INTRODUCTION

Top Previous Next References

The purpose of this review is to examine the information available on archaeal stress genes and proteins, particularly those of the Hsp70 and Hsp60 families, while critically discussing the data in comparison with what is known for the bacterial and eukaryotic equivalents. The aim was to treat the specific topics of the review embedded in the framework of closely related areas of science. Cross-fertilization between research with archaea and research with bacteria and eukaryotes is highlighted to show how the study of archaea has contributed, and will continue to contribute, to other fields, both basic and applied.

A deliberate effort has been made to simplify the text and make it readable to a general audience. Consequently, terms are explained and the data and theories are presented within a historical perspective. A minimal amount of overlapping between related sections distant from one another in the body of the review is included to enhance the flow, particularly when a later section expands on an earlier one.

A comprehensive search of printed literature and databases was attempted. Colleagues were consulted. The majority of the data are displayed in tables and figures, but only illustrative cases are explained in the text. Reviews rather than original reports are cited for topics related to but not strictly dealing with archaeal genes, proteins, or organisms, to reduce the number of references and save space while providing access to a wealth of published information.

Archaea have been found in a wide variety of ecosystems with very different characteristics, very hot or very cold, temperate, anoxic, oxygenated, etc. (9, 12, 50, 59, 178, 297). Thus, what represents a stressor for a species may be a condition required for the optimal growth of another species. The term "stressor," therefore, must be understood in relation to a particular species or group of organisms that share similar living conditions, for example temperature. In this regard, organisms are classified into psychrophiles (optimal temperature for growth [OTG] 15°C or lower), psychrotolerant organisms (OTG, 20 to 30°C), mesophiles (OTG, 35 to 40°C), thermophiles (OTG, 50 to 70°C), and hyperthermophiles (OTG, 80°C or higher) (186). Temperatures higher or lower than the optimal may cause stress and induce a stress response. A temperature upshift causes a heat shock response (47, 90, 120, 128, 177, 252, 292), whereas a temperature downshift induces a cold stress or cold shock response (255, 275, 304). The latter is not dealt with in this review. Likewise, adaptation to high osmotic and barometric pressures and the response of the cell to their changes (33, 68, 77, 121, 138, 190, 201, 204, 232, 266, 271) are not treated in any detail.

We draw attention, though, to the formation of multicellular structures that result in improved cell resistance to physical, mechanical, and chemical stressors. These structures are of various types and have considerable potential for the biotechnology industry and for the exploration and conquest of inhospitable ecosystems, but their relevance to stress resistance is rarely discussed. We highlight the topics that future studies ought to address in relation to the proteins (and their genes) and other molecules that build the intercellular connective material to keep the cells together in a functional, three-dimensional arrangement.

STRESS, STRESS RESPONSE, STRESS GENES AND PROTEINS, HEAT SHOCK, MOLECULAR CHAPERONES, AND CHAPERONINS

Top Previous Next References

Primer

A cell confronted with an abrupt change in its immediate surroundings suffers stress. The cause, or stressor, may be of various types, for example, physical (temperature elevation) or chemical (increase or decrease in pH, salinity, or oxygen concentration) (47, 113, 252, 292).

A key component of stress is protein denaturation (93, 96, 123, 124, 223, 293). Many proteins lose their native, functional configuration and tend to aggregate. The process may be reversible up to a degree, beyond which it becomes irreversible and generalized within the cell, which ultimately dies.

Another main component of stress is the down-regulation of many housekeeping genes, some of which are actually shut off. Whether this is all due to protein denaturation and represents just the breakdown of the cellular machinery or is an active, induced process by which genes are "told" to slow down or stop has not yet been elucidated. Perhaps both mechanisms, gene failure and regulated shutdown, participate.

Protein damage and gene down-regulation are part of the stress response. There is yet another important component of the stress response, i.e., activation of the stress genes (90, 120, 205, 208, 219, 253, 254, 279). The concentrations of the protein products of these genes increase in response to stressors, protecting the cell from the destructive effects of stress and enhancing post-stress recovery by promoting renaturation (refolding) of partially denatured proteins (93, 101, 223, 293).

Thus, stress inactivates or down-regulates many genes but activates others, whose function is to save the cell. Most stress genes also function in the absence of stress, namely, under normal physiological conditions. The proteins encoded by these genes play critical roles in physiological protein biogenesis. They assist in the folding, translocation, and assembly of other proteins (191, 214, 235, 236, 247). This is the reason why many stress proteins are also called molecular chaperones (72). They help other cellular proteins to (i) fold correctly during and after translation; (ii) migrate to the cell's locale, where they will reside and function; and (iii) assemble into the quaternary structure that will make them useful to the cell when the proteins function as polymers.

Furthermore, some stress proteins participate in the degradation of other polypeptides, for example when these are denatured beyond recovery and could pose a serious threat to the cell if they aggregated (39, 92, 97, 98, 122, 123, 133).

In summary, stress proteins, particularly those that are molecular chaperones, aid and protect other cellular proteins from their birth on, but they also contribute to the elimination of polypeptides that are no longer useful and endanger cell viability.

It is important to bear in mind that not all stress proteins are chaperones and, vice versa, that not every molecular chaperone is a stress protein.

The wide spectrum of activities of stress proteins is not limited to the chaperoning of other proteins as described above. These activities also include other functions, for example, modulation of their own synthesis (6, 20, 90), regulation of the stress kinase JNK (85), association with enzymes (for purposes yet to be determined) (43), and participation in signal transduction pathways (175) and in rRNA processing (249). It is therefore clear that stress proteins are multifunctional and ubiquitous. They play their roles in all cells, cell compartments, and organelles and are said to be promiscuous because they interact with a great variety of other molecules.

This diversity of functions is reflected in the structural features of the stress proteins, which are composed of domains and motifs with specific roles. As we discuss below, characterization of these domains and motifs has helped in the classification of newly found genes and proteins, identification of stress proteins in their various anatomical locations, and determination of their evolutionary origins.

hsp and Hsp were first observed in Drosophila exposed to a temperature higher than the optimum for growth (27°C) (reference 177 and references therein). The genes activated by the temperature upshift were called heat shock genes, and their products were called heat shock proteins. In this review, we use the terms "stress" and "heat shock" interchangeably to qualify the words response, gene, and protein, although we favor the use of "stress" rather than "heat shock" and reserve the latter for the specific instances in which the stressor is a temperature elevation.

Hsp (and their genes) are classified into groups or families according to their molecular mass in kilodaltons (Table 1). The proteins of the 55- to 64-kDa group, or Hsp60 family, are also called chaperonins and are included within the molecular chaperones, generally speaking. More specifically, the latter term is applied to the Hsp70 family. The genes and proteins belonging to the Hsp60 and Hsp70 families have been extensively studied in many bacterial and eukaryotic species.

The classification of all living cells into three main evolutionary lines, or phylogenetic domains, Bacteria (eubacteria), Archaea (formerly archaebacteria), and Eucarya (eukaryotes) (11, 297, 298, 300), is still useful despite its limitations and the challenges generated by new findings and contrasting theories (66, 67, 80, 103, 105, 106, 109-111, 188, 189, 192, 198, 234, 286, 298). It helps us to visualize how evolution produced what we see today and to track down genes from the past to the present.

The overwhelming majority of information available on stress genes and proteins comes from studies of bacteria (e.g., Escherichia coli and Bacillus subtilis) and eukaryotes (e.g., Drosophila melanogaster, Saccharomyces cerevisiae, plants, and mammals including humans). The study of stress genes and proteins in organisms of the phylogenetic domain Archaea began only a decade ago and is much less advanced than in the other two domains.

HSP70(DNAK) LOCUS

Top Previous Next References

Structure and Organization

The first hsp70 gene identified by cloning and sequencing within the domain Archaea was reported in 1991 (182). The gene was found in the mesophilic methanogen Methanosarcina mazei S-6 (OTG, 37°C). Shortly thereafter, in 1992, a homolog was cloned and sequenced from another archaeon, Halobacterium (Haloarcula) marismortui (110). This organism is also mesophilic but belongs to a group, the extreme halophiles, different from that of M. mazei S-6.

For a while, the above genes were the only two archaeal hsp70 genes known. In 1994, two additional homologs were reported: one in another mesophilic extreme halophile, Halobacterium cutirubrum (OTG, 45°C), and the other in a thermophile, Thermoplasma acidophilum (OTG, 55°C) (111).

These findings seemed to indicate that the hsp70 gene was present in archaea, confirming the widely held notion that this gene is one of the most highly conserved, occurring in all organisms. This idea was challenged in 1996, when the sequencing of the whole genome of Methanococcus jannaschii (OTG, 85°C) revealed the absence of the hsp70 gene in this archaeon (31), a result that confirmed observations in other laboratories (47, 164).

More recently, in 1997, the sequencing of the thermophilic methanogen Methanobacterium thermoautotrophicum H (OTG, 55°C) was published (263). The hsp70 gene is present in this methanogen, which in this regard is therefore similar to the mesophilic M. mazei S-6 but different from the hyperthermophilic methanogen M. jannaschii.

While the sequencing of the M. thermoautotrophicum H was under way, another hsp70 gene was cloned and sequenced from a second Methanosarcina species, the thermophilic species Methanosarcina thermophila TM-1 (OTG, 50°C) (126). So by this time, it was clear that at least some mesophilic and thermophilic methanogens do have the gene but that some (perhaps all) hyperthermophiles do not.

After the discovery of an archaeal hsp70 gene in M. mazei S-6 in 1991, more sequencing up- and downstream of this gene revealed the other genes that accompany hsp70(dnaK) in bacteria: hsp40(dnaJ) and hsp23(grpE); these discoveries were made in 1993 (181) and 1994 (45), respectively.

It is pertinent to note here, as was done above for hsp70, that the hsp40 gene and its protein Hsp40 are named dnaJ and DnaJ, respectively, when they are from bacteria. Similarly, the bacterial hsp23 gene and its protein, Hsp23, are called grpE and GrpE, respectively.

We use the terms hsp40 and Hsp40 when referring to archaea, for the same reasons we use the terms hsp70 and Hsp70. However, for the hsp23 gene, we use the terms grpE and GrpE, because the alternative hsp23 and Hsp23 may be confusing since there are several different small heat shock proteins with a mass close to 23 kDa (76, 78, 156, 206, 227). Also, the archaeal GrpE molecule has a counterpart in bacteria and the eukaryotic organelles of bacterial origin but apparently not in the eukaryotic cytosol. The latter does not seem to have a GrpE protein but has some other alternative to exercise similar functions, although recent findings suggest that grpE homologs might also occur in the eukaryotic cytosol (211, 212).

As a result of the sequencing of the M. mazei S-6 and M. thermophila TM-1 hsp70 chromosomal regions and the sequencing of the M. thermoautotrophicum H genome, there are today three archaeal hsp70 loci whose structure and organization have been determined (Fig. 1) (31, 126, 179). The gene order 5'-grpE-hsp70-hsp40-3' occurs in the three archaeal loci and is the same as that observed in many bacteria, particularly gram-positive bacteria (Fig. 2) (179). However, there are differences between the three archaeal loci. For example, they differ in the length of the 5'-grpE-hsp70-3' and 5'-hsp40-next gene-3' intergenic regions and in the gene that follows hsp40 downstream. This gene is the same in M. mazei S-6 and M. thermophila TM-1 but different in M. thermoautotrophicum H.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   The hsp70(dnaK) locus genes of the archaea for which sequences are available, including genes up- and downstream of hsp70(dnaK). The genes are represented by rectangular boxes from the 5' to the 3' end (left to right), with their names above their respective boxes in the locus on top [dnaK and dnaJ are used instead of hsp70(dnaK) and hsp40(dnaJ) for clarity]. The numbers within the boxes indicate the number of amino acids encoded. The lines joining the boxes represent the intergenic regions, with their lengths, in base pairs, shown underneath. The sequences of M. thermophila TM-1 grpE and trkA are still incomplete (what is available would encode 53 and 401 amino acids, respectively). Accession numbers and other details are provided in Tables 2, 5, and 10. Reprinted from references 126 and 179 with permission of the publishers.

View larger version (27K): [in this window] [in a new window]   FIG. 2.   The hsp70(dnaK) locus genes of the archaeon M. mazei S-6 and three gram-positive bacteria: B. subtilis (M84964), C. acetobutylicum (M74569), and S. aureus (D30690). Symbols are the same as those described in the legend to Fig. 1. Modified from reference 179 with permission of the publisher.

The length of the intergenic region between hsp70 and hsp40 is conserved in the three loci, particularly in comparison with the other intergenic regions.

The meaning of these structural characteristics is not completely understood. They suggest, for example, that hsp70 and hsp40 may have evolved together, as a unit. This notion is also supported by the conservation of the homologous gene pair in many bacteria (see, for example, Fig. 2).

As discussed later in this review, there are indications that the hsp70 gene in archaea was received via lateral transfer from bacteria. Perhaps it was accompanied by hsp40 and grpE, since both are always present whenever hsp70 is, and the three genes appear next to each other in many bacteria. However, other structural characteristics and experimental results, to be discussed below (see "Occurrence of hsp70 in nature"), tend to make this notion less credible, at least in its simplest formulation. Analyses of the nucleotide sequences between the protein-coding regions of the genes do not reveal obvious similarities, except for the presence of putative archaea-type promoters and bacterial-type termination signals in the expected locations with regard to the translation start and stop codons, respectively (44, 181, 182). Other sequence features vary with the intergenic region and the species, but the regions upstream of hsp70 in two of the methanosarcinas possess a series of repeats and palindromes. They might be cis-acting signals, namely, binding sites for regulatory factors (168). In contrast, the region upstream of hsp70 in M. thermoautotrophicum H is very short and lacks anything that might be a promoter or a cis-acting site.

No bacterial-type promoter sequences (52, 120, 219, 258, 259) are identifiable in these archaeal intergenic regions, nor are there bacterium-type regulatory elements, such as CIRCE (259, 310, 313) or ROSE (208, 209), that one can detect by sequence comparisons.

If one considers the high degree of conservation of these regulatory sequences among bacteria, it is reasonable to conclude that they do not occur in the three known archaeal loci and that regulation of the hsp70 locus genes in these organisms is mediated by factors different from those operating in bacteria. Thus, despite the similarities in organization between the archaeal and bacterial hsp70 loci, their mode of expression and regulatory mechanisms appear to be different.

Also remarkable is that the 5'-grpE-hsp70-3' intergenic region and the distance between grpE and the next gene upstream in M. mazei S-6 and M. thermophila TM-1 are considerably longer than the equivalent regions in bacteria. The latter have their genes closer to one another, in agreement with their polycistronic mode of transcription and their being regulated as a unit, or operon. Instead, the structure of the archaeal loci in the two methanosarcinas shown in Fig. 2 does not suggest the bacterial modes of transcription and regulation but different ones (see also experimental data, given below).

M. mazei S-6 hsp70(dnaK), hsp40(dnaJ), and grpE respond to heat shock by an increase in the production of their transcripts (Fig. 3) (42, 44), as one would expect for stress genes. The transcripts are monocistronic as in eukaryotes (302) and in contrast to bacteria (10, 95, 120, 131, 132, 294, 313). Likewise, the peak response in terms of transcript levels is reached after heat shocks longer than those that would induce a peak response in bacteria (Fig. 4) (164), also in agreement with what is observed in eukaryotes. The transcription initiation sites map to positions reminiscent of the eukaryotic initiation sites with respect to the promoter (Fig. 5) (42, 44, 179). Furthermore, the genes respond to temperatures ranging from 45 to 60°C (Fig. 6) (164) and to other stressors such as cadmium (Cd²⁺) (Fig. 7) (179) and ammonia (165) as expected for heat shock genes. Thus, the data show that the archaeal hsp70 locus genes are stress or heat shock genes but have mixed bacterial and eucaryal characteristics.

View larger version (67K): [in this window] [in a new window]   FIG. 3.   (A to D) Northern blots with M. mazei S-6 total RNA (10 µg/lane) showing an increase in the levels of transcripts of hsp70(dnaK) (A), hsp40(dnaJ) (B), and grpE (C), and a decrease in the level of the transcript of orf16 (D) in response to heat shock. (E) Dot blot showing a decrease in the level of the transcript of orf11-trkA in response to heat shock. Hybridizations were done in all cases with radiolabelled probes specific for the respective genes. In panels A, B, and D, I is the gel stained with ethidium bromide showing the RNAs, 23S and 16S while II is the corresponding Northern blot. Lanes: A, total RNA from M. mazei S-6 cells maintained at the optimal growth temperature of 37°C, i.e., non-heat-shocked cells; B and C or B to D, total RNA from cells heat shocked at 45°C for increasing time periods, from 15 to 60 min. The sizes of the transcripts in panels A to D are indicated in kilobases. Transcripts were detected for all the genes in non-heat-shocked cells. Heat shock caused an increase in the levels of the transcripts of hsp70, hsp40, and grpE. The reverse occurred for orf16, and orf11-trkA. The latter two genes overlap and are cotranscribed, whereas the other genes are transcribed monocystronically. Reprinted from references 42, 44, 49, and 184, with permission of the publishers.

View larger version (76K): [in this window] [in a new window]   FIG. 4.   Response of the M. mazei S-6 hsp70(dnaK) gene to heat shocks of various durations. Northern blots of total RNA (10 µg/lane) extracted from M. mazei S-6 cells before heat shock (lane 0 in both panels) or after a heat shock at 45°C for the times indicated in the horizontal axis, in minutes (min) or hours (h). Hybridization was done with a probe for dnaK. The size of the hybridization bands in kilobases is indicated to the right. Reprinted from reference 164 with permission of the publisher.

View larger version (27K): [in this window] [in a new window]   FIG. 5.   M. mazei S-6 promoters for the hsp70(dnaK) locus genes. Bases with asterisks are identical to those in the consensus sequence for promoters in methanogens, and bases with dots denote positions without base preference (25); underlined bases represent the archaeal box A (reference 312 and references therein). The consensus sequence was derived from comparative analysis of promoters for many non-heat-shock genes (25). There is no information on promoters for archaeal grpE, hsp70(dnaK), or hsp40(dnaJ), except that shown here and in Tables 2, 5, and 10. Therefore, no consensus sequence is available for these archaeal heat shock genes. Note that while the promoters for the non-heat-shock genes orf16 and orf11-trkA match the consensus 100%, the grpE, hsp70(dnaK), and hsp40(dnaJ) promoters do not match it to the same extent. Reprinted from reference 179 with permission of the publisher.

View larger version (18K): [in this window] [in a new window]   FIG. 6.   Response of the M mazei S-6 genes grpE, hsp70(dnaK), and hsp40(dnaJ) to heat shock at various temperatures demonstrated by slot-blotting with M. mazei S-6 RNA. The levels of mRNA for grpE, hsp70(dnaK), and hsp40(dnaJ) (top three panels) are represented by vertical bars expressed in the optical density (OD) × millimeter units given by the densitometer. The respective slot blots (10 µg of total RNA from S-6 cells per slot) are shown at the foot of the bars, while the heat shock temperatures are indicated in the horizontal axis at the bottom of the figure (°C). Hybridization was done with the respective labelled probes. The culture density is shown in the bottom panel. The OD660 was determined at time zero (open bars) and at 30 min (hatched bars) in cultures maintained at 37°C or heat shocked during this 30-min period at the temperatures indicated at the foot of the bars. Reprinted from reference 164 with permission of the publisher.

View larger version (29K): [in this window] [in a new window]   FIG. 7.   Response of the M. mazei S-6 genes grpE, hsp40(dnaJ), and hsp70(dnaK) to the stressors cadmium (Cd2+) and heat. The bars represent levels of mRNA determined by slot blotting with probes for the grpE, hsp40(dnaJ), and hsp70(dnaK) genes. The total RNAs were from cells grown at 37°C (i.e., the optimal temperature for growth of M. mazei S-6) in medium without Cd2+ (a) and in medium with 5 or 27 mM CdCl2 (b and c, respectively) and from cells grown in medium without Cd2+ but heat shocked at 45°C for 30 min (d). Note that the levels of the mRNAs from the three genes increased after heat shock by comparison with the levels before heat shock (constitutive or basal levels; compare a and d). Likewise, the presence of Cd2+ in the medium also induced an increase in the levels of the three mRNAs. This effect was more marked with 27 mM than with 5 mM CdCl2 (compare a with b and c; and compare b with c). Reprinted from reference 179 with permission of the publisher.

The body of structural and functional data available at present suggests that the mechanism of transcription initiation for the archaeal hsp70 locus genes differs from those known to operate in bacteria (26, 27, 120, 132, 208-210, 242, 259, 310, 311, 313) and must involve factors which are not of a bacterial type, i.e., different from  factors (180). These data, as well as the fact that all transcription initiation studies with archaeal systems (albeit none involving heat shock genes and practically all done with hyperthermophilic systems) have demonstrated transcription factors of the eucaryal type (16, 51, 94, 117, 118, 264, 274, 276, 312), force the prediction that initiation for the M. mazei S-6 hsp70, hsp40, and grpE genes involves eucaryal-type factors.

The archaeal homologs of the eucaryal TATA-binding protein (TBP) and the transcription factor IIB (TFIIB) (aTFB and aTFA, respectively [117, 118]), have been identified and shown to be required for the transcription of archaeal, non-heat shock genes in vitro (57, 94, 117, 118, 276). There is no comparable information for archaeal stress genes, but one may hypothesize, based on the observations described in the previous section, that these genes will also require TBP and TFIIB as basal factors. Moreover, it is likely that other factors would also be necessary to induce the response to stressors and preferentially, or even specifically, start transcription of hsp70 and its teammates, hsp40 and grpE.

The tbp gene of M. mazei S-6 has been cloned and sequenced (51). The deduced amino acid sequence of the protein possesses some of the expected archaeal characteristics, but it also shows unique features. For example, like all archaeal TBPs known (reviewed in reference 264), the M. mazei S-6 protein is shorter than most eucaryal homologs, amounting to what is the C-terminal domain in eucaryal molecules. Also, the M. mazei S-6 protein is acidic, like the other known archaeal proteins, but it differs from them in that its N-terminal third is basic, not acidic. The direct, imperfect repeats found in all TBPs, archaeal and eucaryal, are also present in the M. mazei S-6 molecule. Repeats of approximately 42 amino acids separated by a spacing segment of 51 residues on average can be identified (51). The repeats are better conserved in the archaeal than in the eucaryal TBPs, and this is also true for the M. mazei S-6 homolog. A few archaeal TBPs have an acidic tail composed of a series of Glu residues in the C-terminal end. This acidic tail is not present in the M. mazei S-6 molecule.

The overall and regional characteristics of the M. mazei S-6 protein most probably determine its functional properties in what pertains to the binding to DNA at the promoter and to the potential interaction with other transcription factors, such as TFIIB and perhaps stress-specific factors (139a). These structure-function aspects of M. mazei S-6 TBP are being investigated at present. Purified TBP binds to the M. mazei S-6 hsp70 promoter, as demonstrated by the electrophoretic mobility shift assay (EMSA) (57a).

Research to determine how transcription initiation starts and proceeds under constitutive (basal) conditions and in response to stress (heat shock) is under way. In experiments with cell lysates from M. mazei S-6, it was demonstrated that TBP present in the lysates binds to the hsp70 promoter (51a). The phenomenon is observed with lysates from both unstressed and stressed cells. In the latter, a protein appears or increases in concentration or in its ability to bind DNA or TBP, which causes an additional shifted band in EMSA. The nature and role of the protein are under investigation. It might be a regulatory factor that binds near the hsp70 promoter.

ARCHAEAL HSP70 AND HSP70

Top Previous Next References

The Gene

The salient characteristics of the archaeal genes sequenced thus far are described in Table 2. The promoters, terminators, and ribosome-binding sequences or sites (RBS) are putative, except for the M. mazei S-6 promoter, for which preliminary experimental evidence supports the promoter shown (Fig. 5).

The archaeal Hsp70 proteins are quite similar to each other and, most remarkably, equally so to proteins from gram-positive bacteria (Table 3). Among the archaeal proteins, the most similar pairs are those from the two methanosarcinas, the two extreme halophiles, and the two thermophiles (T. acidophilum and M. thermoautotrophicum H) (see "Hsp70-based phylogenetic trees" below).

The archaeal proteins all have the universal markers for Hsp70 and DnaK and the bacterial markers (Table 4). However, they do not have any of the markers typical of eucaryal molecules. Thus, the archaeal Hsp70 is of bacterial type in sequence and in structural features that reflect its function.

A remarkable feature that appears to be distinctive for the archaeal Hsp70 is the absence of a stretch of 23 to 25 amino acids in the N-terminal quadrant, which became evident when the sequences were aligned with those of proteins from gram-negative bacteria (Fig. 8) (182). This major marker is shared with the DnaK proteins from gram-positive bacteria and is not present in eukaryotic homologs (109-111).

View larger version (23K): [in this window] [in a new window]   FIG. 8.   Amino acid sequences (single-letter symbols) of six archaeal Hsp70(DnaK) proteins and of four bacterial homologs, two from gram-negative bacteria (E. coli and C. crescentus) and two from gram-positive bacteria (C. acetobutylicum and B. subtilis) between positions 41 and 120, aligned with the program PileUp (Genetics Computer Group, University of Wisconsin, Madison, Wis.). The absence of 23 residues in the proteins from the archaea and gram-positive bacteria compared with those from gram-negative bacteria is shown by dots. Organisms and accession numbers are as follows: H.m., Haloarcula (Halobacterium) marismortui (Q01100); H.c., Halobacterium cutirubrum (P42372); E.c., Escherichia coli (P04475); C.c., Caulobacter crescentus (P20442); M.t. (H), Methanobacterium thermoautotrophicum H (O27351); T.a., Thermoplasma acidophilum (P50023); M.m., Methanosarcina mazei S-6 (P27094); M.t. (TM-1), Methanosarcina thermophila TM-1 (Y17862); C.a., Clostridium acetobutylicum (P30721); B.s., Bacillus subtilis (P13343). Modified from references 179 and 182 with permission of the publishers.

The evolutionary and functional significance of this sequence gap in archaea and gram-positive bacteria has not been elucidated. However, it has given support to a phylogenetic classification that places the archaea closer to gram-positive bacteria than to eukaryotes (105, 106, 109-111), in contrast to the classical 16S-18S rRNA-based tree (11, 299, 300). In the Hsp70-based tree, the extant gram-negative bacteria would have separated from their ancestors, i.e., the ancestors of today's gram-positive bacteria, early in evolution. As this happened, or shortly thereafter, the gram-negative line acquired the 23 to 25 extra amino acids that characterize its Hsp70. Also, within the framework of this hypothesis, the eukaryotic nucleus would have arisen from a fusion of a primitive archaeon with a gram-negative ancestor. The gene that ultimately became established in the eukaryotic line was that which came from the bacterial partner.

These are speculations based on sequence comparisons and other data that are not completely satisfactory in view of all the information available today. Alternative explanations have been put forward and are discussed in some detail in subsequent sections of this review.

ARCHAEAL HSP40 AND HSP40

Top Previous Next References

The Gene

The four archaeal hsp40(dnaJ) genes sequenced thus far are described in Table 5. They are remarkably similar to each other, as are the proteins they encode (see below).

The four archaeal Hsp40(DnaJ) proteins known at present are similar to one another and to their bacterial homologs (Table 6). As is the case for the Hsp70, the most similar pair is that of the two proteins from methanosarcinas. The universal motifs and signatures that characterize the Hsp40 molecule, whether from eukaryotes or from bacteria, also occur in the archaeal homologs, except those that are distinctive for the eucaryal molecules (Table 7).

The Gly-rich domain of the H. cutirubrum Hsp40 is longer and has a higher percentage of Gly than those of the three molecules from methanogens (Table 8). The H. cutirubrum molecule also shows a different pattern of distribution of the CxxCxGxG motif from the molecules from the methanogens (Table 9). Motifs 1 and 2 (counting from the N to the C terminus) are separated by 9 amino acids, motifs 2 and 3 are separated by 18 amino acids, and motifs 3 and 4 are separated by 6 amino acids in the four molecules. However, motif 1 begins farther away from the N terminus in the H. cutirubrum molecule than in the others. In consequence, motif 4 is the closest to the C terminus in the molecule from the extreme halophile compared with those from the methanogens. The question remains open whether these seemingly unique features of the molecule from the extreme halophile reflect an adaptation to life under high-salinity conditions and/or to cope with salinity changes.

Motif 1 in the M. thermoautotrophicum H molecule is, barring a sequencing error, aberrant in that the last residue is Arg (R) instead of Gly (G).

ARCHAEAL GRPE AND GRPE

Top Previous Next References

The Gene

The two archaeal grpE genes whose sequences have been determined are described in Table 10. The genes differ considerably in length; the M. mazei S-6 gene encodes a molecule 35 amino acids longer than that encoded in the M. thermoautotrophicum H homolog. This disparity confirms the poor degree of conservation of grpE and predicts that it will be very difficult to identify homologs in nature on the basis of sequence comparisons alone. The failure to detect GrpE in the eukaryotic-cell cytosol for example, may be due to its diversity. Methods other than structural analyses may be necessary to unveil the true spectrum of this molecule, as suggested by recent work (212) and by the data in Table 10 (see also below).

The amino acid sequence of GrpE is not as highly conserved as that of Hsp70 or even Hsp40 (Table 11). However, if discrete regions, for example regions I and II (294), are compared, the similarity increases (Table 12). These regions and the GrpE motifs (45) are shown in Fig. 9. The functions of these structural features have not been determined. It has been suggested that they might be important portions of the molecule, involved in the interaction of GrpE with the other members of the chaperone machine, Hsp70 and Hsp40 (45, 294).

View this table: [in this window] [in a new window]   TABLE 12.   Comparison of the GrpE amino acid sequences from M. mazei S-6 and M. thermoautotrophicum H: entire molecule, and regions I and II

View larger version (45K): [in this window] [in a new window]   FIG. 9.   Amino acid sequences (single-letter symbols) of the archaeal GrpE proteins from M. mazei S-6 (M.m. S-6; P42367) and M. thermoautotrophicum H (M.t. H; O27350) aligned with the program PileUp. Regions I and II (294), in that order from the N terminus, are underlined. Motifs 1, 2, and 3 (45), also from the N to the C terminus, are shaded, with their respective consensus sequences on top (light and dark shades show hydrophobic and hydrophilic residues, respectively). The M. thermoautotrophicum H molecule is shorter than the M. mazei S-6 protein, with amino acids missing at the beginning and the end (tildes) and inside (dots).

OCCURRENCE OF HSP70 IN NATURE

Top Previous Next References

The Archaeal Puzzle

The absence of the hsp70 gene in some archaeal species has been noted since the early 1990s (47) and was also found later, when it could not be detected in the hyperthermophiles Methanothermus fervidus, Sulfolobus sp., and M. jannaschii or in the mesophile Methanospirillum hungateii (47, 164). These reports, however, were based on negative results obtained by Northern, Southern, and Western blots with heterologous probes. Consequently, they could not be taken as proof of the absence of the gene.

A definitive confirmation came in 1996, when the sequencing of the M. jannaschii genome did indeed reveal that this organism does not contain hsp70 or the other two genes of the chaperone machine triad, hsp40 and grpE (31). Although this finding helped to give credence to previous negative results obtained by blotting procedures with heterologous nucleic acid and antibody probes and to reaffirm the idea that some organisms may indeed lack hsp70, it raised questions about the earlier finding of the gene in M. mazei S-6. This organism is a methanogen like M. jannaschii. Why is it, then, that the former contains hsp70 while the latter does not? Was the reported M. mazei S-6 gene real or artifactual?

There were additional data confirming the occurrence of hsp70 in other methanosarcinas, different from M. mazei S-6, from before the M. jannaschii genome had been sequenced (42). However, once again, these data had been obtained by Northern and Southern blots with a probe for the M. mazei S-6 gene, and the possibility of nonspecific hybridizations could not be ruled out.

The situation was finally clarified when the full genome sequence of another methanogen, M. thermoautotrophicum H was reported in 1997 (263). Like M. mazei S-6, this methanogen contains hsp70, as well as hsp40 and grpE (Fig. 1).

As things stand today, it is clear that hsp70 occurs in some but not all methanogens. It also occurs in extreme halophiles, but it is not known whether there are organisms in this group that lack the genethis remains to be demonstrated. The gene does not occur in any of the extreme thermoacidophilic archaea investigated up until now. This had been suggested, as mentioned above, by results obtained by blotting procedures (47, 164) and was confirmed for Archaeoglobus fulgidus (151) and other species by whole-genome sequencing (Table 13). In addition, a search for the hsp70(dnaK) relative hsc66, found in Escherichia coli and other bacteria (260), in the genomes of A. fulgidus, Pyrococcus horikoshii, M. jannaschii, and M. thermoautotrophicum did not reveal its presence (180a).

Several important conclusions may be derived from the data available at present: (i) the absence of hsp70 seems to be a characteristic of archaeal species that live at very high temperatures (hyperthermophiles); (ii) in sharp contrast, no hyperthermophilic bacterium has been found yet that lacks the gene; (iii) hsp70 is scattered among methanogenic archaea that are either mesophiles or thermophiles, like M. mazei S-6 and M. thermophila TM-1, but is absent in other methanogens; (iv) whenever hsp70 was present in a genome, hsp40 and grpE were also found if enough sequencing was done; (v) conversely, genome sequencing has demonstrated that if the hsp70 gene is absent, hsp40 and grpE are also absent; and (vi) the gene has been found in two extreme halophiles, but there are no reports of full-genome sequences for this group of organisms, and so it is not possible to be certain about the situation with them. Do they all have hsp70, and also hsp40 and grpE? We know that at least one of them, H. cutirubum, has hsp40 in addition to hsp70 (32). It may be argued that archaea did not have the genes to begin with and that some of them received the genes via lateral gene transfer. However, the observations listed above, particularly (iv) and (v), and other data challenge the lateral-gene-transfer hypothesis, at least in its simplest form.

One must assume that the three genes jumped as a block, or unit, from a bacterium into an archaeon, particularly the pair hsp70 and hsp40. This implies that the unit also carried the intergenic regions. In agreement with the hypothesis, the proteins encoded by the genes, particularly Hsp70 and Hsp40, are of a bacterial type. Against this hypothesis stands the fact that there are no signal sequences of the bacterial type in the intergenic regions. They are typically archaeal. Hence, more data are needed to determine the origin of the archaeal hsp70 locus genes, i.e., archaeal or bacterial, or prearchaeal or prebacterial; and if the origin is archaeal or prearchaeal, more research is necessary to elucidate the mechanism by which these genes came to be in today's species that have them.

Several groups of investigators have used Hsp70 to make phylogenetic trees (22, 99, 105-111, 142, 239). Most of these do not agree with the classical tree based on comparisons of 16-18S rRNA sequences (11, 299, 300). In the rRNA-based tree, archaea and eucarya have a common line up to a point at which they diverge. The archaeal-eucaryal and the bacterial lines are shown to branch off earlier from a primitive, common line. Hsp70-based trees do not support the archaea-eucarya sisterhood or the monophyletic character of archaea suggested by the rRNA-based tree. Some Hsp70 trees suggest a close relationship between archaea and gram-positive bacteria on one side and between eucarya and gram-negative bacteria on the other (105, 106).

The reliability of the trees has been questioned lately. This applies to both rRNA- and protein-based trees (66, 67, 80, 99, 105, 106, 188, 189, 192, 286, 298).

Evidence showing that lateral gene transfer events are more frequent and widespread than was previously realized has been accumulating in the last couple of years (1, 4, 66, 67, 213, 298). Thus, finding that the Hsp70 proteins, for example, of two organisms are very similar does not necessarily mean that the organisms are phylogenetically close. It only means that their Hsp70 molecules are close to each other and have a common ancestor. It does not prove that the ancestor molecule was present in a common ancestor of the two organisms. One of the two organisms may have received its Hsp70 via lateral gene transfer and thus mistakenly appears to be a close relative of the donor's ancestral lineage.

In summary, by studying amino acid sequences, one can follow the natural history of genes and their occurrence in nature, namely, their itinerary, as it were, along the series of organisms in which they are found.

A recent study addresses these points, taking advantage of the fact that considerably more Hsp70 sequences are known now than when previous studies were carried out (99). A systematic search for hsp70 among archaea was performed, and the gene was cloned from Aquifex pyrophilus and Thermotoga maritima. These two bacterial species represent the deepest offshoots in the rRNA-based tree. The gene was not found in 8 of the 10 archaea investigated (Table 13). Alignments of 70 Hsp70 sequences, including the 2 new ones from A. pyrophilus and T. maritima, confirmed the previous observations that the M. mazei S-6 protein clusters with the proteins from the low-G+C gram-positive bacteria while the proteins from the extreme halophiles cluster with the proteins from the high-G+C gram-positive bacteria (Fig. 10). Remarkably, the Hsp70 proteins from T. acidophilum and M. thermoautotrophicum H clustered together (Table 3), along with those from the Aquifexales, Thermotogales, and green nonsulfur bacteria (Fig. 10). The two archaeal Hsp70 proteins in this group appeared to have an ancestor in common with T. maritima. In brief, the Hsp70 from the archaeal species T. acidophilum and M. thermoautotrophicum H did not cluster with the proteins from gram-positive species, as suggested by others, but with bacteria unrelated to the gram positive ones. Another unexpected finding was that the Hsp70 from T. maritima did not have the 23-amino-acid insert characteristic, it was believed, of gram-negative bacteria. Thus, T. maritima Hsp70 possesses a structural feature (i.e., a 23-amino-acid gap in its N-terminal quadrant [Fig. 8]) that is assumed to be distinctive for gram-positive bacteria and archaea, despite the fact that this organism is not a gram-positive bacterium or an archaeon.

View larger version (5343K): [in this window] [in a new window]   FIG. 10.   Maximum-parsimony (A) and evolutionary-distance (B) phylogenetic trees based on Hsp70(DnaK) sequences. Both trees show essentially the same clustering of the archaeal molecules with those from gram-positive bacteria and a group formed by the Aquifexales, Thermotogales, and green nonsulfur bacteria. Numbers represent bootstrap confidence levels calculated from 100 bootstraps (only those that were 30% or higher are shown). Asterisks indicate the newly described genes-proteins (see reference 99). Abbreviations: ac, acetobutylicum; am, amazonensis; av, avium; br, brucei; ca, capricolum; chl, chloroplasts; co, coelicolor; cr, cruzis; cu, cutirubrum; ge, genitalium; gr, griseus; in, infantum; le, leprae; ma, marismortui or major (Leishmania); me, megaterium; mt, mitochondria; NS, nonsulfur; pa, paratuberculosis; pe, perfringens; pn, pneumonia; py, pyogenes; st, stearothermophilus; su, subtilis; tr, trachomatis; and tu, tuberculosis. Reprinted from reference 99 with permission of the American Society for Microbiology.

There are several possibilities to explain these observations. For example, if one assumes that the eucarya and archaea had a common ancestor that contained hsp70, it is possible that both received the gene but some archaea lost it afterwards. A second possibility is that there was no hsp70 in the common ancestor and the gene was acquired after the three lineages separated, with some archaea being excluded. A third possibility is that there was a common ancestor which contained hsp70 and that the three lineages received the gene vertically but the archaeal lineage lost it very early (the species that have the gene today received it via lateral transfer). Finally, if one disregards the common-ancestor idea, another possibility is that the archaea never had the gene while the bacteria and eucarya had it from the beginning. Here, also, archaea that have the gene today must have acquired it by lateral transfer.

The above possibilities and others one might easily think of are more or less improbable depending on (i) how one interprets available data from other studies; (ii) what molecule(s) and criteria were used to generate these data; (iii) what methods were applied to obtain, study, and statistically validate the data; and (iv) what classification scheme was adopted as a master scaffolding to compare with the Hsp70-based tree.

In any case, all the possibilities mentioned share an important characteristic: they stimulate research in this fascinating area of biology and evolution. Molecular phylogeny and detailed analyses of proteins and other macromolecules have already demonstrated their enormous value as tools for research. They are instrumental in uncovering relationships between organisms, the origins of the eukaryotic cell components, the functional meaning of structural motifs, and the role of domains in large proteins of eukaryotes whose ancestors are smaller proteins in more primitive organisms.

FUNCTIONS OF ARCHAEAL MOLECULAR CHAPERONES

Top Previous Next References

Biochemistry

There is little information on the functions of the archaeal Hsp70, Hsp40, and GrpE molecules, as assessed in vitro or in vivo. Since they are so similar in sequence and/or structural features to the homologs from bacteria (Tables 4 and 7; Fig. 9), it may be assumed that both groups of proteins have the same functions and participate in the same mechanisms as molecular chaperones and regulators of their own synthesis.

For example, the bacterial Hsp70(DnaK) is an ATPase and binds ATP and unfolded polypeptides (substrate) (191, 214, 236). Hsp70(DnaK)-ATP has a lower affinity for substrate than Hsp70(DnaK)-ADP. Thus, ATP hydrolysis, which is enhanced by Hsp40(DnaJ) via interaction of its J domain with at least two sites on the Hsp70(DnaK) molecule (87, 270), promotes substrate binding, and the polypeptide is maintained in an extended form, avoiding aggregation. Interaction with GrpE, or nucleotide exchange factor, regenerates the Hsp70(DnaK)-ATP complex, lowers the affinity for the substrate, and releases it (the polypeptide may then be taken by the chaperonin system for final folding (see "Chaperonins" below). Hsp40(DnaJ) is thought to also bind the substrate, before Hsp70(DnaK) does, and to tag the polypeptide so that the Hsp70(DnaK)-ATP complex "sees" it and binds it (176, 191, 236, 272). Also, E. coli DnaK interacts with ribosome-bound trigger factor (62). There is no experimental information on whether the archaeal Hsp70 system operates like that of bacteria and, if so, to what extent the details described above are similar or dissimilar. This is an area that requires investigation and deserves to be explored. It has the potential for revealing how a bacterial-like molecular machine works in a cell whose genome bears eucaryal-like features and probably encodes accessory (regulatory, auxiliary) factors of the eucaryal type while lacking the complementary Hsp60 system of the bacterial type.

It would be of particular interest to explore whether a self-regulating circuit similar to that described for E. coli (references 6, 20, 90, and 259 and references therein), or some variation of it, also operates in archaea. Hsp70(DnaK) in some bacteria participates along with Hsp40(DnaJ) and perhaps also GrpE in the degradation of ³² as a way to down-regulate hsp70(dnaK) transcription. How much of this mechanism operates in archaea? We know that archaea do not have  factors, and so regulatory circuits for Hsp70(DnaK) synthesis cannot include this factor. Does it include another?

We also know that in eukaryotes, the Hsp70 protein intervenes to prevent trimerization of the heat shock factor (HSF) and thus precludes induction of the hsp70 gene (205, 253). Archaea do not have an identifiable HSF, or heat shock element (75), in the hsp70 promoter region (180a). How, then, is the archaeal hsp70 gene regulated? Does Hsp70 participate in this process? If so, how? Does it interact with an archaeal equivalent of the eucaryal HSF or with another kind of regulator?

These are but some of the fascinating questions posed by recent research with the archaeal hsp70 locus genes. The answers to these questions will shed light on the details of the transcription initiation machinery for stress-inducible genes in archaea and will help us to understand the evolution and principles of the transcription mechanisms in the three phylogenetic domains, not just in the Archaea.

CHAPERONINS

Top Previous Next References

Chaperonin Systems I and II

The bacteria have the "bacterial" type (group) I chaperonins, i.e., the genes/proteins groEL/GroEL and groES/GroES (191, 214, 235, 247). A three-dimensional (3-D) view of the barrel-like GroEL/S complex is shown in Fig. 11. What makes the lack of this chaperonin system in archaea very intriguing is that these organisms, at least some of them, have an Hsp70(DnaK) molecular chaperone machine very similar to the bacterial homolog, as described in previous sections. The absence of type I chaperonins and the presence of the Hsp70(DnaK) chaperone machine in a single organism is per