OUP user menu

Alcohol dehydrogenases from thermophilic and hyperthermophilic archaea and bacteria

Helia Radianingtyas , Phillip C. Wright
DOI: http://dx.doi.org/10.1016/S0168-6445(03)00068-8 593-616 First published online: 1 December 2003

Abstract

Many studies have been undertaken to characterise alcohol dehydrogenases (ADHs) from thermophiles and hyperthermophiles, mainly to better understand their activities and thermostability. To date, there are 20 thermophilic archaeal and 17 thermophilic bacterial strains known to have ADHs or similar enzymes, including the hypothetical proteins. Some of these thermophiles are found to have multiple ADHs, sometimes of different types. A rigid delineation of amino acid sequences amongst currently elucidated thermophilic ADHs and similar proteins is phylogenetically apparent. All are NAD(P)-dependent, with one exception that utilises the cofactor F420 instead. Within the NAD(P)-dependent group, the thermophilic ADHs are orderly clustered as zinc-dependent ADHs, short-chain ADHs, and iron-containing/activated ADHs. Distance matrix calculations reveal that thermophilic ADHs within one type are homologous, with those derived from a single genus often showing high similarities. Elucidation of the enzyme activity and stability, coupled with structure analysis, provides excellent information to explain the relationship between them, and thermophilic ADHs diversity.

Keywords
  • Enzyme activity
  • Thermostability
  • Protein structure
  • Directed evolution
  • Rational design

1 Introduction

Thermophilic and hyperthermophilic organisms have the ability to thrive at high temperatures. By inference, the internal components must also be able to withstand this extreme environment. Thermophiles are microorganisms that grow fastest above 40°C [1, 2]. They are found within natural and anthropogenic biotopes such as hot waters, sun-heated oils, geothermal areas, heated waste dumps and cooling waters. Most thermophiles known are moderate, and show an upper temperature range of growth between 50 and 70°C [3], although they are still able to grow slowly at 25–40°C. In general, thermophiles are defined as microorganisms that grow optimally at 50–80°C. Owing to their close phylogenetic relationship to mesophiles and their modest thermophily, moderate thermophiles may be secondarily adapted to hot environments. Thermophilic representatives are known within a wide range of microbial genera.

Microorganisms that grow better at temperatures of 80°C or above are described as hyperthermophiles [4]. These organisms are able to grow not only at high temperatures, but also extremes of pH, redox potential, pressure, and salinity. They have been isolated mainly from water-containing terrestrial and marine high-temperature areas [3]. Hyperthermophiles are present within the archaea and bacteria, where they represent entire deep short lineages [5]. For practical reasons, from this point forward, thermophilic and hyperthermophilic microorganisms will be referred to as thermophiles unless otherwise stated.

Studies of thermophiles have important implications in regard to evolution, the diversity of both microbial metabolism and prokaryotic ecosystems, and the nature of protein stability. The understanding that proteins from thermophiles are inherently thermostable, and active at the organisms’ maximum temperature, was started in the early 1970s with the discovery of thermostable enzymes in Thermus aquaticus [6]. Since then, the interest in enzymes from thermophilic microorganisms is justified by the fact that many enzymes isolated from these microorganisms are thermostable, and capable of acting at high temperatures. Furthermore, these enzymes have also been found to be resistant toward common protein denaturants and organic solvents. These remarkable features from thermophilic enzymes have generated great interest from an industrial perspective, especially concerning their potential biotechnological applications, with many reviews published on thermophiles and biotechnology (e.g. [7, 8]). However, despite their extraordinary stability, many thermophilic enzymes have been shown to have narrow substrate specificities. In search of a more stable enzyme with broader substrate specificities, many studies have been carried out to determine factors and mechanisms that control and affect both activity and stability. It is noteworthy to mention here that one of the most remarkable studies on thermophilic enzymes is the demonstration of hydrogen tunnelling contribution to Bacillus stearothermophilus alcohol dehydrogenase (ADH) chemical reactivity at 65°C [9, 10], which enabled the study of hydrogen tunnelling effects across a sizeable temperature range.

One of the most interesting enzyme groups from thermophilic microorganisms are the ADHs. The considerable interest in their molecular microbiology depends both on their potential biotechnology, and on the possibility of elucidating the mechanism action, the regulation, and the evolutionary relationships of the various enzymes from different sources. ADHs play considerable process and production roles, for example in generation of potable alcohol [11], solvents and acetic acid [12]. ADHs also support the growth of methylotrophs, oxidise alcohols and catalyse lignin degradation [12]. ADHs from thermophiles are of interest for industrial alcohol and enzyme production because these microbes are uniquely suited for direct biomass fermentation to ethanol via reduced-pressure distillation, and yield active thermostable enzymes of commercial interest [13]. There is also considerable interest in the use of such enzymes in the chemical synthesis industry, particularly the pharmaceutical industry, where the production of chiral synthons is an increasingly important step in the synthesis of chirally pure pharmaceutical agents [14, 15]. Thermophilic ADHs are also suitable for synthesising cofactors NAD+ and NADP, used in various processes [16].

Owing to microbial whole genome sequence analysis that allows for information on the putative function of a gene product to be identified by sequence comparisons, there is a lot of information on thermophiles with putative ADHs, complementing studies on experimentally characterised thermophilic ADHs. These data on putative and characterised thermophilic ADHs are valuable for further studies into their diversity. Furthermore, these data can be used to understand more about their stability properties, using both directed evolution and rational design approaches to eventually elucidate structure to function mechanisms.

The current study aims to compare and contrast factors that control and affect stability in thermophilic ADHs, and how these factors further influence enzyme activities, with a view to future potential for increased stability and activity for applied purposes.

2 ADHs from thermophilic bacteria and archaea

To date, there are 20 thermophilic archaeal and 17 thermophilic bacterial strains known to have ADHs or similar enzymes, including the hypothetical proteins. As can be seen in Table 1, these thermophiles have been isolated from natural biotopes such as hydrothermal vents, volcanic marine sediments, hot springs, deep sea chimney walls and solfataric fields, as well as from anthropogenic biotopes such as petroleum reservoirs, canned peas, manure piles, apple waste digestors, beet sugar factories, and coal refuse piles. Some of these thermophiles are known to have multiple ADHs, sometimes of different types. For example, there is one well-characterised and 12 putative ADHs of one type found in Sulfolobus solfataricus [17], while Pyrococcus furiosus has been found to have two different types of ADHs [18]. This is not unusual, considering many organisms contain multiple ADHs and their physiological roles can sometimes be difficult to disentangle [19]. The presence of multiple ADHs within one organism conceivably reflects the environment in which the organism has been exposed and adapted. Therefore, each ADH may have a different role and/or specificity for survival.

View this table:
Table 1

Thermophiles and hyperthermophiles known to have alcohol dehydrogenases

SpeciesOriginTopt (°C)Ae/anIdentified proteinsAcc. no.Abbv.Ref.
Archaea–Crenarchaeota
Aeropyrum pernix strain K1 (ATCC 700893)hot sediment and venting water at a coastal solfataric vent, Kodakara-Jima Island, Japan, 199390strict aeZn-ADHBAA81251AePxADHI[69]
long hypotheticalBAA80556AePxADHII[70]
Zn-ADHBAA80235AePxADHIII
long hypothetical Fe-ADHBAA80973AePxADHIV[70]
Pyrobaculum aerophilum IM2 (DSM 7253)marine waterhole, Maronti Beach, Ischia, Italy100aehypothetical Zn-ADHAAL64544PyAeADHI[71, 72]
AAL64366PyAeADHII
AAL64117PyAeADHIII
AAL63909PyAeADHIV
AAL63819PyAeADHV
Sulfolobus solfataricus P2 (ATCC 35092)volcanic hot spring, Italy75aehypothetical Zn-ADHAAK40795SuSoADH I[17]
AAK41061SuSoADHII
AAK41467SuSoADHIII
AAK41537SuSoADHIV
AAK41859SuSoADHV
AAK42487SuSoADHVI
AAK42584SuSoADHVII
AAK42630SuSoADHVIII
AAK42635SuSoADHIX
AAK42827SuSoADHXI
AAK42913SuSoADHXII
AAK42983SuSoADHXIII
NAD-dependent Zn-ADHAAK42665SuSoADHX[64, 73]
Sulfolobus tokodaii strain 7 (JCM10545)Beppu hot spring, Japan80aelong hypothetical Zn-ADHBAB65565SuToADH I[74]
BAB65008SuToADH II
BAB67688SuToADH III
BAB65032SuToADH IV
BAB64993SuToADH V
BAB67154SuToADH VI
BAB67719SuToADH VII
BAB67362SuToADH VIII
BAB65473SuToADH IX
BAB64994SuToADH X
Sulfolobus sp. RC3solfataric field, Pisciarelli, Italy75aeNAD-dependent Zn-ADHCAA87591SuRC3ADH[40]
Sulfolobus solfataricus MT4 (DSM 5833)acidic hot spring, Agnano, Naples87aeNAD-dependent ADHNA[39]
Sulfolobus solfataricussolfataric field, Pisciarelli, Italy75aeNAD-dependent ADHNA[41]
Archaea–Euryarchaeota
Archaeoglobus fulgidus VC-16 (ATCC 49558)submarine hot spring, Volcano Island, Italy83anhypothetical Zn-ADHAAB89145ArFuADHI[75]
hypothetical Fe-ADHAAB91203ArFuADHII[75]
AAB90896ArFuADHIII
AAB89231ArFuADHIV
Methanothermobacter thermoautotrophicus (ATCC 29096)sewage sludge60anhypothetical propanediol Fe dehydrogenaseAAB85481MeThPdDH[76]
Methanocaldococcus janaschii strain JAL-1 (ATCC 43067)submarine hydrothermal vent; East Pacific Rise85anhypothetical glycerol-1-phosphate dehydrogenaseAAB98707MeJaGDH[77]
Methanoculleus thermophilicus (DSM 3915)apple waste digestor; USA55anF420-dependent ADHCAA77275McThADH[78, 79]
Pyrococcus abyssi strain Orsaydeep sea hydrothermal vent100anhypothetical short-chain dehydrogenase/reductaseCAB49186PyAbADHI[80]
hypothetical Fe-ADHCAB50216PyAbADHII[80]
hypothetical threonine 3 Zn dehydrogenaseCAB50292PyAbThDH[80]
Pyrococcus furiosus (ATCC 43587)volcanic marine sediment, Italy100anNADP-dependent short-chain 2° ADHAAC25556PyFuADHI[18, 81]
NADP(H)-dependent 1° Fe-ADHAAC25557PyFuADHII[81]
O2-sensitive 1° NADP-dependent Fe- and Zn-ADH[45, 46, 82]
Pyrococcus horikoshii OT3 (ATCC 700860)marine hydrothermal vent at a depth of 1395 m in the Okinawa Trough, Pacific Ocean, 199495anlong hypothetical Zn-DHBAA29746PyHoADHI[83]
hypothetical short-chain DHBAA31026PyHoADHII[83]
long hypothetical Fe-ADHBAA29834PyHoADHIII[83]
Thermococcus hydrothermalis AL 662chimney wall, East Pacific Rise85strictly anNADP-dependent 1° Fe-ADHCAA74334THydADH[47]
Thermococcus litoralis (ATCC 51850)marine thermal spring; Italy, Naples, Bay of Lucrino85strictly anNADP-specific 1° ADH fragmentAAB30263ThLiADH[46, 84]
Thermococcus zilligii (previously T. AN1) (ATCC 700529)freshwater hot spring, Kuirau Park, Rotorua, New Zealand, 197675anNADP(H)-dependent 1° Fe-ADHAAB63011ThZiADH[48, 85]
Thermococcus ES-1black smoker chimney polychaete worm (Paralvinella sp.) at Juan de Fuca Ridge91ansulfur-dependent 1° ADH fragmentAAB35052TES1ADH[46, 86]
Thermoplasma acidophilum (DSM 1728)coal refuse pile60aeZn-ADH-related proteinCAC12437TpAcADHI[87]
CAC11970TpAcADHII
ADH-related proteinCAC11176TpAcADHIII[87]
NP394293TpAcADHIV
Thermoplasma volcanium strain GSS1 (DSM 4299)solfatara, Vulcano Island, Italy60anpredicted Zn-ADHBAB60114TpVoADHI[88]
BAB59409TpVoADHV
predicted short-chain ADHBAB59588TpVoADHII[88]
BAB59216TpVoADHIII
predicted Fe-ADHBAB60449TpVoADHVI[88]
BAB59540TpVoADHIV
Bacteria–Aquificae–Aquificales–Aquificaceae
Aquifex aeolicus VF5thermal springs, Yellowstone National Park, WY, USA90fac aehypothetical Zn-ADHAAC07327AqAeADH[89]
hypothetical 1,3-propanediol Fe dehydrogenaseAAC07174AqAePdDH[89]
Bacteria–Firmicutes–Bacillales
Alicyclobacillus acidocaldariusunknown60aeNAD-dependent alcohol-aldehyde oxidoreductaseNA[90]
Geobacillus stearothermophilus NCA 1503 (ATCC 12976)canned peas65anNAD-dependent Zn-ADHBAA14411BS1503ADH[36]
Geobacillus stearothermophilus (DSM 2334)manure pile55fac aeNAD-dependent Zn-ADHCAA80989BSADH[36]
Geobacillus stearothermophilus LLD-R (NCIMB 12403)unknown70fac anNAD-dependent Zn-ADHCAA81612BSLLD-RADH[37, 91]
Geobacillus thermoleovorans B23deep subterranean petroleum reservoirs, Minami-aga and Yabase oil fields70aeNAD-dependent short-chain ADHBAA94092BTlADHI[92]
NAD-dependent Zn-ADHBAB16599BTlADHII[93, 94]
Bacteria–Firmicutes–Clostridia–Thermoanaerobacteriales–Thermoanaerobacteriaceae
Thermoanaerobacter brockii HTD 4 (ATCC 53556)thermal springs, Yellowstone National Park, WY, USA60anNADP-dependent 2° Zn-ADHCAA46053ThBrADH[11, 13, 57, 60, 95, 96]
Thermoanaerobacter brockiihot spring, New Zealand70anNADP-linked 2° ADHNA[97, 98]
Thermoanaerobacter ethanolicus JW200 (ATCC 31937)mud, alkaline hot spring; USA, Yellowstone National Park60anNADP-dependent 1° Fe-ADHAAG01186TEJW200ADH[44]
NADP-dependent 1° Zn-ADHNA[29, 30]
NADP-dependent 2° Zn-ADHNA[29, 30]
Thermoanaerobacter ethanolicus 39E (ATCC 53033)algal–bacterial mat, Octopus Spring, Yellowstone National Park60anNADP(H)-linked Zn-ADHAAB06720TE39EADH[25, 31, 61, 99]
NADP/NADH 1° ADHNA[25, 99]
Thermoanaerobacter tengcongensis MB4T (JCM 11007T)hot spring, Tengcong, China75anuncharacterised Fe-ADHAAM23605ThTeADHI[32]
AAM23958ThTeADHIV
related Zn dehydrogenaseAAM24795ThTeADHIII[32]
AAM23957ThTeADHII
related short-chain ADHAAM23359ThTeADHV[32]
AAM23376ThTeADHVI
AAM24694ThTeADHVII
AAM24982ThTeADHVIII
AAM25384ThTeADHIX
Thermoanaerobacter thermohydrosulfuricus (ATCC 35045)extraction juices of beet sugar factory60anNADP-linked 2° alcohol-aldehyde/ketone oxidoreductaseNA[13]
Thermoanaerobacter thermohydrosulfuricus (ATCC 53016)mud hot springs, New Zealand60anNADP-linked 2° ADHNA[97]
Bacteria–Cyanobacteria–Chroococcales–Thermosynechococcus
Thermosynechococcus elongatus BP-1hot spring50aeprobable Fe-ADHBAC07780TsElADHI[100]
probable dehydrogenaseBAC08522TsElDHII[100]
Bacteria–Thermomicrobia–Thermomicrobiales–Thermomicrobiaceae
Thermomicrobium roseumhot spring70aeNAD-dependent 1° ADHNA[101]
Bacteria–Thermotogales
Fervidobacterium nodosum (ATCC 35602)thermal spring, New Zealand70anNADP-linked 2° ADHNA[97]
Thermotoga maritima strain MSB8 (ATCC 43589)anaerobic marine mud, Vulcano Island, Italy, 198280anprobable Zn-ADHAAD35386ThMaADHI[102]
AAD35497ThMaADHII
AAD35521ThMaADHIII
probable Fe-ADHAAD35205ThMaADHIV[102]
Fe-ADHAAD36001ThMaADHV[50]
probable Fe-butanol dehydrogenaseAAD35902ThMaBuDH[102]
probable short-chain dehydrogenaseAAD35412ThMascDH[102]
  • Topt (°C)=optimum growth temperature in °C; Ae/an=aerobic or anaerobic; strict=strictly; fac=facultative; NA=not available; Acc no=accession number at NCBI (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi); Abbv.=abbreviation of designated protein name; Ref.=references.

Several categories of ADHs can be distinguished based on their cofactor specificity, these being: (i) NAD or NADP, (ii) the pyrrolo-quinoline quinine, haem or cofactor F420, and (iii) FAD. The NAD(P)-dependent ADHs can further be sub-divided into zinc-dependent ADHs, short-chain ADHs, and iron-activated ADHs [12]. Most of the known ADHs from thermophiles are NAD(P)-dependent, with the exception of the Methanoculleus thermophilicus ADH that uses cofactor F420 in the place of NAD(P) (Table 1). This similarity of cofactor utilisation with their mesophilic counterparts indicates that they share a common catalytic mechanism [20, 21].

In order to analyse diversity amongst thermophilic ADHs and related proteins, all the thermophilic ADH amino acid sequences available from public databases (Swiss-Prot, TrEMBL, OWL, PIR, NCBI, PDB and PRF) were compiled and aligned using ClustalW version 1.8 EBI, for the purpose of constructing a phylogenetic tree using the average distance method and bootstrap analysis. The resulting phenogram (Fig. 1) revealed that within the NAD(P)-dependent group, the thermophilic ADHs are orderly clustered according to enzyme type, i.e. zinc-dependent ADHs, short-chain ADHs, and Fe-containing/activated ADHs, irrespective of archaeal or bacterial origin, location of isolation or growth conditions. Using this method, ungrouped thermophilic ADHs can be classified via phylogenetic relatedness into their respective ADH types. For example, Sulfolobus tokodaii ADH sequences predicted from its whole genome sequence, previously ungrouped, can be clustered with known Zn-dependent ADHs. In this manner, previously ungrouped ADHs were able to be assigned, allowing for potential extrapolation of associated physiological and structural properties. It is worth mentioning at this point that ADHs from closely related species are often clustered together (Fig. 1). Interestingly, nucleic acid alignments of the coding region from genes encoding thermophilic ADHs that clustered together showed a comparable degree of relatedness with the amino acid alignments (data not shown), although analysis of the complete data set of thermophilic adh genes did not result in a matching phenogram as the protein. This result implies that only adh genes with similar functions from closely related species can be clustered together.

Figure 1

Phylogenetic tree derived from thermophilic ADH and related protein amino acid sequences extracted from protein databases (Swiss-Prot, TrEMBL, OWL, PIR, NCBI, PDB, PRF). The sequences were aligned using ClustalW and adjusted by gap insertion before being clustered using the average distance method with 1000 bootstrap. Abbreviations are explained in Table 1.

Analysis of thermophilic ADH properties in relation to phenotypic origin, i.e. terrestrial and marine, shows that most marine thermophiles have at least one Fe-ADH, with 50% having more than one type of ADH (Table 2). In contrast, 78% of all terrestrial thermophiles have Zn-dependent ADHs, with 67% of them having only one type of ADH. This variation might be caused by differences between the two environments, for example salinity, pH, and nutrients. In their natural habitats, thermophiles form complex food webs, consisting of primary producers and consumers of organic materials. Their metabolic potential includes various types of aerobic and anaerobic respiration and different modes of fermentation. Marine biotopes, such as hydrothermal systems located at shallow and abyssal depth and at active seamounts, are characterised by relatively high concentrations of salt (about 3% v/v) and by pH values which range from slightly acidic to slightly alkaline (pH 5.0–8.5) [22]. Thermophiles known to have ADHs from marine biotopes were all isolated from similar kinds of environments, were anaerobic, and have optimum growth temperatures ranging from 80 to 100°C. On the other hand, natural terrestrial biotopes are water-containing volcanic areas [4] such as hot springs, mud holes and solfataric fields, which are characterised by low salinity (0.1–0.5%) and pH values ranging from about 0.5 to 9.0 [23]. Anthropogenic biotopes, such as sewage sludges, manure and coal refuse piles, however, are more variable environments. The physical proximity of the organisms, in addition to their phylogenetic proximity [24], may help to explain why most thermophiles from marine biotopes have Fe-ADHs with little observed variety amongst them, while those from terrestrial biotopes mainly have Zn-ADHs and higher variability of ADH types. It is interesting to note, however, that Pyrobaculum aerophilum is the only aerobic Archaeon isolated from marine biotopes, and thus it is the only species in the group that does not contain Fe-ADH. On the other hand, the anaerobic archaea Methanothermobacter thermoautotrophicus, M. thermophilicus, and Thermococcus zilligii AN1 are the only three species isolated from terrestrial biotopes that do not contain Zn-ADHs. This ADH variation within thermophiles from different environments provides evidence that the genetic and metabolic diversity present in high-temperature environments reflects the stress components that characterise the environments, such as the range of pH, oxidation/reduction states, solute concentration, gas composition and mineralogy.

View this table:
Table 2

Thermophiles known to have ADHs isolated from marine and terrestrial biotopes

SpeciesOriginZn-ADHFe-ADHsc ADHF420 ADH
Marine biotopes
Archaeoglobus fulgidus VC-16submarine hot spring
Methanocaldococcus janaschii JAL-1submarine vent
Pyrococcus abyssi st. Orsaydeep sea vent
Pyrococcus furiosusvolcanic marine
Pyrococcus horikoshii OT3marine vent
Pyrobaculum aerophilum IM2marine water hole
Thermococcus hydrothermalis AL662chimney wall
Thermococcus litoralismarine spring
Thermococcus sp. ES-1black smoker
Thermotoga maritima MSB8marine mud
Terrestrial biotopes
Aeropyrum pernix K1coastal solfataric vent
Sulfolobus solfataricus P2hot spring
Sulfolobus solfataricus MT4hot spring
Sulfolobus solfataricussolfataric field
Sulfolobus tokodaii 7hot spring
Sulfolobus sp. RC3solfataric field
Methanothermobacter thermoautotrophicussewage sludge
Methanoculleus thermophilicusapple waste digestor
Thermococcus zilligii AN1freshwater hot spring
Thermoplasma acidophilumcoal refuse pile
Thermoplasma volcanium GSS1solfatara fields
Aquifex aeolicus VF5thermal spring
Geobacillus stearothermophilus DSM 2334mannure piles
Geobacillus stearothermophilus NCA 1503canned peas
Geobacillus thermoleovorans B23oil reservoir
Thermoanaerobium brockii HTD-4hot spring
Thermoanaerobacter ethanolicus JW200hot spring
Thermoanaerobacter ethanolicus 39Ehot spring
Thermoanaerobacter tengcongensis MB4Thot spring
Thermosynechococcus elongatus BP-1hot spring

2.1 Zinc-dependent thermophilic ADHs

Using the clustering method mentioned above, it can be observed in Fig. 1 that more than 50% of known thermophilic ADHs are grouped into the Zn-dependent ADH type, dominated by Sulfolobus spp. ADHs. When the amino acid sequences of the thermophilic ADHs within the Zn-dependent family were compared with the amino acid sequence from a representative of the mesophilic Zn-dependent ADH Zymomonas mobilis (ZyMoADH) (NCBI accession no. A35260) using Sequence Comparator (v 2.0) after being aligned and adjusted by gap insertion, similarities ranging from 23.6% to 65.4% were observed. This indicates that Zn-dependent ADHs from thermophilic and mesophilic prokaryotes are homologous (Table 3). The comparison also shows that thermophilic ADHs derived from a single genus often show very high similarity. PyHoADHI and PyAbThDH, ThBrADH, TE39EADH and ThTeADHII, SuSoADHX, SuRC3ADH and SuToADHIII are the thermophilic Zn-ADHs from one genus with similarities above 90% respectively. This implies that the amino acid sequences, and to some extent the genes responsible for ADH production, are highly conserved during evolution due to phylogenetic proximity and thus seemingly necessary for the organisms’ survival.

Many thermophilic ADHs within this group have been studied comprehensively ever since Lamed and Zeikus [11] in 1980 successfully isolated a Zn-dependent ADH from an anaerobic thermophilic Clostridia Thermoanaerobacter brockii (ThBrADH). The enzyme was found to be most active toward 2° alcohols in the presence of NADP, and retained its activity at high temperature. In a subsequent experiment, it was found that the amino acid sequence of ThBrADH shares 86.9% similarity with a Zn-ADH from the mesophile Clostridium beijerinckii (ClBeADH), despite the 26°C difference in thermal stability and the similarity in activity [25, 26]. This interesting finding led to a series of research studies to investigate the factors that determine the thermal stability of ThBrADH in comparison to ClBeADH. Peretz et al. [27] proposed that the presence of eight additional proline residues in ThBrADH versus ClBeADH (Fig. 2), and the exchange of hydrophilic and large hydrophobic residues in ClBeADH for the small hydrophobic amino acids proline, alanine, and valine in ThBrADH might contribute to the higher thermostability of the T. brockii enzyme. Bogin et al. [26], however, found that only proline substitution for Ser-24, Leu-316 and Ala-347 in ClBeADH increased the enzyme's thermostability, while substituting proline for His-222, Leu-275, and Thr-149 reduced thermal stability parameters. The authors also observed that the thermal stability of each chimeric enzyme was approximately exponentially proportional to the content of the amino acid sequence, indicating that the amino acid residues contributing to the thermal stability of ThBrADH are distributed along the whole protein molecule. In addition to that, comparison of ThBrADH with yeast ADH, regarding their catalytic activity and stability in the presence of a number of miscible and immiscible organic solvents, revealed that non-polar solvents were able to enhance the thermal stability of both proteins [28].

Figure 2

Amino acid alignment of ADHs from T. brockii (ThBrADH), T. ethanolicus 39E (Te39EADH), and T. tengcongensis (ThTeIIADH) in comparison with C. beijerinckii (ClBeADH). The secondary structures of ClBeADH and ThBrADH are also shown, together with catalytic Zn+ binding domain (▾) and residues that interact with ligand (⋅) and metal (*). Accession no. of each protein within NCBI is available in Table 1.

Another anaerobic thermophilic Clostridia, Thermoanaerobacter ethanolicus, was found to have two different types of Zn-dependent ADHs. One was most active toward 1° alcohols, and the other toward 2° alcohols [25, 29, 30]. T. ethanolicus JW200 1° ADH, however, had preferences toward NAD as a cofactor, while the T. ethanolicus 39E 1° ADH preferred NADP, which was also preferred by both 2° ADHs. This finding led to the conclusion that in nature, T. ethanolicus 2° ADHs play a role in ethanol formation, while T. ethanolicus 1° ADHs convert alcohols to aldehydes [29, 30], with differences in the hydrophobic clusters present in the overall 1° ADH and 2° ADH sequences suggesting significant differences in the overall structure of these enzymes [25]. The negatively charged Asp-223, -195 and -215 residues were found to be consistent with the NAD(H) dependence of the horse liver, B. stearothermophilus, and Alcaligenes eutrophus enzymes respectively, whereas the presence of an uncharged residue Gly-198 at the analogous position within the T. ethanolicus, T. brockii, and C. beijerinckii 2° ADH is consistent with the NADP(H) dependence of these enzymes [25]. T. ethanolicus 39E Zn-dependent 2° ADH (TE39EADH) shares 99.4% similarity with ThBrADH and 86.6% with ClBeADH, and was found to be very resistant to thermal or chemical denaturation [31]. Amongst the 12 non-conservative sequence substitutions between the mesophilic ClBeADH and the thermophilic TE39EADH, nine correspond to the introduction of prolines (22, 24, 149, 177, 222, 275, 313, 316 and 347) in the TE39EADH (Fig. 2). All but Pro-313 are also present in ThBrADH [25], suggesting that they have a similar function regarding thermostability, as in ThBrADH. Furthermore, TE39EADH, ThBrADH and ClBeADH are known to have highly similar nucleic acid sequences within both the coding and the preceding untranslated regions [27], suggesting the three genes are homologous. Based on this conclusion, it is very tempting to hypothesise that ThTeADHII, a predicted Zn-dependent dehydrogenase from the whole genome sequence of another anaerobic thermophilic Clostridia Thermoanaerobacter tengcongensis [32], has a similar specificity and thermostability to ThBrADH and TE39EADH, seeing as though the coding region of the gene is very similar (data not shown), the amino acid sequence only differs by eight and 11 residues, and all the proline substitutions are conserved (Fig. 2). The significant involvement of proline residues in the thermostability mechanism within these particular enzymes, which is in agreement with what has been proposed previously [33], leads one to question whether it is a universal protein stabilisation mechanism within thermophilic ADHs, particularly within the Zn-dependent group. However, the result of all thermophilic ADH (and related protein) amino acid sequence alignments using ClustalW has shown that these prolines are not conserved throughout the whole data set, and the number of proline residues within one protein sequence is variable, ranging from 1.21% to 8.22%). Furthermore, it has also been shown previously that insertions, deletions and substitutions of proline are not consistent between thermophilic and mesophilic proteins in general [34]. Thus, phylogenetic proximity and function diversity appears to greatly influence the unique combination of different stabilisation mechanisms within one protein family.

However, similar proteins from similar species, but different strains, do not always adopt similar mechanisms for activity and stability. For example, Geobacillus stearothermophilus strains DSM 2334, NCA 1503 and LLD-R all contain NAD-dependent Zn-ADHs (BSADH, BS1503ADH and BSLLD-RADH respectively) that share 78.9–93.4% similarities amongst them (Table 3). BSADH and BS1503ADH can oxidise a range of alcohols, but only BSADH is active towards methanol [35]. Moreover, Sheehan et al. [35] also revealed that the two enzymes did not have cross-reactivity as shown by enzyme-linked immunosorbent assay. The difference between the two ADHs from different G. stearothermophilus strains was also confirmed by the comparison of their ADH genes as demonstrated by Robinson et al. [36]. BSADH and BS1503ADH protein sequences are identical at 244 out of 349 positions, and the catalytic Zn2+ ligands, the structural Zn2+ ligands and residues associated with the NAD+-dependent proton release step are conserved (Fig. 3). However, although the protein-coding DNAs of the two ADH genes are closely related, it was found that the G. stearothermophilus NCA 1503 adh gene has different terminator and promoter regions, and there was no apparent relationship between the sequences adjacent to the coding regions. The authors [36] suggested that the two proteins have evolved to carry out opposite metabolic steps which require that they are encoded in very different regions of the respective genomes. Regarding protein stability, it has been suggested, based on the amino acid sequence comparison between BSLLD-RADH with BS1503ADH, that the higher stability to temperature, organic solvents and denaturing agents of BSLLD-RADH might be due to amino acid substitutions of Glu-11 Lys within a salt bridge at the N-terminus and Pro-242 Ala at coenzyme binding domain [37, 38]. The difference in optimum growth temperatures and conditions among the three G. stearothermophilus strains (see Table 1) probably explains the diversity among their ADHs. Strain DSM 2334 is a facultative aerobe, whilst strain NCA 1503 is an anaerobe. Strain LLD-R grows optimally at 70°C, whilst the optimum growth temperature for strain NCA 1503 is 65°C.

Figure 3

Amino acid alignment of ADHs from G. stearothermophilus strain DSM 2334 (BSADH), NCA 1503 (BS1503ADH), and LLD-R (BSLLD-RADH) showing conserved region (grey area) and theoretical binding domains for catalytic Zn+ (▾) and structural Zn+ (▿). Accession no. of each protein within NCBI is available in Table 1.

ADHs from Sulfolobus sp., particularly S. solfataricus, are among the most well-characterised ADHs from thermophilic archaea to date. Several different strains of Sulfolobus sp., especially S. solfataricus, have been studied, providing information about diversity of this particular protein inter and intra species. All of the known ADHs from Sulfolobus sp. are of the Zn-dependent type requiring NAD to oxidise a wide range of 1° and 2° alcohols, although the enzymes have preference toward 1° alcohols [3941]. Their activity is inducible, as demonstrated by S. solfataricus strain Gθ which exhibited highest ADH activity when grown in Brock's salt with 0.1% glucose and medium DSM 182 with 1 mM benzaldehyde [41]. Moreover, the expression of its ADH RNA at very early exponential phase, i.e. after only one duplication, was shown to be five-fold higher for cells grown in the presence of benzaldehyde [41].

ADHs from Sulfolobus sp. often show high inter and intra species similarity. Highly conserved ADH genes with some nucleotide substitutions were also observed within the Sulfolobus genus. By using the nucleotide substitutions within the coding sequences and their flanking regions, and various enzyme features, Cannio et al. [40, 41] were able to assign S. solfataricus MT3, MT4, S. solfataricus P2 and S. solfataricus Gθ in a group distinct from S. shibatae, S. acidocaldarius, and Sulfolobus sp. RC3. The amino acid sequence of SuSoADHX from S. solfataricus P2 (DSM 1617) is 96% identical to SuRC3ADH from Sulfolobus sp. strain RC3. However, SuSoADHX was shown to have higher affinity towards benzyl alcohol than ethanol, while SuRC3ADH was shown to have identical activity towards both compounds [40]. Furthermore, the two enzymes were found to exhibit identical thermal resistance, with a half life of 3 h at 85°C, but catalytic efficiency of SuSoADHX was found to be considerably higher at all temperatures below 90°C [40]. These differences in thermal activities and substrate specificities between the two enzymes were hypothesised to correspond to only 17 amino acid replacements out of 347 residues (Fig. 4), seven of which are located in the amino-terminal region that contributes to decreased flexibility and potential of coil-flexible turn formation [40]. It will be interesting to see if the hypothetical ADH from S. tokodaii SuToADHIII also has different thermal and metabolic activities from SuSoADHX and SuRC3ADH, considering that its amino acid sequence shares more than 91% similarity with the latter two enzymes (Fig. 4). Comparison of the amino acid sequences in correlation to the stability and activities of these three enzymes might provide valuable information on residues that determine their diversity.

Figure 4

Amino acid alignment of ADHs from S. solfataricus P2 (SuSoADHX), Sulfolobus sp. RC3 (SuRC3ADH), and S. tokodaii (SuToADHIII). The secondary structure of SuSoADHX is also shown, together with binding domains for catalytic Zn+ (▾) and structural Zn+ (▿) and residues that interact with metal (*). Accession no. of each protein within NCBI is available in Table 1.

2.2 Short-chain thermophilic ADHs

In contrast to the Zn-dependent thermophilic ADH, only 11 thermophilic ADHs are clustered into short-chain ADHs, as shown in Fig. 1. Using a similar comparison method as the thermophilic Zn-dependent ADHs, similarities of 30.5–64.0% were noted in short-chain thermophilic ADHs in comparison to Drosophila melanogaster (NCBI accession no. AAA88817), Clostridium acetobutylicum (NCBI accession no. AE007758), and Pseudomonas aeruginosa short-chain ADH (NCBI accession no. AE004592) except those from Thermotoga maritima (ThMaADHIII) and Thermoplasma volcanium (TpVoADHII) which show lower similarity (most probably because they are about 100 amino acid residues longer than other short-chain ADHs) (Table 4). Within this group, only enzymes from Pyrococcus horikoshii (PyHoADHII) and Pyrococcus abyssi (PyAbADHI) show a high similarity of 91.3%. Short-chain ADHs, especially from prokaryotes, are not as well understood as Zn-dependent ADHs [12]. Not all of the identified enzymes that clustered within this group are ADHs, but rather they have a wide range of substrate specificities and metabolic roles [12]. The only thermophilic short-chain ADH that has been characterised is that from P. furiosus (PyFuADHI). The enzyme was found to contain a well-conserved NADP binding site at the N-terminus as well as residues involved in catalysis, and have the highest specificity toward 2-pentanol [18]. The responsible gene for the P. furiosus short-chain ADH is located in the celB locus that encodes an intracellular β-glucosidase (CelB) and an extracellular β1,3-endoglucanase (LamA) that constitutes an operon with 2 additional genes that encode AdhA, the short-chain ADH, and AdhB, the iron-containing ADH [18]. PyFuADHI, however, appears to be different from the short-chain ADHs from P. horikoshii (PyHoADHII) and P. abyssi (PyAbADHI). The amino acid sequences from the latter two only share 42.3% and 47.5% similarities with PyFuADHI respectively (Table 4). Moreover, while nucleic acid sequence of genes encoding PyHoADHII and PyAbADHI show a high level of similarity, CelB, LamA or AdhB are not observed within the two microorganisms [18], suggesting they are a different type of short-chain ADH from PyFuADHI. Recently, 15 additional genes from P. furiosus that might potentially encode ADHs have been isolated and expressed in Escherichia coli [42]. When exposed to a range of alcoholic compounds, the previously characterised short-chain ADH PyFuADHI was found to have a broader specificity than the other 15, suggesting that each of them might have different metabolic roles [42].

View this table:
Table 4

Similarities between thermophilic short-chain ADHs

DrMescADHClAcscADHPsAescADHThMaADHIIITpVoADHIITpAcADHIIIPyFuADHIPyHoADHIIPyAbADHITpVoADHIIIBTlADHIThMascDH
DrMescADH100.041.045.011.316.330.543.837.041.343.844.546.3
ClAcscADH41.0100.051.811.819.834.846.546.550.853.052.351.5
PsAescADH45.051.8100.011.820.332.350.844.849.356.857.864.0
ThMaADHIII11.311.811.8100.012.512.013.512.013.59.812.312.3
TpVoADHII16.319.820.312.5100.09.017.318.519.018.319.822.3
TpAcADHIII30.534.832.312.09.0100.033.833.536.032.833.533.3
PyFuADHI43.846.550.813.517.333.8100.042.347.546.551.051.3
PyHoADHII37.046.544.812.018.533.542.3100.091.347.042.847.8
PyAbADHI41.350.849.313.519.036.047.591.3100.051.347.351.8
TpVoADHIII43.853.056.89.818.332.846.547.051.3100.049.854.3
BTlADHI44.552.357.812.319.833.551.042.847.349.8100.056.0
ThMascDH46.351.564.012.322.333.351.347.851.854.356.0100.0
  • Sequence data set compiled using ClustalW, sequence length was made consistent by the addition of gaps. Similarity matrix was derived using Sequence Comparator (2.01). Sequence alignment was made via the BLOSUM62 scoring matrix with no penalty for gaps. Short-chain alcohol dehydrogenases from Drosophila melanogaster (DrMescADH), Clostridium acetobutylicum (ClAcscADH), and Pseudomonas aeruginosa (PsAescADH) are included in the table as comparison.

2.3 Fe-containing/activated thermophilic ADHs

Results from amino acid sequence comparison, using the Sequence Comparator (v 2.0), show that the 22 thermophilic ADHs assigned to the Fe-containing/activated group are homologous and compare well with the mesophilic Z. mobilis Fe-ADH II (NCBI accession no. AAA27683), indicating that they share a common ancestor (Table 5). As in the Zn-dependent thermophilic ADHs, those from related species also show high similarities. For example, ADHs from Thermococcus hydrothermalis (THydADH) and Thermococcus ziligii (ThZiADH) are 93.3% similar, ADHs from T. ethanolicus JW200 (TEJW200ADH) and T. tengcongensis (ThTeADHIV) are 81.8% similar, whilst ADHs from P. horikoshii (PyFuADHIII) and P. abyssi (PyAbADHII) share 87.6% similarity. Interestingly, the extent of similarity between Fe-ADHs from different genera, even superfamily, is exceptionally high. PyFuADHII from the Archaeon P. furiosus with ThMaADHIV from the bacterium T. maritima, for example, share a 77.7% similarity, while Fe-ADHs from Thermoanaerobacter sp. (TEJW200 and ThTeADHIV) and Thermococcus sp. (ThZiADH and THydADH) share a 70.5–72.4% similarity. Considering the nucleic acid sequence of the coding regions of these proteins are also similar, it is likely that horizontal gene transfer is responsible for these similarities [43]. Nevertheless, it has been discussed above that high similarity in amino acid sequences, like those among the G. stearothermophilus strains or between SuSoADHX and SuRC3ADH, does not always imply similar thermal and/or metabolic activities. TEJW200ADH, in fact, is a zinc-containing, NADP(H)-dependent homotetramer with the highest activity toward pentanol [44]. TEJW200ADH is 67–69% identical with the Fe-ADH from Thermococcus sp., and clustered together with other thermophilic Fe-ADHs in the phylogenetic tree (see Fig. 1), despite the fact that it contains zinc instead of iron. Indeed, mesophilic Fe-ADHs do not contain Fe after purification; rather Fe must be present in the assay mixture to obtain ADH activity. These enzymes are therefore termed Fe-activated ADHs. In contrast, the hyperthermophilic ADHs from Thermococcus sp. strain ES-1 (TES1ADH) and T. litoralis (ThLiADH) have been shown to contain Fe after purification and therefore they are referred to as Fe-containing ADHs [45].

The amino acid sequences at the N-terminal of the four Thermococcus ADHs, namely THydADH, ThZiADH, ThLiADH and TES1ADH, were shown to be very closely related and genus-specific [46]. Therefore, Ma et al. subsequently hypothesised that the 5′ end of Thermococcus ADH genes could be utilised as a specific genetic marker for the Thermococcus genus [46]. THydADH and ThLiADH were shown to be strictly specific for NADP as an electron carrier, while TES1ADH and ThZiADH were able to use NAD with much less preference [4648]. All Thermococcus ADHs were reported as able to catalyse the oxidation of a range of aliphatic and aromatic primary alcohols, but not secondary alcohols. However, THydADH was found to be the only one that oxidised methanol and glycerol [47]. Thermophilicities of Thermococcus ADHs were shown to be similar, i.e. from 80°C to 95°C, but their thermostabilities are quite different [4648]. TES1ADH was found to be the most stable enzyme, requiring about 35 h at 85°C and 15 min at 96°C for a 50% loss in activity [46], followed by THydADH and ThZiADH with a half life of about 15 min at 80°C [47, 48]. However, the oxidation reaction was reported to have a higher thermostability than the reduction reaction [46]. Therefore, the authors concluded that the more polymerised state of the protein at alkaline pH partially overcomes the thermal destabilisation of the enzyme molecule, as reflected by the pH optimum for alcohol oxidation, i.e. 8.8 for ThLiADH, 8.8–11.1 for TES1ADH, 6.8–7.0 for ThZiADH, and 10.5 for THydADH [4648]. Interestingly, unlike other Thermococcus ADHs, the enzyme structure of THydADH is pH-dependent, being a non-active 197-kDa tetramer at pH 10.5, the pH optimum for alcohol oxidation, and an active 80.5-kDa dimer at pH 7.5, the pH optimum for aldehyde reduction [47]. The affinity of the enzyme was found to be greater for the aldehyde substrates and NADPH cofactor, suggesting that the dimeric form of the enzyme is probably the active form in vivo [47].

The Thermococcus ADHs were also shown to have an affinity towards aldehydes and NADPH, and thus it was hypothesised that aldehyde reduction must be the biological action of Thermococcus ADHs [46] in agreement with Fernandez et al. [49], who previously emphasised that type III ADHs are involved in aldehyde detoxication in microorganisms rather than in alcohol turnover. TES1ADH and ThZiADH were proved to require elemental sulphur to be active [46, 48], while ThLiADH activity was reported to be unaffected by the presence or absence of elemental sulphur, despite the high similarity between ThLiADH and TES1ADH amino acid sequences at the N-terminal [46]. It was also observed that the iron in TES1ADH was bound tightly, while the iron from ThLiADH was lost more readily [46]. It is very interesting to find that ThLiADH is clustered together with thermophilic Zn-dependent ADHs, although it did not contain Zn. Further studies to elucidate the diversity of the Thermococcus ADHs would be required to understand the responsible factors.

P. furiosus was found to have a unique oxygen sensitive Fe- and Zn-containing ADH with a significant sequence similarity to the ADHs from the Thermococcus genus, and catalysed the oxidation of a range of aliphatic and aromatic primary alcohols at 80°C with NADP as the electron acceptor [45]. Ma and Adams [45] hypothesised that the physiological role of the P. furiosus oxygen sensitive ADH was more likely to be in aldehyde reduction than alcohol oxidation with NADP(H) as the preferred cofactor, as with the Thermococcus ADHs. The authors also noted the high similarity between the N-terminal amino acid sequences of the ADHs from Thermococcus species and P. furiosus. Two potential sites to bind metal ions were found within the P. furiosus oxygen sensitive ADH, predictably one for Fe and the other for Zn that was expected to have a structural role that might enhance the thermostability [45]. The presence of a binding domain to iron and NADP was also observed within an ADH from T. maritima, ThMaADHV [50], confirming that the thermophilic Fe-ADHs are more Fe-containing ADHs, rather than Fe-activated as with the mesophilic counterparts, and NADP is the more preferred cofactor by thermophilic ADHs within this group.

3 Structures of thermophilic ADHs

It has been hypothesised that thermophilic enzyme structures are similar to their mesophilic counterparts, excluding phylogenetic variations, for several reasons:

  1. The amino acid sequences from thermophilic proteins are typically 40–85% similar to mesophilic proteins [25], showing homologous proteins (with sequence identity higher than 40% for long alignments) to have similar conformations [51, 52] with exchanges at certain positions being involved in substrate and coenzyme binding, and roles in the proton relay system [12].

  2. Three-dimensional (3D) structures of thermophilic and mesophilic proteins are superposable [53].

  3. Thermophilic and mesophilic proteins share the same catalytic mechanisms [53].

Thus, scientists are left with the hypothesis that thermostability properties of the proteins from thermophilic organisms are not as a result of major structural differences from their mesophilic counterparts. Several factors have been suggested to contribute to the thermostability property of thermophilic enzymes, and this topic has been extensively reviewed elsewhere [54]. In general, these stabilising factors may arise due to amino acid composition and intrinsic properties [33, 34, 55] and/or structural organisation [34, 53, 5659].

Nevertheless, each thermozyme is stabilised by a unique combination of different mechanisms [53]. Different protein families adapt to higher temperatures by different sets of structural devices. For example, Szilagyi and Zavodszky [58] found that the only significant differences between mesophilic and thermophilic proteins are found in the number of ion pairs, because most properties are utilised for stabilisation only in some protein families, and when the differences in a property are evaluated for the whole unified data set then the stabilising changes found in these families are compensated for by opposite changes in other families. The authors also found that proteins from extreme thermophiles are stabilised in different ways to moderately thermophilic ones. For example, a significant decrease in the number of cavities was observed in extremely thermophilic proteins but not in moderately thermophilic ones [58]. On the other hand, moderately thermophilic proteins were shown to have a significant increase in the polarity of their exposed surface, which was not observed in extremely thermophilic ones. Furthermore, it was observed that proteins from moderately thermophilic microorganisms contained significant increases in α-helices, whilst those from extremely thermophilic ones contained significant increases in β-strands. Their amino acid composition was also shown to be considerably different: in moderately thermophilic species, protein lysine content decreases and arginine content increases, while in extremely thermophilic proteins there is an increase in the percentage of all charged residues, including lysine [58].

To date, there are five thermophilic ADH crystal structures deposited within the Protein Data Bank (Table 6). The most well-characterised are those from T. brockii (PDB 1YKF and 1BXZ) (Fig. 5) as they have been extensively analysed and compared with the crystal structure of the highly homologous ADH from the mesophile C. beijerinckii (PDB 1JQB, 1KEV, 1PED) [56, 59, 60]. The specificity of both enzymes towards NADP(H) was found to be determined by residues Gly-198, Ser-199, Arg-200, and Tyr-218, with the latter three making hydrogen bonds with the 2′-phosphate oxygen atoms of the cofactor. While either ThBrADH or ClBeADH contain structural Zn, the catalytic Zn in apo ClBeADH was found to be tetracoordinated by side chains of residues Cys-37, His-59, Asp-150, and Glu-60. In holo ClBeADH the latter residue was found to be retracted from Zn in three of the four monomers, whereas in holo ThBrADH Glu-60 was not found to participate in Zn coordination. It was also observed that in the two holo enzymes residue Ser-39 and Ser-113 were in the second coordination sphere of the catalytic zinc [60]. Interestingly, the highly homologous thermophilic ADH to ThBrADH, TE39EADH, was proposed to bind its catalytic Zn atom differently, i.e. using a sorbitol dehydrogenase-like Cys-His-Asp motif [61]. In relation to that, Heiss et al. [62] performed a mutation of Cys-295 to alanine in TE39EADH to give C295A 2° ADH, on the basis of molecular modeling studies utilising the X-ray crystal structure coordinates of ThBrADH. The C295A mutation was found to cause a significant shift of enantioselectivity toward the (S)-configuration in the reduction of some ethynylketones to the corresponding chiral propargyl alcohols, which confirmed the prediction that Cys-295 was part of a small alkyl group binding pocket whose size determines the binding orientation of ketone substrates, and hence, the stereochemical configuration of the product alcohol. The ADH from C295A mutation was also found to have a much higher activity towards t-butyl and some α-branched ketones than the wild-type TE39EADH [62].

View this table:
Table 6

Thermophilic ADH crystal structures within the Protein Data Bank

Acc. no.SpeciesAbbrev.FormStructureBinding domainThermostabilising factorRef.
1YKFT. brockiiThBrADHholo; chain A, B, C, Dhomotetramercatalytic Zn and NADPadditional ion pairs, hydrogen bonds, improved stability of α-helices, tighter molecular packing, Pro at strategic location[56]
1BXZT. brockiiThBrADHholo; chain A, B,homotetramercatalytic Zn and NADPmore hydrophilic exterior and hydrophobic interior, smaller[59]
1JVBS. solfataricusSuSoADHXapo; chain Ahomotetramercatalytic Zn, structural Zn and NADnetwork of ion pairs, additional aromatic clusters, Glu at metal binding site, high Arg/Lys ratio; lower Cys residues[64]
1J5RT. maritimaThMaADHVapo; chain A, Bcatalytic Fe and NADP[50]
1H2BaA. pernixAePyADHIapo[69]
  • Acc. No.=accession number in Protein Data Bank; Abbrev.=abbreviation of the protein name as described in Table 1; Ref.=references.

  • aOn hold.

Figure 5

3D structure of ADHs from (A) T. brockii (ThBrADH) and (B) S. solfataricus (SuSoADHX) as shown in the Protein Data Bank (http://www.rcsb.org/pdb) with accession code 1YKF and 1JVB respectively. A: Four chains of homotetramer structure of ThBrADH in holo form are shown, each chain containing catalytic Zn+. B: One chain of homotetramer structure of SuSoADHX in apo form is shown, containing catalytic and structural Zn+.

Several factors such as a more hydrophilic exterior and hydrophobic interior, a smaller surface area, more prolines and alanines but fewer serines, more salt bridges, hydrogen bonds, and hydrophobic interactions, [59], and improved stability of α-helices and tighter molecular packing [56] have been suggested to contribute to the thermal stability of ThBrADH (Table 6). Mutational analysis of structural elements at critical locations revealed that enhanced thermal stability of ThBrADH is mainly due to the strategic placement of structural determinants at positions that strengthen the interface between its subunits [56]. This, in particular, is the proline residues located at a P-turn and a terminating external loop in the polypeptide chain [63].

The crystal structure of SuSoADHX (PDB 1JVB) (Fig. 5) was shown as a NAD(H)-dependent homotetrameric enzyme that uniquely contains two zinc atoms per subunit, with a catalytic and structural role respectively (Table 6) [64]. Despite the rather low level of sequence identity with other medium-chain ADHs, it was revealed that the structurally and functionally important residues and adjacent regions are conserved or conservatively substituted [65, 66]. Surprisingly, the closest structural homologue of SuSoADHX was found to be the NADP(H)-dependent ketose reductase from Bemisia argentifolii (PDB 1E3J), which shares only 26% sequence identity. Reaction with iodoacetate was observed to increase the oxidation rate of aliphatic and aromatic alcohols 25-fold and decreases the reduction rate of aromatic aldehydes, but the modified enzyme observed was found to acquire some mesophilic character, i.e. being more active at low temperatures and showing a temperature optimum at about 65°C, where its specific activity reaches a maximum value before decreasing abruptly [67].

A number of structural differences between SuSoADHX and other ADHs were significantly observed [64], i.e.:

  1. The number of Cys residues in SuSoADHX subunit is lower than in horse liver or yeast ADHs.

  2. Unlike ClBeADH or ThBrADH, SuSoADHX contains structural zinc, and SuSoADHX thermostability is partially due to the presence of glutamate in its structural metal binding site.

  3. Arg/Lys ratio was found to be about twice as high as mesophilic ADHs.

  4. Loop 46–63 was found to be significantly longer and protrudes towards the coenzyme binding domain, forming interactions with the shorter opposite loop (residues 270–275).

  5. The structural lobe in the amino-terminal part of the catalytic domain was observed to have a different orientation.

  6. In the Rossmann fold, the loops connecting the secondary structural elements and at the region 276–288 were found to contain strands βS and helix αF which are involved in the dimer formation.

  7. Location of the α4 helix was found to be closer to the coenzyme binding domain.

  8. The enzyme was found to have three pairwise interacting aromatic residues close to the active site and form an aromatic cluster, which supposedly was one of the major factors that contribute to thermal stability and thermophilicity of thermophilic enzymes [68].

More recently, the crystal structure of the Zn-ADH from Aeropyrum pernix (AePxADHI) has been determined, and was found to be 39% identical to SuSoADHX in apo form [69]. Like SuSoADHX, the enzyme was also found to contain two Zn ions, each was involved in catalytic and structural roles respectively [69]. The only thermophilic Fe-containing/activated ADH of which the crystal structure has been determined is that from T. maritima (ThMaADHV) [50]. However, the detailed analysis of the enzyme is currently unavailable.

4 Conclusions and perspectives

It is apparent from the discussion above that diversity within thermophilic ADHs was influenced by differences in physical environment where the microorganisms were originally isolated, apart from phylogenetic proximity. Both factors determine the metal and cofactor utilisation, metabolic activity, and stabilisation strategy adopted by these enzymes, despite highly homologous amino acid sequences. Within the Zn-dependent ADHs, three types of ADHs have been found: (i) NAD-dependent 1° ADHs; (ii) NADP-dependent 1° ADH; and (iii) NADP-dependent 2° ADHs, with the latter being the most well-characterised. In contrast, short-chain ADHs within prokaryotes, especially from thermophiles, are understudied with a lot more research required to elucidate their characteristics. It is noteworthy, however, that despite the wide range of substrate specificities and metabolic roles of the protein clustered within this group, the NADP binding site and some residues that are involved in catalysis are well-conserved. Within the Fe-ADH group, a wider variety has been observed. In contrast to mesophilic Fe-ADHs, most of the thermophilic Fe-ADHs contain iron, although one contains zinc, and another contains both iron and zinc with a unique feature of extremely oxygen sensitivity. However, in line with the mesophilic Fe-ADHs, all of them were found to have a preferential selection towards NADP, and are most active towards 1° alcohols.

Currently, limited information is available regarding the genes encoding thermophilic ADHs, making their analysis much more complicated than the protein. Whereas there has been some evidence that coding and flanking regions of adh genes within comparable loci could be utilised as a specific gene marker in a genus, further study involving bench experiments as well as bioinformatics analyses is necessary to further determine the physiological function of the protein.

Characterisation of crystal structures of Zn-ADHs from T. brockii and S. solfataricus suggested that while some stabilising strategies were generally adopted, such as networks of ion pairs, other strategies may not be adopted by all thermophilic ADHs, and thus there is no apparent universal thermostabilising strategy observed within one group. More studies on thermophilic ADHs protein neighbourhoods are still required to elucidate their function and stability mechanisms. The availability of a more thorough structure database of highly thermostable proteins and, in particular, of enzymes belonging to the same family inter and intra genus, even superfamilies, would provide further information on the different routes towards macromolecular stabilisation explored in the course of evolution.

Acknowledgements

The authors thank the UK Biotechnology and Biological Sciences Research Council (BBSRC) for financial support (97/E11124). P.C.W. thanks the Engineering and Physical Sciences Research Council (EPSRC) for provision of an Advanced Research Fellowship (GR/A11311/01).

Footnotes

  • 1 Up to August 2002.

  • 2 Comparison of amino acid residue composition from all thermophilic ADH amino acid sequences using the property profile option in Omiga bioinformatics software package version 2.0.1 (Accelrys, Boston, MA, USA

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
  77. [77].
  78. [78].
  79. [79].
  80. [80].
  81. [81].
  82. [82].
  83. [83].
  84. [84].
  85. [85].
  86. [86].
  87. [87].
  88. [88].
  89. [89].
  90. [90].
  91. [91].
  92. [92].
  93. [93].
  94. [94].
  95. [95].
  96. [96].
  97. [97].
  98. [98].
  99. [99].
  100. [100].
  101. [101].
  102. [102].
View Abstract