The evolution of putative starch-binding domains

The present bioinformatics analysis was focused on the starch-binding domains (SBDs) and SBD-like motifs sequentially related to carbohydrate-binding module (CBM) families CBM20 and CBM21. Originally, these SBDs were known from microbial amylases only. At present homologous starch- and glycogen-binding domains (or putative SBD sequences) have been recognised in various plant and animal proteins. The sequence comparison clearly showed that the SBD-like sequences in genethonin-1, starch synthase III and glucan branching enzyme should possess the real SBD function since the two tryptophans (or at least two aromatics) of the typical starch-binding site 1 are conserved in their sequences. The same should apply also for the sequences corresponding with the so-called KIS-domain of plant AKINbetagamma protein that is a homologue of the animal AMP-activated protein kinase (AMPK). The evolutionary tree classified the compared SBDs into three distinct groups: (i) the family CBM20 (the motifs from genethonins, laforins, starch excess 4 protein, beta-subunits of the animal AMPK and all plant and yeast homologues, and eventually from amylopullulanases); (ii) the family CBM21 (the motifs from regulatory subunits of protein phosphatase 1 together with those from starch synthase III); and (iii) the (CBM20+CBM21)-related group (the motifs from the pullulanase subfamily consisting of pullulanase, branching enzyme, isoamylase and maltooligosyl trehalohydrolase).     

Domain evolution in the GH13 pullulanase subfamily with focus on the carbohydrate-binding module family 48

Glycoside hydrolase (GH) family 13 comprises about 30 different specificities. Four of them have been proposed to form the GH13 pullulanase subfamily: pullulanase, isoamylase, maltooligosyl trehalohydrolase and branching enzyme forming the seven CAZy GH13 subfamilies: GH13 8-GH13 14. Recently, a new family of carbohydrate-binding modules (CBMs), the family CBM48 has been established containing the putative starch-binding domains from the pullulanase subfamily, the β-subunit of AMP-activated protein kinase and some other GH13 enzymes with pullulanase and/or α-amylase-pullulanase specificity. Since all of these enzymes are multidomain proteins and the structure for at least one representative of each enzyme specificity has already been determined, the main goal of the present study was to elucidate domain evolution within this GH13 pullulanase subfamily (84 real enzymes) focusing on the CBM48 module. With regard to CBM48 positioning in the amino acid sequence, the N-terminal end of a protein appears to be a predominant position. This is especially true for isoamylases and maltooligosyl trehalohydrolases. Secondary structure-based alignment of CBM modules from CBM48, CBM20 and CBM21 revealed that several residues known as consensus for CBM20 and CBM21 could also be identified in CBM48, but only branching enzymes possess the aromatic residues that correspond with the two tryptophans forming the evolutionary conserved starch-binding site 1 in CBM20. The evolutionary trees constructed for the individual domains, complete alignment, and the conserved sequence regions of the α-amylase family were found to be comparable to each other (except for the C-domain tree) with two basic parts: (i) branching enzymes and maltooligosyl trehalohydrolases; and (ii) pullulanases and isoamylases. Taxonomy was respected only within clusters with pure specificity, i.e. the evolution of CBM48 reflects the evolution of specificities rather than evolution of species. This is a feature different from the one observed for the starch-binding domain of the family CBM20 where the starch-binding domain evolution reflects the evolution of species.     

A new clan of CBM families based on bioinformatics of starch-binding domains from families CBM20 and CBM21

Approximately 10% of amylolytic enzymes are able to bind and degrade raw starch. Usually a distinct domain, the starch-binding domain (SBD), is responsible for this property. These domains have been classified into families of carbohydrate-binding modules (CBM). At present, there are six SBD families: CBM20, CBM21, CBM25, CBM26, CBM34, and CBM41. This work is concentrated on CBM20 and CBM21. The CBM20 module was believed to be located almost exclusively at the C-terminal end of various amylases. The CBM21 module was known as the N-terminally positioned SBD of Rhizopus glucoamylase. Nowadays many nonamylolytic proteins have been recognized as possessing sequence segments that exhibit similarities with the experimentally observed CBM20 and CBM21. These facts have stimulated interest in carrying out a rigorous bioinformatics analysis of the two CBM families. The present analysis showed that the original idea of the CBM20 module being at the C-terminus and the CBM21 module at the N-terminus of a protein should be modified. Although the CBM20 functionally important tryptophans were found to be substituted in several cases, these aromatics and the regions around them belong to the best conserved parts of the CBM20 module. They were therefore used as templates for revealing the corresponding regions in the CBM21 family. Secondary structure prediction together with fold recognition indicated that the CBM21 module structure should be similar to that of CBM20. The evolutionary tree based on a common alignment of sequences of both modules showed that the CBM21 SBDs from alpha-amylases and glucoamylases are the closest relatives to the CBM20 counterparts, with the CBM20 modules from the glycoside hydrolase family GH13 amylopullulanases being possible candidates for the intermediate between the two CBM families.     

Starch-binding domains as CBM families–history,occurrence, structure,function and evolution

The term “starch-binding domain” (SBD) has been applied to a domain within an amylolytic enzyme that gave the enzyme the ability to bind onto raw, i.e. thermally untreated, granular starch. An SBD is a special case of a carbohydrate-binding domain, which in general, is a structurally and functionally independent protein module exhibiting no enzymatic activity but possessing potential to target the catalytic domain to the carbohydrate substrate to accommodate it and process it at the active site. As so-called families, SBDs together with other carbohydrate-binding modules (CBMs) have become an integral part of the CAZy database (http://www.cazy.org/). The first two well-described SBDs, i.e. the C-terminal Aspergillus-type and the N-terminal Rhizopus-type have been assigned the families CBM20 and CBM21, respectively. Currently, among the 85 established CBM families in CAZy, fifteen can be considered as families having SBD functional characteristics: CBM20, 21, 25, 26, 34, 41, 45, 48, 53, 58, 68, 69, 74, 82 and 83. All known SBDs, with the exception of the extra long CBM74, were recognized as a module consisting of approximately 100 residues, adopting a β-sandwich fold and possessing at least one carbohydrate-binding site. The present review aims to deliver and describe: (i) the SBD identification in different amylolytic and related enzymes (e.g., CAZy GH families) as well as in other relevant enzymes and proteins (e.g., laforin, the β-subunit of AMPK, and others); (ii) information on the position in the polypeptide chain and the number of SBD copies and their CBM family affiliation (if appropriate); (iii) structure/function studies of SBDs with a special focus on solved tertiary structures, in particular, as complexes with α-glucan ligands; and (iv) the evolutionary relationships of SBDs in a tree common to all SBD CBM families (except for the extra long CBM74). Finally, some special cases and novel potential SBDs are also introduced.     

Two structurally related starch-binding domain families CBM25 and CBM26

Starch binding domains (SBDs) are able to bind to and facilitate the degradation of raw starch and starchy substrates. In general, in the CAZy database they have been classified among the carbohydrate-binding module (CBM) families. The two families CBM25 and CBM26 together with families CBM20, 21, 34, 41, 45, 48, 53, 58, 68 and 69 belong to twelve SBD CAZy families. They represent a group of closely related modules exhibiting some sequence similarity, although each of the two families possesses its own features. Both CBM25 and CBM26 adopt a typical SBD fold of distorted β-barrel as recognized in the modules present in the maltohexaose-producing amylase from Bacillus halodurans. With regard to catalytic domains, most members are α-amylases and maltooligosaccharide-producing amylases from the α-amylase glycoside hydrolase (GH) family GH13, but also some β-amylases (GH14) and hypothetical proteins (e.g. from the family GH31) are known. The main goal of this review was to compare the available amino acid sequences of SBDs from both families CBM25 and CBM26 and to reveal, if possible, SBD(s) with the character “intermediary” between the CBM25 and CBM26. Emphasis was also given on a structural comparison of the identified intermediary SBD with the CBM25 and CBM26 representatives and a detailed evolutionary division of both CBM families that can be utilized for defining the future subfamilies.     

Dynamics of starch granule biogenesis – the role of redox-regulated enzymes and low-affinity carbohydrate-binding modules

Abstract

The deposition and degradation of starch in plants is subject to extensive post-translational regulation. To permit degradation of B-type crystallites present in tuberous and leaf starch these starch types are phosphorylated by glucan, water dikinase (GWD). At the level of post-translational redox regulation, ADPglucose pyrophosphorylase, β-amylase (BAM1), limit dextrinase (LD), the starch phosphorylator GWD and the glucan phosphatase dual-specificity phosphatase 4 (DSP4), also named starch excess 4 (SEX4), are reductively activated in vitro. Redox screens now suggest the presence of a substantially more extensive and coordinated redox regulation involving a larger number of enzymes. Noticeably several of these enzymes contain a new type of low-affinity carbohydrate-binding module that we term a low-affinity starch-binding domain or LA-SBD. These are present in the CBM20, CBM45 and CBM53 families and can enable diurnal dynamics of starch–enzyme recognition. Such diurnal changes in starch binding have been indicated for the redox-regulated GWD and SEX4.     

An Unusual Mode of Galactose Recognition by a Family 32 Carbohydrate-Binding Module

Carbohydrate-binding modules (CBMs) are ancillary modules commonly associated with carbohydrate-active enzymes (CAZymes) that function to mediate the adherence of the parent enzyme to its carbohydrate substrates. CBM family 32 (CBM32) is one of the most diverse CBM families, whose members are commonly found in bacterial CAZymes that modify eukaryotic glycans. One such example is the putative μ-toxin, CpGH84A, of the family 84 glycoside hydrolases, which comprises an N-terminal putative β-N-acetylglucosaminidase catalytic module and four tandem CBM32s. Here, we report a unique mode of galactose recognition by the first CBM32, CBM32-1 from CpGH84A. Solution NMR-based analyses of CpGH84A CBM32-1 indicate a divergent subset of residues, located in ordered loops at the apex of the CBM, conferring specificity for the galacto-configured sugars galactose, GalNAc, and LacNAc that differs from those of the canonical galactose-binding CBM32s. This study showcases the impressive variability in ligand binding by this CBM family and offers insight into the growing role of these modules in the interaction of CAZymes with eukaryotic glycans.     

A CBM20 low-affinity starch-binding domain from glucan, water dikinase

The family 20 carbohydrate-binding module (CBM20) of the Arabidopsis starch phosphorylator glucan, water dikinase 3 (GWD3) was heterologously produced and its properties were compared to the CBM20 from a fungal glucoamylase (GA). The GWD3 CBM20 has 50-fold lower affinity for cyclodextrins than that from GA. Homology modelling identified possible structural elements responsible for this weak binding of the intracellular CBM20. Differential binding of fluorescein-labelled GWD3 and GA modules to starch granules in vitro was demonstrated by confocal laser scanning microscopy and yellow fluorescent protein-tagged GWD3 CBM20 expressed in tobacco confirmed binding to starch granules in planta.     

Starch-binding domains in the CBM45 family--low-affinity domains from glucan, water dikinase and α-amylase involved in plastidial starch metabolism

Starch-binding domains are noncatalytic carbohydrate-binding modules that mediate binding to granular starch. The starch-binding domains from the carbohydrate-binding module family 45 (CBM45, http://www.cazy.org) are found as N-terminal tandem repeats in a small number of enzymes, primarily from photosynthesizing organisms. Isolated domains from representatives of each of the two classes of enzyme carrying CBM45-type domains, the Solanum tuberosumα-glucan, water dikinase and the Arabidopsis thaliana plastidial α-amylase 3, were expressed as recombinant proteins and characterized. Differential scanning calorimetry was used to verify the conformational integrity of an isolated CBM45 domain, revealing a surprisingly high thermal stability (T(m) of 84.8 °C). The functionality of CBM45 was demonstrated in planta by yellow/green fluorescent protein fusions and transient expression in tobacco leaves. Affinities for starch and soluble cyclodextrin starch mimics were measured by adsorption assays, surface plasmon resonance and isothermal titration calorimetry analyses. The data indicate that CBM45 binds with an affinity of about two orders of magnitude lower than the classical starch-binding domains from extracellular microbial amylolytic enzymes. This suggests that low-affinity starch-binding domains are a recurring feature in plastidial starch metabolism, and supports the hypothesis that reversible binding, effectuated through low-affinity interaction with starch granules, facilitates dynamic regulation of enzyme activities and, hence, of starch metabolism.     

Chain-length specificities of maize starch synthase I enzyme: studies of glucan affinity and catalytic properties

It is widely known that some of the starch synthases and starch-branching enzymes are trapped inside the starch granule matrix during the course of starch deposition in amyloplasts. The objective of this study was to use maize SSI to further our understanding of the protein domains involved in starch granule entrapment and identify the chain-length specificities of the enzyme. Using affinity gel electrophoresis, we measured the dissociation constants of maize SSI and its truncated forms using various glucans. The enzyme has a high degree of specificity in terms of its substrate-enzyme dissociation constant, but has a greatly elevated affinity for increasing chain lengths of alpha-1, 4 glucans. Deletion of the N-terminal arm of SSI did not affect the Kd value. Further small deletions of either N- or C-terminal domains resulted in a complete loss of any measurable affinity for its substrate, suggesting that the starch-affinity domain of SSI is not discrete from the catalytic domain. Greater affinity was displayed for the amylopectin fraction of starch as compared to amylose, whereas glycogen revealed the lowest affinity. However, when the outer chain lengths (OCL) of glycogen were extended using the phosphorylase enzyme, we found an increase in affinity for SSI between an average OCL of 7 and 14, and then an apparently exponential increase to an average OCL of 21. On the other hand, the catalytic ability of SSI was reduced several-fold using these glucans with extended chain lengths as substrates, and most of the label from [14C]ADPG was incorporated into shorter chains of dp < 10. We conclude that the rate of catalysis of SSI enzyme decreases with the OCL of its glucan substrate, and it has a very high affinity for the longer glucan chains of dp approximately 20, rendering the enzyme catalytically incapable at longer chain lengths. Based on the observations in this study, we propose that during amylopectin synthesis shorter A and B1 chains are extended by SSI up to a critical chain length that soon becomes unsuitable for catalysis by SSI and hence cannot be elongated further by this enzyme. Instead, SSI is likely to become entrapped as a relatively inactive protein within the starch granule. Further glucan extension for continuation of amylopectin synthesis must require a handover to other SS enzymes which can extend the glucan chains further or for branching by branching enzymes. If this is correct, this proposal provides a biochemical basis to explain how the specificities of various SS enzymes determine and set the limitations on the length of A, B, C chains in the starch granule.     

A single residue mutation abolishes attachment of the CBM26 starch-binding domain from <Emphasis Type="Italic">Lactobacillus amylovorus</Emphasis> α-amylase

Starch is degraded by amylases that frequently have a modular structure composed of a catalytic domain and at least one non-catalytic domain that is involved in polysaccharide binding. The C-terminal domain from the Lactobacillus amylovorus α-amylase has an unusual architecture composed of five tandem starch-binding domains (SBDs). These domains belong to family 26 in the carbohydrate-binding modules (CBM) classification. It has been reported that members of this family have only one site for starch binding, where aromatic amino acids perform the binding function. In SBDs, fold similarities are better conserved than sequences; nevertheless, it is possible to identify in CBM26 members at least two aromatic residues highly conserved. We attempt to explain polysaccharide recognition for the L. amylovorus α–amylase SBD through site-directed mutagenesis of aromatic amino acids. Three amino acids were identified as essential for binding, two tyrosines and one tryptophan. Y18L and Y20L mutations were found to decrease the SBD binding capacity, but unexpectedly, the mutation at W32L led to a total loss of affinity, either with linear or ramified substrates. The critical role of Trp 32 in substrate binding confirms the presence of just one binding site in each α-amylase SBD.     

Role of the N-terminal starch-binding domains in the kinetic properties of starch synthase III from Arabidopsis thaliana

Starch synthase III (SSIII), one of the SS isoforms involved in plant starch synthesis, has been reported to play a regulatory role in the synthesis of transient starch. SSIII from Arabidopsis thaliana contains 1025 amino acid residues and has an N-terminal transit peptide for chloroplast localization which is followed by three repeated starch-binding domains (SBDs; SSIII residues 22-591) and a C-terminal catalytic domain (residues 592-1025) similar to bacterial glycogen synthase. In this work, we constructed recombinant full-length and truncated isoforms of SSIII, lacking one, two, or three SBDs, and recombinant proteins, containing three, two, or one SBD, to investigate the role of these domains in enzyme activity. Results revealed that SSIII uses preferentially ADPGlc, although UDPGlc can also be used as a sugar donor substrate. When ADPGlc was used, the presence of the SBDs confers particular properties to each isoform, increasing the apparent affinity and the V max for the oligosaccharide acceptor substrate. However, no substantial changes in the kinetic parameters for glycogen were observed when UDPGlc was the donor substrate. Under glycogen saturating conditions, the presence of SBDs increases progressively the apparent affinity and V max for ADPGlc but not for UDPGlc. Adsorption assays showed that the N-terminal region of SSIII, containing three, two, or one SBD module have increased capacity to bind starch depending on the number of SBD modules, with the D23 protein (containing the second and third SBD module) being the one that makes the greatest contribution to binding. The results presented here suggest that the N-terminal SBDs have a regulatory role, showing a starch binding capacity and modulating the catalytic properties of SSIII.     

A starch-binding domain identified in α-amylase (AmyP) represents a new family of carbohydrate-binding modules that contribute to enzymatic hydrolysis of soluble starch

A novel starch-binding domain (SBD) that represents a new carbohydrate-binding module family (CBM69) was identified in the α-amylase (AmyP) of the recently established alpha-amylase subfamily GH13_37. The SBD and its homologues come mostly from marine bacteria, and phylogenetic analysis indicates that they are closely related to the CBM20 and CBM48 families. The SBD exhibited a binding preference toward raw rice starch, but the truncated mutant (AmyP_ΔSBD) still retained similar substrate preference. Kinetic analyses revealed that the SBD plays an important role in soluble starch hydrolysis because different catalytic efficiencies have been observed in AmyP and the AmyP_ΔSBD.     

Molecular and Biochemical Analyses of CbCel9A/Cel48A,a Highly Secreted Multi-Modular Cellulase by Caldicellulosiruptor bescii during Growth on Crystalline Cellulose

During growth on crystalline cellulose, the thermophilic bacterium Caldicellulosiruptor bescii secretes several cellulose-degrading enzymes. Among these enzymes is CelA (CbCel9A/Cel48A), which is reported as the most highly secreted cellulolytic enzyme in this bacterium. CbCel9A/Cel48A is a large multi-modular polypeptide, composed of an N-terminal catalytic glycoside hydrolase family 9 (GH9) module and a C-terminal GH48 catalytic module that are separated by a family 3c carbohydrate-binding module (CBM3c) and two identical CBM3bs. The wild-type CbCel9A/Cel48A and its truncational mutants were expressed in Bacillus megaterium and Escherichia coli, respectively. The wild-type polypeptide released twice the amount of glucose equivalents from Avicel than its truncational mutant that lacks the GH48 catalytic module. The truncational mutant harboring the GH9 module and the CBM3c was more thermostable than the wild-type protein, likely due to its compact structure. The main hydrolytic activity was present in the GH9 catalytic module, while the truncational mutant containing the GH48 module and the three CBMs was ineffective in degradation of either crystalline or amorphous cellulose. Interestingly, the GH9 and/or GH48 catalytic modules containing the CBM3bs form low-density particles during hydrolysis of crystalline cellulose. Moreover, TM3 (GH9/CBM3c) and TM2 (GH48 with three CBM3 modules) synergistically hydrolyze crystalline cellulose. Deletion of the CBM3bs or mutations that compromised their binding activity suggested that these CBMs are important during hydrolysis of crystalline cellulose. In agreement with this observation, seven of nine genes in a C. bescii gene cluster predicted to encode cellulose-degrading enzymes harbor CBM3bs. Based on our results, we hypothesize that C. bescii uses the GH48 module and the CBM3bs in CbCel9A/Cel48A to destabilize certain regions of crystalline cellulose for attack by the highly active GH9 module and other endoglucanases produced by this hyperthermophilic bacterium.     

X-ray crystal structure of a non-crystalline cellulose-specific carbohydrate-binding module: CBM28

Natural cellulose exists as a composite of different forms, which have historically been broadly characterized as "crystalline" or "amorphous". The recognition of both of these forms of cellulose by the carbohydrate-binding modules (CBM) of microbial glycoside hydrolases is central to natural and efficient biotechnological conversion of plant cell wall biomass. There is increasing evidence that, at least some, individual binding modules target distinct and different regions of non-crystalline "amorphous" cellulose. Competition experiments show that CBM28 modules do not compete with CBM17 modules when binding to non-crystalline cellulose. The structure of the BspCBM28 (http://afmb.cnrs-mrs.fr/CAZY/) module from the Bacillus sp. 1139 family GH5 endoglucanase, comprising a 191 amino acid protein, has therefore been determined at 1.4A resolution using single isomorphous replacement with anomalous scattering methods. The structure reveals a "beta-jelly roll" topology, with high degree of similarity to the structure of CBM17 domains. Sequence and structural conservation strongly suggests that these two families of domains have evolved through gene duplication and subsequent divergence. The ligand-binding site "topographies" of CBMs from families 28, 17 and 4 begins to shed light on the differential recognition of non-crystalline cellulose by multi-modular plant cell wall-degrading enzymes.     

Potential mechanism(s) involved in the regulation of glycogen synthesis by insulin

Stimulation of glycogen synthesis is one of the major physiological responses modulated by insulin. Although, details of the precise mechanism by which insulin action on glycogen synthesis is mediated remains uncertain, significant advances have been made to understand several steps in this process. Most importantly, recent studies have focussed on the possible role of glycogen synthase kinase-3 (GSK-3) and glycogen bound protein phosphatase-1 (PP-1G) in the activation of glycogen synthase (GS) - a key enzyme of glycogen metabolism. Evidence is also accumulating to establish a link between insulin receptor induced signaling pathway(s) and glycogen synthesis. This article summarizes the potential contribution of various elements of insulin signaling pathway such as mitogen activated protein kinase (MAPK), protein kinase B (PKB), and phosphatidyl inositol 3-kinase (PI3-K) in the activation of GS and glycogen synthesis.     

Dividing the Large Glycoside Hydrolase Family 43 into Subfamilies: a Motivation for Detailed Enzyme Characterization

The rapid rise in DNA sequencing has led to an expansion in the number of glycoside hydrolase (GH) families. The GH43 family currently contains α-l-arabinofuranosidase, β-d-xylosidase, α-l-arabinanase, and β-d-galactosidase enzymes for the debranching and degradation of hemicellulose and pectin polymers. Many studies have revealed finer details about members of GH43 that necessitate the division of GH43 into subfamilies, as was done previously for the GH5 and GH13 families. The work presented here is a robust subfamily classification that assigns over 91% of all complete GH43 domains into 37 subfamilies that correlate with conserved sequence residues and results of biochemical assays and structural studies. Furthermore, cooccurrence analysis of these subfamilies and other functional modules revealed strong associations between some GH43 subfamilies and CBM6 and CBM13 domains. Cooccurrence analysis also revealed the presence of proteins containing up to three GH43 domains and belonging to different subfamilies, suggesting significant functional differences for each subfamily. Overall, the subfamily analysis suggests that the GH43 enzymes probably display a hitherto underestimated variety of subtle specificity features that are not apparent when the enzymes are assayed with simple synthetic substrates, such as pNP-glycosides.     

The carbohydrate-binding domain of overexpressed STBD1 is important for its stability and protein–protein interactions

STBD1 (starch-binding domain-containing protein 1) belongs to the CBM20 (family 20 carbohydrate binding module) group of proteins, and is implicated in glycogen metabolism and autophagy. However, very little is known about its regulation or interacting partners. Here, we show that the CBM20 of STBD1 is crucial for its stability and ability to interact with glycogen-associated proteins. Mutation of a conserved tryptophan residue (W293) in this domain abolished the ability of STBD1 to bind to the carbohydrate amylose. Compared with the WT (wild-type) protein, this mutant exhibited rapid degradation that was rescued upon inhibition of the proteasome. Furthermore, STBD1 undergoes ubiquitination when expressed in COS cells, and requires the N-terminus for this process. In contrast, inhibition of autophagy did not significantly affect protein stability. In overexpression experiments, we discovered that STBD1 interacts with several glycogen-associated proteins, such as GS (glycogen synthase), GDE (glycogen debranching enzyme) and Laforin. Importantly, the W293 mutant of STBD1 was unable to do so, suggesting an additional role for the CBM20 domain in protein–protein interactions. In HepG2 hepatoma cells, overexpressed STBD1 could associate with endogenous GS. This binding increased during glycogenolysis, suggesting that glycogen is not required to bridge this interaction. Taken together, our results have uncovered new insights into the regulation and binding partners of STBD1.     

Rice has two distinct classes of protein kinase genes related to SNF1 of Saccharomyces cerevisiae , which are differently regulated in early seed development

We have isolated five cDNA clones (osk1–5) for protein kinases from rice which are related to SNF1 protein kinase of Saccharomyces cerevisiae. Based on the sequence homology, these cDNAs can be classified into two groups, group 1 (osk1) and group 2 (osk2–5). The products of these genes were demonstrated to be functional SNF1-related protein kinases by in vitro and in vivo experiments. Recombinant proteins expressed from both groups of genes were fully active as protein kinases and could phosphorylate SAMS peptide, a substrate specific for the SNF1/AMPK family, as well as themselves (autophosphorylation). Moreover, expression of osk3 cDNA in yeast snf1 mutants restored SNF1 function. Northern blot analyses showed differential expression of these two gene groups; group 1 is expressed uniformly in growing tissues (young roots, young shoots, flowers, and immature seeds), whereas group 2 is strongly expressed in immature seeds. SNF1-related protein kinases have been reported from different plant species, such as rye, barley, Arabidopsis, tobacco, and potato, while the type of gene strongly expressed in immature seeds is known only in cereals such as rye, barley, and, from our findings, in rice. Expression levels of the group 2 genes were further analyzed in seeds during seed maturation. Expression is transiently increased in the early stages of seed maturation and then decreases. The expression peak precedes those of the sbe1 and waxy genes, which are involved in starch synthesis in rice. Taken together, these findings suggest that group 2 OSK genes play important roles in the early stages of endosperm development in rice seeds. Received: 30 April 1998 / Accepted: 20 August 1998