lir-2, lir-1 and lin-26 Encode a New Class of Zinc-Finger Proteins and Are Organized in Two Overlapping Operons Both in Caenorhabditis elegans and in Caenorhabditis briggsae

Corresponding author: Michel Labouesse, IGBMC, CNRS/INSERM/ULP, BP163, 1, rue Laurent Fries, 67404 Illkirch Cedex, France., lmichel{at}igbmc.u-strasbg.fr (E-mail)

Communicating editor: R. K. HERMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

lin-26, which encodes a unique Zn-finger protein, is required for differentiation of nonneuronal ectodermal cells in Caenorhabditis elegans. Here, we show that the two genes located immediately upstream of lin-26 encode LIN-26-like Zn-finger proteins; hence their names are lir-1 and lir-2 (lin-26 related). lir-2, lir-1, and lin-26 generate several isoforms by alternative splicing and/or trans-splicing at different positions. On the basis of their trans-splicing pattern, their intergenic distances, and their expression, we suggest that lir-2, lir-1, and lin-26 form two overlapping transcriptional operons. The first operon, which is expressed in virtually all cells, includes lir-2 and long lir-1 isoforms. The second operon, which is expressed in the nonneuronal ectoderm, includes short lir-1 isoforms, starting at exon 2 and lin-26. This unusual genomic organization has been conserved in C. briggsae, as shown by cloning the C. briggsae lir-2, lir-1, and lin-26 homologs. Particularly striking is the sequence conservation throughout the first lir-1 intron, which is very long in both species. Structural conservation is functionally meaningful as C. briggsae lin-26 is also expressed in the nonneuronal ectoderm and can complement a C. elegans lin-26 null mutation.

WHEN the embryo develops from a single pluripotent cell (the zygote) to a multicellular organism, decisions become more and more complex as they involve more and more different cell types. A commonly used genetic strategy to achieve this complexity is to use a combination of different transcription factors in different cells. As a source of diversity for transcription factors, animals have sometimes evolved partially homologous and redundant genes that can be either unlinked (e.g., the retinoic acid receptors in mammals, CHAMBON 1996 ) or linked (e.g., the proneural genes in Drosophila, SKEATH and CARROLL 1994 ). As another source of diversity, animals have also evolved mechanisms to generate multiple isoforms from a single gene through the use of different promoters or alternative splicing (e.g., the retinoic acid receptors, CHAMBON 1996 ).

In this study, we focus on the developmental decision that leads an ectodermal cell to become neuronal, glial or epidermal. We have previously described a gene in Caenorhabditis elegans, lin-26, that is expressed in and required for nonneuronal ectodermal cells to acquire their normal fates (LABOUESSE et al. 1994 , LABOUESSE et al. 1996 ). lin-26 is also expressed in the somatic gonad and is required for the formation of the somatic gonad epithelium (DEN BOER et al. 1998 ). Strong and null lin-26 alleles lead to embryonic lethality because of the degeneration of most hypodermal (in C. elegans, epidermal-like cells are termed hypodermal cells) and neuron-associated, glial-like cells (LABOUESSE et al. 1994 , LABOUESSE et al. 1996 ), while a weak loss-of-function mutation, lin-26(n156), causes a set of hypodermal precursors to adopt a neural fate (FERGUSON et al. 1987 ; LABOUESSE et al. 1994 ).

Because the decision to become neuronal vs. nonneuronal certainly depends on more than one gene, we decided to determine if there are genes partially homologous to lin-26 that could be involved in the same decision as lin-26. The gene lin-26 encodes a putative transcription factor (LABOUESSE et al. 1994 ) with two C2H2 motifs that are related to but different from the classical TFIIIA Zn-fingers (RHODES and KLUG 1986 ). In particular, the rst putative Zn-finger motif is characterized by the presence of five additional amino acids between the second cysteine and the first histidine and by the absence of the conserved phenylalanine normally located four amino acids after the second cysteine. The second putative Zn-finger motif is more canonical, except that it also misses the conserved phenylalanine.

Here we describe two genes that encode several isoforms with the same unusual C2H2 motif as LIN-26. These genes, which we named lir-1 and lir-2 (lin-26-related genes), map immediately upstream of lin-26. We report their structure and expression pattern. In addition, we describe the structure of the homologous lin-26 and lir genes in the nematode C. briggsae and show that the C. briggsae lin-26 homolog can substitute for the C. elegans lin-26 gene.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and general methods:
Methods for maintaining animals and genetic analysis were as described by BRENNER 1974 . The wild-type reference strain was N2 (BRENNER 1974 ); the C. briggsae strain used was a wild-type isolate provided by the Caenorhabditis Genetic Center (CGC); ML581, lin-26(mc15) unc-4(e120)/mnC1 [dpy-10(e128) unc-52(e444)] was described by DEN BOER et al. 1998 .

Isolation of cDNAs:
lir-2 and lir-1 cDNAs were isolated from two libraries prepared from size-selected cDNAs (1- to 2- and 2- to 3-kb ranges) primed with oligo(dT) (WATERSTON et al. 1992 ). Plaques (2 x 10⁶) were screened with a mixture of lir-2 and lir-1 probes obtained after reverse transcriptase followed by PCR amplification using the primers 5'AATGAGCATTGGGCTCGAATCGCTG (exon 2) and 5'ATGAGTGGAGTGTTTCATCATTAC (exon 7) for lir-2, 5'CAGTGAATTCAACCAAGGTGTCGTCTTCG (exon 1) and 5'GGTGATTTTAAGTCAATTATCG (exon 3/4 boundary) for lir-1. The structure of cDNAs was determined by PCR analysis and by sequencing.

lir-2: Seven lir-2 cDNAs were isolated. The longest cDNA (pML124) contains 1781 bp. Four clones were found to contain the first exon, among which three had additional nucleotides (nts) at their 5' end that were identical to the final nts of the spliced leader type 1 (SL1); a smaller cDNA starting at the beginning of the second exon had three nts matching the last three nts of SL1.

lir-1: Forty-two lir-1-positive clones were isolated. We found that cDNAs can start either within the first exon (9 clones) with a member of the SL2 family of spliced leaders (1 clone had nine additional nts identical to the final nts of the SL2 variant, called SL4, and another had five additional nts identical to the final nts of SL4 and SL5, see ROSS et al. 1995 ) or within the second exon with the SL1 spliced leader (4 clones had nine nts identical to the final nine nts of SL1). We also found that they can have three different structures at their 3' ends: 15 clones had a short size (referred to as lir-1A or lir-1D, depending on their 5' end; see Figure 2), 8 had an intermediate size (referred to as lir-1B), and 19 had a long size (referred to as lir-1C or lir-1E, depending on their 5' end); among cDNAs starting within the first exon, 4 were lir-1A-like, 1 lir-1B-like, and 4 lir-1C-like. The sequenced type A (pML127) and type C (pML129) cDNAs are 1131 and 1204 bp long, respectively.

View larger version (50K): In this window In a new window Download PPT slide   Figure 1. Alignment of the predicted C. elegans LIN-26-related proteins. The sequences of the predicted C. elegans LIN-26B, LIR-1A, LIR-2A, LIR-3 proteins and of their C. briggsae homologs (see text) were aligned using the program ClustalX (THOMPSON et al. 1997 ). The top part shows the alignment of the C2H2 motifs using the one-letter code for amino acids (similarities are not meaningful in more N-terminal parts). The bottom part shows the alignment of the region immediately C-terminal to the C2H2 motifs. A consensus sequence was determined and shown alongside the TFIIIA-type consensus zinc-finger motif (RHODES and KLUG 1986 ; note that in TFIIIA-like zinc fingers, the spacing between the two cysteines or between the two histidines can vary). Conserved amino acids, as defined by the program Prettyplot from GCG, (DEVEREUX et al. 1984 ) are boxed: amino acids present in four of seven or five of seven proteins are in lowercase letters; those present in more than six of seven proteins are in uppercase letters. +, basic residue (K or R); -, acidic residue (D or E); %, large hydrophobic residue (L, M, V, I); #, aromatic residue (F, Y, W); ., optimal alignment prescribed that the residue be considered missing. Spaces in the consensus lines indicate no conservation. Bold numbers above LIN-26B indicate the positions of point mutations found in lin-26 mutants: 1, lin-26(mc1), G to E (LABOUESSE et al. 1994 ); 2, lin-26(mc2), S to F (LABOUESSE et al. 1994 ); 3, lin-26(ga91), R to H (this work); 4, lin-26(n156), L to F (LABOUESSE et al. 1994 ). Pairwise comparisons between C. elegans LIN-26 and LIR proteins using the Bestfit program from GCG (DEVEREUX et al. 1984 ) show that identities and similarities over the entire length of the proteins are as follows: LIN-26B/LIR-1A (27.86, 50.16); LIN-26B/LIR-2A (21.93, 44.65); LIN-26B/LIR-3 (26.07, 47.85); LIR-1A/LIR-2A (22.38, 46.93); LIR-1/LIR-3 (21.4, 44.64); LIR-2/LIR-3 (26.43, 46.74).

View larger version (39K): In this window In a new window Download PPT slide   Figure 2. Genomic organization of the lir-2, lir-1, and lin-26 genes in C. elegans and C. briggsae. (A) The top line shows a partial restriction map of the BamHI fragment spanning nts 12,013–32,601 of cosmid F18A1 (GenBank accession no. U41535). The genomic structure of lir-2, lir-1, and lin-26 isoforms, as inferred from F18A1 and cDNA sequence comparisons, is schematized below. The general structure of these three genes was as predicted by Genefinder, except for the first exon. Boxes represent exons and open triangles represent introns; ORFs within transcripts are shown in dark, untranslated regions, (UTRs) are shown in white. The nature of the trans-spliced leader (SL) found at the 5' end of the different transcripts is symbolized by the symbol (members of the SL2 family of spliced leaders are collectively referred to as SL2). Inverted triangles show the positions of potential methionine initiation codons. lir-1 isoforms starting at exon 1 can have three possible 3' ends: in lir-1C, which contains four exons and can code for a potential protein of 296 aa, the last exon is not spliced; in lir-1A and lir-1B, which can code for proteins of 309 and 300 aa, respectively, the fourth exon is alternatively spliced from the same donor site, but onto different acceptor sites (see C). lir-1 isoforms starting at the beginning of exon 2 undergo alternative splicing similarly to lir-1A and lir-1C and were called lir-1D and lir-E, respectively; we did not identify any lir-1B-like cDNA among cDNAs starting at exon 2, possibly because we did not isolate enough cDNAs. RT-PCR experiments using an SL1 primer identified an isoform starting at lir-1 exon 3 (lir-1F); we have not determined the structure of the lir-1F 3' end (two possible structures are shown by analogy with lir-1D-E). (B) Detailed nt coordinates of exon boundaries for lir-2, lir-1, and lin-26 within the F18A1 sequence. (C) Enlargement of the region where lir-1 and lin-26 overlap, focusing on the alternative splicing pattern (C. elegans in the top part and C. briggsae in the bottom part). Below, the alternative C-terminal region of the different LIR-1 proteins is shown, starting at the level of the fourth exon; vertical arrows show the positions of introns, and an asterisk symbolizes the stop codon. Note that C.e.lin-26A, which is the cDNA described in LABOUESSE et al. 1994 , results from the use of the same splice donor site as C.e.lir-1B, but of a unique acceptor site (compared to lir-1A-C). The putative initiation codon of LIN-26A (ATG) and the stop codon (TGA) of LIR-1A overlap (ATGA). Although we detected only two C.b.lir-1 splice variants, we cannot exclude the existence of a third one that would result from the use of an alternative acceptor site located 9 nts downstream of the acceptor site described here. (D) Exon-intron map of lir-2, lir-1, lin-26, and four other genes (F18A1.6, F18A1.7, F18A1.5, and the beginning of F18A1.1) predicted to be encoded by the C. briggsae sequence a3d1 (top) and the C. elegans cosmid F18A1 (bottom). Horizontal lines represent intergenic regions, and arrows indicate the direction of transcription (other symbols are as described in A). A single isoform is shown for each gene. Note that the order, size, and direction of transcription of all genes in this part of the genome have been conserved between C. briggsae and C. elegans. The horizontal lines marked pML303 and pML208 represent the extent of C. briggsae or C. elegans genomic sequences, respectively, cloned in plasmids pML303 and pML208.

lin-26: Twenty-three partial and full-length lin-26 cDNA clones were isolated after screening 4 x 10⁵ plaques of an oligo(dT) cDNA library (BARSTEAD and WATERSTON 1989 ) probed with a lin-26 fragment prepared by PCR using the primers 5'TTAGTTGATCACCCTTCTTTCAAG (exon 4) and 5'TACGGATCCACAGAGTGCACACTTGTA (exon 4). All cDNAs started at or after the beginning of the third exon of the original lin-26 cDNA (LABOUESSE et al. 1994 ). The longest cDNA (pML500) contained the last seven nts of SL2, while two other clones contained additional nts at their 5' end matching the last nts of the spliced leader, SL2.

Isolation of C. briggsae lir-2, lir-1, and lin-26 homologs:
Southern blots of gels containing restriction digests of C. briggsae genomic DNA (prepared according to HESCHL and BAILLIE 1990 ) were hybridized at 50° in 5x SCP/2.5% Sarkosyl NL30/8% Dextran sulfate (20x SCP is 2 M NaCl/0.6 M Na₂HPO₄/20 mM EDTA, pH 6.2) and washed with 0.1% SDS/0.5x SCP at 50°. Probes were obtained after PCR amplification from cDNAs using the primers 5'AGCGCCTAGACTCGATACGGT (exon 6) and 5'TCGAAGCTTAATCATATTCACAGGTTCGTG (exon 7) for lir-2, 5'TACGGATCCAACGGATACAGATGGCTCCAGGAC (exon 2) and 5'GGTGATTTTAAGTCAATTAT (exon 3/4 boundary) for lir-1, 5'GGCTAACACCAGAAT (exon 4) and 5'CGTTGGATTTTATTCCG (exon 5) for lin-26. All three probes could detect fragments in the C. briggsae genome that appeared to be very tightly linked upon further analysis (data not shown). A Charon4A C. briggsae genomic library (HESCHL and BAILLIE 1990 ) was hybridized under the same conditions and with the same lir-2 and lin-26 probes as above (we did not screen with lir-1, as we expected the lir-1 homolog to be found on lin-26-positive clones). Six lir-2-positive and three lin-26-positive clones were found. Restriction analysis of the longest lir-2-positive clone (named #a3) and the longest lin-26-positive clone (named #d1) showed that they overlap.

DNA sequencing and sequence analysis:
DNA (~20 µg) from recombinant clones #a3 and #d1 was sheared by sonication. The ends were repaired with Klenow enzyme and T4 DNA polymerase (SAMBROOK et al. 1989 ). DNA fragments ranging in size from 1 to 2.5 kb were gel purified and cloned at the EcoRV site of the pBSKII vector (Stratagene, La Jolla, CA). Plasmid DNA hybridizing to the EcoRI fragments corresponding to the inserts of clones #a3 and #d1 was prepared with Nucleobond AX20 cartridges (Macherey-Nagel, Düren, RFA) and sequenced with the pBSKII primers 5'CCTCGAGGTCGACGGTATCG and 5'GCCGCTCTAGAACTAGTGGATC. Assembly of sequences was processed using the Autoassembler program (Perkin Elmer, Norwalk, CT). Each strand was sequenced an average of 3.3 times for #a3 and 4.1 times for #d1. The resulting sequence was named a3d1. The program Genefinder (P. Green and L. Hillier, Washington University) was used to make predictions that were refined by comparing nt sequences from both species using programs from the Wisconsin GCG package (DEVEREUX et al. 1984 ).

In the remainder of the text, nts will be numbered starting at the +1 nt of cosmid F18A1 (GenBank accession no. U41535) for C. elegans genes and at the +1 nt of the a3d1 sequence for C. briggsae genes.

Northern blot analysis:
C. elegans and C. briggsae RNA were isolated from mixed-stage populations, as described by KRAUSE 1995 . For Northern analysis, 3 µg of poly(A)⁺ RNA was electrophoresed and transferred as described by SAMBROOK et al. 1989 . The resulting membrane was hybridized with the same probes described above.

Reverse transcriptase-PCR experiments:
Reverse transcriptase reactions (RT) were performed using the first-strand Synthesis kit from Stratagene. To determine the respective ratios of certain C. elegans lir-1 and lin-26 isoforms, we performed PCR experiments for only 20 cycles (to avoid becoming rate limiting for one component of the PCR reaction) starting from different amounts of RT products. RT-PCR products were detected by Southern blot analysis with a probe that was PCR generated using the same primers as for the RT-PCR reaction. RT-PCR experiments involving C. briggsae transcripts were carried out for 35 cycles; RT-PCR products were detected by ethidium bromide staining after agarose gel electrophoresis. We used the primers described below.

• C.e.lir-1: 5'TACGGATCCGACGCCGTCTCGGAAAGCTGAAG (exon 3) and 5'TCTAAGCTTTAATAACCTGATTAAATTTATTATTTG (3' UTR);

• C.e.lin-26: 5'TCTAAGCTTGAAGAGAGTCGATCA (exon 4) and 5'GGTTGAGCCATTGGCTCGAG (3' UTR);

• C.e.act-123: 5'GAGTCATGGTCGGTATGGGAC and 5'GGGAGAGGACAGCTTGGATGG;

• C.b.lir-2: 5'TGACTACAAATGGCCCACGG (exon 1) and 5'CAAGTATTGCTCGGCAGGGC (exon 4), or 5'GAAAATGGGCACTCACAGGGCC (exon 4) and 5'AAATTGAAAACATCGCGACC (3' UTR);

• C.b.lir-1: 5'GAACGATGAAGCAGCAACCA (exon 1) and 5'TATGGCTTGATGTGCGAGTA (exon 3), or 5'CACAGACTTTGGCTATGGAA (exon 2) and 5'ATTAAATTTGCACACGGCAG (3' UTR);

• C.b.lin-26: 5'TGAAGTGAGCAATACCAACA (exon 1) and 5'TTCTTTGATGGCACACTTCC (exon 2), or 5'CAGGAAGTGTGCCATCAAAG (exon 2) and 5'AATCTCGAGTTTCTGATGGG (3' UTR).

We also used the primers 5'TCTAGAATTCCGCGGTTTAATTACCCAAGT and 5'TCTAGAATTCCGCGGTTTTAACCCAGTTAC, which are derived from the sequences of the C. elegans SL1 and SL2 splice leaders, respectively, as alternatives to the exon 1 primers listed above; note that the SL2-specific primer cannot discriminate between SL2 and its variants in RT-PCR experiments.

lacZ reporter and other constructs:
Standard protocols (SAMBROOK et al. 1989 ) were used for all recombinant DNA work. Plasmids were microinjected at 50 ng/µl, except pML303, which was used at 1 ng/µl with 100 ng/µl of the plasmid pRF4 (MELLO et al. 1991 ). At least three stable lines were examined for each construct.

lir-2::lacZ construct: A 4781-nt genomic fragment was generated by PCR from the cosmid F18A1 using the primers 5'TACACTGCAGAAGCTCCAGCTACTATTGCG and 5'GCAGGATCCGGTTCGTGTTCCGAATTGTG and was subcloned into the PstI andBamHI sites of pPD21.28 (FIRE et al. 1990 ) to generate plasmid pML121.

lir-1::lacZ constructs: Plasmid pML169 carries the F18A1 BamHI/EcoRI fragment spanning nts 12,013–27,160 (an 8-mer BglII linker was added at the EcoRI site after making it blunt), the BamHI/StuI fragment containing the ß-galactosidase coding sequence from plasmid pPD16.43 (FIRE et al. 1990 , a BglII linker was added at the StuI site), and, finally, a fragment obtained after ExoIII nuclease digestion that starts at the EcoRI site at position 27,504 (transformed into a BglII site with a linker) and ends at position 28,254. Plasmid pML165 has a structure similar to pML169, except that it starts at the EagI site at position 17,709, and the lacZ is inserted at the NdeI site at position 27,501 of F18A1 instead of the EcoRI site at position 27,160.

lin-26::lacZ construct: Plasmid pML20 carries the F18A1 PstI/NheI fragment spanning nts 19,873–28,789, the SmaI/SpeI fragment containing the ß-galactosidase coding sequence from plasmid pPD16.43 (with an 8-mer NheI linker at the SmaI site), and the F18A1 NheI/BssHII fragment spanning nts 28,789–36,308.

Other lin-26 constructs: Plasmid pML224 is a derivative of pML301 (DEN BOER et al. 1998 ), which contains the AatII/NheI fragment from the lin-26A cDNA (see Figure 2) instead of the genomic AatII/NheI fragment spanning nts 27,899–28,789; it lacks the sequence corresponding to the second lin-26A intron and, thus, the acceptor site at which the SL2 is added onto lin-26B. Plasmid pML225 is a derivative of pML301 in which a stop codon was introduced between the potential initiation methionines of LIN-26A and LIN-26B isoforms at the AatII site located at position 27,899 by addition of a BglII linker (the new sequence reads TCG ACG TAG ATC TAC instead of TCG ACG TCT).

C. briggsae construct: Plasmid pML303 contains the SacII/SpeI fragment of C. briggsae genomic DNA spanning nts 14,043–25,168.

Immunostaining and in situ hybridization:
Animals and embryos were fixed and stained as described previously (CHANAL and LABOUESSE 1997 ) using LIN-26 antiserum at a 1/2000-fold dilution (LABOUESSE et al. 1996 ), the anti-adherens junction monoclonal antibody MH27 at a 1/200-fold dilution (FRANCIS and WATERSTON 1991 ) and a polyclonal anti-ß-galactosidase antibody (Cappel-Organon Teknika, Turnhout, Belgium), as suggested by the supplier. In situ hybridization was performed according to SEYDOUX and FIRE 1994 using the antisense RNA probes corresponding to the probes used to screen cDNA libraries, except that the lin-26 probe was prepared from the lin-26 cDNA plasmid pML500 linearized at the NheI site (exon 4 of lin-26A).

The accession numbers of the sequences described in this work are AJ130959 (lir-2A), AJ130960 (lir-2B), AJ012485 (lir-1A), AJ012486 (lir-1C), AJ130957 (lin-26B), and AJ131134 (C. briggsae a3d1 sequence).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

LIN-26 and three LIN-26-related proteins define a new group of Zn-finger proteins:
Sequencing of an 8-kb fragment surrounding lin-26 (LABOUESSE et al. 1994 ; M. LABOUESSE, unpublished results) and sequence analysis of the cosmid F18A1 (GenBank accession no. U41535, WILSON et al. 1994 ) revealed the existence of two open reading frames encoding putative proteins with C2H2 motifs related to the two LIN-26 C2H2 motifs. We named these putative genes lir-1 and lir-2 (for lin-26-related genes); they correspond to the candidate genes F18A1.3 and F18A1.4, respectively. A search through the C. elegans genome database identified a third lin-26-related gene mapping to another part of chromosome II within the cosmid F37H8 (GenBank accession no. Z81534), which will be referred to as lir-3 and corresponds to F37H8.1. As described below, we have also cloned the lin-26, lir-1, and lir-2 homologs from the nematode C. briggsae.

Figure 1 shows that the predicted proteins encoded by lir-1, lir-2 and lin-26 in both species and by C. elegans lir-3 share the same unique features within their C2H2 motifs as LIN-26. In particular, there are four (LIR-1 and LIR-3) or five (LIR-2 and LIN-26) additional amino acids (aa) between the second cysteine and the first histidine of the first motif compared to the canonical TFIIIA zinc finger (RHODES and KLUG 1986 ), and both motifs lack the conserved phenylalanine. In the second C2H2 motif, a valine generally replaces the conserved leucine two residues upstream of the first histidine. Interestingly, four out of the six known lin-26 point mutations affect the stretch GSRWNL in the first C2H2 motif, which is well conserved in LIR proteins (Figure 1). We propose that LIN-26 defines a new group of zinc-finger motifs within the TFIIIA family of Zn-finger proteins. We did not find any other C2H2 motif with 16 or 17 aa instead of 12 aa between the second cysteine and the first histidine, nor with the sequence GSR/KWNL in protein sequence databases.

The overall similarity among the C. elegans LIR proteins is rather low (the highest pairwise similarity is 28% identity, as determined by the Bestfit program, see Figure 1 legend). Besides the C2H2 motifs, another region, within the C end (PEL/M/VR/KAEWxxxL/MxxCFP, see Figure 1), is conserved between all proteins and does not show any similarity with proteins in the databases. A region rich in acidic residues, as can be found in the transactivation domain of certain transcription factors (MA and PTASHNE 1987 ), can be recognized in all proteins, at the C terminus of LIN-26 and LIR-1 or upstream of the zinc-finger motifs in LIR-2 and LIR-3, suggesting that all four proteins might be transcription factors.

Genomic organization of the lir-2/lir-1/lin-26 gene cluster:
To confirm the predicted structures of lir-2 and lir-1, we isolated cDNAs (see MATERIALS AND METHODS) and compared their sequences to the sequence of cosmid F18A1 (Figure 2). We performed RT-PCR experiments to confirm the existence of cDNAs (Figure 3 and data not shown).

View larger version (25K): In this window In a new window Download PPT slide   Figure 3. Transcript analysis. (A and B) Northern blot analysis of poly(A)+ RNA from (A) wild-type C. elegans or (B) C. briggsae mixed-stage populations was performed as described in MATERIALS AND METHODS and hybridized with the probe indicated above each lane. The sizes of the main isoforms including untranslated regions but excluding the spliced leader and the poly(A) tail are as follows: C.e.lir-2A, 1769 nts; C.e.lir-2B, 1712 nts; C.e.lir-1A, 1124 nts; C.e.lir-1B, 1157 nts; C.e.lir-1C, 1201 nts; C.e.lir-1D, 1025 nts; C.e.lir-1E, 1102 nts; C.e.lin-26A, 2176 nts; C.e.lin-26B, 1895 nts; C.b.lir-2A, 1743 nts; C.b.lir-1A, 1096 nts; C.b.lir-1C, 1155 nts; C.b.lir-1D, 974 nts; C.b.lir-1E, 1033 nts; C.b.lin-26B, 1812 nts. On the basis of these predicted sizes and on RT-PCR experiments (see C), we suggest that lir-2A, lir-1A, lir-1B, lir-1C and/or lir-1E, and lin-26B correspond to the most abundant isoforms of each gene. (C) RT-PCR experiments were performed as described in MATERIALS AND METHODS. Three different amounts of the reverse transcriptase reaction product, differing by a threefold factor (symbolized above each set of three lanes for embryos, L1 larvae and adults by 9, 3, and 1) were used to initiate the PCR reactions. Densitometric scannings of the middle autoradiogram show that lin-26C is 20–25 times less abudant than lin-26B at the L1 stage, 50 times less abundant in embryos, and not detectable at other stages (not shown for L2, L3, and L4 stages) and that the relative and absolute abundance of the three major classes of alternatively spliced lir-1 isoforms varies very little during development. Transcipts from the act-123 locus, which encodes actin (FILES et al. 1983 ), provide a loading control.

lir-2 structure: We found a major lir-2 isoform (lir-2A), as well as a minor isoform (lir-2B, see Figure 2A) that misses the first lir-2A exon. Both isoforms are trans-spliced to SL1 (KRAUSE and HIRSH 1987 ; NILSEN 1993 ). The lir-2A isoform can code for a 548-aa protein. Northern blot analysis shows that the major lir-2 mRNA is ~1.9 kb, which is the expected size of lir-2A (Figure 3A).

lir-1 structure: A complex pattern of trans-splicing and alternative splicing could potentially lead to nine different lir-1 isoforms (Figure 2, A–C) that would code for proteins ranging in size between 141 and 309 aa. The first and second lir-1 exons are separated by a 9147-bp intron, which is very striking, as most C. elegans introns are <60 nts (BLUMENTHAL and STEWARD 1997 ). We found that lir-1 transcripts can start either at exon 1 (isoforms lir-1A, B and C), in which case they are trans-spliced to SL2 variants (ROSS et al. 1995 ; EVANS et al. 1997 ) or at exon 2 (isoforms lir-1D and E) or exon 3 (isoform lir-1F), in which cases they are trans-spliced to SL1. Alternative splicing at the 3' end of lir-1 generates three classes of isoforms that can encode proteins with a distinct C-terminal domain enriched in acidic residues (Figure 2C). By Northern blot analysis, we detected a large band in the 1.2-kb range (Figure 3A) that can be resolved upon longer migration into two main bands with estimated sizes of 1.2 and 1.3 kb (data not shown), the expected sizes of lir-1A and/or lir-1E and of lir-C. RT-PCR experiments confirmed the existence of the three alternative lir-1 splice variants and indicated that their relative abundance remains stable during development (Figure 3C).

lin-26 structure: We had previously identified a long lin-26 cDNA (LABOUESSE et al. 1994 , now referred to as lin-26A) that overlaps with the 3' end of all lir-1 isoforms although its splicing pattern differs (Figure 2C). We have now identified a second lin-26 isoform (referred to as lin-26B) that is trans-spliced to SL2 at the third lin-26A exon (Figure 2C). The lin-26B isoform can code for a protein of 438 aa. RT-PCR experiments revealed a third lin-26 isoform (referred to as lin-26C, see Figure 2A), that lacks the fifth lin-26A exon that specifically encodes the second zinc-finger motif. Absence of this exon leads to the use of a different reading frame in the last coding exon (the LIN-26C protein would contain 358 aa instead of 438 aa in LIN-26B). The lin-26C isoform is at least 20-fold less abundant than the lin-26B isoform in L1 larvae, 50-fold less abundant in embryos, and essentially not detectable at other stages (Figure 3C and data not shown), indicating that it is a minor isoform, probably active only in a subset of cells (see DISCUSSION). Further evidence that LIN-26C is likely to play a minor role comes from the position of the mutation lin-26(mc4), a very strong or null allele that is located in the last lin-26 coding exon (LABOUESSE et al. 1994 ). Because lin-26B and lin-26C use different reading frames in that last exon, lin-26(mc4) creates an amber stop codon in lin-26B (TGG to TAG, nts 29,663–29,665 in F18A1), whereas it transforms a methionine codon into an isoleucine codon in lin-26C (ATG to ATA), which is likely to be less severe. Northern blots using a lin-26 probe identified a major band of 2.1 kb (Figure 3A) that is not detected by lin-26A-specific exons (data not shown), suggesting that lin-26A is a minor isoform or possibly a transcription intermediate (see also below).

The intergenic regions separating the probable polyadenylation signal (AATAAA) of lir-2 and the beginning of lir-1A-C or that of lir-1 and the beginning of lin-26B are short (125 and 136 bp, respectively). In addition, the major lir-1 and lin-26 isoforms start with SL2 or its variants. These criteria indicate that lir-2, lir-1 and lin-26 have characteristics typical of genes organized as transcriptional operons in C. elegans (ZORIO et al. 1994 ; BLUMENTHAL and STEWARD 1997 ).

Synteny of the lir-2, lir-1, and lin-26 region in C. briggsae:
To gain further insights into the function of the genes lir-2, lir-1, and lin-26 and to identify potential sequences important to control their expression, we cloned their homologs from the nematode C. briggsae using low-stringency hybridization conditions. We discovered that the C. briggsae lir-2, lir-1, and lin-26 homologs are also tightly linked (Figure 2D). This led us to sequence a unique 27,242-nt fragment (which will be referred to as a3d1, see MATERIALS AND METHODS). Comparison of cDNA (see above), F18A1, and a3d1 sequences showed that a3d1 contains six complete putative genes corresponding to the homologs of lir-2, lir-1, lin-26 and of three genes located upstream of lir-2 in C. elegans; a3d1 might also contain part of a gene located downstream of lin-26. Each predicted gene in a3d1 was named after its homolog in C. elegans (as referred to in the ACeDB databank; EECKMAN and DURBIN 1995 ), and C.b. was added in front of the name (e.g., C.b.F18A1.6; unless there is no ambiguity, C. elegans genes and proteins will similarly be preceded by the letters C.e.).

We observed that the gene order and the intron/exon positions and sizes are strikingly similar in this region (Figure 2D and Figure 4). This is not the first reported example of synteny between C. briggsae and C. elegans (KUWABARA and SHAH 1994 ). In agreement with others (see KENNEDY et al. 1993 ; MADURO and PILGRIM 1996 ), we noticed that the C. briggsae genome is more compact, containing fewer introns (as in F18A1.6 and F18A1.5) and often shorter introns (as for C.b.F18A1.6 introns 4 and 6 and C.b.lir-2 introns 1 and 5; however, the converse is true for the C.b.lir-2 intron 4, see Figure 2D).

View larger version (31K): In this window In a new window Download PPT slide   Figure 4. Graphic representation of the conserved areas in the syntenic region surrounding lir-2, lir-1, and lin-26. (A–E) Graphs were obtained using the GCG programs Compare and Dotplot requesting a conservation of 21 nts in a window of 30 nts: horizontal sequence, a3d1 (A, nts 1–7,600; B, nts 7,600–12,500; C, nts 12,000–22,300; D, nts 22,000–25,000; E, nts 25,000–27,242); vertical sequence, F18A1 (A, nts 4,000–12,800; B, nts 12,500–18,000; C, nts 17,600–28,100; D, nts 28,000–31,000; E, nts 31,000–34,000). C. briggsae genes (see Figure 2) are shown under each graph and named according to their C. elegans homologs. Conserved areas generally correspond to coding regions. In addition, a few intergenic or intronic sequences (denoted CSE1–CSE7 in A, B and D), the 3'-UTR of lir-1 and lin-26 (arrows) and most of lir-1 intron 1 (denoted CISa–CISh) have been conserved. We have not specifically looked for a lin-26C-like isoform in C. briggsae L1 larvae; on the basis of the fact that the exon that is spliced out in C.e.lin-26C is conserved in position and sequence in C. briggsae, and that short intron sequences have been conserved on both sides of that exon (CSE6 and CSE7), we would expect that a lin-26C isoform can be made in C. briggsae. The beginning of F18A1.1, as predicted in ACeDB, appears to have been poorly conserved, because the first well-conserved C.e.F18A1.1 exon corresponds to exon 5 (downstream exons fall outside of the a3d1 sequence); part of exon 3 has also been conserved and could correspond to the first exon in C. briggsae; however, the putative ATG initiation codon in C. briggsae is, instead, a GTA codon in C. elegans. The isolation of specific cDNAs should solve this problem. (F) The graph, which further illustrates the sequence conservation of lir-1 intron 1, was obtained with the programs Bestfit and Plotsimilarity (taking a window of 100 nts for Plotsimilarity). Horizontal axis: distance from the beginning of the sequence taken for comparison (a3d1: nts 12,100–22,100; F18A1: nts 17,600–27,900). Vertical axis: similarity score calculated by Plotsimilarity; the broken line corresponds to the average similarity score. Notice that long stretches of intronic sequences have been better conserved than exons, which on average show >78% identity between both species (Table 1). Notice also that the beginning of lir-1 intron 1 (up to CISb) has been less well conserved, as materialized by the fact that Compare and Bestfit reveal different alignments of lower similarity scores; it is not clear which alignment is more significant. The small region CISc corresponds to a sequence necessary for normal lin-26 expression in the somatic gonad (DEN BOER et al. 1998 ).

As shown in Figure 3B, C.b.lir-2, C.b.lir-1, and C.b. lin-26 transcripts have approximate sizes of 1.8, 1.2, and 2.0 kb, respectively, in agreement with expected sizes and similar to sizes observed in C. elegans. Sequencing of RT-PCR products (see MATERIALS AND METHODS) confirmed the predicted intron/exon boundaries and showed that the trans- and alternative splicing patterns have been conserved, although the number of isoforms that can be produced in C. briggsae seems smaller. We specifically found (i) a single SL1 trans-spliced C.b.lir-2 equivalent to C.e.lir-2A; (ii) four potential C.b.lir-1 isoforms equivalent to C.e.lir-1A, C.e.lir-1C, C.e.lir-1D, and C.e.lir-1E that, as in C. elegans, originate from trans-splicing at the beginning of exon 1 with SL1 or at the beginning of exon 2 with one of the SL2 variants and from alternative splicing at the 3' end (Figure 2C); (iii) one SL2 trans-spliced C.b.lin-26 isoform equivalent to C.e.lin-26B. We did not find the C.b.lir-1 isoform corresponding to C.e.lir-1B (see also Figure 2 legend) nor the C.b.lin-26 isoform corresponding to C.e.lin-26A. Although we did not directly search for the C.e.lin-26C homolog, its generation in C. briggsae is possible given the precise conservation of intron/exon positions; the LIN-26C isoforms would be conserved, even within their unique terminal exon. Absence of these isoforms in C. briggsae coincides with poor sequence conservation in the lir-1/lin-26 intergenic region and suggests that these forms might not be essential for lir-1 and lin-26 function in C. elegans.

lir-2 is less conserved than lir-1 and lin-26:
The six predicted proteins encoded within a3d1 have been strongly conserved. Whereas C.b.LIR-2 is only 62% identical to C.e.LIR-2, all other predicted proteins show >77% identity with their C. elegans counterparts (Table 1 and Figure 4). This weaker conservation of LIR-2 extends to its zinc fingers, which have accumulated the most changes in comparison to LIR-1 and LIN-26 (Figure 1). We also estimated the evolutionary divergence using the parameters K_S and K_A, which count the number of substitutions per synonymous and nonsynonymous sites, respectively (LI 1993 ). A high K_S value (>1.5) indicates first that the silent substitution sites are saturated, second that the codon bias imposed by the relative abundance of the different tRNA cognates for a given amino acid was not a constraint, and, finally, that the corresponding gene is not strongly expressed (STENICO et al. 1994 ). On the basis of these parameters (Table 1), all the genes described here have quite a high K_S, although lin-26, lir-1 and F18A1.5 (which is homologous to the yeast replication factor-A protein 1, LUCHE et al. 1993 ) have a lower K_S than the others. As for the other criteria, lir-2 is different from the other genes in that it has a higher K_A. Thus, on the basis of three criteria, lir-2 appears less conserved.

Analysis of noncoding regions:
Noncoding regions have generally been less conserved during evolution than coding regions (WAY et al. 1991 ; HOOPER et al. 1992 ; KENNEDY et al. 1993 ; KRAUSE et al. 1994 ; LUKOWITZ et al. 1994 ). Conservation is often indicative of a functional requirement (WAY et al. 1991 ). Alignment of the a3d1 and F18A1 sequences detected seven small blocks of conservation within intergenic regions and introns (denoted CSE1-CSE7 in Figure 4). These blocks, which might control transcription or alternative splicing, range between 45 and 290 nts and show 70–80% identity. The 3' UTR of lin-26 is well conserved (70% similarity over 390 nts, Figure 4), indicating that it could be folded similarly in both species to control RNA stability and/or expression. The 3' UTR of lir-1, including a potential polyadenylation signal (AATAAA), is partially conserved (70% similarity over 135 nts). By contrast, the 3' UTR of lir-2 is not conserved.

The most remarkable homologies, however, reside in the first intron of lir-1, which is unusually long in both species (almost 9 kb). As shown in Figure 4C and Figure F, several long blocks (>500 nts) that are spread over the entire intron display up to 88% identity and are interspersed with smaller blocks of nonconserved sequences. This is very different from other noncoding areas in the a3d1 sequence (see the lin-26/F18A1.1 intergenic region in Figure 4) or from noncoding regions belonging to other genes that have been sequenced in nematodes (WAY et al. 1991 ; HOOPER et al. 1992 ; KENNEDY et al. 1993 ; KRAUSE et al. 1994 ; MADURO and PILGRIM 1996 ) and in other species (DEN DUNNEN et al. 1989 ; HOOPER et al. 1992 ; LUKOWITZ et al. 1994 ), where only small blocks of homologies exist in globally nonconserved regions. Hypotheses to account for this striking conservation will be examined in the DISCUSSION.

LIN-26B is the major functional isoform:
To assess the functional requirement for the C.e.LIN-26B and C.e.LIN-26A isoforms, we generated two constructs that should allow the expression of only one isoform. Compared to plasmid pML301, which can rescue the null mutation lin-26(mc15) (DEN BOER et al. 1998 ) and should produce both isoforms, the construct pML225 should allow only the expression of C.e.LIN-26B, whereas the construct pML224 should allow only the expression of C.e.LIN-26A (see MATERIALS AND METHODS). We observed that pML225 but not pML224 could rescue the lethality and sterility caused by lin-26(mc15) (data not shown). We conclude that the C.e.LIN-26B isoform is sufficient to rescue the null phenotype of lin-26, and that the C.e.LIN-26A isoform plays only a minor role, if any.

Further support in favor of the idea that LIN-26A is nonfunctional comes from the fact that a plasmid containing C.b.lin-26 with its promoter (pML303, note that there is no lin-26A-like isoform in C. briggsae, see above) rescued all phenotypes conferred by lin-26(mc15) (Table 2). Specifically, pML303 rescued the lethality, which assays lin-26 function in hypodermal cells; the amphid and phasmid defects, which assay lin-26 function in support cells; and partially the sterility conferred by lin-26(mc15), which assays lin-26 function in the somatic gonad (LABOUESSE et al. 1994 , LABOUESSE et al. 1996 ; DEN BOER et al. 1998 ). These results demonstrate that C.b.lin-26 is a functional homolog of C.e.lin-26.

lir-2, lir-1, and lin-26 expression patterns are overlapping:
To analyze the lir-2 and lir-1 expression patterns, we performed in situ hybridization and analyzed the expression of lacZ reporter constructs (Figure 5 and Figure 6).

View larger version (96K): In this window In a new window Download PPT slide   Figure 5. In situ hybridization of embryos with lir-2, lir-1 and lin-26 probes. Embryos were fixed and stained by in situ hybridization (A–L) using a lir-2 probe (A–D), a lir-1 probe (E–H) or a lin-26 probe (I–L). Animals are oriented anterior to the left and dorsal up. (A, E and I) Two-cell stage embryos. lir-2 and lir-1 transcripts were detected in both cells, lin-26 transcripts predominantly in P1. (B, F, and J) 50-cell stage embryos. lir-2 and lir-1 transcripts were detected in all cells, lin-26 transcripts only in the germline precursors (J). (C, G, and K) 350-cell stage embryos. lir-1 and lin-26 transcripts were detected posteriorly and dorsally in hypodermal cells; lir-2 transcripts were no longer detected at this stage. (D, H, and L). Comma-stage embryos. lir-1 and lin-26 continue to show similar expression patterns in support cells and in minor hypodermal cells of the head (arrow), hypodermal cells of the excretory system (triangle), as well as the tail and anus (open arrow). Bar, 10 µm.

View larger version (136K): In this window In a new window Download PPT slide   Figure 6. Expression of lir-2::lacZ, lir-1:lacZ, lin-26::lacZ reporter constructs and expression of the C.b.LIN-26 protein. Embryos or larvae were fixed and stained with (A, E–J, O) polyclonal antibodies against ß-galactosidase or (C, K, and L) C.e.LIN-26. (B and D) DAPI staining of the embryos shown in A and C, respectively. (M and N) Staining with the MH27 monoclonal antibody (mAb) of the embryos shown in K and L, respectively. Anterior is to the left. (A and B) A 50-cell stage lir-2::lacZ (pML121) transgenic embryo; notice that the expression pattern is similar to that obtained by in situ hybridization. (C and D) A 50-cell stage C. briggsae embryos; antibodies against C.e.LIN-26 recognize a protein in C. briggsae embryos that is expressed in all cells, as lir-2 is in C. elegans. (E) A 350-cell stage lir-1::lacZ (pML169) transgenic embryo, dorsal view; expression is detected in hypodermal cells. (F and G) Comma stage lir-1::lacZ (pML169) transgenic embryo, (F) lateral view or (G) central view; expression is detected in (F) hypodermal and support cells more weakly in most other cells, e.g. pharyngeal cells (open arrow and G) and intestinal cells (arrows). (H) A 350-cell stage lir-1D-E::lacZ (pML165) transgenic embryo, dorsal view; expression is detected in hypodermal cells. (I) A 350-cell stage lin-26::lacZ (pML20) transgenic embryo; expression is detected in hypodermal cells. (J) Comma stage lin-26::lacZ (pML20) transgenic embryo; expression is detected in hypodermal and support cells, as can be seen with antibodies against C.e.LIN-26. (K and M) Wild-type C. elegans comma-stage embryo. (L and N) Wild-type C. briggsae embryo; the pattern is identical to that observed in C. elegans embryos, both with LIN-26 antiserum and the MH27 mAb, which detects a conserved epitope located at the adherens junctions separating hypodermal cells in both species. (O) lir-1::lacZ (pML169) transgenic L4 larva, ventral view; expression is detected in most cells, except in the germline and the gut (n, neurons; ph, pharynx; m, body wall muscle; vc, ventral cord neuron; sg, somatic gonad; v, vulva). Note that because of mosaicism, few animals truly showed expression in all cells. Bar, 20 µm (A–N); 50 µm (O).

lir-2 transcripts were detected by in situ hybridization in all cells between the 2- and 200-cell stages, the signal being strongest at the 50-cell stage (Figure 5, A–D). Expression of a lir-2::lacZ construct was also detected ubiquitously between the 50- and 200-cell stages (Figure 6, A–B) and disappeared after the 350-cell stage. Presence of ß-galactosidase after the 200-cell stage probably reflects a difference in stability between the fusion protein and native lir-2 transcripts.

lir-1 transcripts were detected by in situ hybridization in all cells between the 2- and 50-cell stages (Figure 5E and Figure F), disappearing after that stage. They reappeared between the 200- and 350-cell stages, but they were too weak to allow a clear identification of cells. However, we presume that cells expressing lir-1 at the 200-cell stage are precursors of the major hypodermal cells (e.g., P, seam, and hyp4-7 cells), as these were clearly labeled at the 350-cell stage (Figure 5G). At the comma stage (Figure 5H), a signal was detected mostly in the anterior and ventral parts of the head and in the tail. Because this pattern resembles the pattern observed at the same stage for lin-26 (Figure 5L), we believe that it could correspond to head and tail support cells and to all minor hypodermal cells (hyp1-3, hyp8-11, and cells of the anus and excretory system). Expression of the lir-1::lacZ construct pML169, which should reflect the expression of all lir-1 isoforms, and of the lir-1::lacZ construct pML165, which should reflect only the expression of lir-1 isoforms starting after the second exon, were first very weakly detected at the 200-cell stage and then clearly detected at the 350-cell stage in major hypodermal cells (Figure 6E and Figure H) and at the early comma stage in those same hypodermal cells, as well as in support cells and in all minor hypodermal cells (Figure 6F and data not shown). Starting at the late comma stage, expression of the lir-1::lacZ construct pML169, but not that of pML165, was detected in all cells (i.e., in intestinal, pharyngeal, muscle, neuronal, hypodermal, and support cells, Figure 6G). Similarly, during larval development, expression of the lir-1::lacZ pML169 was detected in virtually all cells, including cells of the somatic gonad, but not in intestinal cells after the late L1 stage nor in germ cells (Figure 6O).

The expression pattern of lin-26 as determined by in situ hybridization and lacZ reporter constructs correlates perfectly with that reported for the LIN-26 protein (LABOUESSE et al. 1996 ). lin-26 transcripts were first weakly detected by in situ hybridization in the germline precursors (P₃, P₄ and Z2/Z3) until the 50-cell stage and occasionally in their sisters (Figure 5I and Figure J). After the 100-cell stage, lin-26 transcripts and the ß-galactosidase encoded by the lin-26::lacZ construct were detected in hypodermal precursors, then in differentiating hypodermal and support cells (Figure 5K and Figure L). The only difference between in situ hybridization and antibody staining is that lin-26 transcripts could not be detected in body hypodermal cells after the 350-cell stage nor in any cell after the comma stage, whereas LIN-26 protein can be detected until the end of embryogenesis in all hypodermal cells and in all support cells. One possible explanation is that transcription of lin-26 is prevalent in precursors and newly born cells and that lin-26 transcripts have a shorter half-life in quiescent cells than does LIN-26 protein.

To assess the expression pattern of C.b.lin-26, we reasoned that the strong conservation of LIN-26 proteins in both species meant that the antiserum raised against C.e.LIN-26 was likely to detect C.b.LIN-26. We found that the C.e.LIN-26 antiserum does detect a nuclear protein in C. briggsae embryos (larvae were not examined). Assuming that this staining is as specific as with C. elegans animals, C.b.LIN-26 was detected at very low levels in all or nearly all cells of embryos with <200 cells (Figure 6C and Figure D), whereas C.e.LIN-26 is detected only in the germline precursors until the 100-cell stage and only in hypodermal precursors between the 100- and 200-cell stages (LABOUESSE et al. 1996 ). However, from the lima bean stage onward, both C.e.LIN-26 and C.b.LIN-26 have indistinguishable expression patterns when visualized using the C.e.LIN-26 antiserum; i.e., they are expressed in support cells and in hypodermal cells [to identify hypodermal cells, we used the antiadherens junction antibody MH27 (FRANCIS and WATERSTON 1991 ), which we found to detect a conserved epitope in both species (Figure 6, K–N)]. Interestingly, the expression pattern detected for C.b.LIN-26 before the 200-cell stage corresponds to the pattern described above for C.e.lir-2. This observation might be linked to the fact that lir-2 genes in C. elegans and C. briggsae are diverging (see above), and it raises the possibility that C.b.lin-26 might be taking over the early function that is normally fulfilled by C.e.lir-2.

In summary, we have observed that lir-2 and lir-1A-C isoforms are expressed in all or most cells, whereas lir-1D-F and lin-26 are expressed in the nonneural ectoderm.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Multiple lir-2, lir-1, and lin-26 isoforms define a new zinc-finger motif:
This work reports the analysis of the structure and the expression of the C. elegans genes lir-2, lir-1, and lin-26, which span a 20-kbp region on chromosome II, as well as their conservation in the nematode C. briggsae. Several features characterize lir-2, lir-1, and lin-26. First, they encode proteins that define a new C2H2 motif related to that of TFIIIA zinc fingers. Their first C2H2 motif deviates from the canonical TFIIIA zinc nger by the presence of four or five additional amino acids between the second cysteine and the first histidine, and by the absence of a conserved phenylalanine close to the second cysteine. Second, each gene can generate multiple isoforms by alternative splicing and trans-splicing. Trans-splicing is a mechanism that involves the transfer of a 22-nt leader sequence from either of two classes of spliced leader RNAs (SL1 or one of the SL2 variants) to the 5' ends of mRNAs (BLUMENTHAL and STEWARD 1997 ). As a result, each gene can generate multiple protein isoforms. We found specifically that C.e.lir-2 can generate two isoforms (C.e.lir-2A-B) and C.b.lir-2, one; that C.e.lir-1 can generate at least six and possibly nine isoforms (C.e.lir-1A-F) and C.b.lir-1, four; that C.e.lin-26 can generate three isoforms (C.e.lin-26A-C) and C.b.lin-26, at least one.

Although C. elegans and C. briggsae appear to have diverged a long time ago (current estimates put the divergence at 20–50 mya; HESCHL and BAILLIE 1990 ; LEE et al. 1992 ; KENNEDY et al. 1993 ), the general organization of the genes lir-2, lir-1, and lin-26 and their transcription patterns remain conserved in both species. It is noteworthy that C. briggsae genes tend to be more compact (regarding the size and number of introns, as noted by several authors) and to generate fewer isoforms (this is the case for C.b.lir-2, C.b.lir-1, and C.b.lin-26).

Possible function of lir-2, lir-1, and lin-26 isoforms:
The presence of C2H2 motifs related to TFIIIA zinc fingers and the presence of potential transcription transactivation domains suggest that LIR-2, LIR-1, and LIN-26 could be transcription factors. Because the LIR-1A/LIR-1D and LIR-1C/LIR-1E specific exons correspond to the potential transactivation domains, these isoforms could have different activation specificity and/or strength. The fact that lir-2 and lir-1 are expressed in most cells raises the possibility that they could act as cofactors of other transcription factors. Determining the function of lir-1 and lir-2 will require the isolation of specific mutations. We would be surprised if lir-1 had no function because it has been strongly conserved in C. briggsae. By contrast, lir-2 has been less conserved, both at the nucleotide and amino acid levels. Either selective pressures are not high enough to maintain strict conservation, or its function has diverged between the two species.

As far as lin-26 is concerned, we show that C.e.lin-26B, but not C.e.lin-26A, can rescue the null phenotype of lin-26, and that C.e.lin-26C is much less abundant than C.e.lin-26B. Taken together, these data suggest that lin-26B is the major isoform, that lin-26A might be dispensable, and, on the basis of its relative abundance, that lin-26C is required only in a subset of cells. We are currently exploring the possibility that LIN-26C is required during the L1 stage in the Pn.p cells, which are hypodermal precursors. The rationale for this hypothesis is threefold: first, lin-26C is mainly detected at the L1 stage; second, the Pn.p cells are transformed into neuroblasts at the L1 stage by the mutation lin-26(n156); third, the Pn.p cells are the only cells to be transformed with a high penetrance into neuronal cells by any lin-26 mutation (FERGUSON et al. 1987 ; LABOUESSE et al. 1994 ). We have also shown that the strong LIN-26 sequence conservation in C. briggsae is matched by an equally strong conservation in expression pattern and function, because the C. briggsae lin-26 homolog is expressed in the nonneural ectoderm, as C.e.lin-26 is in C. elegans, and can rescue a lin-26 null mutation in C. elegans.

lir-2, lir-1, and lin-26 form two overlapping operons:
Although eukaryotic genes are usually transcribed individually, ~25% of C. elegans genes appears to be organized in transcriptional operons that resemble bacterial operons (ZORIO et al. 1994 ; BLUMENTHAL and STEWARD 1997 ). Three criteria characterize genes in C. elegans operons: (1) they are only 100–400 bp apart; (2) generation of the downstream RNA is achieved by coupling polyadenylation of the upstream RNA with trans-splicing at the 5' end of the downstream RNA; (3) upstream transcripts can get trans-spliced, but only to the spliced leader SL1, while downstream transcripts receive an SL2 or a close relative. Although the function and expression of genes organized in operons is known for only a very small number of operons, this form of organization is thought to assure coordinate regulation of genes whose products appear to be expressed in the same cells or to perform related functions (reviewed in BLUMENTHAL and STEWARD 1997 ).

The organization and expression of lir-2, lir-1, and lin-26 bear many features characteristic of C. elegans operons. First, the intergenic distances between lir-2 and lir-1 and between lir-1 and lin-26 are short: the first lir-1 exon is located at 125 bp from the putative lir-2 polyadenylation signal; the first lin-26 exon of lin-26B, the major isoform, is located at 136 bp from the putative lir-1 polyadenylation signal. Second, we found that lir-2 is trans-spliced by SL1, while lir-1 transcripts, starting at the first exon, are trans-spliced by members of the SL2 family. In addition, we found that lir-1 transcripts starting at the beginning of the second or the third exons are trans-spliced by SL1, while major lin-26 isoforms are trans-spliced by SL2. Third, as further discussed below, the expression domains of lir-2, lir-1, and lin-26 partially overlap. Finally, this organization has been conserved in C. briggsae.

Two features of the lir-2, lir-1, and lin-26 genes suggest that these genes actually form a complex of two overlapping operons: a lir-2/lir-1A-C operon followed by a lir-1D-F/lin-26B operon. Indeed, transcripts from the most upstream genes (lir-2 and lir-1D-F) get trans-spliced to SL1, and the intergenic distances between all three genes are short. Although we cannot formally exclude the possibility that these three genes form a single operon, we strongly favor the two-operon model mainly because lir-2 and lir-1A-C are nearly ubiquitously expressed, whereas lir-1D-F and lin-26 are expressed in the nonneuronal ectoderm (see below). A unique operon, whose transcription would be under the control of a unique promoter, would not easily account for these differences. One would have to postulate some specific and complex mechanism that prevents the accumulation of lir-1D-F and lin-26 transcripts in many cells. By contrast, the two different expression patterns are easily explained in the context of two operons transcribed from distinct promoters. Further support in favor of the two-operon model is provided by our characterization of the lin-26 promoter. We have shown that it is possible to fully rescue the lin-26 null phenotype using plasmids that lack the first half of what would be a single operon (lir-2 and part of lir-1 intron 1, LABOUESSE et al. 1994 ; DEN BOER et al. 1998 ). In addition, assays for enhancer elements using sequences located upstream of lin-26 have revealed multiple elements capable of driving expression of a GFP reporter in cells that normally express lin-26. However, these experiments did not reveal any element that would drive GFP expression in all cells (as would be expected for lir-2 and lir-1A-C; S. QUINTIN, A. DAVIES and M. LABOUESSE, unpublished results). Thus, it seems very unlikely that the promoter of the putative unique operon is located within the plasmids that can rescue lin-26. Collectively, these arguments lead us to propose the model schematized in Figure 7. To our knowledge, lir-2, lir-1, and lin-26 would be the first transcription factors known to be organized in operons (see BLUMENTHAL and STEWARD 1997 ).

View larger version (13K): In this window In a new window Download PPT slide   Figure 7. Model for the organization of the lir-2/lir-1/lin-26 complex. We propose that the genes lir-2, lir-1 and lin-26 form two overlapping operons, one that includes lir-2 and long lir-1 isoforms (operon 1) and another that includes short lir-1 isoforms and lin-26 (operon 2). As further discussed in the text, we believe that an alternative model postulating that lir-2, lir-1 and lin-26 form a single operon is much less likely. In this figure, the SL2 symbol refers to SL2 and its variants.

Strong sequence similarity of a 9-kb noncoding intron:
On the basis of in situ hybridization, we found that lir-2 and lir-1 are maternally expressed and present in all cells until the 50-cell stage. On the basis of the expression of lacZ reporter fusions, we conclude that lir-2 and lir-1 are expressed in nearly all cells, lir-2 until the 200-cell stage and lir-1 starting at the comma stage (except in the gut of larvae). It is important to note that ubiquitous lir-1::lacZ expression was observed only for a construct containing 5.6 kb upstream of the first lir-1 exon (pML169) but not for constructs starting within the first exon (pML165). Thus, we suggest that there is a promoter element, which could correspond to the conserved sequences called CSE4 and CSE5 in Figure 4B, that would drive expression of the lir-2/lir-1A-C operon in the hermaphrodite adult germline and subsequently in all cells. The fact that we failed to detect ubiquitous lir-1::lacZ expression in early embryos might reflect a difficulty of the rapidly dividing cells of the early embryo in transcribing and splicing out, in a short time range, the first lir-1 intron, which is very long by C. elegans standards (9 kb). Such time constraints have previously been suggested to account for the apparent absence of function during early Drosophila embryogenesis of Dwnt-4 and invected, which are homologs of wingless and engrailed, respectively, because of the presence of large introns (GRABA et al. 1995 ; GUSTAVSON et al. 1996 ). Conversely, the fact that lir-2 expression was not detected after the lima bean stage might reflect a specific downregulation and/or instability of lir-2 transcripts once differentiation starts.

On the basis of in situ hybridization and lacZ reporter fusions, we found that lir-1 and lin-26 are expressed at midembryogenesis in all hypodermal and support cells. Because the construct pML165 reflects the expression of lir-1D-F isoforms and was expressed in these cells, we suggest that common promoter elements drive the expression of the lir-1D-F/lin-26B operon in hypodermal and support cells. The most likely position of these promoter elements is within the first lir-1 intron. Interestingly, this intron has been very well conserved in size and sequence in C. briggsae. Genefinder predictions, careful examination, and C. elegans/C. briggsae sequence comparisons did not reveal any significant and conserved open reading frame within this 9-kb intron. Because most of this intron is required for normal lin-26 expression (DEN BOER et al. 1998 ; S. QUINTIN, A. DAVIES and M. LABOUESSE, unpublished results), we suggest that conserved sequences within the first lir-1 intron represent regulatory elements for lin-26 (which is located downstream) and perhaps also for the lir-2/lir-1A-C operon. This proposed requirement for a long promoter is reasonable given that lin-26 is expressed in several different cell types (hypodermal, support, somatic gonad, germline) originating from four blastomeres (AB, C, MS, and P₄) and starting at different periods in development. We know of at least one other example of striking sequence similarity in noncoding DNA, namely at the human and mouse / T-cell receptor loci (KOOP and HOOD 1994 ), which challenges the view that noncoding DNA constitutes so-called "junk DNA."

In conclusion, the lir-2/lir-1/lin-26 complex has revealed some unique features in terms of genomic organization. We believe that it constitutes two separate and overlapping operons that contain an unusually large intron. Our initial motivation for characterizing lir-2 and lir-1 was to determine if they would be involved in the same process as lin-26, by analogy, for instance, with the four proneural genes from the achaete-scute complex in Drosophila (SKEATH and CARROLL 1994 ). On the basis of lir-2 and lir-1 expression patterns, we would argue that it is unlikely or at least that the situation is more complex. Although some lir-1 isoforms could have a role related to that of lin-26 because they are expressed in the same cells, the function of other lir-1 isoforms and of lir-2 is probably unrelated in terms of biological process (i.e., differentiation of the nonneural ectoderm). A clear goal for the future will be to isolate lir-2 and lir-1 mutants and to determine whether this organization is required to achieve normal expression of lin-26, the most downstream gene in the complex.

	FOOTNOTES

¹ These authors have equally contributed to this work.
² Present address: Harvard Medical School, Department of Pathology, 200 Longwood Ave., Boston, MA 02115.

	ACKNOWLEDGMENTS

We are very grateful to David Baillie for the C. briggsae genomic library, to David Eisenman for the mutant lin-26(ga91) and to Chad Nusbaum for the gift of RNA samples prepared from staged populations. The C. briggsae strain was provided by the CGC, which is supported by the National Institutes of Health and the National Center for Research Resources. We thank Satis Sookhareea for expert technical assistance. We also thank Thomas Bürglin for sharing cDNA filters, Elisabeth Georges for useful comments about the manuscript, as well as Sandra Metz and Bernard Boulay for help in the preparation of figures. P.D. was supported by fellowships from the Association de Recherche contre le Cancer, and P.C. received support from the Ministère de l'Education Nationale and the Ligue Nationale de Recherche contre le Cancer. This work was funded by the CNRS, INSERM, Hôpital Universitaire de Strasbourg, Human Frontier Science Program Organization, Association pour la Recherche contre le Cancer, Groupement de Recherches et d'Etudes sur les Génomes, Action Concertée Coordonnée des Sciences du Vivant (thème no. 4) du Ministère de la Recherche et de l'Education Nationale.

Manuscript received December 2, 1998; Accepted for publication February 1, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BARSTEAD, R. J. and R. H. WATERSTON, 1989  The basal component of the nematode dense-body is vinculin. J. Biol. Chem. 264:10177-10185[Abstract/Free Full Text].

BLUMENTHAL, T., and K. STEWARD, 1997 RNA processing and gene structure, pp. 117–145 in C. elegans II, edited by D. L. RIDDLE, T. BLUMENTHAL, B. J. MEYER and J. R. PRIESS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

BRENNER, S., 1974  The genetics of Caenorhabditis elegans.. Genetics 77:71-94[Abstract/Free Full Text].

CHAMBON, P., 1996  A decade of molecular biology of retinoic acid receptors. FASEB J. 10:940-954[Abstract].

CHANAL, P. and M. LABOUESSE, 1997  A screen for genetic loci required for hypodermal cell and glial-like cell development during Caenorhabditis elegans embryogenesis. Genetics 146:207-226[Abstract].

DEN BOER, B., S. SOOKHAREEA, P. DUFOURCQ, and M. LABOUESSE, 1998  A tissue specific knock-out strategy reveals that lin-26 is required for the formation of the somatic gonad epithelium in C. elegans. Development 125:3213-3224[Abstract].

DEN DUNNEN, J. T., J. W. VAN NECK, R. P. M. CREMERS, N. H. LUBSEN, and J. G. SCHOENMAKERS, 1989  Nucleotide sequence of the rat –crystallin gene region and comparison with an orthologous human region. Gene 78:201-213[Medline].

DEVEREUX, J., P. HAEBERLI, and O. SMITHIES, 1984  A comprehensive set of sequence analysis programs for VAX. Nucleic Acids Res. 12:387-395.

EECKMAN, F. H., and R. DURBIN, 1995 ACeDB and Macace, pp. 583–605 in Methods in Cell Biology, Vol. 48, edited by H. F. EPSTEIN and D. C. SHAKES. Academic Press, San Diego.

EVANS, D., D. ZORIO, M. MACMORRIS, C. E. WINTER, and K. LEA et al., 1997  Operons and SL2 trans-splicing exist in nematodes outside the genus Caenorhabditis.. Proc. Natl. Acad. Sci. USA 94:9751-9756[Abstract/Free Full Text].

FERGUSON, E. L., P. W. STERNBERG, and H. R. HORVITZ, 1987  A genetic pathway for the specification of the vulval cell lineages of Caenorhabditis elegans.. Nature 326:259-267[Medline].

FILES, J. G., S. CARR, and D. HIRSH, 1983  Actin gene family of Caenorhabditis elegans.. J. Mol. Biol. 164:355-375[Medline].

FIRE, A., S. W. HARRISON, and D. DIXON, 1990  A modular set of lacZ fusion vectors for studying gene expression in Caenorhabditis elegans.. Gene 93:189-198[Medline].

FRANCIS, R. and R. H. WATERSTON, 1991  Muscle cell attachment in Caenorhabditis elegans.. J. Cell Biol. 114:465-479[Abstract/Free Full Text].

GRABA, Y., K. GIESELER, D. ARAGNOL, P. LAURENTI, and M. C. MARIOL et al., 1995  DWnt-4, a novel Drosophila Wnt gene acts downstream of homeotic complex genes in the visceral mesoderm. Development 121:209-218[Abstract].

GUSTAVSON, E., A. S. GOLDSBOROUGH, Z. ALI, and T. B. KORNBERG, 1996  The Drosophila engrailed and invected genes: partners in regulation, expression and function. Genetics 142:893-906[Abstract].

HESCHL, M. F. P. and D. L. BAILLIE, 1990  Functional elements and domains inferred from sequence comparisons of a heat-shock gene in two nematodes. J. Mol. Evol. 31:3-9[Medline].

HOOPER, J. E., M. PEREZ-ALONSO, J. R. BERMINGHAM, M. PROUT, and B. A. ROCKLEIN et al., 1992  Comparative studies of Drosophila Antennapedia genes. Genetics 132:453-469[Abstract].

KENNEDY, B. P., E. J. AAMODT, F. L. ALLEN, M. A. CHUNG, and M. F. HESCHL et al., 1993  The gut esterase gene (ges-1) from the nematodes Caenorhabditis elegans and Caenorhabditis briggsae.. J. Mol. Biol. 229:890-908[Medline].

KOOP, B. F. and L. HOOD, 1994  Striking sequence similarity over almost 100 kilobases of human and mouse T-cell receptor DNA. Nature Genet. 7:48-53[Medline].

KRAUSE, M., 1995 Techniques for analyzing transcription and translation, pp. 513–529 in Methods in Cell Biology, Vol. 48, edited by H. F. Epstein and D. C. Shakes. Academic Press, San Diego.

KRAUSE, M. and D. HIRSH, 1987  A trans-spliced leader sequence on actin mRNA in C. elegans.. Cell 49:753-761[Medline].

KRAUSE, M., S. W. HARRISON, S. Q. XU, L. CHEN, and A. FIRE, 1994  Elements regulating cell- and stage-specific expression of the C. elegans MyoD family homolog hlh-1.. Dev. Biol. 166:133-148[Medline].

KUWABARA, P. E. and S. SHAH, 1994  Cloning by synteny: identifying C. briggsae homologues of C. elegans genes. Nucleic Acids Res. 21:4414-4418[Free Full Text].

LABOUESSE, M., S. SOOKHAREEA, and H. R. HORVITZ, 1994  The Caenorhabditis elegans gene lin-26 is required to specify the fates of hypodermal cells and encodes a presumptive zinc-finger transcription factor. Development 120:2359-2368[Abstract/Free Full Text].

LABOUESSE, M., E. HARTWIEG, and H. R. HORVITZ, 1996  The Caenorhabditis elegans LIN-26 protein is required to specify and/or maintain all non-neuronal ectodermal cell fates. Development 122:2579-2588[Abstract].

LEE, Y. H., X. Y. HUANG, D. HIRSH, G. E. FOX, and R. M. HECHT, 1992  Conservation of gene organization and trans-splicing in the glyceraldehyde-3-phosphate dehydrogenase-encoding genes of Caenorhabditis briggsae.. Gene 121:227-235[Medline].

LI, W. H., 1993  Unbiased estimation of the rates of synonymous and nonsynonymous substitution. J. Mol. Evol. 36:96-99[Medline].

LUCHE, R. M., W. C. SMART, T. MARION, M. TILLMAN, and R. A. SUMRADA et al., 1993  Saccharomyces cerevisiae BUF protein binds to sequences participating in DNA replication in addition to those mediating transcriptional repression (URS1) and activation. Mol. Cell Biol. 13:5749-5761[Abstract/Free Full Text].

LUKOWITZ, W., C. SCHRODER, G. GLASER, M. HULSKAMP, and D. TAUTZ, 1994  Regulatory and coding regions of the segmentation gene hunchback are functionally conserved between Drosophila virilis and Drosophila melanogaster.. Mech. Dev. 45:105-115[Medline].

MA, J. and M. PTASHNE, 1987  Deletion analysis of GAL4 defines two transcriptional activating segments. Cell 48:847-853[Medline].

MADURO, M. and D. PILGRIM, 1996  Conservation of function and expression of unc-119 from two Caenorhabditis species despite divergence of non-coding DNA. Gene 183:77-85[Medline].

MELLO, C. C., J. M. KRAMER, D. STINCHCOMB, and V. AMBROS, 1991  Efficient gene transfer in C. elegans: extrachromosomal maintenance and integration of transforming sequences. EMBO J. 10:3959-3970[Medline].

NILSEN, T.W., 1993  Trans-splicing of nematode premessenger RNA. Annu. Rev. Microbiol. 47:413-440[Medline].

RHODES, D. and A. KLUG, 1986  An underlying repeat in some transcriptional control sequences corresponding to half a double helical turn of DNA. Cell 46:123-132[Medline].

ROSS, L. H., J. H. FREEDMAN, and C. S. RUBIN, 1995  Structure and expression of novel spliced leader RNA genes in Caenorhabditis elegans.. J. Biol. Chem. 270:22066-22075[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SEYDOUX, G. and A. FIRE, 1994  Soma-germline asymmetry in the distributions of embryonic RNAs in Caenorhabditis elegans.. Development 120:2823-2834[Abstract].

SKEATH, J. B. and S. B. CARROLL, 1994  The achaete-scute complex: generation of cellular pattern and fate within the Drosophila nervous system. FASEB J. 8:714-721[Abstract].

STENICO, M., A. T. LLOYD, and P. M. SHARP, 1994  Codon usage in Caenorhabditis elegans: delineation of translational selection and mutational biases. Nucleic Acids Res. 22:2437-2446[Abstract/Free Full Text].

THOMPSON, J. D., T. J. GIBSON, F. PLEWNIAK, F. JEANMOUGIN, and D. G. HIGGINS, 1997  The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tool. Nucleic Acids Res. 25:4876-4882[Abstract/Free Full Text].

WATERSTON, R., C. MARTIN, M. CRAXTON, C. HUYNH, and A. COULSON et al., 1992  A survey of expressed genes in Caenorhabditis elegans.. Nature Genet. 1:114-123[Medline].

WAY, J. C., L. WANG, J. Q. RUN, and A. WANG, 1991  The mec-3 gene contains cis-acting elements mediating positive and negative regulation in cells produced by asymmetric cell division in Caenorhabditis elegans.. Genes Dev. 5:2199-2211[Abstract/Free Full Text].

WILSON, R., R. AINSCOUGH, K. ANDERSON, C. BAYNES, and M. BERKS et al., 1994  2.2 Mb of contiguous nt sequence from chromosome III of C. elegans.. Nature 368:32-38[Medline].

ZORIO, D. A., N. N. CHENG, T. BLUMENTHAL, and J. SPIETH, 1994  Operons as a common form of chromosomal organization in C. elegans.. Nature 372:270-272[Medline].

This article has been cited by other articles:

				D. M. Eisenmann and S. K. Kim Protruding Vulva Mutants Identify Novel Loci and Wnt Signaling Factors That Function During Caenorhabditis elegans Vulva Development Genetics, November 1, 2000; 156(3): 1097 - 1116. [Abstract] [Full Text]

				H. M. Robertson The Large srh Family of Chemoreceptor Genes in Caenorhabditis Nematodes Reveals Processes of Genome Evolution Involving Large Duplications and Deletions and Intron Gains and Losses Genome Res., February 1, 2000; 10(2): 192 - 203. [Abstract] [Full Text]

				J. M. Bosher, P. Dufourcq, S. Sookhareea, and M. Labouesse RNA Interference Can Target Pre-mRNA: Consequences for Gene Expression in a Caenorhabditis elegans Operon Genetics, November 1, 1999; 153(3): 1245 - 1256. [Abstract] [Full Text]