3' end and represented the 3' end of the gene. Following S1 nuclease hydrolysis of unprotected regions of the hybrids, the products were sized with neutral 1.2% agarose gel (Fig. 8). The approximate size of the fragment defining the transcriptional start site to the sole ClaI site in the coding sequence was 790 nucleotides, while the size of the fragment from the same ClaI site to the transcriptional termination point was about 370 nucleotides. These results are compatible with single transcriptional start and stop sites for the gene being expressed under nutrient starvation conditions.

DISCUSSION
The amino acid sequence of Achlya calmodulin gene was deduced from the DNA sequence shown in Fig. 4. When it is compared to the amino acid sequence of calmodulins from trypanosome (19), electric eel (16), and the uniform sequence found in mammals such as human (14), chicken (15), and rat (17), and the amphibian Xenopus (18), it is seen that the fungal protein is 92% identical to the mammalian calmodulin and 93% identical to that of trypanosome (Fig. 4). Whereas Achlya and trypanosome calmodulins have 11 amino acid replacements, in each case, compared to the mammalian protein, 5 of the  are identical replacements in the fungus and protozoan proteins.
Calmodulin gene from Drosophila melanoguster has been sequenced (21). Unlike the fungal (this report) and protozoan (19) genes, it is segmented into introns and exons. The amino acid sequence of the insect calmodulin differs from the mammalian protein a t only 3 residue positions: 99, 143, and 147. Interestingly, the replacements at these positions are the same as found in the fungal and protozoan calmodulins. As Smith et al. (21) have noted, the presence of serine at position 147 seems, therefore, to be a characteristic of practically all invertebrates.
All calmodulins studied so far are composed of four similar calcium binding domains. Each domain is defined by a stretch of amino acids running from residues 12 to 39,48 to 75,85 to 112, and 121 to 148 (43), and each fits the helix-loop-helix test sequence motif proposed by Kretsinger (44). The deduced amino acid sequence of Achlya calmodulin can also be divided into four structural calcium binding domains (Fig. 6). Analysis of the homology that exists between the four domains (calcium   Table I. Domains I and I11 are nearly 70% is the general pattern that has been observed for calcium homologous whereas domains I1 and IV are less than 56% binding domain homologies in other calmodulin genes (16,21, homologous  scripts. S1 nuclease protection reactions were carried out with approximately 20 pg of total RNA and either 5' end-labeled 2-kb ClaI DNA fragment representing the 5' end of the calmodulin gene or 3' end-labeled 1.7-kh ClaI DNA fragment representing the 3' end of the calmodulin gene. Lane 1, Rethesda Research Laboratories 1-kb DNA size markers electrophoresed in 1.2% neutral agarose gel, Southern blotted, and probed with '"P-labeled 1-kb ladder DNA. Lane 2, 2-kh ClaI DNA fragment. Lane 3, S1 nuclease protected hybrid of total RNA and 2-kh ClaI DNA. Lane 4, 1.7-kb ClaI DNA fragment. Lane 5 , SI nuclease protected hybrid of total RNA and 1.7-kb ClaI DNA.   (Fig. 2). Hybridization patterns of these same DNA fragments to the 0.19 kb (5' end) and 0.75 kb (3' end) of electric eel calmodulin cDNA used as probes gave identical results, albeit slightly different from those presented in Fig. 2 (data not shown). Second, S1 nuclease mapping results (Fig. 8) show that the cloned gene is probably transcribed in uiuo, and there is no indication that there exists another calmodulin transcript with significant homology to the one detected in S1 nuclease mapping. Third, the two calmodulin clones isolated from five haploid genome equivalents Achlya library were identical. The occurrence of a single calmodulin gene in the genomes of widely disparate organisms such as Drosophila (21), chicken (15), and Dictyostelium (20) has been reported. As well, multiple tandem copies and two non-allelic calmodulin genes have been reported for trypanosome (19) and Xenopus (18), respectively.
But with the exception of trypanosome and Achlya, the gene appears to be segmented into introns and exons in every other case, including humans.' The DNA sequence results presented includes about 500 nucleotides of presumed 5'-untranslated region, 447 nucleotides encoding the calmodulin protein and about 320 nucleotides of presumed 3"untranslated region (Fig. 4). The 5'untranslated region has several TATA-and CAAT-like sequence homologies, but it is not known which (if any) function in transcription. An interesting arrangement of two CAATlike sequence homologies in direct repeat mode is present between -441 and -457. Within the 5'-untranslated region are several in frame stop codons, one of which is only 6 nucleotides upstream from the assigned translation start codon. Such structural organization implies that ( a ) the protein is unlikely to be synthesized as a secretable entity, ( b ) Achlya calmodulin is not synthesized as a precursor protein, and (c) the translation start codon designated is probably correct. The 3"untranslated end is replete with in-frame translation stop codons, three of them occurring within 40 nucleotides of the assigned translation stop codon. Two possible closely set consensus sequences that may serve as polyadenylation signals are about 300 nucleotides downstream of the last translation codon. S1 nuclease maps of the transcript derived from this gene sequence show that transcription may start about 400 nucleotides upstream of the translation start codon and cease some 300 nucleotides downstream of the translation stop codon (Fig. 4). This gene, therefore, may encode the transcript identified in Fig. 7. Suryanarayana et al. (28) concluded from amino acid analysis of A. ambisexualis calmodulin that the protein is 181 amino acid residues long (33 residues longer than the average), but with a molecular weight that is 4,000 smaller than the average size of 17,000. The amino acid composition of this protein (shown in Table 11) indicates significant differences in the relative number of several amino acid residues (notably, arginine, glycine, asparagine/aspartic acid, glutamine/glutamic acid) when compared to calmodulins from human, electric eel, and trypanosome. By contrast, the chemical properties of A. klebsiana calmodulin deduced from the gene sequence show that the polypeptide is typically 148 amino acid residues long with an amino acid composition that is similar to that of other calmodulins (Table 11).
In summary, this communication has presented evidence showing that A, klebsiana has a single type of calmodulin gene that does not have any introns but does have transcription signals sequence homologies similar to those of eukaryotic genes. Evidence is also presented to show that the gene is probably expressed in uiuo as a single mRNA species. The protein encoded by this gene is quite typical of calmodulins in general.