Size Comparisons among Integral Membrane Transport Protein Homologues in Bacteria, Archaea, and Eucarya

Transcription

1 JOURNAL OF BACTERIOLOGY, Feb. 2001, p Vol. 183, No /01/$ DOI: /JB Copyright 2001, American Society for Microbiology. All Rights Reserved. Size Comparisons among Integral Membrane Transport Protein Homologues in Bacteria, Archaea, and Eucarya YONG JOON CHUNG, CHRISTEL KRUEGER, DAVID METZGAR, AND MILTON H. SAIER, JR.* Department of Biology, University of California at San Diego, La Jolla, California Received 7 July 2000/Accepted 3 November 2000 Integral membrane proteins from over 20 ubiquitous families of channels, secondary carriers, and primary active transporters were analyzed for average size differences between homologues from the three domains of life: Bacteria, Archaea, and Eucarya. The results showed that while eucaryotic homologues are consistently larger than their bacterial counterparts, archaeal homologues are significantly smaller. These size differences proved to be due primarily to variations in the sizes of hydrophilic domains localized to the N termini, the C termini, or specific loops between transmembrane -helical spanners, depending on the family. Within the Eucarya domain, plant homologues proved to be substantially smaller than their animal and fungal counterparts. By contrast, extracytoplasmic receptors of ABC-type uptake systems in Archaea proved to be larger on average than those of their bacterial homologues, while cytoplasmic enzymes from different organisms exhibited little or no significant size differences. These observations presumably reflect evolutionary pressure and molecular mechanisms that must have been operative since these groups of organisms diverged from each other. * Corresponding author. Mailing address: Department of Biology, University of California at San Diego, La Jolla, CA Phone: (858) Fax: (858) msaier@ucsd.edu. Permanent address: Department of Life Science, Jeonju University, Chonju, Korea. The three largest classes of transporters found in nature are channels, secondary carriers, and primary active transporters (8, 10). Channel proteins facilitate passive diffusion of their substrates across membranes through aqueous pores, while secondary carriers generally utilize electrochemical gradients of H,Na, and solutes to drive the active accumulation or efflux of their primary substrates, and primary active transporters couple transport to the expenditure of a primary source of energy such as ATP hydrolysis or electron flow (10, 16). While channel proteins frequently span the membrane only a few times and form oligomeric complexes, secondary carriers and primary active transporters span the membrane multiple times and usually function as monomers or dimers in the absence of accessory proteins (4). Higher complexes of primary and secondary active transporters can provide regulatory (7, 11) or targeting and stability functions (15). Recently we have classified transport proteins according to a functional and phylogenetic system called the transporter classification (TC) system (8 10). While many of the identified families of transport proteins are found in only one of the three domains of living organisms (Bacteria, Archaea, or Eucarya), others are ubiquitous, being found in all three domains. Our studies have led to the conclusion that these ubiquitous families are ancient families that existed prior to the divergence of Eucarya and Archaea from Bacteria and that little horizontal transfer of genetic material encoding transport proteins between these three domains of life has occurred at least during the past 2 to 3 billion years (8, 9). In this study we compared the sizes of homologues of the ubiquitous families in the three domains of living organisms. We showed that while the eucaryotic homologues are consistently larger than their bacterial counterparts, the archaeal homologues are almost always smaller. Moreover, within the Eucarya domain, plant homologues are consistently smaller than the fungal and animal homologues, which are of similar sizes. These observations apparently do not apply to extracellular receptors and cytoplasmic enzymes, which exhibit the reverse size tendencies or no significant differences. The size differences observed for secondary carriers of homologues from the three domains of life proved to be due primarily to variations in the sizes of specific hydrophilic domains within these proteins, and the locations of these size-variable domains appear to be characteristic of specific families. MATERIALS AND METHODS The PSI-BLAST database search method ( /psiblast.cgi) was used to identify homologous proteins. Multiple alignments were generated using the CLUSTAL X program (13), and hydropathy and putative transmembrane spanner (TMS) analyses were conducted using the TMPred program (2). Positions of size variation among homologues were identified using a combination of programs for multiple alignment (CLUSTAL X) and topological analysis (TMPred). To test for statistically significant differences in protein length, the data were analyzed using two-tailed Sign tests (17). RESULTS Size variation in integral membrane transport protein homologues in Bacteria, Archaea, and Eucarya. Table 1 presents the average sizes, in numbers of amino acyl residues, of the integral membrane protein homologues of 15 families of secondary carriers, 3 families of channel proteins, and 4 families of primary active transporters present in the archaeal, bacterial, and eucaryotic domains. The number of homologues examined is presented in parentheses. The average sizes of the archaeal and eucaryotic homologues relative to the average sizes of the bacterial proteins are also provided. All of the archaeal homologues available in the SwissProt, GenBank, and PIR databases at the time these studies were conducted were included in the analysis. When limited numbers of bacterial or eucaryotic homologues comparable to the number of archaeal 1012

2 VOL. 183, 2001 MEMBRANE PROTEINS OF BACTERIA, ARCHAEA, AND EUCARYA 1013 TABLE 1. Comparison of membrane transport homologue sizes between Archaea, Bacteria, and Eucarya Avg size Family TC no. a proteins were identified, all of these were also included. However, when the numbers of bacterial and/or eucaryotic homologues considerably exceeded the number of archaeal family members, several proteins from the former two groups were generally selected at random from various organisms. In some cases, many eucaryotic proteins were included so that proteins within specific Eucarya kingdoms (animals, plants, and fungi) could be compared (see below). Examination of the results presented in Table 1 reveals that of the 22 protein families studied, the average sizes of the eucaryotic homologues are always substantially greater than those of the procaryotic homologues. Moreover, with only three exceptions (the amino acid-polyamine-organocation [APC] and formate-nitrite transporter [FNT] families of secondary carriers and the SecY proteins of the type II protein secretion pathway family of primary active protein secretory systems), the average sizes of the archaeal homologues are always less than those of the bacterial homologues. All of the size difference values, obtained when the archaeal or eucaryotic homologues for the various families were compared with the bacterial homologues (Table 1), were averaged. The average archaeal protein size for all 22 families examined was 92% of that of the bacterial homologues, while the average eucaryotic protein size for all 20 families examined was 140% AAs b Archaea Bacteria Eucarya % Relative size c AAs b AAs b % Relative size c Carriers Sugar porter (major facilitator superfamily) 2.A (4) (6) 527 (18) 124 Amino acid-polyamine-organocation 2.A (6) (6) 602 (5) 131 Cation diffusion facilitator 2.A (4) (4) 491 (8) 165 Resistance-nodulation-division 2.A (3) (56) 1,296 (4) 130 SecDF 2.A (3) (13) d Ca 2 :cation antiporter 2.A (4) (7) 649 (24) 174 Inorganic phosphate transporter 2.A (6) (5) 581 (10) 122 Monovalent cation:proton antiporter-1 2.A (3) (5) 702 (15) 136 Monovalent cation:proton antiporter-2 2.A (5) (13) 793 (5) 162 K transporter 2.A (3) (5) 758 (8) 157 Nucleobase:cation symporter-2 2.A (3) (9) 566 (13) 125 Formate-nitrite transporter 2.A (2) (7) 547 (2) 200 Divalent anion:na symporter 2.A (4) (8) 681 (12) 132 Ammonium transporter 2.A (7) (7) 503 (12) 108 Multi-antimicrobial extrusion 2.A (5) (5) 636 (6) 138 Channels Major intrinsic protein 1.A (2) (11) 278 (33) 111 Chloride channel 1.A (5) (5) 827 (17) 180 Metal ion transporter 9.A (3) (19) 692 (9) 210 Primary active transporters P-type ATPase 3.A (9) (85) 1,096 (62) 150 Arsenite-antimonite efflux 3.A (7) (22) 693 (11) 169 Type II secretory pathway (SecY) 3.A (12) (44) 455 (26) 104 Na -transporting carboxylic acid decarboxylase ( ) 3.B (3) (9) d a For further information on TC numbers, see reference 10 or msaier/transport/. b The average number of amino acyl residues (AAs) per protein homologue is reported, with the number of proteins examined appearing in parentheses. c Average percent size of the archaeal or eukaryotic homologue relative to the bacterial homologue is presented for each of the families examined. When the values for all families were averaged, the archaeal proteins proved to be 8% smaller than their bacterial homologues while the eukaryotic homologues proved to be 40% larger. d No protein homologues were identified in this domain. of that of the bacterial homologues. Thus, while the archaeal proteins are 8% smaller than the bacterial proteins, on average, the eucaryotic proteins are 40% larger. Size variation in integral membrane transport protein homologues in fungi, plants, and animals. Within the Eucarya domain, animal, plant, and fungal (including yeast) homologues were analyzed separately (Table 2). In all but three of the families of transport proteins analyzed, the plant proteins exhibited average sizes that were substantially smaller than the animal or fungal homologues. The exceptions were the sugar porter family of the major facilitator superfamily, the ammonium transporter family, and the SecY family within type II protein secretion pathway systems. In the sugar porter family of the major facilitator super family, animal homologues proved to be slightly smaller on average than the plant homologues. All of the size difference values, obtained when the animal or plant homologues for the various families were compared with the fungal homologues (Table 2), were averaged. The average animal protein size for all 14 families examined was 105% of that of the fungal homologues, while the average plant protein size for the 13 families examined was 83% of that of the fungal homologues. Thus, while the animal proteins are

3 1014 CHUNG ET AL. J. BACTERIOL. TABLE 2. Comparison of membrane transport homologue sizes between animals, plants, and fungi Avg size Family TC no. 5% larger than the fungal proteins, on average, the plant proteins are 17% smaller. Size variation in homologous constituents of the ABC-type transport system in Bacteria and Archaea. The ABC superfamily of uptake permeases is restricted to procaryotes, but it is found in both Bacteria and Archaea. These systems include three constituents: extracytoplasmic receptors, integral membrane proteins, and cytoplasmic ATP-hydrolyzing constituents. Over 20 families of these systems have been identified (10). These types of homologues (receptors, integral membrane constituents, and cytoplasmic ATP-hydrolyzing energizers) were analyzed for size variation (see Tables 3, 4, and 5, respectively). As shown in Table 3, the average archaeal receptor sizes proved to be greater than those of the average bacterial receptor sizes for 11 of the 13 families that have homologues in both domains. Overall, the archaeal receptors are 7% larger, on average, than their bacterial homologues. By contrast, the integral membrane archaeal homologues of ABC systems are usually smaller than the bacterial homologues (Table 4). Thus, of the 20 families examined, 15 proved to have smaller archaeal homologues, on average, than bacterial homologues. The average size difference proved to be 3.5%. Finally, the cytoplasmic ATP-hydrolyzing energizers tend to be somewhat smaller in Archaea than in Bacteria (Table 5). Thus, of the 16 families analyzed, 13 were smaller and 3 were larger, on average. Overall, the archaeal cytoplasmic proteins were 3.5% smaller than their bacterial homologues. Thus, the trend displayed by the ABC membrane proteins (Table 4) agreed with that for other integral membrane transport proteins (Table 1). The size differences for the archaeal extracytoplasmic receptors were opposite to that observed for the integral membrane constituents, with the archaeal receptors being substantially AAs a Animals Plants Fungi % Relative size b AAs a % Relative size b Carriers Sugar porter (major facilitator superfamily) 2.A (6) (6) (6) Amino acid/auxin porter 2.A (13) (16) (9) Ca 2 :cation antiporter 2.A (14) (4) (6) Cation-chloride cotransporter 2.A.30 1,060 (14) (1) 90 1,085 (2) Monovalent cation:proton antiporter-1 2.A (10) (2) (3) Nucleobase:cation symporter-2 2.A (6) (5) (2) K transporter 2.A.43 c 506 (2) 50 1,010 (6) Divalent anion:na symporter 2.A (8) 65 c 891 (4) Ammonium transporter 2.A (4) (5) (3) Channels Major intrinsic protein 1.A (15) (22) (2) Chloride channel 1.A (11) (6) (2) Metal ion transporter 9.A (3) (4) (2) Primary active transporters P-type ATPase 3.A.3 1,348 (21) (29) (12) Arsenite-antimonite efflux 3.A (9) 82 c 809 (2) Type II secretory pathway (SecY) 3.A (6) (16) (3) a The average number of amino acyl residues (AAs) for the protein homologues of each family is reported, with the number of proteins examined in parentheses. Fungal proteins include those from yeast. b Average percent size of the animal or plant homologue relative to the fungal homologue is presented for each of the families examined. When the values for all families were averaged, the plant proteins proved to be 17% smaller than their fungal homologues while the animal homologues proved to be 5% larger. c No protein homologues were identified in this kingdom. AAs a larger than their bacterial homologues. The ATP-hydrolyzing energizers showed minimal size differences. Size variation in homologous cytoplasmic enzymes. Similar analyses were conducted with a variety of catabolic and anabolic cytoplasmic enzymes (Table 6). These proteins showed similar homologue sizes, regardless of the domain or kingdom analyzed. Averaging all of the statistically significant results in Table 6 revealed that, on average, eucaryotic enzymes are only 3% larger than the homologous bacterial enzymes and archaeal enzymes are only 3% smaller than the homologous bacterial enzymes. These average size differences are much less than for the integral membrane transport proteins analyzed (Tables 1 and 4). Moreover, among the Eucarya kingdoms, animal and fungal homologues are essentially the same size while plant homologues are only about 1% smaller on average. This last mentioned average size difference is not statistically significant. Thus, cytoplasmic enzymes do not appear to exhibit the appreciable size differences that were observed for integral membrane proteins. Statistical significance of the observed homologue size differences. To test for statistically significant differences in transport protein lengths between phylogenetic groups, the data were analyzed using two-tailed Sign tests (17). In these analyses, comparisons were made between paired domains of life, or between paired kingdoms within the Eucarya domain, in terms of average lengths of amino acyl sequences within the protein families (Tables 1 to 5). The Sign test is a qualitative, nonparametric paired-sample test that utilizes only the direction of difference ( or ) between paired data. As such, it requires no assumptions regarding the distribution of data either within or between sample groups. We felt that such assumptions might be unwarranted given the size variation observed within

4 VOL. 183, 2001 MEMBRANE PROTEINS OF BACTERIA, ARCHAEA, AND EUCARYA 1015 TABLE 3. Comparison of ABC receptor homologue sizes between Archaea and Bacteria Avg size Family TC no. Archaea Bacteria AAs a % Relative size b AAs a % Relative size b Carbohydrate uptake transporter-1 family 3.A (6) (13) 100 Carbohydrate uptake transporter-2 family 3.A.1.2 c 343 (12) 100 Polar amino acid uptake transporter family 3.A (2) (29) 100 Hydrophobic amino acid uptake transporter family 3.A (4) (14) 100 Peptide-opine-nickel uptake transporter family 3.A (11) (25) 100 Sulfate uptake transporter family 3.A.1.6 c 342 (6) 100 Phosphate uptake transporter family 3.A (6) (18) 100 Molybdate uptake transporter family 3.A (2) (15) 100 Phosphonate uptake transporter family 3.A.1.9 c 300 (5) 100 Ferric iron uptake transporter family 3.A.1.10 c 328 (15) 100 Polyamine-opine-phosphonate uptake transporter family 3.A (2) (20) 100 Quaternary amine uptake transporter family 3.A.1.12 c 421 (7) 100 Vitamin B12 uptake transporter family 3.A (4) (14) 100 Iron chelate uptake transporter family 3.A (2) (20) 100 Manganese-zinc-iron chelate uptake transporter family 3.A (4) (26) 100 Nitrate-nitrite-cyanate uptake transporter family 3.A (4) (10) 100 Taurine uptake transporter family 3.A.1.17 c 333 (11) 100 Putative cobalt uptake transporter family 3.A (4) (3) 100 Thiamine uptake transporter family 3.A (2) (12) 100 Brachyspira iron transporter family 3.A.1.20 c 346 (10) 100 a The average number of amino acyl residues (AAs) per protein homologue is reported, with the number of proteins examined appearing in parentheses. b Average percent size of the archaeal homologues relative to the bacterial homologues is presented for each of the families examined. c No protein homologue or just one such protein was identified in this domain. the taxonomic groups with respect to average length of the proteins within the families represented. Thus, there proved to be more variation between protein families within each domain than there was between domains in any particular protein family. Each pair of domains or kingdoms was therefore compared independently. A P value of 0.05 was considered significant. Tabulated data and associated Sign test P values are presented in Table 7. The results of the statistical analyses strongly support the conclusion that transport protein length differs significantly TABLE 4. Comparison of ABC membrane protein homologue sizes between Archaea and Bacteria Family TC no. Archaea Avg size Bacteria AAs a % Relative size b AAs a % Relative size b Carbohydrate uptake transporter-1 family (MalF) 3.A (4) (24) 100 Carbohydrate uptake transporter-1 family (MalG) 3.A (4) (24) 100 Carbohydrate uptake transporter-2 family (RbsC) 3.A (1) (22) 100 Carbohydrate uptake transporter-2 family (RbsD) 3.A.1.2 c 137 (7) 100 Polar amino acid uptake transporter family (HisM) 3.A (2) (50) 100 Polar amino acid uptake transporter family (HisQ) 3.A (2) (39) 100 Hydrophobic amino acid uptake transporter family (LivH) 3.A (5) (17) 100 Hydrophobic amino acid uptake transporter family (LivM) 3.A (8) (20) 100 Peptide-opine-nickel uptake transporter family (OppB) 3.A (12) (48) 100 Peptide-opine-nickel uptake transporter family (OppC) 3.A (9) (51) 100 Sulfate uptake transporter family 3.A.1.6 c 277 (10) 100 Phosphate uptake transporter family 3.A (7) (23) 100 Molybdate uptake transporter family 3.A (4) (13) 100 Phosphonate uptake transporter family 3.A.1.9 c 246 (5) 100 Ferric iron uptake transporter family 3.A.1.10 c 520 (8) 100 Polyamine-opine-phosphonate uptake transporter family (PotB) 3.A (3) (25) 100 Polyamine-opine-phosphonate uptake transporter family (PotC) 3.A (3) (18) 100 Quaternary amine uptake transporter family 3.A.1.12 c 435 (7) 100 Vitamin B 12 uptake transporter family 3.A (6) (11) 100 Iron chelate uptake transporter family (FecC) 3.A (6) (18) 100 Iron chelate uptake transporter family (FecD) 3.A (4) (22) 100 Manganese-zinc-iron chelate uptake transporter family 3.A (4) (38) 100 Nitrate-nitrite-cyanate uptake transporter family 3.A (5) (16) 100 Taurine uptake transporter family 3.A.1.17 c 328 (7) 100 Putative cobalt uptake transporter family 3.A (30) (11) 100 Thiamine uptake transporter family 3.A (4) (31) 100 Brachyspira iron transporter family 3.A.1.20 c 424 (7) 100 a The average number of amino acyl residues (AAs) per protein homologue is reported, with the number of proteins examined appearing in parentheses. b Average percent size of the archaeal homologues relative to the bacterial homologues is presented for each of the families examined. c No protein homologues were identified in this domain.

5 1016 CHUNG ET AL. J. BACTERIOL. TABLE 5. Comparison of cytoplasmic ABC protein homologue sizes between Archaea and Bacteria Avg size Family TC no. Archaea Bacteria AAs a % Relative size b AAs a % Relative size b Carbohydrate uptake transporter-1 family 3.A (14) (48) 100 Carbohydrate uptake transporter-2 family 3.A (5) (34) 100 Polar amino acid uptake transporter family 3.A (2) (31) 100 Hydrophobic amino acid uptake transporter family 3.A (6) (15) 100 Peptide-opine-nickel uptake transporter family 3.A (9) (25) 100 Sulfate uptake transporter family 3.A (1) (8) 100 Phosphate uptake transporter family 3.A (10) (21) 100 Molybdate uptake transporter family 3.A (1) (8) 100 Phosphonate uptake transporter family 3.A.1.9 c 287 (4) 100 Ferric iron uptake transporter family 3.A (1) (9) 100 Polyamine-opine-phosphonate uptake transporter family 3.A (5) (13) 100 Quaternary amine uptake transporter family 3.A (3) (12) 100 Vitamin B 12 uptake transporter family 3.A (3) (10) 100 Iron chelate uptake transporter family 3.A (2) (24) 100 Manganese-zinc-iron chelate uptake transporter family 3.A (3) (21) 100 Nitrate-nitrite-cyanate uptake transporter family 3.A (3) (18) 100 Taurine uptake transporter family 3.A.1.17 c 332 (11) 100 Putative cobalt uptake transporter family 3.A (9) (11) 100 Thiamine uptake transporter family 3.A.1.19 c 225 (3) 100 Brachyspira iron transporter family 3.A.1.20 c 371 (3) 100 a The average number of amino acyl residues (AAs) per protein homologue is reported, with the number of proteins examined appearing in parentheses. b Average percent size of the archaeal homologue relative to the bacterial homologue is presented for each of the families examined. c No protein homologues were identified in this domain. between domains in all pairwise comparisons. Archaea have significantly shorter transport proteins than Bacteria, and both Archaea and Bacteria have shorter transport proteins than Eucarya. Tests were generally less significant in pairwise comparisons of Eucarya kingdoms. Plants have shorter transport proteins than either animals or fungi, but the difference in protein length between animals and fungi is not significant. Corresponding analyses of the ABC receptors, membrane proteins, and cytoplasmic energizers argued for statistical significance, although the actual size differences between the archaeal and bacterial energizers proved to be minimal. Localization of regions in homologues of secondary transporters responsible for size differences between Bacteria, Archaea, and Eucarya. Five families of secondary carriers were analyzed in detail to determine what portions of these proteins exhibit the greatest size variation. For this purpose, five sequence-divergent members of each family from each of the three domains of living organisms were selected for analysis. These sequences were multiply aligned using the CLUSTAL X program (13), and hydropathy analyses were conducted using the TMpred program (2). The results of these analyses are summarized in Table 8. For each family, the bacterial homologues are presented first, the archaeal homologues are presented second, and the eucaryotic proteins are presented last. Table 8 presents (i) the organismal domains, (ii) the protein abbreviations, (iii) the size of each individual protein, (iv) the database and accession number, allowing easy access to the sequence of that protein, (v) the number of putative TMSs predicted using the TMpred program, (vi) the size of the N-terminal hydrophilic domain (N) in number of amino acyl residues, (vii) the residues predicted to comprise the individual TMSs (1 to 14), and (viii) the size of the C-terminal hydrophilic domain (C). The first family shown is the Ca 2 :cation antiporter (CaCA) family (Table 8). The size differences between the proteins are apparent when examining the data summarized in column 3. In TABLE 6. Comparisons of cytoplasmic enzyme sizes for the domains of organisms and the kingdoms of Eucarya Enzyme EC no. Archaea Bacteria Eucarya All kingdoms Animals Plants Fungi Enolase (6) 430 (23) 438 (30) 434 (13) 444 (8) 438 (9) Phosphoglycerate kinase (10) 398 (21) 414 (27) 416 (11) 401 (4) 417 (12) Glyceraldehyde 3-phosphate dehydrogenase (12) 335 (24) 336 (57) 334 (23) 339 (13) 335 (21) Triose-phosphate isomerase (10) 253 (25) 251 (24) 249 (11) 254 (8) 249 (5) Phosphoglucose isomerase (1) 517 (22) 560 (18) 557 (7) 566 (8) 554 (3) Pyruvate kinase (4) 500 (22) 522 (22) 529 (9) 509 (6) 524 (7) Inosine-5 -monophosphate dehydrogenase (7) 500 (16) 517 (11) 520 (4) 502 (3) 525 (4) Inosine-5 -monophosphate-aspartate ligase (6) 426 (19) 447 (6) 454 (4) a 434 (2) Glutamine synthetase (11) 466 (26) 369 (33) 376 (13) 365 (15) 360 (5) Aspartate aminotransferase (6) 394 (17) 412 (15) 412 (8) 409 (5) 421 (2) Elongation factor-2/elongation factor-g No EC no. 731 (16) 698 (25) 849 (12) 855 (6) 845 (3) 842 (3) a No sequenced homologues of this enzyme were found in plants.

6 VOL. 183, 2001 MEMBRANE PROTEINS OF BACTERIA, ARCHAEA, AND EUCARYA 1017 TABLE 7. Results of statistical analyses of observed homologue size differences Homologue Groups compared No. with A B:A B Sign test P value Correlation Integral membrane transporter Archaea (A) vs Bacteria (B) 3: A B Bacteria (A) vs Eucarya (B) 0: A B Eucarya (A) vs Archaea (B) 19: A B Plants (A) vs animals (B) 2: A B Animals (A) vs fungi (B) 9: None Fungi (A) vs plants (B) 11: A B ABC receptor Archaea (A) vs Bacteria (B) 11: A B ABC membrane protein Archaea (A) vs Bacteria (B) 5: A B ABC cytoplasmic energizer Archaea (A) vs Bacteria (B) 3: A B column 5, it can be seen that there is substantial variation in the predicted number of TMSs. For this family and other families examined, some of this variation may represent experimental error due to limitations of the TMPred program. For the CaCA family, there is little variation in the sizes of the N-terminal and C-terminal hydrophilic domains (between 0 and 40 residues each). However, three of the eucaryotic proteins (all from animals) (CaSA2 Bta, Orfl Cel, and CaSA Dme) show large inter-tms loops between TMSs 6 and 7. These loops are between 455 and 566 amino acyl residues long, accounting for most of the size differences observed for these proteins compared with other homologues examined. These loops are predicted to be of 29 to 38 residues for the bacterial proteins, of 11 to 22 residues for the archaeal proteins, and of 14 and 10 residues for the two remaining eucaryotic proteins, plant and yeast proteins, respectively. Additionally, it can be seen that other eucaryotic inter-tms loops are of somewhat increased size relative to their procaryotic counterparts. For example, loops 1 and 2 (between TMSs 1 and 2) contain 1 to 22 residues in procaryotic proteins but 34 to 99 residues in the eucaryotic homologues; loops 2 and 3 in the procaryotic proteins are 16 to 19 residues long while those in the eucaryotic proteins are 21 to 23 residues long; and loops 3 and 4 in the procaryotic proteins are of 7 to 18 residues while those in the eucaryotic proteins are of 15 to 23 residues. Finally, the program predicts 11 TMSs for the bacterial homologues, 9 or 10 TMSs for the archaeal homologues, and either 10 or 12 TMSs for the eucaryotic proteins. All of these differences, when taken together, account for the observed size variations of the individual proteins of the CaCA family. The second family listed in Table 8 is the inorganic phosphate transporter (Pit) family. All but one of the Pit family members has a small N-terminal hydrophilic region, the one exception being the plant Pit Ath homologue, which has an N-terminal hydrophilic domain of 126 residues. The archaeal proteins generally have shorter hydrophilic N termini than the bacterial proteins. Further, the hydrophilic C termini of all homologues are short (1 to 26 residues). The major size variations observed between the procaryotic and eucaryotic proteins of this family are in loops 7 8 and 8 9. For example, Orf Cel is predicted to have a somewhat large loop 8 9 (46 residues), Glvr Hsa and Nps Sce have a large loop 7 8 (62 and 206 residues, respectively), and Pho4 Ncr has large loops 7 8 and 8 9 (122 and 89 residues, respectively). Differences in the number of putative TMSs predicted are also observed, with Orf Cel and Pho4 Ncr predicted to have more TMSs than the other homologues. The monovalent cation:proton antiporter families (CPA1 and CPA2) show similarly sized N-terminal hydrophilic domains, but their C-terminal hydrophilic domains differ substantially in size (Table 8). Thus, in the CPA1 family, four of the bacterial homologues have hydrophilic extensions of 21 to 32 residues, but one protein, Orf Bsu, has a hydrophilic extension of 131 residues. Similarly, four of the archaeal proteins have C-terminal hydrophilic extensions of 7 to 25 residues, but one protein (Nhe2 Afu) has an extension of 124 residues. Finally, all of the eucaryotic proteins have long C-terminal hydrophilic domains of 102 to 394 residues. In the CPA2 family, the major size differences are also in the C-terminal regions. In this family, the bacterial C-terminal extensions are large (226 to 282 residues), while all but one of the archaeal extensions are short (6 to 21 residues except for that in Orf Mth, which is 174 residues in length). All of the eucaryotic proteins have large hydrophilic C termini (185 to 277 residues). The two yeast proteins additionally exhibit large loops between their final two C-terminal TMSs (308 and 444 residues, respectively). Finally, several proteins of the divalent anion:na symporter family exhibit major size differences in their N-terminal hydrophilic domains (Table 8), although differences in loop sizes and numbers of putative TMSs contribute significantly to the overall protein size differences. Most of these proteins show small C-terminal hydrophilic extensions. In summary, we have found that the positional basis for the size variations observed between secondary carrier homologues from the three domains of life depends primarily on the family and secondarily on the individual proteins within that family. Some families show differences primarily in the N- terminal hydrophilic domains, others show differences in the C-terminal hydrophilic domains, and still others show differences in specific inter-tms loop regions. Most families exhibit size differences between homologues that represent a combination of these effects, with one of these effects predominating. DISCUSSION The average size differences for the various types of protein homologues analyzed are summarized in Table 9. When all family size differences are averaged, integral membrane transport proteins of bacteria are 8% larger than their archaeal homologues and 40% smaller than their eucaryotic homo-

7 1018 CHUNG ET AL. J. BACTERIOL. TABLE 8. Localization of regions of size difference between homologues of five families of secondary carriers Homologue family (TC) Group Name Organism Size a (AAs) No. b Putative TMS c Ca 2 :cation antiporter (2.A.19) Inorganic phosphate transporter (2.A.20) Monovalent cation:proton antiporter-1 (2.A.36) Monovalent cation:proton antiporter-2 (2.A.37) Divalent anion:na symporter (2.A.47) Bacteria ChaA Eco Escherichia coli 366 spp CaHA Syn Synechocystis sp. 372 gbd Orf21 Ype Yersinia pestis 366 embcaa ChaA Mtu Mycobacterium tuberculosis 360 embcaa Y4HA Rhi Rhizobium sp. 367 spp Archaea Orf1 Mth Methanobacterium thermoautotrophicum 322 gi Orf2 Mth Methanobacterium thermoautotrophicum 330 gi Orf Mja Methanococcus jannaschii 302 gbu Orf Pho Pyrococcus horikoshii 325 dbjbaa Orf Pab Pyrococcus abyssi 314 embcab Eucarya CaSA2 Bta Bos taurus 970 spp Orf1 Cel Caenorhabditis elegans 890 gbu CaSA Dme Drosophila melanogaster 950 gbl CAX1 Ath Arabidopsis thaliana 459 gbu Orf3 Sce Saccharomyces cerevisiae 725 spp Bacteria PitA Eco Escherichia coli 499 spp YG04 Hin Haemophilus influenzae 420 spp PitB Mtu Mycobacterium tuberculosis 552 gi Orf Cje Campylobacter jejuni 508 embcab Orf Cpn Chlamydophila pneumoniae 426 gi Archaea Y630 Mja Methanococcus jannaschii 297 spq Orf Afu Archaeoglobus fulgidus 314 gi Orf Mth Methanobacterium thermoautotrophicum 326 gi Orf Pho Pyrococcus horikoshii 406 dbjbaa Orf Pab Pyrococcus abyssi 405 embcab Eucarya Orf Cel Caenorhabditis elegans 516 gbu Glvr Hsa Homo sapiens 679 gbl Pit Ath Arabidopsis thaliana 587 gbx Pho4 Ncr Neurospora crassa 590 spp Npa Sce Saccharomyces cerevisiae 574 spp Bacteria YjcE Eco Escherichia coli 549 spp Orf Syn Synechocystis sp. 527 dbjbaa Orf Mtu Mycobacterium tuberculosis 542 gbz NhaP Pae Pseudomonas aeroginosa 424 dbjbaa Orf Bsu Bacillus subtilis 524 gi Archaea Nhe2 Afu Archaeoglobus fulgidus 494 gi Orf Mth Methanobacterium thermoautotrophicum 399 gi Y057 Mja Methanococcus jannaschii 426 spq Orf Pab Pyrococcus abyssi 390 gi Orf Pho Pyrococcus horikoshii 390 gi Eucarya Orf Cel Caenorhabditis elegans 602 gbu NahB Cmy Oncorhynchus mykiss 759 spq Nhel Hsa Homo sapiens 894 pira Nhx1 Ath Arabidopsis thaliana 538 gbaad Orf Spo Schizosaccharomyces pombe 759 embcab Bacteria KefC Eco Escherichia coli 620 embcaa KefC Mxa Mixococcus xanthus 598 gbaaa KefB Vch Vibrio cholerae 656 gbaaf Orf Pae Pseudomonas aeroginosa 613 gbaag Orf Nme Neisseria meningitidis 658 embcab Archaea Orf Mth Methanobacterium thermoautotrophicum 512 gbaab PhaC Mex Methylobacterium extorquens 276 gbaaa Orf Pab Pyrococcus abyssi 380 embcab Orf Pho Pyrococcus horikoshii 380 dbjbaa Orf Mja Methanococcus jannaschii 388 gbaab Eucarya Orf1 Ath Arabidopsis thaliana 756 embcab Orf2 Ath Arabidopsis thaliana 735 gbaad Orf3 Ath Arabidopsis thaliana 617 embcab Orf Sce Saccharomyces cerevisiae 873 embcaa Orf Spo Schizosaccharomyces pombe 898 embcab Bacteria Orf Reu Ralstonia eutropha 513 spq Orf Hin Haemophilus influenzae 461 piri YbdS Eco Escherichia coli 487 gbu Orf Syn Synechocystis sp. 449 dbjbaa Orf Nme Neisseria meningitidis 471 embcab Archaea Orf Mja Methanococcus jannaschii 432 gbu Orf Mth Methanobacterium thermoautotrophicum 443 gi Orf Afu Archaeoglobus fulgidus 397 gi Orf Pho Pyrococcus horikoshii 368 dbjbaa Orf Ape Aeropyrum pernix 430 dbjbaa Eucarya NadC Hsa Homo sapiens 592 gbu Orf Cel Caenorhabditis elegans 545 spp Orf Ath Arabidopsis thaliana 540 dbjbaa Orf Sce Saccharomyces cerevisiae 894 spp Orf Spo Schizosaccharomyces pombe 867 embcaa a Size in number of amino acyl residues (AAs). b Database and accession number. c Putative number of -helical TMSs predicted using the TMPred program. d N and C are the N-terminal and C-terminal hydrophilic domains (regions) preceding the first predicted TMS and following the last predicted TMS, respectively, based on TMPred analyses. Region number refers to the number of the putative TMS, based on TMPred predictions.

8 VOL. 183, 2001 MEMBRANE PROTEINS OF BACTERIA, ARCHAEA, AND EUCARYA 1019 TABLE 8 Continued Residues in region d : N C

9 1020 CHUNG ET AL. J. BACTERIOL. TABLE 9. Average size differences in various domains for the protein types analyzed Protein type Bacteria Archaea Relative size (%) All kingdoms Eucarya Plants Animals Fungi Integral membrane transporters ABC permeases Receptors Membrane proteins Energizers Cytoplasmic enzymes logues. When the three constituents of procaryotic-specific ABC-type uptake permeases were examined, the archaeal extracytoplasmic receptors proved to be 7% larger than their bacterial homologues, on average, while the membrane and cytoplasmic constituents were 3 to 4% smaller. Homologous cytoplasmic enzymes showed little or no significant difference between the three domains of life (Table 9). Within the Eucarya kingdoms, integral membrane transporters of plants proved to be significantly smaller than those of animals (21%) and fungi (17%), although no corresponding size differences were noted for homologous cytoplasmic enzymes. These observations clearly show that during evolution, integral membrane transport proteins have been subject to different pressures giving rise to size differences that are not paralleled in cytoplasmic proteins or extracytoplasmic receptors. In fact, the latter proteins exhibit significant size differences between Bacteria and Archaea that are opposite to those observed for the integral membrane proteins. These observations must be explainable at the molecular level. The molecular explanation(s) for the protein size differences documented in this report is currently elusive. Several investigators have noted that when plasmidic DNA sequences exhibiting short repetitive elements are transferred to yeast, the repeats tend to increase in number, although the reverse is true in Escherichia coli (1, 5, 6, 12). The molecular basis for this observation is not known, but if operative on chromosomal DNA over an extended period of evolutionary time, it could account for the observed average membrane protein homologue size differences. However, because the cytoplasmic proteins analyzed do not show this trend and because extracytoplasmic receptors show the opposite trend, we disfavor such an explanation. Other explanations may exist. Our domain analyses summarized in Table 8 show that the major size differences in secondary carrier proteins occur primarily in the N- and C-terminal hydrophilic extensions and specific inter-tms loops of these integral membrane proteins, and that the locations where the major size differences occur are family specific. Sometimes the numbers of putative -helical TMSs differ, but these differences may be in part artifactual and do not generally account for the size variations observed. Hydrophilic domains in transporters are known to play regulatory roles in various well-studied procaryotic and eucaryotic transport proteins (3, 14). It is possible that Eucarya have been under greater pressure to evolve regulatory domains controlling transport than have Bacteria and that Bacteria have in turn been under greater pressure to evolve such regions than have the Archaea. If this possibility does account for the observed size differences, then plants must have been under less stringent pressure to evolve protein regulatory sequences than were animals and fungi. Moreover, cytoplasmic enzymes have not been subject to similar constraints. These observations may have predictive value for purposes of annotation. However, one can expect that multiple explanations will account for the size variations observed. It is clear that the studies reported here pose more questions than they have answered. What are the membrane structural features or mechanistic features that promote the observed size differences? Are repeated DNA sequences present in the structural genes for these proteins, and if so, do numbers of repeats contribute to or even account for the size differences observed for their protein products? What are the physiological benefits to organisms in the three domains of life to promote homologue size variation? What accounts for the size differences observed between plant transporters and those from other Eucarya? Further computational experimentation, currently in progress, will be required to provide answers to these interesting questions. ACKNOWLEDGMENTS We thank Donna Yun, Monica Mistry, Milda Simonaitis, and Yolanda Anglin for their assistance in the preparation of this manuscript. Work in the authors laboratory was supported by NIH grant no. 2R01 AI14176 from the National Institute of Allergy and Infectious Diseases and no. 9RO1 GM55434 from the National Institute of General Medical Sciences, as well as by the M. H. Saier, Sr. Memorial Research Fund. REFERENCES 1. Henderson, S. T., and T. D. Petes Instability of simple sequence DNA in Saccharomyces cerevisiae. Mol. Cell. Biol. 12: Hofmann, K., and W. Stoffel TMPred a database of membrane spanning protein segments. Biol. Chem. Hoppe-Seyler 347: Hoischen, C., J. Levin, S. Pitaknarongphorn, J. Reizer, and M. H. Saier, Jr Involvement of the central loop of the lactose permease of Escherichia coli in its allosteric regulation by the glucose-specific enzyme IIA of the phosphoenolpyruvate-dependent phosphotransferase system. J. Bacteriol. 178: Kaback, H. R., and J. Wu From membrane to molecule to the third amino acid from the left with a membrane transport protein. Q. Rev. Biophys. 30: Levinson, G., and G. A. Gutman Slipped-strand mispairing: a major mechanism for DNA sequence evolution. Mol. Biol. Evol. 4: Morel, P., C. Reverdy, B. Michel, S. D. Ehrlich, and E. Cassuto The role of SOS and flap processing in microsatellite instability in Escherichia coli. Proc. Natl. Acad. Sci. USA 95: Persson, B. L., J. Petersson, U. Fristedt, R. Weinander, A. Berhe, and J. Pattison Phosphate permeases of Saccharomyces cerevisiae: structure, function and regulation. Biochim. Biophys. Acta 1422: Saier, M. H., Jr Molecular phylogeny as a basis for the classification of transport proteins from bacteria, archaea and eukarya, p In R. K. Poole (ed.), Advances in microbial physiology. Academic Press, San Diego, Calif. 9. Saier, M. H., Jr Genome archeology leading to the characterization and classification of transport proteins. Curr. Opin. Microbiol. 2: Saier, M. H., Jr A functional-phylogenetic classification system for transmembrane solute transporters. Microbiol. Mol. Biol. Rev. 64: Stevens, B. R., A. Fernandez, B. Hirayama, E. M. Wright, and E. S. Kempner Intestinal brush border membrane Na /glucose cotransporter functions in situ as a homotetramer. Proc. Natl. Acad. Sci. USA 87: Strand, M., T. A. Prolla, R. M. Liskay, and T. D. Petes Destabilization