structural similarities between spinach chloroplast and ... · in contrast, some cross-reactivity...

Plant Physiol. (1989) 91, 1393-14010032-0889/89/91/1 393/09/$01 .00/0

Received for publication April 18, 1989and in revised form July 7, 1989

Structural Similarities between Spinach Chloroplast andCytosolic Class I Fructose 1,6-Bisphosphate Aldolases'

Immunochemical and Amino-Terminal Amino Acid Sequence Analysis

James J. Marsh, Kenneth J. Wilson, and Herbert G. Lebherz*Department of Chemistry, San Diego State University, San Diego, California 92182 (J.J.M., H.G.L.), and Applied

Biosystems Inc., 850 Lincoln Center Drive, Foster City, California 94404 (K.J.W.)

ABSTRACT

Immunochemical studies using polyclonal antisera preparedindividually against highly purified cytosolic and chloroplast spin-ach leaf (Spinacia oleracea) fructose bisphosphate aldolasesshowed significant cross reaction between both forms of spinachaldolase and their heterologous antisera. The individual crossreactions were estimated to be approximately 50% in both casesunder conditions of antibody saturation using a highly sensitiveenzyme-linked immunosorbent assay. In contrast, the class Iprocaryotic aldolase from Mycobacterium smegmatis and theclass 11 aldolase from yeast (Saccharomyces cerevisiae) did notcross-react with either type of antiserum. The 29 residue longamino-terminal amino acid sequences of the procaryotic M. smeg-matis and the spinach chloroplast aldolases were determined.Comparisons of these sequences with those of other aldolasesshowed that the amino-terminal primary structure of the chloro-plast aldolase is much more similar to the amino-terminal struc-tures of class I cytosolic eucaryotic aldolases than it is to thecorresponding region of the M. smegmatis enzyme, especially inthat region which forms the first "beta sheet" in the secondarystructure of the eucaryotic aldolases. Moreover, results of asystematic comparison of the amino acid compositions of anumber of diverse eucaryotic and procaryotic fructose bisphos-phate aldolases further suggest that the chloroplast aldolasebelongs to the eucaryotic rather than the procaryotic "family" ofclass I aldolases.

One of the most striking distinctions between eucaryoticand procaryotic life forms is the complexity of their intracel-lular architectures. In contrast to procaryotes, most eucaryoticcells contain specialized membrane-limited intracellular or-

ganelles in which specific biochemical pathways are compart-mentalized. For example, reactions of the Krebs cycle andthose of the oxidative phosphorylation pathway occur inmitochondria while the "light" and "dark" reactions of pho-tosynthesis take place within chloroplast organelles of photo-synthetic eucaryotes. Many glycolytic enzymes in the leavesof higher plants exist as enzyme "pairs," one member of thepair residing in the cytosol and the other member sequesteredwithin the chloroplast organelle (8, 26). The regulation and

'This work was supported, in part, by Research Grant GM 23045from the National Institutes of Health.

structural features ofa number ofthese plant isoenzymes havebeen recently reviewed (21).We have been using the glycolytic enzyme fructose bis-

phosphate aldolase as a model for studies on protein evolu-tion. Early studies by Anderson indicate that although thetwo forms of pea leaf aldolase differ somewhat in charge (4)and amino acid composition (1), they possess very similarcatalytic (4) and structural (2) properties. Results of geneticexperiments have shown that the pea chloroplast aldolase isinherited in a Mendelian (nonmaternal) fashion (3, 28), sug-gesting that this form of aldolase is coded for by a nucleargene. Previous work by Kruger and Schnarrenberger (12, 22)and ourselves (14) showed that the cytosolic and chloroplastforms of spinach leaf aldolase exhibit similar specific catalyticactivities and appear to possess functional carboxy-terminalstructures. However, they are distinctive on the basis ofcharge, subunit molecular weight, thermal stability, aminoacid composition, tryptic peptide patterns and immunologicalproperties. Hence, the two spinach leaf enzymes are presum-ably the products of different structural genes.

All aldolases are further categorized as being either class Ior class II enzymes depending on the mechanism they utilizefor the cleavage of fructose bisphosphate to triose phosphates(19), and the two classes of fructose bisphosphate aldolase areknown to have a somewhat restricted phylogenetic distribu-tion (19). For example, all fructose bisphosphate aldolases ofhigher eukaryotes, including those of plant chloroplasts, areclass I (Schiff's base) enzymes while, most procaryotic cellscontain class II (metallo) aldolases. More recently, it has beendemonstrated that some procaryotic organisms express a classI aldolase (6, 9), which is often induced under nutritionalconditions that require the synthesis of hexoses via the glu-coneogenic pathway.The present studies were undertaken to better clarify the

structural relationships which may exist between the spinach(Spinacia oleracea) chloroplast aldolase and the procaryoticand eucaryotic forms of class I fructose bisphosphate aldolase.In this report, we present the results of immunological exper-iments, NH2-terminal amino acid sequence analysis and com-parisons of the amino acid compositions of procaryotic andeucaryotic aldolases. Our findings suggest that the chloroplastaldolase is structurally much more similar to the class Ieucaryotic aldolases than it is to the procaryotic class I en-

1393

https://plantphysiol.orgDownloaded on January 8, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

https://plantphysiol.org

Plant Physiol. Vol. 91, 1989

zymes. The implications of these findings concerning theprobable evolutionary origin of the chloroplast aldolase arediscussed.

MATERIALS AND METHODS

Materials

Mycobacterium smegmatis (CDC 46) was grown on You-mans and Carlson medium as previously described (9) usingglycerol as the primary carbon source. Bacteria were harvestedby centrifugation at I0,000g, washed twice with ice-coldwater, and then were stored frozen at -20°C until used. Cubesof Fleishmann's cake yeast (Saccharomyces cerevisiae) andfresh spinach leaves (Spinacia oleracea) were purchased lo-cally. The Bradford Protein Assay reagent was obtained fromPierce Chemical Co. (Rockford, IL). Nitrocellulose paper, theprotein A-horseradish peroxidase conjugate and the peroxi-dase activity color development reagent (4-chloro-1-napthol)were purchased from Bio-Rad (Richmond, CA). Polystyreneassay plates (Falcon 3912) used for quantitative ELISA wereobtained from Becton Dickinson Labware Co. (Oxnard, CA).All biochemicals, other than the aldolases, were purchasedfrom Sigma Chemical Co. Other chemicals used were ofreagent grade.

Purification of Fructose Bisphosphate Aldolases

The cytosolic and chloroplast forms ofspinach leafaldolasewere purified from fresh leaf tissue as previously described(14). Purified preparations of both types of spinach aldolasewere judged to be at least 95% pure as determined by electro-phoretic analysis of enzyme preparations in SDS polyacryl-amide slab gels followed by densitometric analysis of gelsstained for protein with Coomassie blue.The M. smegmatis class I aldolase was purified 150-fold

from 100 g of bacteria by the method of Jayanthi Bai et al.(9), with the following modifications. (a) Instead of usingsonication, cells were mechanically disrupted at 4°C with glassbeads in a Bead-Beater apparatus (Biospec Products, Bartles-ville, OK) according to the manufacturer's instructions usingthree, 1 min pulses. (b) Two, instead of one, DEAE-cellulosechromatography steps were performed, the first on a What-man DE-52 cellulose column (2 x 20 cm) equilibrated in 50mM Tris-HCl, 1 mm 2-mercaptoethanol, 0.5 mm EDTA (pH8.0). The column was developed with a 0 to 0.5 M NaClgradient (400 ml total). The second DEAE step was performedon a 1.4 x 16 cm DE-52 column equilibrated in 10 mM Mes,1 mm 2-mercaptoethanol, 0.5 mM EDTA (pH 6.0) and thiscolumn was developed with a 0 to 0.3 M NaCl gradient (200mL total). (c) We omitted the gel filtration step of the originalprocedure. The M. smegmatis aldolase preparation obtainedin this fashion contained a major 39 kD protein (presumedto be the aldolase subunit) plus several minor lower molecularmass (<35 kD) polypeptides as revealed by electrophoreticanalysis. We positively identified the major 39 kD protein asthe fructose bisphosphate aldolase subunit in the followingway. An aliquot of the M. smegmatis aldolase preparationwas subjected to cellulose polyacetate strip electrophoresis(see ref. 14) and briefly (<5 min) stained for aldolase activity

in order to visualize the position of the aldolase band. Thisband (approximately 2 mm wide) was carefully excised fromeach of four identical strips and placed in a minimal volumeofSDS sample buffer. After soaking for 2 h, the entire mixturewas heated at 100°C for 2 min and the resulting solutionsubjected to SDS gel electrophoresis. Only the 39 kD proteinwas present following this separation.The class II (metallo) aldolase derived from yeast was

purified about 10-fold by ammonium sulfate fractionationusing a procedure similar to that originally described by Rutterand co-workers (20). Yeast cakes were suspended in 50 mM2-mercaptoethanol, 0.1 mm ZnSO4, 0.2 mM PMSF, and thecells were harvested by centrifugation at 10,000g for 10 min.After washing the cells two more times, the final cell pastewas resuspended in two volumes ofthe same solution and thecells were disrupted in the "Bead-Beater" apparatus as above.After centrifugation at 1 5,000g for 20 min to remove partic-ulate material, the supernate was adjusted to 80% ammoniumsulfate saturation by the slow addition of the solid salt. Afterstirring for 1 h, precipitated proteins were collected by cen-trifugation and were discarded. The supernate was then ad-justed to 95% saturation by addition of solid ammoniumsulfate and, after stirring, this aldolase-enriched protein frac-tion was collected by centrifugation. The class II yeast aldolaserepresented about 50% of the total protein in this fraction asjudged by electrophoretic analysis.

All aldolase preparations were stored at 4°C as precipitatesin 95% saturated ammonium sulfate prepared in 10 mM Tris-HCI and 1 mM 2-mercaptoethanol (pH 7.6) until used.

Electrophoretic Procedures and Protein Determinations

Crude extracts and purified aldolase preparations were ana-lyzed by SDS-PAGE in 9% polyacrylamide slab gels using thereagents and buffer system suggested by Laemmli (see ref. 14).Gels were fixed in 10% (v/v) acetic acid, 45% (v/v) methanoland were then stained for protein with 0.2% (w/v) Coomassieblue R250 prepared in fixing solution. Gels were destained infixing solution and densitometric analysis of stained gels wasperformed using the Zehneih "soft-laser" scanner. Proteindeterminations were performed by the method of Bradford(see ref. 14), using BSA as the standard.

Immunological Methods

Mono-specific antisera were raised in rabbits against purepreparations of the chloroplast and cytosolic spinach leafaldolases using native protein emulsified in Freund's adjuvantby the procedure recently described by us (14). More thanone rabbit was used for production of each type of antiserumand, after initial screening for specificity, antisera raisedagainst each type of spinach aldolase were pooled separatelyand stored frozen in small aliquots.The abilities of these antibody preparations to recognize

aldolases from various sources was assessed on a qualitativebasis by probing the enzymes, which were previously boundto nitrocellulose paper, using the procedure recommended byBio-Rad Laboratories. Antigen-antibody complexes were vis-ualized using the protein A-horseradish peroxidase conjugatefollowed by histochemical staining of the paper using the

1 394 MARSH ET AL.



SIMILARITIES BETWEEN CHLOROPLAST AND CYTOSOLIC ALDOLASE

method and peroxidase color development reagent suppliedby Bio-Rad.

Quantitative estimates of the degrees of immunologicalrecognition of various aldolases by antisera directed againstthe two spinach aldolases were determined by quantitativeELISA as follows: Samples of aldolase preparations werediluted to 50 ng aldolase/ml in 50 mm sodium bicarbonate(pH 9.6) and these samples were incubated in wells (200 ,uL/well) of microtiter plates overnight at 25°C. After washing toremove any unbound aldolase, remaining protein bindingsites in the wells were saturated by blocking with a 0.1% (w/v) solution of BSA prepared in the same bicarbonate buffer;after the blocking step, wells of microtiter plates were washedwith "antibody binding" buffer (20 mM NaPO4,150 mM NaCl[pH 7.4] which also contained freshly added Tween-20 andBSA, final concentrations, 0.05% [v/v] and 0.1% [w/v], re-spectively). Increasing concentrations ofantisera (in triplicate)were then added to these wells and, after 2.5 h, unboundprotein was removed from the wells by washing with buffer.Finally, the relative amounts of rabbit immunoglobulin Gbound to each well was determined by incubating the wellsfor 1 h with a 1:5000 dilution of the protein A-horseradishperoxidase conjugate and, after washing, each well was treatedwith 250 uL of peroxidase substrate solution (50 mMNa2HPO4, 25 mm citric acid [pH 5.0] containing 5 mM 0-phenylenediamine and 1 mM H202, both added immediatelybefore use). The reaction was monitored by measuring theabsorbance of each well at 490 nm 20 min after the additionof substrate in an automatic Bio-Tek Model EL-3 10 ELISAreader. In all cases, a linear increase in absorbance was ob-served throughout the 20 min incubation period. A controlconsisting of all reagents except antigen was performed ateach antibody dilution to assess, and correct for, any nonspe-cific binding of immunoglobulin G to the wells.

Amino Acid Sequence Analysis

NH2-terminal amino acid sequence analysis was performedon SDS polyacrylamide gel electroblotted samples of chloro-plast and M. smegmatis aldolase according to the method ofMatsudaira (17) using a model 477A gas-phase sequencerwith an on-line model 120A PTH analyzer (Applied Biosys-tems, Foster, City, CA). Chemicals from the same source wereused in each instrument. Data reduction/quantitation wascarried out using the software developed for sequence callingby the same company.

See the legends of specific figures or tables for any additionalmethods or procedures.

RESULTS

Immunological Cross-Reactivity Exhibited by theCytosolic and Chloroplast Forms of Spinach LeafAldolase

Results ofan electrophoretic analysis of the chloroplast andcytosolic aldolase preparations used in this study are shownin Figure 1. The subunit molecular masses were estimated tobe 38 kD and 40 kD, respectively, as previously described(14). The two forms of aldolase did not cross-react in Ouch-

1 2

38kD- _ m -4OkD

Figure 1. Electrophoretic analysis of spinach leaf aldolases. Purifiedpreparations of spinach chloroplast (4.8 jig, lane 1) and cytosolic (6.3jig, lane 2) aldolases were electrophoresed on an SDS-polyacrylamideslab gel and stained for protein with Coomassie blue as described in"Materials and Methods." Subunit molecular masses were determinedas previously describedi (14).

terlony double-diffusion tests (14), a procedure which, how-ever, only detects insoluble (precipitating) immune com-plexes. In contrast, some cross-reactivity between the twospinach aldolases was reported by Kruger and Schnarrenber-ger (12) using a more sensitive immunological technique.Consequently, we investigated in more detail the immunolog-ical relationships which may exist between the two spinachleaf aldolases and included the procaryotic class I aldolasederived from M. smegmatis and the prototypic class II yeastaldolase in these analyses for comparative purposes. We chosethe M. smegmatis aldolase since, unlike other class I procar-yotic aldolases, the M. smegmatis enzyme is similar to class Ieucaryotic aldolases in terms of its subunit molecular mass(about 40 kD) and its tetrameric structure (9).As shown in Figure 2, both spinach leaf aldolases tested

positive in qualitative "dot blot" experiments when probedwith antisera raised against pure preparations of the chloro-plast and cytosolic aldolases. In contrast, neither type ofantiserum recognized the class I procaryotic aldolase, nor thestructurally and catalytically distinct class II yeast aldolase.Similar results were obtained when several different, inde-pendently generated, antisera were used in dot blot analysis(data not shown).More quantitative estimates of the degrees of immunolog-

ical cross-reactivity exhibited by the two spinach aldolaseswere obtained using a quantitative ELISA. In contrast toOuchterlony analysis, this technique (as well as the dot blotprocedure) will detect all types of immune complexes, includ-ing nonprecipitating ones. Equal amounts of each spinachaldolase, as well as the procaryotic class I enzyme and the

1 395




1 2 3 4

Figure 2. Qualitative "dot blot" analysis of various aldolases probedwith antisera raised against the cytosolic and chloroplast spinach leafaldolases. Samples (about 2 gsg) of spinach chloroplast (lane 1),spinach cytosolic (lane 2), M. smegmatis (lane 3), and yeast (lane 4)aldolases were spotted onto nitrocellulose paper and, after the block-ing reaction, each section of the nitrocellulose paper containing anindividual aldolase was incubated independently in 16 mm plasticwells with 0.5 ml of a 1:50 dilution of either antichloroplast (row A) oranticytosolic (row B) antisera. Finally, the blots were developed asdescribed under "Materials and Methods."

class II yeast aldolase, were incubated with increasing levelsof each antiserum and the relative amounts of antigen-anti-body complexes formed were quantitated using the proteinA-horseradish peroxidase conjugate system as described under"Materials and Methods." Typical immunotitration curvesobtained in these experiments are presented in Figure 3. Asshown, the curves obtained with both spinach aldolases, ti-trated with each type of antiserum, were hyperbolic andreached saturation at high levels of antisera. As expected,neither type of antiserum was able to bind to microtiter wellscontaining either the class I procaryotic aldolase or the classII yeast enzyme. Double-reciprocal plots of the immunotitra-tion data (Fig. 4) were then constructed to calculate thetheoretical maximum degrees of immunological cross-reactiv-ity expressed by the two spinach aldolases. Comparisons ofthe "y-intercepts" in these double-reciprocal plots compiledfrom a number of independent immunotitration experiments(Table I) showed that about 60% of the antibody populationsraised against the chloroplast aldolase cross-reacted with thecytosolic enzyme and that a similar percentage (about 50%)of the antibody populations raised against the cytosolic aldo-lase cross-reacted with the chloroplast enzyme. Thus, the twoantisera exhibit the immunological phenomenon known asreciprocity (25), a property which is taken as evidence thatthe measured cross-reactivity is actually due to the presenceof homologous antibody binding sites on the cross-reactingproteins.

Finally, we ruled out the unlikely possibility that the ob-served immunological cross-reactivity exhibited by the twospinach aldolases could have been due to "cross-contamina-tion" of antibody preparations resulting, for instance, by theimmunization of rabbits with impure preparations of thespinach aldolases. This was done in "competition" experi-ments in which increasing amounts of the two pure aldolaseswere preincubated with a constant amount of their respectivehomologous antisera and then, these preincubated sampleswere tested for the presence of "free antibody" in our standardELISA system. Results obtained in one of these control ex-periments are presented in Figure 5. When antiserum directed

against the cytosolic aldolase was incubated with increasingamounts of pure cytosolic aldolase, the level of free antibodythat could subsequently bind to either aldolase adsorbed ontomicrotiter wells was proportionately decreased. If specificantibody to the chiloroplast aldolase was present in this anti-cytosolic antiserum, then preincubation ofthe antiserum withcytosolic aldolase would have had no effect on the ability ofthe "chloroplast-specific" antibody (if present) to subse-quently bind to the chloroplast aldolase on the microtiterplate, and no competition would have been observed. Thus,the cytosolic aldolase is capable of "neutralizing" antibodypopulations in the anticytosolic antiserum that recognize boththe cytosolic and chloroplast forms of aldolase in a similardose-dependent manner. Similar results were obtained incompetition experiments in which antiserum raised againstthe chloroplast enzyme was preincubated with increasingamounts of pure chloroplast aldolase (data not shown). Re-sults ofthese competition experiments show that the observedimmunological cross-reactivity is due to the presence of ahomologous antigenic determinant(s) on both the chloroplastand cytosolic aldolases.

Structural Similarities between the Cytosolic andChloroplast Forms of Spinach Leaf Aldolase

It is known that some proteins can cross-react immunolog-ically, even though they may be quite different in terms oftheir primary structures (11). Consequently, we wanted toobtain structural data to supplement results of our compara-tive immunological analysis. We employed micro-amino acidsequencing techniques to determine the NH2-terminal se-quences of the class I chloroplast and procaryotic M. smeg-matis aldolases (a total of 29 residues each). Amino terminalsequence analysis of the cytosolic spinach aldolase was notfeasible because, like other class I eucaryotic aldolases, itapparently has a "blocked" NH2 terminus (14). We thencompared the NH2-terminal sequences of the chloroplast andM. smegmatis aldolases with the full length sequences ofeucaryotic class I aldolases obtained by a computer assistedsearch of the National Biomedical Research Foundation pro-tein sequence databases; at present, there is no sequenceinformation available on any class I procaryotic fructosebisphosphate aldolase in these databases. Sequence align-ments of the various aldolases compared were performedusing the Pearson Fast-P program (15). In all cases, optimalalignment of the chloroplast and M. smegmatis aldolase se-quences occurred within the NH2-terminal regions of thealdolase sequences extracted from the database (Fig. 6). It isapparent that the NH2-terminal sequence of the chloroplastaldolase is much more similar to sequences near the NH2terminus of the eucaryotic cytosolic aldolases (52% identityto the maize cytosolic aldolase) than it is to the NH2-terminalsequence of the M. smegmatis aldolase (13% identity). Parti-cularily striking was the observation that the chloroplast al-dolase contains a 10 amino acid long sequence which is veryhighly conserved in the class I cytosolic eucaryotic aldolases(see boxed region, Fig. 6). Only two differences were observedin this region; namely a functionally conservative lysine/arginine difference at position 19 and an alanine/methioninesubstitution at position 24 of the chloroplast aldolase se-

1396 MARSH ET AL.




=2.0 7

0

0

CYT

0-11.0

0

0

YEAST/ M. smeg

100 200 300 400 500RELATIVE [Ab]

ANTI - CYTOSOLICCYT

2.0

° p CHL O212.0

0~~~~

0~~~~

/ ~~~~~~~~~~~YEAST/M.smeg

100 200 300 400 500RELATIVE [At

Figure 3. Typical ELISA immunotitration curves obtained when spinach chloroplast (CHL), spinach cytosolic (CYT), M. smegmatis and yeastaldolases were probed with antibody preparations against the chloroplast (top) or cytosolic (bottom) forms of spinach aldolase. The ordinaterefers to the rate (velocity) of the peroxidase reaction in terms of the change in absorbance at 490 nm per 20 min. During this incubation period,the change in absorbance with time was found to be linear in all cases. The antiserum dilution used for each point has been normalized suchthat a relative antibody concentration (abscissa) of 512 corresponds to a 1 :10 dilution of antiserum.

quence. Recent high resolution x-ray crystallographic analysisshows that the first beta sheet in the secondary structure ofthe prototypic class I rabbit muscle aldolase (24) is containedwithin this highly conserved region. Also, a 13 residue longalpha helical segment is known to precede the first beta sheetin the secondary structure of rabbit muscle aldolase (Fig. 6).Interestingly, the chloroplast aldolase sequence starts preciselywhere this alpha helical region begins in rabbit muscle aldo-lase. In contrast, no clear-cut relationship between the NH2-terminal region ofthe procaryotic M. smegmatis aldolase andthe corresponding segments of the chloroplast or cytosoliceucaryotic aldolases was observed (Fig. 6).To verify the observed similarities in NH2-terminal se-

quences of the chloroplast enzyme and other eucaryotic classI aldolases, we utilized the method described by Lipman andPearson ( 15) to assess the statistical significance of the align-ments by comparing the chloroplast sequence with randomlypermuted versions of the entire sequence of each potentiallyrelated aldolase. Results of this type of analysis are expressedas a "Z" value. The relationship between two sequencesyielding Z values greater than six is regarded as significant,and when Z is found to be greater than 15, the sequences areconsidered to be genuinely related, while unrelated sequenceswill generally have Z values less than three. As shown inFigure 6, the cytosolic maize and trypanosome aldolasesreceived the highest Z value ranking (both greater than 15)

1 397




Figure 4. Double reciprocal plots of the immu-notitration data presented in Figure 3. Linearregression analysis of the experimental datapoints was performed using the Enzpack com-puter software and an Apple 2E computer. The-oretical antibody saturation levels for each im-munotitration curve were obtained from the in-verse of the y-intercepts.

when compared to the chloroplast sequence. The enzyme

from Trypanosoma brucei is known to be posttranslationallysequestered in a glycolytic organelle (5) and provides an

interesting parallel to the chloroplast aldolase which is as-

sumed to be posttranslationally processed as it is transportedinto the chloroplast organelle (14). The muscle aldolases (typeA, Fig. 6) form a group, with corresponding Z values of aboutseven, followed by the liver aldolases (type B, Fig. 6) whichare of borderline significance (Z< 6). Finally, the low Z valueobtained for the M. smegmatis aldolase (Z = 0.7) suggeststhat this procaryotic enzyme is unrelated to the chloroplastaldolase, at least at the NH2-terminal region of the molecule.A final comparative study was performed to determine if

the dissimilarities observed between the chloroplast and M.

smegmatis class I aldolases could be generalized to other classI procaryotic aldolases. Since no amino acid sequence infor-mation on any class I procaryotic fructose bisphosphate al-dolase could be found by computer search, we attempted toestimate the possible similarities between the chloroplast andprocaryotic class I aldolases by making comparisons of theiramino acid compositions utilizing the method ofMarchalonisand Weltman (16). Although not as sensitive as comparativeamino acid sequence analysis, this method can be used tomake reasonable predictions concerning the relative similari-ties in the primary structures of different proteins. Marcha-lonis and Weltman found that comparisons ofthe amino acidcompositions of proteins which have similar primary struc-tures generated "similarity indices" (referred to as SAQ values)

ANTI - CHLOROPLAST

4 8

yYV

_V

N--

12

ANTI - CYTOSOLIC

1 02[b]xl

MARSH ET AL.1 398




Table I. Saturation Levels and Cross-Reactivities of Aldolases UsingAntisera Prepared against Spinach Cytosolic and ChloroplastAldolases

SaturationPercent

Serum a Cross ReactionLevel" at Saturationb

Antichloroplast aldolase serumChloroplast 2.15 ± 0.17 100Cytosolic 1.34 ± 0.03 62M. smegmatis <0.05 0Yeast <0-05 0

Anticytosolic aldolase serumCytosolic 2.52 ± 0.14 100Chloroplast 1.22 ± 0.15 48M. smegmatis <0.05 0Yeast <0.05 0

a Saturation levels are defined as the change in absorbance at 490nm in 20 min observed under conditions of antibody saturation. Thesemaximum levels of antibody binding were obtained from double-reciprocal plots of ELISA data (see Fig. 4). Saturation levels areaverages ±SD for four independent determinations. b Ratio ofsaturation levels (expressed as a percentage) of the potentially cross-reacting protein relative to the protein used as the immunizing agent.

of less than 50, with the theoretical minimum value beingzero for a comparison of identical proteins. Comparisons ofamino acid compositions of proteins which are structurallyunrelated generated SAQ values of greater than 100. Usingthis approach, very low SAQ values (less than 15) were

observed when the amino acid composition of the chloroplastaldolase was compared with those of both the cytosolic spin-ach aldolase and the prototypic class I rabbit muscle aldolase.In contrast, we observed relatively high SAQ values (70-100)when the amino acid composition of the chloroplast aldolasewas compared with those of the class I aldolases derived fromdiverse procaryotes including Staphylococcus aureus, Lacto-bacillus casei, Escherichia coli, and Micrococcus aerogenes.These latter, rather high SAQ values are similar to thoseobserved when the amino acid composition ofthe chloroplastaldolase was compared with those of a number of structurallyunrelated proteins, including two class II (metallo) aldolases(data not shown).

DISCUSSION

Results of the present studies suggest that the chloroplastform of spinach leaf fructose bisphosphate aldolase is struc-turally more related to the class I eucaryotic aldolases than itis to aldolases derived from procaryotic origin. This conten-tion is based on our demonstration that: (a) approximately50% of the populations of polyclonal antibodies in antiserumraised against the chloroplast aldolase recognized a homolo-gous antigenic determinant(s) on the cytosolic enzyme (andvice versa) while neither type of antiserum recognized theprocaryotic class I aldolase derived from M. smegmatis nor

the class II (metallo) yeast aldolase; (b) The NH2-terminalamino acid sequence of the spinach chloroplast aldolase wasfound to be statistically much more similar to the NH2-terminal sequences of animal and plant cytosolic aldolasesthan it was to the NH2-terminal sequence of the procaryotic

100

ANTI- CYTOSOLIC0

!\o50 \@\

* CYTo CHL

0~~~

50 100 150PRE-INC [CYT] (ug/mi)

Figure 5. Competitive inhibition of antibody binding to the two spin-ach aldolases by preincubation of antiserum raised against the cyto-solic aldolase with pure cytosolic aldolase. Aliquots of antiserum,diluted 1:100 in antibody binding buffer (1.2 mL each), were prein-cubated with increasing amounts of highly purified cytosolic aldolasefor 5 d at 40C with occasional mixing. Final concentrations of aldolasein the preincubation mixtures ranged from 0 to 150 Ag/mL. Thesepreincubated mixtures were then substituted for antiserum in thestandard ELISA procedure. Data are plotted as a percentage of thereactivity observed when no competing antigen (cytosolic aldolase)was present in the preincubation mixtures (ordinate) versus theconcentration of cytosolic aldolase present in the preincubation mix-ture (abscissa).

M. smegmatis class I enzyme, particularily in that regionwhich corresponds to the first beta sheet in the secondarystructure of rabbit muscle aldolase; (c) Results of systematicamino acid composition comparisons of diverse eucaryoticand procaryotic fructose bisphosphate aldolases indicated thatthe chloroplast enzyme is more closely related to the eucar-yotic class I aldolases than it is to class I aldolases of procar-yotic origin. Furthermore, the chloroplast aldolase, like othereucaryotic aldolases, possesses a carboxy-terminal structurewhich is essential for maximum catalytic activity (14) and iscomposed of four subunits (12), characteristics which are notexhibited by most procaryotic class I aldolases. Collectively,these observations support the assignment of the chloroplastaldolase to the eucaryotic "family" of class I fructose bisphos-phate aldolases.We have recently demonstrated that a subpopulation of

antibodies raised against the chloroplast aldolase also recog-nizes a homologous domain(s) in cytosolic aldolases derivedfrom animals as well as plants (our unpublished observation)and that these homologous antibody sites consist predomi-nantly of "linear" stretches of amino acids, rather than beingconstructed of three-dimensional "conformational" motifswhich are characteristic of only the native form(s) of theseproteins. That is, the immunological cross-reactivity observedbetween chloroplast and other eucaryotic aldolases is stillevident after the enzymes are subjected to harsh denaturingconditions or after their primary structures are fragmented by

1 399




zResidues ValueaSequence Alignment

SSYADELVKTAKTVAS

MSAYCGK*K***I*N*AYIGT

PGRGILAMDESk

**K** *A** *tGTIFKRL

MSKRVEVLLTQLPAYNRLKTP*EA**IE***KMTA **K*L**A***JGSCSKRF

TTYFNYPSKELQ***REI*QKIVA4 *K****A***PPTHGKRLMPHPYPALTPEQKK**ADI*HRIV4 *K* ***A** *GSIAKRL

MPHSHPALTPEQKK**SDI*HRIVA **K****A***

PYQYPALTPEQKK**SDI*HRIVA **K****A***

MAHRFPALTQEQKK**SEI*QSIVA *K****A***

MAHRFPALT*EQKK**SEI*QRIV *K****A***

VNQQQADKMT *Q*FI*AALC

rGSIAKRL

rGSIAKRL

VGTMGNRL

VGTMGNRL

ADGGELHL

2940 15.25443444443444429

15.67.97.37.67.95.15.40.7

az = (optimized similarity score mean of optimized scores from randomized sequences)/standard deviation of optimized scoresobtained from randomized sequences (15). In each case, 20 comparisons (ktup = 1) were made with shuffled sequences.

Figure 6. Amino acid sequence alignments and corresponding Z values for class I fructose bisphosphate aldolases. Amino acids are representedby their one letter code. The sequence of the cytosolic maize aldolase was taken from ref. 10 and the sequences of the Trypanosoma, Drosophila,rat A, rabbit A, human A, human B, and rat B aldolases were obtained from National Biomedical Research Foundation protein sequencedatabase. A "*" indicates the same amino acid appears in the corresponding position of the chloroplast aldolase. The highly conserved regioncommon to all eucaryotic, class I aldolase is enclosed in the large open box. Underlined residues in the rabbit A sequence correspond to the firsta-helix (left) and the first beta sheet (right) structures in the crystallographic structure of rabbit muscle aldolase (24).

enzymatic (trypsin) or chemical (cyanogen bromide) means.We should be able to use an immunological approach toidentify and characterize one or more domains present inclass I eucaryotic fructose bisphosphate aldolases which havebeen highly conserved during establishment of the large rep-ertoire of class I eucaryotic aldolases currently found innature.

Finally, results of the present work have important impli-cations concerning the probable evolutionary origin of thechloroplast aldolase molecule. Two distinct models have beenproposed to explain the origin of proteins which reside in thechloroplast and mitochondrial organelles of eucaryotic cells.The "autogenic" model (reviewed in ref. 18) speculates thatthese organellar proteins are encoded for by structural geneswhich were created in situ during evolution by duplicationand modification of genetic information already present inthe eucaryotic genome. In contrast, the popular "endosym-biotic" model (reviewed in ref. 26) proposes that these struc-tural genes were originally derived from the genomes ofprocaryotic endosymbionts which long ago took up residencewithin primative eucaryotic cells. The case for an endosym-biotic origin of the chloroplast organelle is considered to be a

strong one and in particular, results of comparative biochem-ical studies have revealed a potentially close relationshipbetween chloroplasts and the procaryotic blue-green algae(cyanobacteria) (26).A common method for investigating the probable evolu-

tionary origin of the chloroplast form of a particular chloro-plast/cytosolic enzyme pair is to perform "three-way" struc-tural comparisons between the two members of the plantenzyme pair and the corresponding enzyme derived from a

suitable procaryotic organism. The rationale is as follows: ifthe chloroplast form ofthe enzyme pair arose by an endosym-

biotic mechanism, it should be structurally more similar tothe procaryotic enzyme while, if it arose by an autogenicmechanism, the chloroplast form should be structurally moresimilar to its cytosolic counterpart. Structural and/or func-tional comparisons of this sort have supported an endosym-biotic origin for a number of nuclear-encoded chloroplastenzymes including glyceraldehyde-3-phosphate dehydrogen-ase (23), phosphoglucose isomerase, fructose 1,6-bisphospha-tase and malate dehydrogenase (reviewed in ref. 27). However,using the same approach, the structural similarities we havefound between the chloroplast and cytosolic forms of spinachleafaldolase are difficult to accomodate within the frameworkof the endosymbiotic model. In addition, it is important tonote that the chloroplast contains a class I (Schiffs base)aldolase in contrast to the mechanistically, structurally andpresumably evolutionarily distinct class II (metallo) aldolase( 19) found in all photosynthetic procaryotes studied thus far,including the blue-green algae (1 9, 29).Our findings, plus the observed distribution of class I and

class II aldolases in nature, seem to be more consistant withthe notion that the chloroplast form of fructose bisphosphatealdolase was derived from an autogenic evolutionary origin.According to this model, it is proposed that the nuclear-localized structural gene which codes for the chloroplast al-dolase arose by the duplication and subsequent modificationofan ancestral gene already present in the eucaryotic nucleus,an established mechanism for the creation of isoenzymes in a

number of other systems (8). Such a mechanism would notrequire the transfer of genetic information from the genomeof a procaryotic endosymboint to the nuclear genome of theeucaryotic "host," a necessary complication of the endosym-biotic alternative. Comparative studies on the chloroplast andcytosolic forms of plant triose phosphate isomerase (13) and

Protein

ChloroplastMaizeTrypanosomeDrosophilaRat ARabbit AHuman AHuman BRat BM. smegmatis

1 400 MARSH ET AL.




superoxide dismutase (reviewed in ref. 27) plus the mitochon-drial and cytoplasmic forms of phenylalanyl-tRNA synthase(7) suggest an autogenic origin for the organellar forms ofthese enzymes as well. These findings, in conjunction withthe results of the present work, raise the intriguing possibilitythat both endosymbiotic and autogenic mechanisms may

have contributed to the creation of nuclear-localized struc-tural genes which code for the organellar form of a particularprotein pair.

ACKNOWLEDGMENTS

We thank Ms. Marie Ayers for excellent secretarial assistance andSylvia Yuen for sequence analysis.

LITERATURE CITED

1. Anderson LE, Heinrikson RL (1979) Chloroplast and cyto-plasmic enzymes. VIII. Amino acid composition of the pealeaf aldolases. Plant Physiol 64: 404-405

2. Anderson LE, Heinrikson RL, Noyes C (1975) Chloroplast andcytoplasmic enzymes. Subunit structure of pea leaf aldolases.Arch Biochem Biophys 169: 262-268

3. Anderson LE, Levin DA (1970) Chloroplast aldolase is controlledby a nuclear gene. Plant Physiol 46: 819-820

4. Anderson LE, Pacold I (1972) Chloroplast and cytoplasmic en-zymes. IV. Pea leaf fructose 1,6-diphosphate aldolases. PlantPhysiol 49: 393-397

5. Clayton CE (1985) Structure and regulated expression of genesencoding fructose biphosphate aldolase in Trypanosoma bru-cei. EMBO J 4: 2997-3003

6. Fischer S, Luezak J, Schleifer KH (1982) Improved methods forthe detection of class I and class II fructose-1,6-bisphosphatealdolases in bacteria. FEMS Micro Lett 15: 103-108

7. Gabius H-J, Engelhardt R, Schroder FR, Cramer F (1983) Evo-lutionary aspects of accuracy of phenylalanyl-tRNA synthe-tase. Accuracy of fungal and animal mitochondrial enzymesand their relationship to their cytoplasmic counterparts and aprokaryotic enzyme. Biochemistry 22: 5306-5315

8. Gottlieb LD (1982) Conservation and duplication of isozymes inplants. Science 216: 373-380

9. Jayanthi Bai N, Ramachandra Pai M, Suryanarayana Murthy P,Venkitasubramanian TA (1975) Fructose diphosphate aldolasefrom Mycobacterium smegmatis. Arch Biochem Biophys 168:230-234

10. Kelly PM, Tolan DR (1986) The complete amino acid sequencefor the anaerobically induced aldolase form maize derived fromcDNA clones. Plant Physiol 82: 1076-1080

11. Kim S-H, de Vos A, Ogata C (1988) Crystal structures of twointensely sweet proteins. Trends Biochem Sci 13: 13-15

12. Kruger I, Schnarrenberger C (1983) Purification, subunit struc-ture and immunological comparison of fructose-bisphosphatealdolases from spinach and corn leaves. Eur J Biochem 136:101-106

13. Kurzok H-G, Feierabend J (1984) Comparison of a cytosolic anda chloroplast triosephosphate isomerase isoenzyme from ryeleaves. II. Molecular properties and phylogenetic relationships.Biochim Biophys Acta 788: 222-233

14. Lebherz HG, Leadbetter MM, Bradshaw RA (1984) Isolationand characterization of the cytosolic and chloroplast forms ofspinach leaf fructose diphosphate aldolase. J Biol Chem 259:1011-1017

15. Lipman DJ, Pearson WR (1985) Rapid and sensitive proteinsimilarity searches. Science 227: 1435-1441

16. Marchalonis JJ, Weltman JK (1971) Relatedness among pro-teins: A new method of estimation and its application toimmunoglobulins. Comp Biochem Physiol 38B: 609-625

17. Matsudaira P (1987) Sequence from picomole quantities ofproteins electroblotted onto polyvinylidene difluoride mem-branes. J Biol Chem 261: 10035-10038

18. Raff RA, Mahler HR (1972) The non symbiotic origin of mito-chondria. Science 177: 575-582

19. Rutter WJ (1964) Evolution of aldolase. Fed Proc 23: 1248-125720. Rutter WJ, Hunsley JR, Groves WE, Calder J, Rajkumar TV,

Woodfin BM (1966) Fructose diphosphate aldolase. I. Yeast.Methods Enzymol 9: 480-486

21. Schnarrenberger C (1987) Regulation and structure of isozymesof sugar phosphate metabolism in plants. Isozymes Cuff TopBiol Med Res 16: 223-240

22. Schnarrenberger C, Kruger I (1986) Distinction between cytosoland chloroplast fructose-bisphosphate aldolases from pea,wheat, and corn leaves. Plant Physiol 80: 301-304

23. Shih M, Lazar G, Goodman HM (1986) Evidence in favor of thesymbiotic origin of chloroplasts: Primary structure and evolu-tion of tobacco glyceraldehyde-3-phosphate dehydrogenases.Cell 47: 73-80

24. Sygusch J, Beaudry D, Allaire M (1987) Molecular architectureof rabbit skeletal muscle aldolase at 2.7-A resolution. Proc NatlAcad Sci USA 84: 7846-7850

25. Thorpe JP (1982) The molecular clock hypothesis; biochemicalevolution, genetic differentiation and systematics. Annu RevEcol Syst 13: 139-168

26. Weeden NF (1981) Genetic and biochemical implications of theendosymbiotic origin of the chloroplast. J Mol Evol 17: 133-139

27. Weeden NF (1983) Evolutionary affinities of plant phosphoglu-cose isomerase and fructose- 1 ,6-bisphosphatase isozymes. Iso-zymes Curr Top Biol Med Res 8: 53-66

28. Weeden NF, Gottlieb LD (1980) The genetics of chloroplastenzymes. J Hered 71: 392-396

29. Willard JM, Gibbs M (1968) Role of aldolase in photosynthesis.II. Demonstration of aldolase types in photosynthetic orga-nisms. Plant Physiol 43: 793-798

1401



structural similarities between spinach chloroplast and ... · in contrast, some cross-reactivity...

Documents