©

New Phytologist

(2001)

151

: 000–000

www.newphytologist.com

1

Review

Blackwell Science Ltd

Research review

The CDPK superfamily of

protein kinases

Alice C. Harmon

1,2

, Michael Gribskov

3

, Erika Gubrium

1

and Jeffrey F. Harper

4

1

Department of Botany, PO Box 118526, University of Florida, Gainesville, Florida 32611–8526, USA;

2

Plant Molecular and Cellular Biology Program, PO Box 118526, University of Florida, Gainesville,

Florida 32611–8526, USA;

3

San Diego Supercomputer Center, University of California San Diego,

9500 Gilman Drive, La Jolla, California 92093–0537, USA;

4

Department of Cell Biology, The Scripps

Research Institute, 10550 N. Torrey Pines Rd. La Jolla, California 92037, USA

Summary

The CDPK superfamily consists of six types of protein kinases, which differ in theegulatory domains they contain. CDPKs (calcium-dependent protein kinases orcalmodulin-like domain protein kinases) are activated by the binding of calcium totheir calmodulin-like regulatory domains. The carboxyl terminal domains of CRKs(CDPK-related kinases) have sequence similarity to the regulatory domains ofCDPKs, but do not bind calcium. PPCKs (PEP carboxylase kinases) contain only acatalytic domain. PRKs (PPCK-related kinases) have a carboxyl-terminal domain thathas no similarity to that of any other member of the superfamily. CCaMKs (calciumand calmodulin regulated kinases) bind both calcium ions and the calcium/calmodulincomplex, whereas CaMKs (calmodulin-dependent protein kinases) bind the calcium/calmodulin complex, but not calcium. Phylogenetic trees constructed from aminoacid sequences of catalytic or regulatory domains show that CDPKs and CRKs areclosely related and might share a common ancestor. Plant CCaMKs and CaMK forma group more closely related to protozoan, than to plant, CDPKs. Intron analysis ofthe 42 CDPK, CRK, PPCK, and PRK genes from

Arabidopsis

supports the structureof the gene trees, the possibility that PPCKs/PRKs belong to the CDPK superfamily,and suggests that several introns have been added during evolution of the family.

©

New Phytologist

(2001)

151

: 000–000

Author for correspondence:

Alice C. Harmon

Tel: +1 (352) 392 3217 Fax: +1 (352) 392 3993

Email: harmon@botany.ufl.edu

Received:

20 September 2000

Accepted:

27 March 2001

Key words:

CDPK, CRK, CCaMK, CaMK, PPCK, PRK, evolution, intron.

Introduction

Plants contain a large number of genes that encode proteinkinases related to CDPKs (Fig. 1). CDPKs (calcium-dependentprotein kinases; gene name CPK) are found in vascular andnonvascular plants, green algae, and certain protozoa (ciliatesand apicomplexans). These enzymes contain four domains,which, listed in order from the amino to the carboxyl terminus,are: an amino terminal domain of variable length and sequence;a protein kinase catalytic domain; an autoinhibitory junction

domain; and a calmodulin-like calcium-binding domain. Manyof these kinases contain N-terminal myristoylation sites, andmyrystoylation and association with membranes of severalCDPKs has been demonstrated (Ellard-Ivey

et al.

, 1999;Martin & Busconi, 2000). Activity of CDPKs is regulated bycalcium ions, and these kinases are thought to function insignal transduction pathways that utilize changes in cellularCa

2

+

concentration to couple cellular responses to extracellularstimuli. In Arabidopsis, CDPKs are encoded by a gene familycontaining 34 members. The structure and function of CDPKs

Research review

www.newphytologist.com

©

New Phytologist

(2001)

151

: 000–000

Review2

have been the subject of recent reviews (Harmon

et al.

, 2000;Hrabak, 2000). Additional information on members of theCDPK superfamily from Arabidopsis may be found at thePlantsP Web Site (http://plantsp.sdsc.edu).

In addition to CDPKs, Arabidopsis and other angiospermscontain a related family of protein kinases called CRKs forCDPK-

r

elated

k

inases. These kinases have catalytic domainsthat are related in sequence to those of CDPKs, but theircarboxyl terminal domains contain apparently degeneratecalcium-binding sites. Examples of CRKs from Arabidopsisand maize do not bind calcium and are not regulated by calcium(Lindzen & Choi, 1995; Furumoto

et al.

, 1996). It is notknown if their junction domains function as autoinhibitorydomains, nor how their activities are regulated

in vivo

.PPCKs phosphorylate and regulate the activity of PEP

carboxylase, an enzyme important in C

4

metabolism. The roleof these kinases in C

3

plants has not yet been identified.PPCKs contain only amino-terminal and catalytic domains.Their activity is regulated at the level of transcription, whichis subject to diurnal variation (Jiao

et al.

, 1991; Hartwell

et al.

,1999; Nimmo, 2000; Taybi

et al.

, 2000). The PPCK-relatedkinases, PRKs, have catalytic domains that are closely relatedto those of PPCKs, but unlike PPCKs they have carboxyl terminal

domains. These domains have no similarity to any othermembers of the CDPK superfamily.

CaMKs are calmodulin-dependent protein

k

inases and theiractivities depend (at least initially) on calmodulin in the formof a calcium/calmodulin complex. These enzymes contain acatalytic domain, a domain that has the dual functions ofautoinhibition and calmodulin-binding, and a carboxyl-terminal domain that is involved in protein–protein interac-tions. A large family of these enzymes is known in animals andyeast, but there is only one example reported for a plant (apple)(Watillon

et al.

, 1995).The final members of the CDPK family are the CCaMKs,

which contain a catalytic domain, autoinhibitory domain,and a calcium-binding domain that is similar to visinin. Theseenzymes are subject to complex regulation by both calciumions and the calcium/calmodulin complex. CCaMKs have beenfound in lily and tobacco (Patil

et al.

, 1995; Liu

et al.

, 1998),but to date none have been reported by the Arabidopsis genomesequencing project (99% complete).

To examine possible evolutionary relationships betweenmembers of the CDPK superfamily, we constructed proteinsequence trees for Superfamily members from all species. Tocomplement this analysis, the exon/intron patterns of all 42

1

CaM-like

Visinin-Like

Autoinhibitor

CDPK and related kinases

CRK (Calcium-independent; means of regulation unknown)

CCaMK (Ca2+ and/or Ca2+/ Calmodulin regulated)

CaMK (Ca2+/Calmodulin-dependent)

PPCK (Regulated at the transcription level)

CDPK (Calcium-dependent)

PRK (Regulation unknown)

NH2 COOHNo functional EF-hands

NH2COOH

NH2 COOHAssociation Domain

COOHNH2

NH2 COOH

NH2Function unknown

COOH

Fig. 1 Diagrams of calcium dependent protein kinases (CDPK) and related protein kinases. Reading left to right, the domains of the protein kinases are: amino-terminal domain, which is variable in length and sequence (thin line); protein kinase catalytic domain (grey box); junction or autoinhibitory domain (thick line); C-terminal domain (black box). The C-terminal domains of CDPKs and calcium and calmodulin regulated kinases (CCaMKs) are calcium-binding, regulatory domains. EF-hands within the calmodulin-like (CaM-like) or visinin-like domains are shown as grey bars. The C-terminal domain of nonplant CaMKs has a protein-association function. It is unknown if the C-terminal domain of CDPK-related-kinases (CRKs), which has some sequence similarity to CaM-like domains, or of PPCK(PEP carboxylase kinases)-related kinases (PRK), which has no sequence similarity to CaM-like domains, has a function.

2

3

Fig. 2

(

opposite

) Neighbour-joining tree based on catalytic domains. The amino acid sequences of the catalytic domains of calcium dependent protein kinases (CDPKs), CDPK-related-kinases (CRKs), calcium and calmodulin regulated kinases (CCaMKs), PEP carboxylase kinases (PPCKs), PPCK(PEP carboxylase kinases)-related kinases (PRKs), calmodulin-dependent protein kinases (CaMKs), and SNF-1 kinases were analysed as described in Methods. The sequence names include the gene name or common name followed by the NCBI Geninfo database number (GI number), or

Arabidopsis

systematic number (e.g. at5g04870). The major branches of plant CDPKs were assigned arbitrary group numbers. Arabidopsis sequences are shown in green.

4

Research review

©

New Phytologist

(2001)

151

: 000–000

www.newphytologist.com

Review 3

Research review

www.newphytologist.com

©

New Phytologist

(2001)

151

: 000–000

Review4

members of the CDPK superfamily from Arabidopsis werecompared.

Materials and Methods

Cloning of AtCPK cDNAs

AtCPK cDNAs were amplified by RT PCR from RNA isolatedfrom Arabidopsis WS roots, shoots, or seedlings. Sense primerscontained a restriction site (

Sal

I or

Xho

1) followed by the first29 nucleotides of the open reading frame. The antisense primerscontained a restriction site (

Spe

I or

Avr

I) followed by nucleotidescomplementary to the last 29 nucleotides of the open readingframe. The cloning vector was a yeast/bacteria shuttle vector(YX-GFP JH1), which adds the coding sequences for GFPand a 6His tag at the 3

′

end of the inserted CPK cDNA.

Construction of trees

Sequences of CDPK family members from all species in theviridiplantae were retrieved from the GenBank database andsubjected to phylogenetic analysis. In cases where the GenBanksequences contained errors predicted from comparison withknown CDPK sequences, hand-corrected sequences were used.Most errors occurred due to incorrect intron splice site predictions.Sequences of Arabidopsis isoforms were also obtained fromThe Institute for Genomic Research (ftp://ftp.tigr.org/pub/data/a_thaliana/ath1/PUBLICATION_RELEASE/) and theMunich Information Center for Protein Sequences (ftp://ftp.mips.biochem.mpg.de/pub/cress/). Protein trees were con-structed using the Neighbour-Joining method as implementedin the ClustalW (Thompson

et al.

, 1994) program. Alignmentswere constructed by a combination of automatic and manualprocedures; briefly, sequences, omitting the most divergent,were initially aligned with ClustalW. This alignment was usedto construct an Evolutionary Profile (Gribskov & Veretnik,1996). The sequences were then aligned to the profile toconstruct the multiple alignment. Finally this alignment washand edited to correct errors. Trees were constructed aftereliminating divergent regions with unreliable local alignmentsfrom the overall alignments. Gaps were included and theClustalW adjustment for multiple mutations was used. Bootstrapanalysis was performed using 1000 iterations of tree building.

Intron pattern analysis

Introns in genes were identified by comparison of mRNA andgenomic sequences for Arabidopsis CPKs 1–12, 15, 16, 21;

CRKs 1 and 5; and PEPCK1; and for

Plasmodium falciparum

CPK1. The intron patterns for CPKs 13, 14, 17–20, and 22–34 were predicted by inspection of genomic sequences forintrons having the flanking sequence, 5

′

gt … ag 3

′

, and byusing the predicted amino acid sequences as a guide.

Results and Discussion

Sequences of CDPK family members from all species,including 34 CDPKs, eight CRKs, two PPCKs, and twoPRKs from Arabidopsis, were retrieved from the GenBankdatabase and subjected to sequence similarity analysis. Theunrooted tree that resulted from analysis of catalytic domainsis shown in Fig. 2. Representative animal and fungal proteinkinases were included in the analysis so that similarities ofCDPK subfamilies these groups could be determined. Previousanalysis of protein kinase sequences indicated that CDPKsbelong to the CAMK group, which includes animal and yeastcalmodulin-dependent protein kinases, SNF1-related kinasesand AMP-dependent protein kinase (Hanks & Hunter, 1995).The SNF-1 kinases, including two from Arabidopsis, and thenonplant CaMKs clustered together, indicating that these kinasegroups are distinct. The single available example of a plantCaMK, mdccamk from apple, was located with the CCaMKs.

Plant CDPKs and CRKs clustered on the other end of thetree, suggesting that they may share a common evolutionaryorigin. While the CPKs can be divided into four subgroupsbased on their positions on the tree, there are no remarkablefeatures or sequence differences that distinguish the groupsfrom each other. Only a few CDPKs have been characterizedin detail, so it remains to be determined whether there are anyfunctional differences between the subgroups.

The PPCKs and PRKs clustered together between CCaMKsand the CPKs/CRKs. The inclusion of PPCKs in the analysiscaused a shift in the positions of groups II and III of CDPKs relativeto the tree shown in Harmon

et al

. (2000). The reason for thischange is due to the presence of insertions in the PEPCK sequences,and the exclusion of areas containing insertions in the analysis.

A second tree was constructed from the analysis ofC-terminal domains of CDPKs, CRKs, and CCaMKs (Fig. 3).This tree differs from the one based on catalytic domains inthe relative positioning of the protozoan CDPKs, which arenot as tightly clustered, and CCaMKs relative to CRKs andgroup IV CDPKs. This change in positions reflects the lack ofEF-hand motifs in CRKs and the fact that the C-terminaldomains of CCaMKs are more similar to visinin than to cal-modulin. The CPKs generally fall into the same groups as inFig. 2, but are less tightly clustered.

5

Fig. 3

(

opposite

) Neighbour-joining tree based on c-terminal domains. The amino acid sequences of the C-terminal domains of calcium dependent protein kinases (CDPKs), CDPK-related-kinases (CRKs), and calcium and calmodulin regulated kinases (CCaMKs) were analysed as described in Methods. The sequence names include the gene name or common name followed by the NCBI Geninfo database number (GI number), or

Arabidopsis

systematic number (e.g. at5g04870) The major branches of plant CDPKs were given the same Group number as in Fig. 2. Arabidopsis sequences are shown in green. Note that due to the considerably greater divergence between the calmodulin-like domains, that this tree is less reliable the that for the catalytic domain.

6

Research review

©

New Phytologist

(2001)

151

: 000–000

www.newphytologist.com

Review 5

Research review

www.newphytologist.com

©

New Phytologist

(2001)

151

: 000–000

Review6

Research review

©

New Phytologist

(2001)

151

: 000–000

www.newphytologist.com

Review 7

The intron patterns for 34 CDPKs, eight CRKs, two PPCKs,and two PRKs from Arabidopsis were determined from experi-mental data or by prediction. The sizes of the introns variedgreatly (Fig. 4) but there was conservation in their locations(Fig. 5). Introns were found at each of the seven locations

shown as black triangles in Fig. 5 in more than half of the 46genes. Three of these introns were phase 0 (i.e. the intron wasinserted between two complete codons), three were phase 1(inserted after the first nucleotide of a codon), and one wasphase 2 (inserted after the second nucleotide of a codon). Four

Fig. 5 Summary of intron positions within the coding regions. Intron patterns for each group of kinases are indicated as triangles pointing to sites of insertion. The Roman numerals to the left indicate the group number in Figs 2 and 3. The roman numerals within the bars refer to the 12 conserved subdomains of protein kinases, J refers to the junction (autoinhibitory) domain, and EF1–4 are EF hands. The number within the triangle indicates the phase of the intron. The seven introns common to most members of Groups I–III are indicated by black triangles and additional introns are white. In the Group IV/CRK genes, introns at similar positions to those in Groups I–III are black. Gray triangles mark the positions of conserved introns that are ‘missing’ in a particular gene. Note that introns split EF4 in most genes, EF3 in Group II genes, and subdomain VII of Group IV/CRK genes. *Intron patterns for these genes were predicted from genomic sequences; 1The complete genomic sequence for CPK 7 is not available (February, 2001); 2CPKs 14 and 32 may have an addition intron in the coding region C-terminal to EF hand 4.

7

8

Fig. 4

(

opposite

) Scaled diagram of the coding regions of Arabidopsis CPK, CDPK-related-kinases (CRK), and PEPCK Genes. Exons (boxes) and introns (lines) were determined from comparison of mRNA and genomic sequences, or predicted (*) from genomic sequences. CPK genes are named by their gene numbers. Sequences are ordered according to the groups determined from the phylogenetic trees. The beginning of each catalytic domain is aligned, so that variation in the lengths of amino-terminal domains is easily seen. Protein coding regions are: open boxes, variable amino-terminal domain; narrow bars coloured maroon to purple; the 12 conserved subdomains of protein kinase catalytic domains; blue box, autoinhibitory domain; bars shaded in yellow to orange, the four EF-hands of the calmodulin-like domain. Some of these EF-hands may be nonfunctional due to sequence differences. The putative autoinhibitory domain and the degenerate EF hands of CRKs are shown in lighter shades of blue, yellow and orange. The question mark at the end CPK7 indicates that full-length genomic sequence is unavailable. The sequence for CPK25 ends after the second EF hand.

Research review

www.newphytologist.com

©

New Phytologist

(2001)

151

: 000–000

Review8

of the conserved intron positions were located betweenregions coding for functional motifs, but one interrupted sub-domain XI of the catalytic domain and two others interruptedEF hands 3 and 4, respectively. The only position at which anintron was found in every gene was the one that interruptssubdomain XI.

The number of introns in Groups I–III varied from five toeight, and one typical intron pattern dominated in each group.The Group IV CPKs and CRKs have 11 and 10 introns,respectively. For 10 of these introns, the positions and phases(seven phase 0, two phase 1, and one phase 2) were identical.The positions and phases of four introns were identical to onesin the Group I–III CPKs. These observations support the ideathat genes within a given subgroup have a common origin.

The similarities in sequence and intron patterns of the CPKsin Groups I–III suggest they proliferated by gene duplication.This idea is also supported by occurrence of four highlyrelated CPKs (CPKs 21, 23, 27, and 31 in Group II) in acluster on Chromosome IV. The differences in sequences andintron patterns between the Group IV/CRK genes and thosein Groups I–III, suggests that these two groups may have haddifferent origins. The similarity of the Group IV CPKs andCRKs raises the possibility that CRKs may have derived fromthe Group IV CPKs.

PPCKs have only one intron, and its phase and locationnear the end of the subdomain XI are identical to those forthe other genes. PRKs have introns at two positions conservedin other family members. Introns are not found at these loca-tions in any of the 41 SNF-1 related kinases (SNRKs) fromArabidopsis (data not shown), and this supports the idea thatPPCKs are more closely related to the CPK family than theSNRKs.

The similar pattern of intron insertion in regions encodingsubdomains within the catalytic and CaM-like domains, ofGroup I–III CPKs supports the ‘intron-early’ hypothesis(Gilbert, 1978), which states that introns are the descendentsof spacers between ancestral minigenes. The presence of anintron at the junction domain/CLD border in all CPKs,except 12 and 14, supports the idea that CDPKs may havearisen by the fusion of a kinase gene and a CaM-like proteingene. However, several observations support the ‘intron-late’hypothesis (Cavalier-Smith, 1991): introns vary in length atall positions; introns occur at several sites in addition to thebasic pattern; introns disrupt conserved catalytic subdomainsand EF hands; and many phase 1 and 2 introns are present.These observations suggest these introns may have been addedafter the ancestral gene was formed. Similar findings resultedfrom the analysis of 135 Arabidopsis P450 genes (Paquette

et al.

, 2000).The sequences of the N-terminal regions are highly variable

despite being encoded by the same exon that bears a large partof the conserved catalytic domain. The function of this regionof the protein has not yet been determined. It is possible thatits function may differ between isoforms and that the mutation

rate in the coding sequence may be greater due to a lack ofevolutionary pressure to maintain it.

Acknowledgements

This work was supported by grants (MCB-9973770 to A.C.H.and DBI-9975808 to J.C. Walker

et al

.) from the NationalScience Foundation. The PlantsP database and web site areunder the direction of M.G. and D. Smith and are supportedby the National Science Foundation grant (DBI-9975808 toJ.C. Walker

et al

.). We thank Estelle Hrabak and CatherineChan for providing data prior to publication and for helpfuldiscussions.

References

Cavalier-Smith T. 1991.

Intron phylogeny: a new hypothesis.

Trends in Genetics

7

: 145–148.

Ellard-Ivey M, Hopkins RB, White TJ, Lomax TL. 1999.

Cloning, expression and N-terminal myristoylation of CpCPK1, a calcium-dependent protien kinase from zucchini (

Curbita pepo

L.).

Plant Molecular Biology

39

: 199–208.

Furumoto T, Ogawa N, Hata S, Izui K. 1996.

Plant calcium-dependent protein kinase-related kinases (CRKs) do not require calcium for their activities.

FEBS Letters

396

: 147–151.

Gilbert W. 1978.

Why genes in pieces?

Nature

271

: 501.

Gribskov M, Veretnik S. 1996.

Identification of sequence pattern with profile analysis.

Hanks SK, Hunter T. 1995.

The eukaryotic protein-kinase superfamily kinase (catalytic) domain-structure and classification.

FASEB Journal

9

: 576–596.

Harmon AC, Gribskov M, Harper JF. 2000. CDPKs – a kinase for every Ca2+ signal? Trends in Plant Science 5: 154–159.

Hartwell J, Gill A, Nimmo GA, Wilkins MB, Jenkins GI, Nimmo HG. 1999. Phosphoenolpyruvate carboxylase kinase is a novel protein kinase regulated at the level of expression. Plant Journal 20: 333–342.

Hrabak EM. 2000. Calcium-dependent protein kinases and their relatives. In: Kreis X, Walker X, eds. Plant protein kinases, 32. San Diego, CA, USA: Academic Press, 185–223.

Jiao JA, Vidal J, Echevarria C, Chollet R. 1991. In vivo Regulatory Phosphorylation Site in C4-Leaf Phosphoenolpyruvate Carboxylase from Maize and Sorghum. Plant Physiology 96: 297–301.

Lindzen E, Choi JH. 1995. A carrot cDNA encoding an atypical protein kinase homologous to plant calcium-dependent protein kinases. Plant Molecular Biology 28: 785–797.

Liu Z, Xia M, Poovaiah BW. 1998. Chimeric calcium/calmodulin-dependent protein kinase in tobacco: differential regulation by calmodulin isoforms. Plant Molecular Biology 38: 889–897.

Martin ML, Busconi L. 2000. Membrane localization of a rice calcium-dependent protein kinase (CDPK) is mediated by myristoylation and palmitoylation. Plant Journal 24: 429–435.

Author X. Year. Title of article. Methods in Enzymology 266: 198–212.Nimmo HG. 2000. The regulation of phosphoenolpyruvate carboxylase

in CAM plants. Trends in Plant Science 5: 75–80.Paquette SM, Bak S, Feyereisen R. 2000. Intron-Exon Organization and

Phylogeny I a Large Superfamily, the Paralogous Cytochrome P450 Genes of Arabidopsis thaliana. DNA Cell Biology 19: 307–317.

Patil S, Takezawa D, Poovaiah BW. 1995. Chimeric plant calcium/calmodulin-dependent protein-kinase gene with a neural visinin-like calcium-binding domain. Proceedings of the National Academy of Sciences, USA 92: 4897–4901.

9

10

11

12

13

14

Research review

© New Phytologist (2001) 151: 000–000 www.newphytologist.com

Review 9

Taybi T, Patil S, Chollet R, Cushman JC. 2000. A Mininal Ser/Thr Protein Kinase Circadianly Regulates Phosphoenolpyruvate carboxylase activity in CAM-induced Leaves of Mesembryanthemum crystallinum. Plant Physiology 123: 1471–1482.

Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improving

the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673–4680.

Watillon B, Kettmann R, Boxus P, Burny A. 1995. Structure of a calmodulin-binding protein kinase gene from apple. Plant Physiology 108: 847–848.

Author Query Form

Journal: New Phytologist

Article: 171

Dear Author,

During the preparation of your manuscript for publication, the questions listed below have arisen. Please attend to these matters and return this form with your proof.

Many thanks for your assistance.

Query Refs.

Query Remarks

1 Furomoto et al., 1996 has been changed to Furumoto et al., 1996 so that this citation matches the list. Which is correct?

2 Please check all figure legends carefully as they have been changed slightly for journal style. please ensure any queries are answered

3 Watilon et al., 1995 has been changed to Watillon et al., 1995 so that this citation matches the list. Which is correct?

4 analysed as described in Methods – please give a brief description of the analysis here as the figure must stand alone without reference to the text.

5 WS – please give in full

6 analysed as described in Methods – please give a brief description to the analysis here as the figure must stand alone without reference to the text

7 EF – again, please give in full

8 EF – please give in full

9 EF – please give in full

10 Paquette, 2000 has been changed to Paquette et al., 2000 so that this citation matches the list. Which is correct?

11 Please check expanded journal titles

12 Gribskov M, Veretnik S. 1996. Identification of sequence pattern with profile analysis. - Please complete this reference. Journal title, volume number and page range are needed

13 Kreis X, Walker X, eds. – please provide editor’s initial(s)

14 Methods in Enzymology Methods in Enzymology 266: 198–212. Please complete this reference. The author(s) names are required as well as the year and the title of the article. Please ensure that this reference is also cited in the text

Query Refs.

Query Remarks