Gap Junctional Proteins of Animals: The Innexin/Pannexin Superfamily

Summary

There has been some controversy as to whether vertebrate pannexins are related to invertebrate innexins. Using statistical, topological and conserved sequence motif analyses, we establish that these proteins belong to a single superfamily. We also demonstrate the occurrence of large homologues with C-terminal proline-rich domains that may have arisen by gene fusion events. Phylogenetic analyses reveal the orthologous and paralogous relationships of these homologues to each other. We show that different sets of orthologous paralogues underwent sequence divergence at markedly different rates, suggesting differential pressures through evolutionary time promoting or restricting sequence divergence. We further show that the first 2 TMS-containing halves of these homologues underwent sequence divergence more slowly than the second 2 TMS-containing halves and analyze these differences. These bioinformatic analyses should serve as useful guides for future studies of structure, function and evolutionary aspects of this important superfamily.

1. Introduction

In 2003, we published sequence, topological and phylogenetic analyses of junctional proteins from animals (Hua et al., 2003). We described four previously recognized families as follows: (1) two gap junction protein families, the connexins (found only in chordates) (Alexopoulos et al., 2004; Evans and Martin, 2002) and the innexins (found primarily in invertebrates but also in vertebrates and other chordates) (Bauer et al., 2005; Panchin, 2005; Phelan, 2005; Sasakura et al., 2003), as well as (2) two tight junctional constituents, the claudins (Van Itallie and Anderson, 2006) and the occludins (Feldman et al., 2005). The members of all four families have similar topologies with four α-helical transmembrane segments (TMSs), and all exhibit well-conserved extracytoplasmic cysteines that either are known to or potentially can form disulfide bridges.

A multiple alignment of the sequences of each family was used to derive average hydropathy and similarity plots as well as phylogenetic trees. The data generated led to the following evolutionary, structural and functional suggestions: (1) In all four families, the most conserved regions of the proteins are the four TMSs although the extracytoplasmic loops between TMSs 1 and 2, and TMSs 3 and 4 are usually well conserved. (2) The phylogenetic trees revealed sets of orthologues except for the innexins where phylogeny primarily reflects organismal source, possibly due to a lack of relevant invertebrate genome sequence data. (3) The two halves of the connexins exhibit similarities suggesting that they were derived from a common origin by an intragenic duplication event. (4) Conserved cysteyl residues in the connexins and innexins pointed to a similar extracellular structure involved in the docking of hemi-channels to create intercellular communication channels. (5) We suggested a similar role in homomeric interactions for conserved extracellular residues in the claudins and occludins.

The lack of sequence or motif similarity between the four different families indicated that if they did evolve from common ancestral genes, they had diverged considerably to fulfill separate, novel functions. We suggested that internal duplication was a general evolutionary strategy used to generate new families of channels and junctions with unique functions (see Saier, 2003a,b).

In this report, we statistically analyze sequence comparisons and show first, that the innexins and pannexins are members of a single superfamily as originally suggested by Panchin et al. (2000), but disputed by Bruzzone et al. (2003) and Phelan (2005), second, that the innexins and pannexins exhibit common conserved motifs as well as common topologies, and third, that one group of pannexins, the pannexin 2 orthologues, possess large, unique, C-terminal, proline-rich, hydrophilic extensions that may have arisen by a gene fusion event. Our average hydropathy and average similarity plots reveal the probable topological properties of the members of this superfamily, and a phylogenetic tree shows the relatedness of representative sets of pannexins and innexins. We identify four conserved residues plus three well-conserved motifs common to all members of this superfamily. We use the term “superfamily” to describe the homologues of innexins because of the large number of constituent members, and because of their extensive sequence divergence.

2. Innexins and Pannexins are Members of a Single Superfamily

The proteins included in this study are presented in Table 1. As can be seen, they derive from a variety of invertebrates and vertebrates. The protein sizes vary considerably, but most full-length homologues are of about 400 (300–500) residues. Some are smaller, but most of these are believed to be fragments or putative products of pseudogenes (see Table 1). Others are larger and possess extra hydrophilic domains (as for the pannexins 2). These last mentioned domains are not represented in the CDD or Pfam database, and their function seems to be unknown.

Table 1.

Members of the Innexin/Pannexin superfamily included in this study¹

Abbreviation	Organism	Protein Size	GI #
Pannexins
Bta1	Bos taurus	425	76674501
Bta2	Bos taurus	639	76617339
Bta3	Bos taurus	392	76657274
Cfa1	Canis familiaris	426	73987804
Cfa2	Canis familiaris	582	73968865
Cfa3	Canis familiaris	392	73954928
Dre1	Danio rerio	417	41055476
Dre2	Danio rerio	472	94733329
Dre3	Danio rerio	397	68367600
Hsa1	Homo sapiens	426	39995064
Hsa2	Homo sapiens	633	16418341
Hsa3	Homo sapiens	392	16418453
Mmu1	Mus musculus	426	86262134
Mmu2	Mus musculus	667	68522178
Mmu3	Mus musculus	392	86262155
Rno1	Rattus norvegicus	426	45476996
Rno2	Rattus norvegicus	664	45476997
Rno3	Rattus norvegicus	392	45476998
Tni1	Tetraodon nigroviridis	345	47211113
Tni2	Tetraodon nigroviridis	609	47225783
Tni3	Tetraodon nigroviridis	343	47223826
Innexins
Annelida
Hme1	Hirudo medicinalis	414	25264684
Hme2	Hirudo medicinalis	398	25264687
Hme3	Hirudo medicinalis	479	77997501
Hme4	Hirudo medicinalis	421	77997503
Hme5	Hirudo medicinalis	413	77997505
Hme6	Hirudo medicinalis	480	77997507
Hme7	Hirudo medicinalis (fragment)	141	77997509
Hme8	Hirudo medicinalis (fragment)	221	77997511
Hme9	Hirudo medicinalis	412	77997513
Hme10	Hirudo medicinalis (fragment)	272	77997515
Hme11	Hirudo medicinalis	420	77997517
Hme12	Hirudo medicinalis	381	77997519
Cnidaria
Hvu1	Hydra vulgaris	396	86769609
Insecta
Aga1	Anopheles gambiae str. PEST	393	55240888
Aga2	Anopheles gambiae str. PEST	441	55239046
Aga3	Anopheles gambiae str. PEST	376	55233868
Aga4	Anopheles gambiae str. PEST	361	55233857
Aga5	Anopheles gambiae str. PEST	356	55233856
Aga6	Anopheles gambiae str. PEST	386	116116396
Aga7	Anopheles gambiae str. PEST	373	21288531
Ame1	Apis mellifera (triplication)	765	66561454
Bmo1	Bombyx mori	337	40949811
Bmo2	Bombyx mori	359	45775782
Bmo3	Bombyx mori	371	54111988
Hvi2	Heliothis virescens	307	48926842
Dme1	Drosophila melanogaster	362	129075
Dme2	Drosophila melanogaster	367	10720056
Dme3	Drosophila melanogaster	395	10720057
Dme4	Drosophila melanogaster	367	11386891
Dme5	Drosophila melanogaster	419	41019525
Dme6	Drosophila melanogaster	481	12643925
Dme7	Drosophila melanogaster	438	10720055
Dme8	Drosophila melanogaster	372	12644213
Sam1	Schistocerca americana	361	10720059
Sam2	Schistocerca americana	359	10720060
Sfr1	Spodoptera frugiperda	359	37781375
Tci3	Toxoptera citricida	393	52630963
Nematoda
Cel2	Caenorhabditis elegans	419	21264468
Cel3	Caenorhabditis elegans	420	12643734
Cel4	Caenorhabditis elegans	554	74960058
Cel5	Caenorhabditis elegans	447	44889063
Cel6	Caenorhabditis elegans	389	12643866
Cel7	Caenorhabditis elegans	556	10720049
Cel8	Caenorhabditis elegans	382	21264467
Cel9	Caenorhabditis elegans	382	74966908
Cel10	Caenorhabditis elegans	559	10720050
Cel11	Caenorhabditis elegans	465	10720052
Cel12	Caenorhabditis elegans	408	67460980
Cel13	Caenorhabditis elegans	385	74958810
Cel14	Caenorhabditis elegans	434	12643626
Cel15	Caenorhabditis elegans	382	74959920
Cel16	Caenorhabditis elegans	372	74959921
Cel17	Caenorhabditis elegans	362	50400811
Cel18	Caenorhabditis elegans	412	75017319
Cel19	Caenorhabditis elegans	454	74959869
Cel20	Caenorhabditis elegans	483	74961824
Cel21	Caenorhabditis elegans	481	75022996
Cel22	Caenorhabditis elegans	508	75022997
Cel23	Caenorhabditis elegans	423	10719983
Cel24	Caenorhabditis elegans	522	418153
Cel25	Caenorhabditis elegans	386	10720325
Mollusca
Aca1	Aplysia californica	405	54398898
Aca2	Aplysia californica	416	54398902
Aca3	Aplysia californica	406	54398904
Aca4	Aplysia californica	413	54398900
Aca5	Aplysia californica	406	60550112
Aca6	Aplysia californica	424	60550114
Aca7	Aplysia californica	423	89118243
Aca8	Aplysia californica	373	60550118
Cli1	Clione limacina	426	8515128
Cli2	Clione limacina (fragment)	120	14210377
Cli3	Clione limacina (fragment)	104	14210379
Cva1	Chaetopterus variopedatus	399	15706257
Platyhelminthes
Dja1	Dugesia japonica (fragment)	236	86355173
Dja2	Dugesia japonica	466	86355153
Dja3	Dugesia japonica	483	86355155
Dja4	Dugesia japonica	445	86355157
Dja5	Dugesia japonica	399	86355159
Dja6	Dugesia japonica (fragment)	51	86355175
Dja7	Dugesia japonica	407	86355161
Dja8	Dugesia japonica	434	86355163
Dja9	Dugesia japonica	439	86355165
Dja10	Dugesia japonica	415	86355167
Dja11	Dugesia japonica	438	86355169
Dja13	Dugesia japonica	399	86355171

¹Select innexins and pannexins are included in this study. All proteins included were identified in the SwissProt database. The table presents the abbreviations of the proteins used in this study (column 1), their organismal sources (column 2), the protein sizes in numbers of amino acyl residues (column 3) and the GenBank Index numbers (GI #) (column 4).

Using the GAP program (Devereux et al., 1984) with default settings and 100 random shuffles, results obtained are recorded in Table 2. A sequence alignment upon which a representative result is based is shown in Figure 1. In Table 2, the three human pannexins are compared with invertebrate innexins, and the three zebra fish pannexins are similarly compared with invertebrate innexins. In all cases, excellent comparison scores were obtained. Thus, the percent identities were between 25 and 34%, the percent similarities ranged between 36 and 46%, and the comparison scores (expressed in standard deviations (S.D.)) ranged from 15 to 23.

Table 2.

Comparison of vertebrate pannexins with invertebrate innexins¹

Pannexin	Innexin	# Residues compared	% Identity	% Similarity	Comparison score (S.D.)
Hsa1	Dja1	81	33	45	15
Hsa2	Dja1	151	26	36	23
Hsa3	Hme2	255	25	36	17
Dre1	Hme5	88	34	46	21
Dre2	Hme3	106	29	42	16
Dre3	Cel11	227	28	40	20

¹The GAP program (Devereux et al., 1984) was run with default settings and 100 random shuffles. For two of these comparisons (Hsa3 with Hme2 and Dre3 with Cel11), most of the two sequences were compared. In other cases, the regions of greatest sequence similarity (including TMSs 1 and 2) were selected to optimize the comparison (see Saier, 1994). Hsa, Homo sapiens (humans); Dre, Danio rerio (zebrafish); Dja, Dugesia japonica (flatworm); Hme, Hirudo medicinalis (leech); Cel, Caenorhabditis elegans (roundworm).

Fig. 1.

Alignment of almost all of Hsa3 (pannexin 3 of man; top sequence) with almost all of Hme2 (innexin 2 of the leech; bottom sequence). The GAP program (Devereux et al., 1984) was used to generate the alignment which gave a comparison score of 17 S.D. (see Table 2). Numbers at the beginning and end of each line indicate residue numbers in the proteins. *, identity; :, close similarity, ●, more distant similarity as defined by the GAP program. Motifs 1 and 2 which align are boxed and shaded. Motif 3 which does not align in the two sequences is overlined (top sequence) or underlined (bottom sequence).

Our extensive studies on integral membrane proteins and protein domains have shown that when two proteins exhibit a comparison score of 9 S.D. or greater, they are related by common descent (Saier, 1994; 2003a,b). A value of 9 S.D. corresponds to a probability of 10⁻¹⁹ that these proteins could have exhibited this degree of sequence similarity purely by chance (Dayhoff et al., 1983). We have never come across an example where two homologues exhibiting a comparison score of 9 S.D. or better have been shown to have evolved independently. The comparison scores obtained when vertebrate pannexins are compared with invertebrate innexins are far in excess of what is required to establish homology. Earlier investigators failed to establish homology because they provided no equivalent statistical analyses.

In Figure 1, the actual alignment of a human pannexin with a leech innexin is shown. In our opinions, there is no question that the proteins compared are homologous. Employing the superfamily principle (Doolittle, 1986), we therefore consider it established that the proteins that have been designated as “pannexins” are in the innexin superfamily (see the Innexin Family, TC #1.A.25, in the Transporter Classification Database (TCDB), www.tcdb.org) (Busch and Saier, 2002; Saier et al., 2006). The next section identifies and describes the most conserved motifs in these proteins.

3. Three Conserved Motifs Shared by Innexins and Pannexins

A well-conserved motif in the innexins has been identified and suggested to be the signature sequence for this family (Barnes, 1994). This motif, YYQWV, can be found in the alignment shown in Figure 1 (Motif II, boxed). Although the YYQWV is not well conserved between all innexins and pannexins included in this study, this motif plus the flanking region (23 residues included) exhibit 39% identity and 61% similarity (Fig. 1), the highest observed for any extended region in these proteins.

Preceding this region are the two fully conserved cysteyl residues, present in all innexins and pannexins included in our study (see Fig. S1 on our website: http://biology.ucsd.edu/~msaier/supmat/InnPan). The sequence alignment is shown in Figure 1 (Motif I, boxed, 24 residues included). This conserved motif shows 37% identity and 58% similarity. Motif I is known to occur in the first extracellular loop separating TMSs 1 and 2 (Phelan, 2005; Phelan and Starich, 2001). Motifs I and II are the best-conserved regions of these proteins.

Innexins have two well-conserved pairs of cysteyl residues in extracellular loop 2, between TMSs 3 and 4, but only one of these pairs is fully conserved. The two fully conserved cysteyl residues of the innexins are misaligned in both Figure 1 and Figure S1 with those of the pannexins. However, the fully conserved pair of cysteines in the innexins are separated most frequently by 17 residues, and the fully conserved pair of cysteines in the pannexins are also separated by 17 residues (underlined or overlined motifs III in Fig. 1). When the consensus sequences of these two regions were optimally aligned, the alignment showed 20% identity and 45% similarity (9 out of 20 similarities plus identities; see Fig. 2). Although these two regions do not align in the binary or multiple alignment, we suggest that they share a common origin. Insertions or deletions introduced during evolutionary divergence of the innexins relative to the pannexins presumably account for this misalignment.

Fig. 2.

Alignment of the second fully conserved pair of cysteyl residues in the pannexins (Hsu3, Pan, top) with the fully conserved pair of cysteyl residues in the innexins (Hmo2, Inn, bottom). The alignment, taken from Figure 1, shows 20% identity and 45% similarity. Vertical lines represent identities while colons represent similarities.

4. A Lack of Significant Sequence Similarity Between Connexins and Innexins

All of the sequences of the innexins and pannexins included in this study were compared with 250 connexin sequences using the IC program (Zhai and Saier, 2002). In no case did a resultant comparison score exceed 5 S.D. Although this result does not exclude the possibility that the innexins and connexins arose from a common ancestor, it provides no evidence for this possibility. It also shows that pannexins are far more similar to innexins than either of these two protein types are to connexins.

5. Short and Large Innexin Sequences

Among the innexins are virally derived innexins (e.g., from the Hyposoter didymator virus (363 amino acyl residues, gi# 27475782), not reported in our earlier review (Hua et al., 2003)). Insect virus-encoded innexins have been noted previously (Phelan, 2005). Proteins annotated as pannexins now can be found in the NCBI database from, for example, Aplysia californica and Clione limacina. Moreover, half-sized protein sequences with only two putative TMSs are now available in the NCBI database (e.g., pannexin 2 and pannexin 3 from Clione limacina; 120 aas, gi# 14210377 and 109 aas; gi# 14210379, respectively). They may be due to incomplete sequencing, sequencing errors, errors in exon identification or to the presence of pseudogenes.

All pannexin-2s are substantially larger than the other members of the innexin superfamily. Thus, pannexins-2 are larger than pannexins-1 and -3 by 200-300 residues (see Table 1). This is due to a large hydrophilic extension at the C-termini of pannexin-2s. This C-terminal domain proved to be proline (and alanine) rich and is present only in the pannexins-2 within the innexin superfamily. However, it was also found in many proteins of dissimilar function from bacteria and eukaryotes. These include a C-terminal region of the large tegument-like protein, UL36 of Gallid herpesvirus-2 (3342 aas; gi# 10180741), a C-terminal region of the surface antigen, PHGST#5 of Trypanosoma cruzi (796 aas; gi# 2209012), an N-terminal region of transcription factor TFIID of Homo sapiens (1083 aas; gi# 2058326), and a C-terminal region of the alginate regulatory protein AlgR3 of Pseudomonas syringae (318 aas; gi# 28867376). In these proteins, repeat units of 20–24 residues could be identified in the regions of sequence similarity with the pannexin-2 hydrophilic domains.

6. Innexin/Pannexin Residue Identity

A multiple alignment of 97 representative invertebrate innexins and vertebrate pannexins was generated using the Clustal X program (Thompson et al., 1997) (see Fig. S1 on our website; http://biology.ucsd.edu/~msaier/supmat/InnPan). Examination of the multiple alignment for the proteins included in this study revealed 3 residues that are fully conserved (see Fig. S1 and Phelan, 2005). These are (1) two cysteyl residues at alignment positions 201 and 221 (see Fig. S1) and (2) a prolyl residue at alignment position 348. Additionally, a tryptophanyl residue at alignment position 352 is fully conserved in all homologues except the Hydra vulgaris protein, Hvu1. This is contrary to the work reported by Phelen (2005) who reported an alignment where the tryptophan residues in the innexins did not align with those in the pannexins. This difference presumably reflects the proteins selected for study as well as the different assumptions made in designing the two programs used to generate the multiple alignments. Several residue types are also well conserved in the same regions of the multiple alignment, providing further evidence for homology between innexins and pannexins (see motif analysis above).

7. Innexin/Pannexin Phylogeny

The phylogenetic tree for the innexins and pannexins is shown in Figure 3. This tree, based on the CLUSTAL X multiple alignment (see Fig. S1), was drawn using the TREEVIEW program (Zhai et al., 2002). As can be seen, the vertebrate pannexins cluster separately from the invertebrate homologues, and the three isoforms (pannexins 1, 2 and 3) fall into three well-defined subclusters. Interestingly, mammalian pannexins 3 have diverged in sequence from each other less than the mammalian pannexins 1 and 2 have diverged from each other. However, the non-mammalian pannexins do not show the same trend. Dre2 and Tni2 have diverged from each other the least; Dre1 and Tni1 have diverged the most, and Dre3 and Tni3 have apparently diverged from each other at an intermediate rate. Divergence between mammalian and non-mammalian pannexins appears to have occurred at still different relative rates: isoforms 3 the most, isoforms 2 the least, and isoforms 1 at intermediate rates. Thus, for the mammalian orthologues, the relative divergence rates were 1 > 2 > 3; for the non-mammalian representatives they were 1 > 3 > 2, and between these two groups, they were 3 > 1 > 2. This unexpected observation suggests that these proteins have been under differential evolutionary pressure since the time when these groups of organisms began to diverge from each other.

Fig. 3.

Phylogenetic tree of representative vertebrate pannexins and invertebrate innexins which cluster separately as labeled. The three pannexin isoforms (pannexins 1, 2 and 3) are so designated. The innexins cluster according to organismal source type. The pannexins and innexins are shaded separately.

The innexins in Figure 3 fall into groups that in general correlate with organismal type. Thus, mollusks, annelids, flatworms and roundworms appear in different parts of the tree, although it should be noted that four of the Aplysia californica paralogues (Aca5 and 6 as well as Aca7 and 8) branch from points distant from branching points for the other 4 paralogues. Further, seven of the C. elegans paralogues branch from points distant from the branch that includes the 19 other paralogues shown. The latter proteins fall into three diffuse subclusters (Fig. 3). Finally, the Hydra vulgaris innexin (Hvu1) is found on its own branch, distant from all other homologues. Altogether, there are 16 deeply rooted branches, each of which can be considered to represent a distinct family within the innexin/pannexin superfamily.

8. Hydropathy Profiles of Innexins versus Pannexins

The average hydropathy (top) and similarity (bottom) plots for the innexins (Fig. 4A) and pannexins (Fig. 4B) reveal strikingly similar profiles. These plots are based on the multiple alignments shown in Figures S2 and S3 on our website. Both show four dominant hydrophobic peaks that are well conserved. In both plots, TMSs 1 and 2 are better conserved than TMSs 3 and 4. This is particularly apparent with the pannexins. These peaks and the extracytoplasmic domains between TMSs 1 and 2 and TMSs 3 and 4 are well conserved as noted before for the entire family (Hua et al., 2003). It is also worthy of note that regions preceding the odd-numbered TMSs and following the even-numbered TMSs are fairly well conserved, suggesting important structural and functional roles for these cytoplasmic regions. The similarities of peaks 1 and 2 with peaks 3 and 4 suggest (as we have proposed; Hua et al., 2003) that these proteins arose by an internal gene duplication event from a 2 TMS precursor. However, we could not prove this suggestion to be true using our statistical tools. The results presented in Figure 4 provide further support that innexins and pannexins are related by common descent (Sobczak and Lolkema, 2005a,b).

Fig. 4.

Average hydropathy (dark line; top) and average similarity (thin line, bottom) for the Innexins (A) and the Pannexins (B). The AveHAS program (Zhai and Saier, 2001) was used to derive the plots. The CLUSTAL X-derived multiple alignments (Figs. S2 and S3) presented on our website (http://biology.ucsd.edu/~msaier/supmat/InnPan) were used to derive these plots.

9. Conclusions

In this report we have presented unequivocal statistical evidence showing that innexins and pannexins belong to a single superfamily and therefore arose from a single precursor peptide. Three fully conserved residues as well as a tryptophan residue conserved in all but one of the homologues were identified. Moreover, two well-conserved motifs were identified among all of these proteins providing further evidence for homology. One of these motifs contains two fully conserved extracellular cysteines (between TMSs 1 and 2). We also identified regions in the C-terminal halves of the innexins and the pannexins which contain two fully conserved cysteines and exhibit sequence similarities between the innexins and the pannexins that suggest homology. The fact that the N-terminal halves are better conserved than the C-terminal halves suggests that the former segments share an essential, universal function while the latter segments have diverged for more specialized functions. It is possible that these 4 TMS proteins arose from 2 TMS precursor peptides (Saier, 2003a,b), but at present, there is no direct evidence for this hypothesis. Future studies will be required to identify these proposed subfunctions.

We have identified apparent domain fusions in pannexin 2 proteins, suggesting an additional function for these hydrophilic domains. Such a function could be regulation, targeting or macromolecular interaction. This observation suggests that pannexin 2s have a subfunction that is unique to this set of proteins. Further studies should clarify this issue.

The results presented in this paper provide guides for future studies. Such studies will be required to establish the structural and functional significance of the many observations reported here. They may also lead to a more detailed understanding of the evolutionary origins of this important class of transcellular communication channels.