New to the CAZy classification? Read this first.
Want to learn more about CAZypedia? Read the CAZypedia 10th anniversary article in Glycobiology.

Difference between revisions of "Glycoside Hydrolase Family 57"

From CAZypedia
Jump to navigation Jump to search
m (Text replacement - "\^\^\^(.*)\^\^\^" to "$1")
 
(134 intermediate revisions by 3 users not shown)
Line 1: Line 1:
<!-- CURATORS: Please replace the {{UnderConstruction}} tag below with {{CuratorApproved}} when the page is ready for wider public consumption -->
+
{{CuratorApproved}}
{{UnderConstruction}}
+
* [[Author]]: [[User:Stefan Janecek|Stefan Janecek]]
* [[Author]]: ^^^Stefan Janecek^^^
+
* [[Responsible Curator]]:  [[User:Stefan Janecek|Stefan Janecek]]
* [[Responsible Curator]]:  ^^^Stefan Janecek^^^
 
 
----
 
----
  
Line 18: Line 17:
 
|-
 
|-
 
|'''Active site residues'''
 
|'''Active site residues'''
|known/not known
+
|known
 
|-
 
|-
 
|{{Hl2}} colspan="2" align="center" |'''CAZy DB link'''
 
|{{Hl2}} colspan="2" align="center" |'''CAZy DB link'''
 
|-
 
|-
| colspan="2" |http://www.cazy.org/fam/GH57.html
+
| colspan="2" |{{CAZyDBlink}}GH57.html
 
|}
 
|}
 
</div>
 
</div>
Line 28: Line 27:
  
 
== Substrate specificities ==
 
== Substrate specificities ==
The family GH57 was established in 1996 (Henrissat & Bairoch 1996) based on the existence of the sequences of two “α-amylases” that were dissimilar to typical family GH13 α-amylases (MacGregor et al. 2001). The two were the heat-stable eubacterial amylase from Dictyoglomus thermophilum known from 1988 (Fukusumi et al. 1988) and the extremely thermostable archaeal amylase from Pyrococcus furiosus determined in 1993 (Laderman et al. 1993a).
+
[[Glycoside hydrolase]] family 57 was established in 1996 <cite>Henrissat1996</cite> based on the existence of the sequences of two “α-amylases” that were dissimilar to typical family [[GH13]]  α-amylases <cite>MacGregor2001</cite>. The two were the heat-stable eubacterial amylase from ''Dictyoglomus thermophilum'' known from 1988 <cite>Fukusumi1988</cite> and the extremely thermostable archaeal amylase from ''Pyrococcus furiosus'' determined in 1993 <cite>Laderman1993a</cite>.
  
The family has expanded mainly due to running genome sequencing projects. Nowadays it contains more than 400 members; all originating from prokaryotes (http://www.cazy.org/fam/GH57.html). With regard to the enzyme specificities, the family GH57 covers the α-amylase (EC 3.2.1.1), α-galactosidase (EC 3.2.1.22), amylopullulanase (EC 3.2.1.1/41), branching enzyme (EC 2.4.1.18) and 4-α-glucanotransferase (EC 2.4.1.25). It is worth mentioning that the two constituent members, i.e. the “α-amylases” from D. thermophilum and P. furiosus are rather the 4-α-glucanotransferases since the former was later proven to have the transglycosylating activity (Nakajima et al. 2004), whereas the latter was shown already in 1993 to exhibit the 4-α-glucanotransferase activity (Laderman et al., 1993b). And it is also of interest that the real enzymes form only about 5% of the family members. The vast majority of the GH57 are hypothetical proteins.
+
The family has expanded mainly due to the results of genome sequencing projects. More than a thousand of members are known <cite>Cantarel2009,Lombard2014</cite>, all from prokaryotes (with ~1:4 ratio for Archaea:Bacteria) <cite>Blesak2012</cite>. The enzyme specificities of family GH57 includes α-amylase (EC [{{EClink}}3.2.1.1 3.2.1.1]), α-galactosidase (EC [{{EClink}}3.2.1.22 3.2.1.22]), amylopullulanase (EC [{{EClink}}3.2.1.41 3.2.1.41]), branching enzyme (EC [{{EClink}}2.4.1.18 2.4.1.18]) and 4-α-glucanotransferase (EC [{{EClink}}2.4.1.25 2.4.1.25]).
 +
 
 +
An archaeal GH57 amylopullulanase from ''Staphylothermus marinus'' has been described exhibiting also the activity of cyclodextrinase (EC [{{EClink}}3.2.1.54 3.2.1.54]) <cite>Li2013</cite>. Based on a preliminary observation that the PF0870 protein encoded in the ''Pyrococcus furiosus'' genome produced maltose <cite>Comfort2008</cite>, a group of GH57 members with proposed specificity of maltogenic amylase (EC [{{EClink}}3.2.1.133 3.2.1.133]) was predicted <cite>Blesak2013</cite> together with another group of non-specified amylases following the partial characterization of an amylolytic enzyme from an uncultured bacterium <cite>Wang2011</cite>. The maltogenic specificity (or maltose-forming amylase) has already been definitively confirmed by both biochemical <cite>Jung2014,Jeon2014</cite> and structural <cite>Park2014</cite> studies. The family GH57 contains further a group of probably non-enzymatic members, i.e. the so-called α-amylase-like proteins having a substitution in one or both catalytic residues <cite>Janecek2011</cite>.
 +
 
 +
It is worth mentioning that the two founding members, i.e. the “α-amylases” from ''Dictyoglomus thermophilum'' and ''Pyrococcus furiosus'' are 4-α-glucanotransferases; the former was proven to have transglycosylating activity <cite>Nakajima2004</cite>, whereas the latter was already known to exhibit also the 4-α-glucanotransferase activity <cite>Laderman1993b</cite>.
 +
 
 +
A detailed in silico study has indicated a close relatedness of family GH57 to the family GH119 concerning catalytic domain fold, catalytic machinery and conserved sequence regions <cite>Janecek2012</cite>.
  
 
== Kinetics and Mechanism ==
 
== Kinetics and Mechanism ==
Content is to be added here.
+
A direct evidence for the stereochemical outcome of the enzyme-catalyzed reaction has been presented for the branching enzyme from ''Thermus thermophilus'' <cite>Palomo2011</cite> by the NMR analysis of reaction products. The previous observation of a trapped covalent glycosyl-enzyme intermediate <cite>Imamura2001</cite> was also a strong evidence that the GH57 enzymes operate with [[retaining|retention]] of anomeric configuration through a [[classical Koshland retaining mechanism]] with retention of anomeric configuration. The X-ray structure of the 4-α-glucanotransferase from ''Thermococcus litoralis'' revealed average distances of 6.72 Å and 6.97 Å between the [[catalytic nucleophile]] (Glu123) and [[general acid/base]] (Asp214) in the free and acarbose-bound forms of the enzyme, respectively, which also supports a retaining mechanism for this family <cite>Imamura2003</cite> (see also <cite>Davies1995</cite>).
 +
 
 +
Detailed kinetic studies have been performed on several GH57 enzymes, including the 4-α-glucanotransferases from ''Thermococcus litoralis'' <cite>Imamura2001,Imamura2003</cite> and ''Pyrococcus furiosus'' <cite>Tang2006</cite>, amylopullulanases from ''Thermococcus hydrothermalis'' <cite>Zona2004</cite> and ''Pyrococcus furiosus'' <cite>Kang2005</cite>, and branching enzymes from ''Thermococcus kodakaraensis'' <cite>Murakami2006</cite> and ''Thermus thermophilus'' <cite>Palomo2011</cite>.
  
 
== Catalytic Residues ==
 
== Catalytic Residues ==
In addition, the sequences of GH57 members are extremely diversified. Certain sequences are shorter than 400 residues whereas others are longer than 1,500 residues (Janecek 2005). This complicated the previous efforts to align the GH57 sequences using the routine alignment programs. Based on a detailed bioinformatics study focused on all available GH57 sequences at that time, five conserved sequence regions in the family GH57 (Fig. 1) were identified and proposed by Zona et al. (2004). This was possible to achieve since the catalytic nucleophile (Glu123) in the GH57 4-α-glucanotransferase from Thermococcus litoralis (Imamura et al. 2001) was known together with its three-dimensional structure (Imamura et al. 2003; PDB: 1k1w) that revealed also the proton donor (Asp214). Thus the first catalytic machinery and the first three-dimensional structure for a GH57 member (Fig. 2) were those of the archaeal 4-α-glucanotransferase from T. litoralis (Imamura et al. 2001, 2003).
+
The sequences of GH57 members are very diverse. Some sequences are shorter than 400 residues whereas others are longer than 1,500 residues <cite>Janecek2005</cite>. This complicated early efforts to align the GH57 sequences using standard alignment methods. A detailed bioinformatics study by Zona ''et al.'' <cite>Zona2004</cite> focused on all available GH57 sequences at that time and identified five conserved sequence regions. This study used knowledge of the identity of the [[catalytic nucleophile]] (Glu123) in the GH57 4-α-glucanotransferase from ''Thermococcus litoralis'' <cite>Imamura2001</cite> together with the three-dimensional structure <cite>Imamura2003</cite> (PDB ID [{{PDBlink}}1k1w 1k1w]) that revealed the [[general acid/base]] residue (Asp214).
  
The catalytic nucleophile (a glutamate) and proton donor (an aspartate) are located in the conserved sequence regions 3 and 4, respectively (Fig. 1). In addition to T. litoralis 4-α-glucanotransferase, they were identified also in the amylopullulanases from Thermococcus hydrothermalis (Zona et al. 2004) and P. furiosus (Kang et al. 2005). The catalytic nucleophile was confirmed also in the α-galactosidase from P. furiosus although without success to find the catalytic proton donor (van Lieshout et al. 2003). It should be taken into account, however, that some GH57 members, which are only hypothetical enzymes/proteins without any biochemical characterization, may lack one or even both catalytic residues (Zona et al. 2004).
+
The [[catalytic nucleophile]] (a glutamate) and [[general acid/base]] (an aspartate) are located in the conserved sequence regions 3 and 4, respectively. In addition to assignments in ''Thermococcus litoralis'' 4-α-glucanotransferase discussed above, these residues were identified in the amylopullulanases from ''Thermococcus hydrothermalis'' <cite>Zona2004</cite> and ''Pyrococcus furiosus'' <cite>Kang2005</cite>. The [[catalytic nucleophile]] was also identified in the α-galactosidase from ''Pyrococcus furiosus'' although no success was achieved in assigning the [[general acid/base]] <cite>vanLieshout2003</cite>. It should be taken into account that some GH57 members, which are only hypothetical enzymes/proteins without any biochemical characterization, may lack one or even both catalytic residues <cite>Zona2004,Janecek2011</cite>.
  
Based on the five identified conserved sequence regions (Fig. 1), the residues His13, Glu79, Glu216, Asp354 together with the Trp120, Trp221 and Trp357 (T. hydrothermalis 4-α-glucanotransferase numbering) were postulated (Zona et al. 2004) as eventually important for the individual GH57 enzyme specificities (Fig. 3). Of these, the Trp221 has already been confirmed to contribute to the transglycosylation activity of 4-α-glucanotransferase since the mutant W229H of the enzyme from P. furiosus exhibited markedly decreased transglycosylation activity in comparison with the wild-type counterpart (Tang et al. 2006).
+
Based on the five identified conserved sequence regions, the residues His13, Glu79, Glu216 and Asp354 together with Trp120, Trp221 and Trp357 (''Thermococcus hydrothermalis'' amylopullulanase numbering) were postulated <cite>Zona2004</cite> as important determinants of the individual GH57 enzyme specificities <cite>Janecek2011,Blesak2012,Blesak2013</cite>. Of these, the Trp221 has already been confirmed to contribute to the transglycosylation activity of 4-α-glucanotransferase since the mutant W229H of the enzyme from ''Pyrococcus furiosus'' exhibited markedly decreased transglycosylation activity in comparison with the wild-type counterpart <cite>Tang2006</cite>.
  
 
== Three-dimensional structures ==
 
== Three-dimensional structures ==
The structure of the catalytic domain adopts a (β/α)7-barrel, i.e. the irregular (β/α)8-barrel called also a pseudo TIM-barrel (Fig. 2) that, in the case of the T. litoralis 4-α-glucanotransferase (Imamura et al. 2003) is succeeded by the C-terminal non-catalytic domain consisting of β-strands only adopting a twisted β-sandwich fold (Fig. 2). In the three-dimensional structure of the α-amylase AmyC from Thermotoga maritima (Dickmanns et al. 2006; PDB: 2b5d), the corresponding catalytic (β/α)7-barrel is followed by a five-helix domain C and a small helical domain B protruding out of the catalytic pseudo TIM barrel in the place of the loop 2 (i.e. succeeding the strand β2). This structure was found to be most closely similar to that of the GH57 member of unknown function from Thermus thermophilus (PDB: 1ufa). In all cases, the catalytic glutamic acid and aspartic acid residues are located near the C-terminal ends of the strands β4 and β7 of the barrel, respectively (Imamura et al. 2003, Dickmanns et al. 2006). There was also a crystallization report in 1995 on a probable GH57 amylopullulanase from Pyrococcus woesei (Knapp et al. 1995), but the detailed crystallographic analysis of this protein has not been published as yet.
+
The structure of the catalytic domain adopts a (β/α)7-barrel, i.e. the irregular (β/α)8-barrel called a pseudo TIM-barrel. In the case of the ''Thermococcus litoralis'' 4-α-glucanotransferase <cite>Imamura2003</cite> the catalytic domain is succeeded by a C-terminal non-catalytic domain consisting of only β-strands that adopts a twisted β-sandwich fold. In the three-dimensional structure of the α-amylase AmyC from ''Thermotoga maritima'' <cite>Dickmanns2006</cite> (PDB ID [{{PDBlink}}2b5d 2b5d]) (note that the sequence of this enzyme strongly resemles that of a branching enzyme <cite>Blesak2012</cite>), the corresponding catalytic (β/α)7-barrel is followed by a five-helix domain C, a small helical domain B being protruded out of the catalytic pseudo TIM barrel in the place of the loop 2 (i.e. succeeding the strand β2). This structure was found to be closely similar to that of the GH57 member of unknown function from ''Thermus thermophilus'' (PDB ID [{{PDBlink}}1ufa 1ufa]). Two structures of biochemically characterized branching enzymes were solved and published: one from ''Thermus thermophilus'' <cite>Palomo2011</cite> (PDB ID [{{PDBlink}}3p0b 3p0b]) and the other one from ''Thermococcus kodakaraensis'' <cite>Santos2011</cite> (PDB ID [{{PDBlink}}3n8t 3n8t]). The former study <cite>Palomo2011</cite>, importantly, is the first report to prove that a retaining mechanism is employed in the family GH57. Also, the 3-D structure is available for the maltogenic (maltose-forming) amylase from ''Pyrococcus'' sp. ST04 <cite>Park2014</cite> (PDB ID [{{PDBlink}}4cmr 4cmr]). In all cases, the catalytic glutamic acid and aspartic acid residues are located near the C-terminal ends of the strands β4 and β7 of the barrel, respectively <cite>Imamura2003,Dickmanns2006</cite>. There was also a crystallization report in 1995 on a probable GH57 amylopullulanase from ''Pyrococcus woesei'' <cite>Knapp1995</cite>, but the detailed crystallographic analysis of this protein has not been published.
  
It is clear that the C-terminal domain cannot be present in some GH57 members with shorter amino acid sequences, e.g., in the α-galactosidases containing less than 400 residues (van Lieshout et al. 2003). On the other hand, some other GH57 members, especially the extra-long amylopullulanases with more than 1,300 residues (Erra-Pujada et al. 1999) have to contain even additional domains. One of them is a longer version of a typical SLH motif (surface layer homology) (Lupas et al. 1994) that was named as the so-called SLH motif-bearing domain in the amylopullulanase from T. hydrothermalis (Erra-Pujada et al. 1999). This domain was found also in the GH15 glucodextranase from Arthrobacter globiformis (Mizuno et al. 2004). Remarkably, within the family GH57, the presence of this SLH motif-bearing domain is restricted only for amylopullulanases (Zona & Janecek 2005).
+
It is clear that the C-terminal domain cannot be present in some GH57 members with shorter amino acid sequences, e.g., in the α-galactosidases containing less than 400 residues <cite>vanLieshout2003</cite>. On the other hand, some other GH57 members, especially the extra-long amylopullulanases with more than 1,300 residues <cite>Erra-Pujada1999</cite> have to contain even additional domains. One of them could be a longer version of a typical surface layer homology (SLH) motif <cite>Lupas1994</cite> that was named as the so-called SLH motif-bearing domain in the amylopullulanase from ''Thermococcus hydrothermalis'' <cite>Erra-Pujada1999</cite>. This domain was found also in the [[GH15]] glucodextranase from ''Arthrobacter globiformis'' <cite>Mizuno2004</cite>. Remarkably, within the family GH57, the presence of this SLH motif-bearing domain is restricted only for amylopullulanases <cite>Zona2005</cite>.
  
It is also worth mentioning that, especially prior the first three-dimensional structure of a GH57 member was available, there were some efforts to join the family GH57 with the main α-amylase family GH13, i.e. the present clan GH-H consisting of the families GH13, GH70 and GH77 (MacGregor et al. 2001). Those efforts were focused mainly on looking for some remote homology at the sequence level only (Dong et al. 1997; Janecek 1998). Although both GH57 and GH-H employ the same retaining reaction mechanism (Imamura et al 2003; Matsuura et al. 1984) the independence of the family GH57 with regard to GH-H clan is at present based not only on differences in the catalytic domain, but more importantly, due to differences in the catalytic machineries and conserved sequence regions (Zona et al. 2004; Janecek 2002). As far as other GH families are concerned, the family GH38 α-mannosidase from Drosophila melanogaster (van den Elsen et al. 2001) was revealed to share some structural similarities within the catalytic domain with the GH57 4-α-glucanotransferase from T. litoralis (Imamura et al. 2001, 2003) indicating an eventuality of originating from a common ancestor.
+
It is also worth mentioning that, especially prior the first three-dimensional structure of a GH57 member was available, there were some efforts to join the family GH57 with the main α-amylase family [[GH13]], i.e. the present clan [GH-H] consisting of the families [[GH13]], [[GH70]] and [[GH77]] <cite>MacGregor2001</cite>. Those efforts were focused mainly on looking for some remote homology at the sequence level only <cite>Dong1997,Janecek1998</cite>. Although both GH57 and GH-H employ the same retaining reaction mechanism <cite>Imamura2003,Matsuura1984</cite> the independence of the family GH57 with regard to GH-H clan is at present based not only on differences in the catalytic domain, but more importantly, on the differences in the catalytic machineries and conserved sequence regions <cite>Zona2004,Janecek2002</cite>. As far as other GH families are concerned, the family [[GH38]] α-mannosidase from ''Drosophila melanogaster'' <cite>vandenElsen2001</cite> was revealed to share some structural similarities within the catalytic domain with the GH57 4-α-glucanotransferase from ''Thermococcus litoralis'' <cite>Imamura2001,Imamura2003</cite> indicating an eventuality of originating from a common ancestor.
  
 
== Family Firsts ==
 
== Family Firsts ==
;First sterochemistry determination: Family GH57 retaining enzymes, as first documented by the X-ray crystallography on the 4-α-glucanotransferase from T. litoralis complexed with acarbose (Imamura et al. 2003). Kinetic studies have been performed with the 4-α-glucanotransferases from T. litoralis (Imamura et al. 2001, 2003), P. furiosus (Tang et al. 2006), amylopullulanases from T. hydrothermalis (Zona et al. 2004) and P. furiosus (Kang et al. 2006) and branching enzyme from Thermococcus kodakaraensis (Murakami et al. 2006).
+
;First sterochemistry determination: The first direct evidence of the retaining reaction stereochemistry by product analysis has been delivered by 1H-NMR analysis of the product mixture that confirmed the α-anomeric configuration of the α-1,6-glucosidic bond formed after the incubation of ''Thermus thermophilus'' branching enzyme with amylose <cite>Palomo2011</cite>. Previously, family GH57 enzymes have been indicated to be retaining by trapping of a covalent glycosyl-enzyme intermediate <cite>Imamura2001</cite> and a 6.7 Å distance between the catalytic nucleophile and acid/base <cite>Imamura2003</cite>, both of which are consistent with a two-step, double-displacement mechanism.  
;First amino acid sequence determination: Cite some reference here, with a ''short'' (1-2 sentence) explanation.
+
;First amino acid sequence determination: The first amino acid sequence of the family GH57 was that of the heat stable amylase from an anaerobic thermophilic bacterium ''Dictyoglomus thermophilum'' <cite>Fukusumi1988</cite>. This "α-amylase" was later characterized as 4-α-glucanotransferase <cite>Nakajima2004</cite>. In fact, the family contains many α-amylase-like proteins, i.e. those that exhibit a clear sequence homology with α-amylases, but possess a substitution in one or both catalytic residues <cite>Janecek2011</cite>.
;First conserved sequence regions determination: Cite some reference here, with a ''short'' (1-2 sentence) explanation.
+
;First conserved sequence regions determination: The five sequence stretches characteristic as conserved regions for the family GH57 were first determined by the bioinformatics study by Zona et al. (2004) <cite>Zona2004</cite>.
;First catalytic nucleophile identification: Cite some reference here, with a ''short'' (1-2 sentence) explanation.
+
;First [[catalytic nucleophile]] identification: The catalytic nucleophile was fist identified by Imamura et al. (2001) <cite>Imamura2001</cite> as Glu123 in the 4-α-glucanotransferase from ''Thermococcus litoralis'' using the 3-ketobutylidene-β-2-chloro-4-nitrophenyl maltopentaoside as a donor.
;First general acid/base residue identification: Cite some reference here, with a ''short'' (1-2 sentence) explanation.
+
;First [[general acid/base]] residue identification: Asp214 of the 4-α-glucanotransferase from ''Thermococcus litoralis'' as indicated by the X-ray crystallography and supported by site-directed mutagenesis <cite>Imamura2003</cite> since the D214N mutant exhibited a 10,000-fold decrease of specific activity in comparison with the wild-type enzyme.
;First 3-D structure: The first 3-D structure of a GH57 member was that of the 4-alpha-glucanotransferase from ''Thermococcus litoralis'' <cite>Imamura2003</cite>.
+
;First 3-D structure: The first 3-D structure of a GH57 member was that of the 4-α-glucanotransferase from ''Thermococcus litoralis'' <cite>Imamura2003</cite>.
  
 
== References ==
 
== References ==
 
<biblio>
 
<biblio>
 +
#Henrissat1996 pmid=8687420
 +
#MacGregor2001 pmid=11257505
 +
#Fukusumi1988 pmid=2453362
 +
#Laderman1993a pmid=8226990
 +
#Cantarel2009 pmid=18838391
 +
#Lombard2014 pmid=24270786
 +
#Blesak2012 pmid=22527043
 +
#Li2013 pmid=23001056
 +
#Comfort2008 pmid=18156337
 +
#Blesak2013 pmid=24109595
 +
#Wang2011 pmid=21455739
 +
#Jung2014 pmid=23884203
 +
#Jeon2014 pmid=24835094
 +
#Park2014 pmid=24914977
 +
#Janecek2011 pmid=21786160
 +
#Nakajima2004 pmid=15564678
 +
#Laderman1993b pmid=8226989
 +
#Janecek2012 pmid=22819817
 +
#Palomo2011 pmid=21097495
 
#Imamura2003 pmid=12618437
 
#Imamura2003 pmid=12618437
 +
#Imamura2001 pmid=11591160
 +
#Davies1995 pmid=8535779
 +
#Tang2006 pmid=17035108
 +
#Zona2004 pmid=15233783
 +
#Kang2005 pmid=15599521
 +
#Murakami2006 pmid=16885460
 +
#Janecek2005 Janecek S ''Amylolytic families of glycoside hydrolases: focus on the family GH-57.'' Biologia 2005; 60(Suppl. 16) 177-84. ([http://biologia.savba.sk/Suppl_16/Janecek_S.pdf PDF])
 +
#vanLieshout2003 van Lieshout JFT, Verhees CH, Ettema TJG, van der Saar S, Imamura H, Matsuzawa H, van der Oost J, de Vos WM. ''Identification and molecular characterization of a novel type of α-galactosidase from Pyrococcus furiosus.'' Biocatal Biotransform 2003; 21(4-5) 243-52. ([http://dx.doi.org/10.1080/10242420310001614342 DOI: 10.1080/10242420310001614342])
 +
#Dickmanns2006 pmid=16510973
 +
#Santos2011 pmid=21104698
 +
#Knapp1995 pmid=8749857
 +
#Erra-Pujada1999 pmid=10322035
 +
#Lupas1994 pmid=8113161
 +
#Mizuno2004 pmid=14660574
 +
#Zona2005 Zona R, Janecek S. ''Relationships between SLH motifs from different glycoside hydrolase families.'' Biologia 2005; 60(Suppl. 16): 115-21. ([http://biologia.savba.sk/Suppl_16/Zona_R.pdf PDF])
 +
#Dong1997 pmid=9293009
 +
#Janecek1998 pmid=9721603
 +
#Matsuura1984 pmid=6609921
 +
#Janecek2002 Janecek S. ''How many conserved sequence regions are there in the α-amylase family?'' Biologia 2002; 57(Suppl. 11): 29-41. ([http://biologia.savba.sk/Suppl_11/Janecek.pdf PDF])
 +
#vandenElsen2001 pmid=11406577
 
</biblio>
 
</biblio>
  
 
[[Category:Glycoside Hydrolase Families|GH057]]
 
[[Category:Glycoside Hydrolase Families|GH057]]

Latest revision as of 14:18, 18 December 2021

Approve icon-50px.png

This page has been approved by the Responsible Curator as essentially complete. CAZypedia is a living document, so further improvement of this page is still possible. If you would like to suggest an addition or correction, please contact the page's Responsible Curator directly by e-mail.


Glycoside Hydrolase Family GH57
Clan not assigned yet
Mechanism retaining
Active site residues known
CAZy DB link
http://www.cazy.org/GH57.html

Substrate specificities

Glycoside hydrolase family 57 was established in 1996 [1] based on the existence of the sequences of two “α-amylases” that were dissimilar to typical family GH13 α-amylases [2]. The two were the heat-stable eubacterial amylase from Dictyoglomus thermophilum known from 1988 [3] and the extremely thermostable archaeal amylase from Pyrococcus furiosus determined in 1993 [4].

The family has expanded mainly due to the results of genome sequencing projects. More than a thousand of members are known [5, 6], all from prokaryotes (with ~1:4 ratio for Archaea:Bacteria) [7]. The enzyme specificities of family GH57 includes α-amylase (EC 3.2.1.1), α-galactosidase (EC 3.2.1.22), amylopullulanase (EC 3.2.1.41), branching enzyme (EC 2.4.1.18) and 4-α-glucanotransferase (EC 2.4.1.25).

An archaeal GH57 amylopullulanase from Staphylothermus marinus has been described exhibiting also the activity of cyclodextrinase (EC 3.2.1.54) [8]. Based on a preliminary observation that the PF0870 protein encoded in the Pyrococcus furiosus genome produced maltose [9], a group of GH57 members with proposed specificity of maltogenic amylase (EC 3.2.1.133) was predicted [10] together with another group of non-specified amylases following the partial characterization of an amylolytic enzyme from an uncultured bacterium [11]. The maltogenic specificity (or maltose-forming amylase) has already been definitively confirmed by both biochemical [12, 13] and structural [14] studies. The family GH57 contains further a group of probably non-enzymatic members, i.e. the so-called α-amylase-like proteins having a substitution in one or both catalytic residues [15].

It is worth mentioning that the two founding members, i.e. the “α-amylases” from Dictyoglomus thermophilum and Pyrococcus furiosus are 4-α-glucanotransferases; the former was proven to have transglycosylating activity [16], whereas the latter was already known to exhibit also the 4-α-glucanotransferase activity [17].

A detailed in silico study has indicated a close relatedness of family GH57 to the family GH119 concerning catalytic domain fold, catalytic machinery and conserved sequence regions [18].

Kinetics and Mechanism

A direct evidence for the stereochemical outcome of the enzyme-catalyzed reaction has been presented for the branching enzyme from Thermus thermophilus [19] by the NMR analysis of reaction products. The previous observation of a trapped covalent glycosyl-enzyme intermediate [20] was also a strong evidence that the GH57 enzymes operate with retention of anomeric configuration through a classical Koshland retaining mechanism with retention of anomeric configuration. The X-ray structure of the 4-α-glucanotransferase from Thermococcus litoralis revealed average distances of 6.72 Å and 6.97 Å between the catalytic nucleophile (Glu123) and general acid/base (Asp214) in the free and acarbose-bound forms of the enzyme, respectively, which also supports a retaining mechanism for this family [21] (see also [22]).

Detailed kinetic studies have been performed on several GH57 enzymes, including the 4-α-glucanotransferases from Thermococcus litoralis [20, 21] and Pyrococcus furiosus [23], amylopullulanases from Thermococcus hydrothermalis [24] and Pyrococcus furiosus [25], and branching enzymes from Thermococcus kodakaraensis [26] and Thermus thermophilus [19].

Catalytic Residues

The sequences of GH57 members are very diverse. Some sequences are shorter than 400 residues whereas others are longer than 1,500 residues [27]. This complicated early efforts to align the GH57 sequences using standard alignment methods. A detailed bioinformatics study by Zona et al. [24] focused on all available GH57 sequences at that time and identified five conserved sequence regions. This study used knowledge of the identity of the catalytic nucleophile (Glu123) in the GH57 4-α-glucanotransferase from Thermococcus litoralis [20] together with the three-dimensional structure [21] (PDB ID 1k1w) that revealed the general acid/base residue (Asp214).

The catalytic nucleophile (a glutamate) and general acid/base (an aspartate) are located in the conserved sequence regions 3 and 4, respectively. In addition to assignments in Thermococcus litoralis 4-α-glucanotransferase discussed above, these residues were identified in the amylopullulanases from Thermococcus hydrothermalis [24] and Pyrococcus furiosus [25]. The catalytic nucleophile was also identified in the α-galactosidase from Pyrococcus furiosus although no success was achieved in assigning the general acid/base [28]. It should be taken into account that some GH57 members, which are only hypothetical enzymes/proteins without any biochemical characterization, may lack one or even both catalytic residues [15, 24].

Based on the five identified conserved sequence regions, the residues His13, Glu79, Glu216 and Asp354 together with Trp120, Trp221 and Trp357 (Thermococcus hydrothermalis amylopullulanase numbering) were postulated [24] as important determinants of the individual GH57 enzyme specificities [7, 10, 15]. Of these, the Trp221 has already been confirmed to contribute to the transglycosylation activity of 4-α-glucanotransferase since the mutant W229H of the enzyme from Pyrococcus furiosus exhibited markedly decreased transglycosylation activity in comparison with the wild-type counterpart [23].

Three-dimensional structures

The structure of the catalytic domain adopts a (β/α)7-barrel, i.e. the irregular (β/α)8-barrel called a pseudo TIM-barrel. In the case of the Thermococcus litoralis 4-α-glucanotransferase [21] the catalytic domain is succeeded by a C-terminal non-catalytic domain consisting of only β-strands that adopts a twisted β-sandwich fold. In the three-dimensional structure of the α-amylase AmyC from Thermotoga maritima [29] (PDB ID 2b5d) (note that the sequence of this enzyme strongly resemles that of a branching enzyme [7]), the corresponding catalytic (β/α)7-barrel is followed by a five-helix domain C, a small helical domain B being protruded out of the catalytic pseudo TIM barrel in the place of the loop 2 (i.e. succeeding the strand β2). This structure was found to be closely similar to that of the GH57 member of unknown function from Thermus thermophilus (PDB ID 1ufa). Two structures of biochemically characterized branching enzymes were solved and published: one from Thermus thermophilus [19] (PDB ID 3p0b) and the other one from Thermococcus kodakaraensis [30] (PDB ID 3n8t). The former study [19], importantly, is the first report to prove that a retaining mechanism is employed in the family GH57. Also, the 3-D structure is available for the maltogenic (maltose-forming) amylase from Pyrococcus sp. ST04 [14] (PDB ID 4cmr). In all cases, the catalytic glutamic acid and aspartic acid residues are located near the C-terminal ends of the strands β4 and β7 of the barrel, respectively [21, 29]. There was also a crystallization report in 1995 on a probable GH57 amylopullulanase from Pyrococcus woesei [31], but the detailed crystallographic analysis of this protein has not been published.

It is clear that the C-terminal domain cannot be present in some GH57 members with shorter amino acid sequences, e.g., in the α-galactosidases containing less than 400 residues [28]. On the other hand, some other GH57 members, especially the extra-long amylopullulanases with more than 1,300 residues [32] have to contain even additional domains. One of them could be a longer version of a typical surface layer homology (SLH) motif [33] that was named as the so-called SLH motif-bearing domain in the amylopullulanase from Thermococcus hydrothermalis [32]. This domain was found also in the GH15 glucodextranase from Arthrobacter globiformis [34]. Remarkably, within the family GH57, the presence of this SLH motif-bearing domain is restricted only for amylopullulanases [35].

It is also worth mentioning that, especially prior the first three-dimensional structure of a GH57 member was available, there were some efforts to join the family GH57 with the main α-amylase family GH13, i.e. the present clan [GH-H] consisting of the families GH13, GH70 and GH77 [2]. Those efforts were focused mainly on looking for some remote homology at the sequence level only [36, 37]. Although both GH57 and GH-H employ the same retaining reaction mechanism [21, 38] the independence of the family GH57 with regard to GH-H clan is at present based not only on differences in the catalytic domain, but more importantly, on the differences in the catalytic machineries and conserved sequence regions [24, 39]. As far as other GH families are concerned, the family GH38 α-mannosidase from Drosophila melanogaster [40] was revealed to share some structural similarities within the catalytic domain with the GH57 4-α-glucanotransferase from Thermococcus litoralis [20, 21] indicating an eventuality of originating from a common ancestor.

Family Firsts

First sterochemistry determination
The first direct evidence of the retaining reaction stereochemistry by product analysis has been delivered by 1H-NMR analysis of the product mixture that confirmed the α-anomeric configuration of the α-1,6-glucosidic bond formed after the incubation of Thermus thermophilus branching enzyme with amylose [19]. Previously, family GH57 enzymes have been indicated to be retaining by trapping of a covalent glycosyl-enzyme intermediate [20] and a 6.7 Å distance between the catalytic nucleophile and acid/base [21], both of which are consistent with a two-step, double-displacement mechanism.
First amino acid sequence determination
The first amino acid sequence of the family GH57 was that of the heat stable amylase from an anaerobic thermophilic bacterium Dictyoglomus thermophilum [3]. This "α-amylase" was later characterized as 4-α-glucanotransferase [16]. In fact, the family contains many α-amylase-like proteins, i.e. those that exhibit a clear sequence homology with α-amylases, but possess a substitution in one or both catalytic residues [15].
First conserved sequence regions determination
The five sequence stretches characteristic as conserved regions for the family GH57 were first determined by the bioinformatics study by Zona et al. (2004) [24].
First catalytic nucleophile identification
The catalytic nucleophile was fist identified by Imamura et al. (2001) [20] as Glu123 in the 4-α-glucanotransferase from Thermococcus litoralis using the 3-ketobutylidene-β-2-chloro-4-nitrophenyl maltopentaoside as a donor.
First general acid/base residue identification
Asp214 of the 4-α-glucanotransferase from Thermococcus litoralis as indicated by the X-ray crystallography and supported by site-directed mutagenesis [21] since the D214N mutant exhibited a 10,000-fold decrease of specific activity in comparison with the wild-type enzyme.
First 3-D structure
The first 3-D structure of a GH57 member was that of the 4-α-glucanotransferase from Thermococcus litoralis [21].

References

  1. Henrissat B and Bairoch A. (1996). Updating the sequence-based classification of glycosyl hydrolases. Biochem J. 1996;316 ( Pt 2):695-6. DOI:10.1042/bj3160695 | PubMed ID:8687420 [Henrissat1996]
  2. MacGregor EA, Janecek S, and Svensson B. (2001). Relationship of sequence and structure to specificity in the alpha-amylase family of enzymes. Biochim Biophys Acta. 2001;1546(1):1-20. DOI:10.1016/s0167-4838(00)00302-2 | PubMed ID:11257505 [MacGregor2001]
  3. Fukusumi S, Kamizono A, Horinouchi S, and Beppu T. (1988). Cloning and nucleotide sequence of a heat-stable amylase gene from an anaerobic thermophile, Dictyoglomus thermophilum. Eur J Biochem. 1988;174(1):15-21. DOI:10.1111/j.1432-1033.1988.tb14056.x | PubMed ID:2453362 [Fukusumi1988]
  4. Laderman KA, Asada K, Uemori T, Mukai H, Taguchi Y, Kato I, and Anfinsen CB. (1993). Alpha-amylase from the hyperthermophilic archaebacterium Pyrococcus furiosus. Cloning and sequencing of the gene and expression in Escherichia coli. J Biol Chem. 1993;268(32):24402-7. | Google Books | Open Library PubMed ID:8226990 [Laderman1993a]
  5. Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, and Henrissat B. (2009). The Carbohydrate-Active EnZymes database (CAZy): an expert resource for Glycogenomics. Nucleic Acids Res. 2009;37(Database issue):D233-8. DOI:10.1093/nar/gkn663 | PubMed ID:18838391 [Cantarel2009]
  6. Lombard V, Golaconda Ramulu H, Drula E, Coutinho PM, and Henrissat B. (2014). The carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Res. 2014;42(Database issue):D490-5. DOI:10.1093/nar/gkt1178 | PubMed ID:24270786 [Lombard2014]
  7. Blesák K and Janeček S. (2012). Sequence fingerprints of enzyme specificities from the glycoside hydrolase family GH57. Extremophiles. 2012;16(3):497-506. DOI:10.1007/s00792-012-0449-9 | PubMed ID:22527043 [Blesak2012]
  8. Li X, Li D, and Park KH. (2013). An extremely thermostable amylopullulanase from Staphylothermus marinus displays both pullulan- and cyclodextrin-degrading activities. Appl Microbiol Biotechnol. 2013;97(12):5359-69. DOI:10.1007/s00253-012-4397-1 | PubMed ID:23001056 [Li2013]
  9. Comfort DA, Chou CJ, Conners SB, VanFossen AL, and Kelly RM. (2008). Functional-genomics-based identification and characterization of open reading frames encoding alpha-glucoside-processing enzymes in the hyperthermophilic archaeon Pyrococcus furiosus. Appl Environ Microbiol. 2008;74(4):1281-3. DOI:10.1128/AEM.01920-07 | PubMed ID:18156337 [Comfort2008]
  10. Blesák K and Janeček Š. (2013). Two potentially novel amylolytic enzyme specificities in the prokaryotic glycoside hydrolase α-amylase family GH57. Microbiology (Reading). 2013;159(Pt 12):2584-2593. DOI:10.1099/mic.0.071084-0 | PubMed ID:24109595 [Blesak2013]
  11. Wang H, Gong Y, Xie W, Xiao W, Wang J, Zheng Y, Hu J, and Liu Z. (2011). Identification and characterization of a novel thermostable gh-57 gene from metagenomic fosmid library of the Juan de Fuca Ridge hydrothemal vent. Appl Biochem Biotechnol. 2011;164(8):1323-38. DOI:10.1007/s12010-011-9215-1 | PubMed ID:21455739 [Wang2011]
  12. Jung JH, Seo DH, Holden JF, and Park CS. (2014). Maltose-forming α-amylase from the hyperthermophilic archaeon Pyrococcus sp. ST04. Appl Microbiol Biotechnol. 2014;98(5):2121-31. DOI:10.1007/s00253-013-5068-6 | PubMed ID:23884203 [Jung2014]
  13. Jeon EJ, Jung JH, Seo DH, Jung DH, Holden JF, and Park CS. (2014). Bioinformatic and biochemical analysis of a novel maltose-forming α-amylase of the GH57 family in the hyperthermophilic archaeon Thermococcus sp. CL1. Enzyme Microb Technol. 2014;60:9-15. DOI:10.1016/j.enzmictec.2014.03.009 | PubMed ID:24835094 [Jeon2014]
  14. Park KH, Jung JH, Park SG, Lee ME, Holden JF, Park CS, and Woo EJ. (2014). Structural features underlying the selective cleavage of a novel exo-type maltose-forming amylase from Pyrococcus sp. ST04. Acta Crystallogr D Biol Crystallogr. 2014;70(Pt 6):1659-68. DOI:10.1107/S1399004714006567 | PubMed ID:24914977 [Park2014]
  15. Janeček S and Blesák K. (2011). Sequence-structural features and evolutionary relationships of family GH57 α-amylases and their putative α-amylase-like homologues. Protein J. 2011;30(6):429-35. DOI:10.1007/s10930-011-9348-7 | PubMed ID:21786160 [Janecek2011]
  16. Nakajima M, Imamura H, Shoun H, Horinouchi S, and Wakagi T. (2004). Transglycosylation activity of Dictyoglomus thermophilum amylase A. Biosci Biotechnol Biochem. 2004;68(11):2369-73. DOI:10.1271/bbb.68.2369 | PubMed ID:15564678 [Nakajima2004]
  17. Laderman KA, Davis BR, Krutzsch HC, Lewis MS, Griko YV, Privalov PL, and Anfinsen CB. (1993). The purification and characterization of an extremely thermostable alpha-amylase from the hyperthermophilic archaebacterium Pyrococcus furiosus. J Biol Chem. 1993;268(32):24394-401. | Google Books | Open Library PubMed ID:8226989 [Laderman1993b]
  18. Janeček S and Kuchtová A. (2012). In silico identification of catalytic residues and domain fold of the family GH119 sharing the catalytic machinery with the α-amylase family GH57. FEBS Lett. 2012;586(19):3360-6. DOI:10.1016/j.febslet.2012.07.020 | PubMed ID:22819817 [Janecek2012]
  19. Palomo M, Pijning T, Booiman T, Dobruchowska JM, van der Vlist J, Kralj S, Planas A, Loos K, Kamerling JP, Dijkstra BW, van der Maarel MJ, Dijkhuizen L, and Leemhuis H. (2011). Thermus thermophilus glycoside hydrolase family 57 branching enzyme: crystal structure, mechanism of action, and products formed. J Biol Chem. 2011;286(5):3520-30. DOI:10.1074/jbc.M110.179515 | PubMed ID:21097495 [Palomo2011]
  20. Imamura H, Fushinobu S, Jeon BS, Wakagi T, and Matsuzawa H. (2001). Identification of the catalytic residue of Thermococcus litoralis 4-alpha-glucanotransferase through mechanism-based labeling. Biochemistry. 2001;40(41):12400-6. DOI:10.1021/bi011017c | PubMed ID:11591160 [Imamura2001]
  21. Imamura H, Fushinobu S, Yamamoto M, Kumasaka T, Jeon BS, Wakagi T, and Matsuzawa H. (2003). Crystal structures of 4-alpha-glucanotransferase from Thermococcus litoralis and its complex with an inhibitor. J Biol Chem. 2003;278(21):19378-86. DOI:10.1074/jbc.M213134200 | PubMed ID:12618437 [Imamura2003]
  22. Davies G and Henrissat B. (1995). Structures and mechanisms of glycosyl hydrolases. Structure. 1995;3(9):853-9. DOI:10.1016/S0969-2126(01)00220-9 | PubMed ID:8535779 [Davies1995]
  23. Tang SY, Yang SJ, Cha H, Woo EJ, Park C, and Park KH. (2006). Contribution of W229 to the transglycosylation activity of 4-alpha-glucanotransferase from Pyrococcus furiosus. Biochim Biophys Acta. 2006;1764(10):1633-8. DOI:10.1016/j.bbapap.2006.08.013 | PubMed ID:17035108 [Tang2006]
  24. Zona R, Chang-Pi-Hin F, O'Donohue MJ, and Janecek S. (2004). Bioinformatics of the glycoside hydrolase family 57 and identification of catalytic residues in amylopullulanase from Thermococcus hydrothermalis. Eur J Biochem. 2004;271(14):2863-72. DOI:10.1111/j.1432-1033.2004.04144.x | PubMed ID:15233783 [Zona2004]
  25. Kang S, Vieille C, and Zeikus JG. (2005). Identification of Pyrococcus furiosus amylopullulanase catalytic residues. Appl Microbiol Biotechnol. 2005;66(4):408-13. DOI:10.1007/s00253-004-1690-7 | PubMed ID:15599521 [Kang2005]
  26. Murakami T, Kanai T, Takata H, Kuriki T, and Imanaka T. (2006). A novel branching enzyme of the GH-57 family in the hyperthermophilic archaeon Thermococcus kodakaraensis KOD1. J Bacteriol. 2006;188(16):5915-24. DOI:10.1128/JB.00390-06 | PubMed ID:16885460 [Murakami2006]
  27. Janecek S Amylolytic families of glycoside hydrolases: focus on the family GH-57. Biologia 2005; 60(Suppl. 16) 177-84. (PDF)

    [Janecek2005]
  28. van Lieshout JFT, Verhees CH, Ettema TJG, van der Saar S, Imamura H, Matsuzawa H, van der Oost J, de Vos WM. Identification and molecular characterization of a novel type of α-galactosidase from Pyrococcus furiosus. Biocatal Biotransform 2003; 21(4-5) 243-52. (DOI: 10.1080/10242420310001614342)

    [vanLieshout2003]
  29. Dickmanns A, Ballschmiter M, Liebl W, and Ficner R. (2006). Structure of the novel alpha-amylase AmyC from Thermotoga maritima. Acta Crystallogr D Biol Crystallogr. 2006;62(Pt 3):262-70. DOI:10.1107/S0907444905041363 | PubMed ID:16510973 [Dickmanns2006]
  30. Santos CR, Tonoli CC, Trindade DM, Betzel C, Takata H, Kuriki T, Kanai T, Imanaka T, Arni RK, and Murakami MT. (2011). Structural basis for branching-enzyme activity of glycoside hydrolase family 57: structure and stability studies of a novel branching enzyme from the hyperthermophilic archaeon Thermococcus kodakaraensis KOD1. Proteins. 2011;79(2):547-57. DOI:10.1002/prot.22902 | PubMed ID:21104698 [Santos2011]
  31. Knapp S, Rüdiger A, Antranikian G, Jorgensen PL, and Ladenstein R. (1995). Crystallization and preliminary crystallographic analysis of an amylopullulanase from the hyperthermophilic archaeon Pyrococcus woesei. Proteins. 1995;23(4):595-7. DOI:10.1002/prot.340230416 | PubMed ID:8749857 [Knapp1995]
  32. Erra-Pujada M, Debeire P, Duchiron F, and O'Donohue MJ. (1999). The type II pullulanase of Thermococcus hydrothermalis: molecular characterization of the gene and expression of the catalytic domain. J Bacteriol. 1999;181(10):3284-7. DOI:10.1128/JB.181.10.3284-3287.1999 | PubMed ID:10322035 [Erra-Pujada1999]
  33. Lupas A, Engelhardt H, Peters J, Santarius U, Volker S, and Baumeister W. (1994). Domain structure of the Acetogenium kivui surface layer revealed by electron crystallography and sequence analysis. J Bacteriol. 1994;176(5):1224-33. DOI:10.1128/jb.176.5.1224-1233.1994 | PubMed ID:8113161 [Lupas1994]
  34. Mizuno M, Tonozuka T, Suzuki S, Uotsu-Tomita R, Kamitori S, Nishikawa A, and Sakano Y. (2004). Structural insights into substrate specificity and function of glucodextranase. J Biol Chem. 2004;279(11):10575-83. DOI:10.1074/jbc.M310771200 | PubMed ID:14660574 [Mizuno2004]
  35. Zona R, Janecek S. Relationships between SLH motifs from different glycoside hydrolase families. Biologia 2005; 60(Suppl. 16): 115-21. (PDF)

    [Zona2005]
  36. Dong G, Vieille C, and Zeikus JG. (1997). Cloning, sequencing, and expression of the gene encoding amylopullulanase from Pyrococcus furiosus and biochemical characterization of the recombinant enzyme. Appl Environ Microbiol. 1997;63(9):3577-84. DOI:10.1128/aem.63.9.3577-3584.1997 | PubMed ID:9293009 [Dong1997]
  37. Janecek S (1998). Sequence of archaeal Methanococcus jannaschii alpha-amylase contains features of families 13 and 57 of glycosyl hydrolases: a trace of their common ancestor?. Folia Microbiol (Praha). 1998;43(2):123-8. DOI:10.1007/BF02816496 | PubMed ID:9721603 [Janecek1998]
  38. Matsuura Y, Kusunoki M, Harada W, and Kakudo M. (1984). Structure and possible catalytic residues of Taka-amylase A. J Biochem. 1984;95(3):697-702. DOI:10.1093/oxfordjournals.jbchem.a134659 | PubMed ID:6609921 [Matsuura1984]
  39. Janecek S. How many conserved sequence regions are there in the α-amylase family? Biologia 2002; 57(Suppl. 11): 29-41. (PDF)

    [Janecek2002]
  40. van den Elsen JM, Kuntz DA, and Rose DR. (2001). Structure of Golgi alpha-mannosidase II: a target for inhibition of growth and metastasis of cancer cells. EMBO J. 2001;20(12):3008-17. DOI:10.1093/emboj/20.12.3008 | PubMed ID:11406577 [vandenElsen2001]

All Medline abstracts: PubMed