Structure of the IC and nmpC Outer Membrane Porin Protein Genes of Lambdoid Bacteriophage*

The IC gene of the lambdoid bacteriophage PA-2 and the nmpC gene located on a defective lambdoid prophage in the 12-min region of the Escherichia coli K12 chromosome have been sequenced. The porin proteins encoded by these two genes were almost identical, with only 4 of the 365 residues of the precursor forms of the proteins being different. The LC and NmpC proteins were strongly homologous to the OmpC, ompF, and PhoE proteins, with greater than 56% of the residues identical in each case. Sequencing of the region flank-ing the IC gene allowed precise positioning of this gene with respect to the rightward cos site of the phage and to sequences which are homologous between PA-2 and X. In wild-type strains of E. coli K12, the nmpC gene is not expressed and contains an IS5 insertion near the 3‘ end of the coding region. This insertion deletes 18 residues from the COOH terminus of NmpC protein and adds 8 residues from an open reading frame ex- tending into IS5 sequence. Expression of this form of the gene in an expression vector plasmid demonstrated that this altered protein is still capable of being trans-located to the outer membrane. Plasmid expression experiments using lc-nmpC hybrid genes show that it is the presence of the IS5 insertion which prevents expression of the porin in wild-type E. coli K12. In the nmpC mutant which expresses the protein, there has been a precise excision of the IS5 which regenerates a COOH terminus of NmpC protein which is identical to that of the LC protein. Blot hybridization detected no

the nmpC gene located on a defective lambdoid prophage in the 12-min region of the Escherichia coli K12 chromosome have been sequenced. The porin proteins encoded by these two genes were almost identical, with only 4 of the 365 residues of the precursor forms of the proteins being different. The LC and NmpC proteins were strongly homologous to the OmpC, ompF, and PhoE proteins, with greater than 56% of the residues identical in each case. Sequencing of the region flanking the IC gene allowed precise positioning of this gene with respect to the rightward cos site of the phage and to sequences which are homologous between PA-2 and X. In wild-type strains of E. coli K12, the nmpC gene is not expressed and contains an IS5 insertion near the 3' end of the coding region. This insertion deletes 18 residues from the COOH terminus of NmpC protein and adds 8 residues from an open reading frame extending into IS5 sequence. Expression of this form of the gene in an expression vector plasmid demonstrated that this altered protein is still capable of being translocated to the outer membrane. Plasmid expression experiments using lc-nmpC hybrid genes show that it is the presence of the IS5 insertion which prevents expression of the porin in wild-type E. coli K12. In the nmpC mutant which expresses the protein, there has been a precise excision of the IS5 which regenerates a COOH terminus of NmpC protein which is identical to that of the LC protein. Blot hybridization detected no mRNA transcripts from the wild-type nmpC gene, although transcripts were readily detected from the IC gene in PA-2 lysogens and from the nmpC mutant which has excised the IS5. This indicates that IS5 affects the production or stability of transcripts from the adjacent nmpC gene.

The nucleotide sequeme(s) reported in this paper has been submitted to the GenBank@/EMBL Data Bank with the accession number 502580.
determined by conditions of growth, with expression controlled primarily by medium osmolarity, temperature, and carbon source (3,4). Under conditions of phosphate starvation, E. coli K12 expresses a third porin, the PhoE protein, which appears better suited to allow the permeation of phosphorylated compounds (5).
A porin gene can also be found in the genomes of certain lambdoid bacteriophage, and this gene is expressed in the lysogenic state. We isolated a new lambdoid phage termed PA-2 from a porcine strain of E. coli and found that strains of E. coli K12 lysogenic for this phage produced a new porin which was distinct from the OmpC, OmpF, and PhoE proteins (6,7). When this protein is expressed, the expression of OmpC and OmpF proteins is reduced substantially (6,8). The phage porin was initialIy called "protein 2." It is now termed LC protein as we have adopted the mnemonic IC for the locus on the phage genome encoding the protein.
Phage PA-2 has a different immunity, host range, and site of chromosomal attachment than X (6,9). When heteroduplexes between X and PA-2 were examined, the two phage were found to be about 70% homologous. Regions of nonhomology were found in the regions corresponding to the immunity region, the int-att region, and the J-gene region which would account for the differences mentioned above (35). An additional region of nonhomology was found near the right cos site in the region of X which includes gene Q and the S and R lysis genes. Genetic mapping indicated that the IC locus of PA-2 lay in this region (lo), and the construction of lc-ompC hybrid genes which expressed hybrid porin proteins showed that this locus included the structural gene for the porin protein. ' These studies also showed that the gene was transcribed in the opposite direction from the late phage genes such as S and R.
Phage PA-2 uses the OmpC protein as its receptor (6), and the ability of this phage to cause expression in the lysogenic state of a new porin which strongly inhibits expression of OmpC protein is probably of value to the phage. When such a lysogen is induced, there is little OmpC protein present on the cell surface to neutralize progeny phage. Other lambdoid phage that use OmpC protein as receptor might also be expected to carry a porin gene similar to the k gene of PA-2.
This is indeed the case. Chang and co-workers (10)

12723
. -which is homologous to the IC region of PA-2, as well as DNA which is unique to this prophage and DNA which is homologous to X. It is this defective prophage which is the source of the DNA substitutions in the Xqin mutants (11). The nmpC gene is not expressed in wild-type strains of E, coli K12. The nmpC locus was originally identified in a single mutant strain ('3384) which expressed NmpC protein in the outer membrane (12). Studies of DNA from the nmpC region from both CS384 and its wild-type parent have shed some light on why the gene is not expressed. Both heteroduplex examination of DNA from Xqin phages derived from CS384 and its parent and Southern blot analysis of chromosomal DNA from CS384 and its parent have shown that, in the wildtype parental strain, there is an IS5 insertion near the 3' end of the nmpC coding region (10). This IS5 insertion is absent in DNA from strain CS384, suggesting that the IS5 insertion may be the reason the gene is not expressed.
In the present report, we describe the cloning and sequencing of the IC region of the phage PA-2 genome and the nmpC region of the E. coli K12 genome. We also show a comparison of the deduced amino acid sequence of the LC and NmpC proteins to the sequences of the OmpC, OmpF, and PhoE proteins.

RESULTS AND DISCUSSION
Sequencing of the IC and nmpC Regions- Fig. 1 is a summary restriction map of the right end of the PA-2 genome including the IC locus and the nmpC locus from the 12-min region of the genome including the IS5 insertion 3' to nmpC. Fig. 2 shows the sequencing strategy used for IC, and areas which were sequenced from both strands are shown by the arrows. The sequencing strategy used for nmpC was similar except that areas where there was good agreement with the corresponding IC sequence were sequenced from only one strand.
The complete sequences are given in the "Appendix." As shown in Fig. 1, the 2800-base pair segment of the PA-2 genome which was sequenced included three complete open reading frames plus a portion of the NH2-terminal end of the Rz gene. The IC coding sequence is slightly longer than 1 kilobase and is read from right to left, as shown in Fig. 1  between the lysis genes R and Rz of X (13) to the similar junction in phage PA-2 between Orf-2, which is not homologous to X sequence, and the PA-2 Rz gene, which is very strongly homologous to that of X. The homology between the A and PA-2 Rz genes begins at the second base of the initiation codon since the X Rz gene initiates with AUG, whereas the PA-2 Rz initiates with the less commonly used GUG. This is analogous to a situation seen in another pair of related phage. The A cistron of MS2 initiates with GUG, whereas the homologous R17 A cistron initiates with AUG (14). The fact that third-base substitutions are seen when the X and PA-2 * "Experimental Procedures" are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814.
Request Document No. 85 "4079, cite the authors, and include a check or money order for $1.60 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. Rz sequences are compared (two are seen in Fig. 3) indicates that this portion of Rz is, in fact, from PA-2 and that the junction is real and not an artifact created by recombination or gene conversion during the construction of the hybrid phage Xlc-1 (IO). The region of homology which begins at this junction extends to cosR.
The sizes of Orf-2 and XR proteins are nearly identical (165 and 157 residues, respectively), which suggests that Orf-2 may be a functional analog of XR protein; however, a comparison of the amino acid sequence of these proteins both visually and with the aid of a computer showed no significant homology. Thus, the function of Orf-2 remains unknown. We also cannot ascribe a function to Orf-1, which begins just prior to the 3' The top portion of the figure shows X sequence in the region encoding the COOH terminus of the R gene product and the NH, terminus of the Rz gene product. The lower portion shows the corresponding region from PA-2 and shows the COOH terminus of Orf-2. Asterisks indicate termination codons. The underlined sequence shows where PA-2 and X are homologous. This region of homology extends to cosR. end of the IC coding sequence.
In agreement with our earlier observations (lo), an IS5 element is inserted near the 3' end of the coding region of nmpC in DNA from wild-type E. coli. Cloning and sequencing of a chromosomal fragment from the wild-type strain CS180 has allowed us to define the location of the IS5 more precisely. Fig. 4 shows that the junction between IS5 and nmpC lies with the codon for threonine at residue 323. The open reading frame extending across the junction into IS5 adds 8 residues not found in LC and deletes 18 residues found at the COOHterminal end of LC protein, resulting in a mature protein of 332 residues as compared to 342 for LC protein.
The IS5 insertion near the 3' end of the nmpC coding region eliminates the NH2-terminal portion of Orf-1 as it is found in IC. However, an open reading frame extending out of IS5 is in-frame with Orf-1, leading to an IS5-nnpC hybrid Orf-1. The hybrid Orf-1 of nmpC is slightly shorter: 57 residues as compared to 69 residues for the Orf-1 of LC. Of these 57 residues, 34 at the COOH terminus are from nmpC and 23 are from IS5.
The IC gene has one feature which is different from the omPC, ompF, and phoE genes in that the coding region is not followed by a typical rho-independent termination signal (I, 15,16). Three large stem-loop structures can be formed 3' to the coding region, and these might function as pause sites for rho-dependent termination.
Sequence 5' to the coding regions is strongly conserved between IC and nmpC. In sequence extending 467 base pairs 5' from the initiation codons, we found only 19 single-base differences, and these were in regions which did not appear to contain any major structural features or consensus sequences. The expressions of both LC and NmpC protein are subject to catabolite repression (6,12). Sequences located 320 and 110 base pairs 5' to the initiation codon of IC are the only sequences in reasonable agreement with the consensus CRP3binding sequence found near other catabolite-repressible genes and operons (17).
The BanII-EcoRI fragment from the 5' end of the IC coding region (see Fig. 1) was end-labeled at the EcoRI site and used as primer to determine the IC transcriptional start site by primer extension. The 5' end of the transcript begins with the sequence 5'-GCAGUG-3' and is located 46 bases upstream from the LC initiation codon. A single labeled band was observed, indicating a single transcriptional start site. Thus, the 5' end of the IC mRNA begins 64 bases downstream from the nearest CRP-binding site.
Homology of the LC and NmpC Proteins to Other Porins- Fig. 5 shows a comparison of the precursor form of the LC protein to the precursor forms of the OmpF, OmpC, and PhoE proteins of E. coli K12. The comparison was done by computer using the FASTP program as described by Lipman and Pearson (18). A comparison against the 3309 protein sequences in the National Biomedical Research Foundation protein library yielded no proteins other than the above which shared significant homology. The LC protein was compared to the recently sequenced yeast mitochondrial outer membrane porin (19). Although a few short overlaps were found, these were not statistically significant enough to indicate a relationship between the two proteins (data not shown).
The results in Fig. 5 confirm and extend the observations 320 330

140
~+ o r J + +~+ S e e G G P + o~o~o~+ + G +~+ " +  The top line shows the LC porin, and the next three lines show OmpF, OmpC, and PhoE, respectively. Capital letters show sequence identity, dashes indicate deletions, and plus signs indicate substitution of a related amino acid (positive relatedness odds or a score of zero or greater on the mutatation data matrix of Barker and Dayhoff (20)). Zeros indicate substitutions of unreIated amino acids (negative relatedness odds). Lower-case letters indicate insertions of unrelated amino acids. Homologous regions (3 of the 4 residues identical at that position) are underlined. The numbers refer to LC residue numbers.

Phage Porin Genes
of Mizuno et al. (1) of strong homology between the amino acid sequences of the E. coli porins. The 342-amino acid LC protein and its 23-amino acid leader peptide showed 58.3% identity to the OmpF precursor, 58.9% identity to the OmpC precursor, and 56.5% identity to the PhoE precursor. A striking feature of the homology between these proteins is its patchy nature, interspersing regions which are very strongly conserved with regions which show little sequence conservation. There is a strongly conserved sequence at the NH, terminus (residues 1-18) and at the COOH terminus (residues 331-342) as well as regions within the protein, for example, residues 38-60 and 286-314, where sequence is strongly conserved. There are also short regions, for example, the regions around residues 120 and 165, where there appears to be little conservation of sequence. It should be noted that the IC coding region exhibits a very strongly biased codon usage, essentially identical to that described by Mizuno et al. (l), indicative of a very strongly translated protein.
As anticipated from previous chemical comparison of the proteins (7) and DNA hybridization results (lo), there was little difference either at the protein or DNA level between the IC and nmpC genes. Within the coding regions, the DNA sequence was more than 95% identical, with the majority of the changes being third-base substitutions. The leader peptide sequences of both proteins are identical. Within the 324 residues of NmpC protein prior to the site of the IS5 insertion, only 4 residues differed from those of LC protein. Three of these, substitution of glutamine for glycine a t residue 126, alanine for threonine a t residue 128, and valine for phenylalanine at residue 131 of the mature protein, fall within a region where sequence is poorly conserved between the other porins. The fourth substitution, lysine for asparagine at residue 302, E H l R v may account for the slight difference in tryptic-chymotryptic peptides noted during the previous chemical comparison of the LC and NmpC proteins.
As noted previously (10,12), there is a single nmpC mutant, strain CS384, which has lost the IS5 insertion and which expresses the NmpC protein in the outer membrane. We have cloned and sequenced (data not shown), a fragment from CS384 which includes the 3' end of the nmpCcodingsequence. The sequence of this fragment indicates that excision of the IS5 in CS384 was precise and that the deduced amino acid sequence from the site of the IS5 insertion to the COOH terminus of the expressed NmpC protein is identical to that of LC protein.
The homology between the various porins at the amino acid level is reflected in extensive homology between the porin genes at the DNA level. Even third-base changes are kept to a minimum by the strong codon bias of these genes. This raises an interesting question. In a cell which contains at least four homologous porin genes, what acts to prevent homologous recombination and subsequent homogenotization between these genes? The observation reported here that minor differences are observed in the sequences of the IC and nmpC genes even though phage PA-2 has been propagated on E. coli K12 for many years suggests that recombination and homogenotization between these genes has not occurred.
Expression of IC and nmpC Cloned into Multicopy Plasmid Vectors-In order to initiate the study of the regulation of the IC gene and to obtain more information about the way in which the IS5 insertion prevents expression of NmpC protein in wild-type E. coli K12, we examined the outer membrane proteins produced by strains carrying the various plasmid constructions which are summarized in Fig. 6. The tm promoter @tm) and the rrnB sequence which provides two terminators are indicated. Restriction site abbreviations are as for Fig. 1.
When outer membrane proteins from a strain carrying plasmid pLc6, which includes the EcoRV-BglII fragment of IC, were examined, the amount of LC protein produced was about the same as the amount produced by a single copy of the intact gene present in a PA-2 lysogen (Fig. 7, lanes B and   D). The amount of LC protein produced by strains carrying this construction is not strongly influenced by temperature or catabolite repression, which are both known to affect expression of LC protein in PA-2 lysogens (6,8). Strains carrying plasmid pLc4, which includes the EcoRV-HpaI fragment of IC, produced considerably more LC protein. When grown at 37 "C (Fig. 7, lane C), the amount of LC protein produced is more than twice that produced by a PA-2 lysogen. There is a reduction in the amount of OmpA protein produced, similar to that reported in strains expressing higher copy numbers of the ompC gene (3). This suggests that sequence between the BglII site and the HpaI site is necessary for full expression of IC. When plasmids which carried an additional 350 base pairs 5' to the HpaI site were examined, there was no additional production of LC protein over that seen with plasmid pLc4. These results are consistent with a regulatory role for one or both of the CRP-binding sites located upstream from the IC transcriptional start site. The BglII site lies within the CRPbinding site closest to the 5' end of IC mRNA; and thus, both CRP-binding sites are absent in pLC6 and present in pLC4.
None of these plasmids resulted in gross overproduction of LC protein. This is in contrast to what was observed with strains carrying plasmids containing the 2.6-kilobase Hind111 fragment including the ompC gene (36).4 When the ompC gene is introduced into E. coli K12 on a multicopy plasmid, there is a vast overproduction of OmpC protein which leads to a complete suppression of the other major proteins including other porins and the OmpA protein. The observation that LC protein is not similarly overproduced indicates that its expression may be self-regulated. This is in agreement with the observations of Fralick and Diedrich (8), who examined LC protein expression as a function of growth temperature in a strain carrying two copies of the IC region of PA-2. They found that, at low temperature, a t which expression of IC is not maximal, the diploid strain produced twice as much Lc protein as a haploid strain; whereas a t 40 "C, which results in maximal expression, the amount of protein produced by haploid and diploid strains was the same. It is quite reasonable that the IC gene should be self-regulated since, during the establishment of lysogeny, a cell may contain many copies of the phage and overexpression of LC protein may be deleterious. Because of the high copy number, LC protein expression from plasmid pLc4 is much greater than that from a PA-2 lysogen a t low temperature or under conditions of catabolite repression, so it was not possible to determine whether the cloned fragment contains all of the information necessary for regulated expression of LC protein.
The expression of NmpC protein in wild-type strains which contain the IS5 insertion is completely null. No protein resembling the NmpC protein is found in the outer membrane, even in strains deficient in the other porins. More significantly, Xqin phage which carry the nmpC locus do not yield porin+ plaques on a porin-deficient indicator strain (10). This is a very sensitive test which will detect porin activity at levels below that at which the protein can be detected in stained gels of outer membrane proteins.
Since the IS5 insertion has deleted residues equivalent to residues 331-342 at the COOH terminus of LC protein which seem to be strongly conserved among the other porins, one E. Click, G. McDonald, and C. Schnaitman, manuscript in preparation. explanation for the null phenotype of nmpC might be that the truncated protein is missing sequence necessary for transport into the outer membrane. In order to test this, we constructed plasmid pBM5.0, which places the nmpC coding region downstream from a strong, inducible tuc (trp-lac hybrid) promoter (Fig. 6). This plasmid was introduced into a derivative of strain JMlOl which carries a strong l a c p mutation to prevent expression in the absence of inducer. Fig. 8 shows a radioautograph of protein from cells which were induced briefly with isopropyl-P-D-thiogalactoside and then given a short pulse label with [35S]methionine. After a short chase, the cells were broken with a French press and separated by centrifugation into an outer membrane fraction and a supernatant which contained the soluble cytoplasmic and periplasmic proteins and most of the cytoplasmic membrane. In the absence of inducer, neither the outer membrane fraction nor the supernatant contained NmpC protein.

C -
After induction, a large amount of new labeled protein was found in the outer membrane fraction. When prepared for electrophoresis by a method which involves boiling briefly (5 min) in a solution containing sodium dodecyl sulfate and 2-mercaptoethanol, about half of the labeled protein migrated slightly faster than LC or NmpC protein, as expected since the protein is slightly truncated. The remainder of the protein migrated more slowly, a t a position on the gel characteristic of undenatured porin trimer (7). Prolonged boiling in sodium dodecyl sulfate/2-mercaptoethanol solution shifted some of the label from the trimer form to the monomeric form, but it was not possible to convert all of the labeled protein into the monomeric form.
These results indicate that the truncated NmpC protein made from the nmpC gene with the IS5 insertion contains the information necessary for export to the outer membrane. However, the protein in the outer membrane is at least partly in an altered form so that trimers become cross-linked or are otherwise modified so that they cannot readily be denatured to the monomeric form. We suggest that the abnormal property of the truncated NmpC protein is due to the loss of a sequence at the COOH terminus which is strongly conserved among the porins. An alternative which cannot be ruled out is that the 8 residues from IS5 and its junction which are at the COOH terminus prevent the truncated protein from assembling properly in the outer membrane. It should be noted (see Fig. 4) that the addition of these residues alters the charge of the COOH terminus with respect to that of the other porins.
Since the truncated form of NmpC protein can be expressed and exported to the outer membrane when the gene is pro- vided with a strong promoter, we are left with two other hypotheses to explain the null phenotype. First, since the truncated NmpC protein resulting from the IS5 insertion is abnormal and may be deleterious to the cell, a secondary mutation might have occurred in the promoter region of nmpC which completely eliminated expression of the gene. A second hypothesis is that the IS5 insertion may have effects other than altering the COOH-terminal protein sequence. For example, a strong promoter in IS5 may result in transcription into nmpC which interferes with transcription from the nmpC promoter or with the translation of nmpC mRNA. The first of these hypotheses seems unlikely since there is so little difference between the sequence of the 5' regions of IC and nmpC. To test these hypotheses further, we constructed the series of plasmids shown in Fig. 6 which carry hybrid inserts consisting of portions of both IC and nmpC. There was no expression of NmpC protein from plasmid pBM8.1 (Fig. 7,  lane E ) , which is entirely nmpC and which carries all of the coding region plus most of the IS5 insertion; nor was there expression from plasmid pBM8.2, which carries the promoter region and the 5' end of the coding region from IC fused to the 3' end of nmpC including the IS5 sequence (Fig. 7, lane  F ) . However, expression was observed from plasmids pBM1l.O and pBM13.0, which carry the promoter regions of nmpC fused to the coding region and 3' end of IC (Fig. 7, lanes   G and H). Expression from these plasmids was comparable to that from pLc6 and pLc4, which include comparable inserts which are entirely from IC.
These results strongly suggest that the null phenotype of nmpC in wild-type strains is solely a consequence of the IS5 insertion, and this insertion prevents expression of the gene rather than preventing translocation of the protein product.
In order to study this further, IC and nmpC transcripts were examined by blot hybridization as shown in Fig. 9. Two strand-specific probes were made by synthesizing "P-labeled DNA from M13 templates carrying nmpC restriction fragments. The labeled restriction fragments were cut out and gel-purified. Fig. 9A shows a blot probed with the internal AccI-EcoRI fragment of nmpC in which the labeled strand was complementary to nmpC mRNA. The blot was then stripped of probe and re-probed with the restriction fragment from the EcoRI site in IS5 to the EcoRI site in nmpC. The labeled strand was the opposite strand to the first probe and thus would detect transcripts originating a t a promoter within IS5 and extending into nmpC in the opposite direction from nmpC mRNA. The results are shown in Fig. 9B.
The first probe hybridized to two closely spaced bands (Fig.   9A, lune B)  indicate the two transcripts from the IC and nmpC genes. Lane A is RNA from CS180, which is the wild-type parent. Lane B is RNA from (3384, which is the nmpC(p+) derivative of CS180 which expresses NmpC protein. Lane C is RNA from CS457 (12), which is a AnmpC derivative of CS384 which is deleted for the entire nmpC locus. Lane D is RNA from CS1385, which is a PA-2 lysogen of an ompR' derivative of CS384. A shows a blot which has been probed with a strand-specific nmpC probe which hybridizes to nmpC and IC mRNA. B shows the same blot, which was stripped and re-probed with a probe specific to the opposite strand. The probes are described in the text. deleted for the nmpC locus. No transcripts were detected with the opposite strand probe (Fig. 9B), although the probe did hybridize to the molecular weight standard (lune M). Thus, there is no indication of an anti-sense transcript originating in IS5 as the explanation of the null phenotype of CS180 and other wild-type strains which have the IS5 insertion into nmpC. The fact that no transcript was detected in RNA from CS180 indicates that the presence of the IS5 insertion affects either the production or the stability of nmpC mRNA and the null phenotype is not the result of rapid protein turnover.
It is likely that the two transcripts from both IC and nmpC have identical 5' ends but different 3' ends. Only a single band was detected when the 5' end of IC mRNA was mapped by primer extension. As noted previously, there are three stem-loop structures 3' to IC which might serve as pause sites, and it is possible that two of these represent rho-dependent termination signals. If the transcripts terminated at the first base after the first and last stem-loop structures, the transcripts would be 1198 and 1244 bases, respectively. The two bands seen in the blots of IC and nmpC mRNA shown in Fig.   9 have sizes in the range of 1200-1400 bases.          CS384. m R156 d ( p 1 , and CS1385.

AACAGTAGACCICGGTATCTIGICCCAAGIAGIACICAGIAGIIGAATGGAAGCGGCIGICACIIAAGICGTCAIICGCGGCAGICTGGICITTICTCIAAAA T T G T C A T C I G G A G C C A T A G A A C A G~G I I C A I C A T G A G T C~T C A A C I I A C C I I C~C C G A C A G T~A A I T C A G C A G I A A~C~C C~T C A~A C C A~A A A A G A G A T I I I T U Q L Y L V P N U M
A(---)