Partial amino acid sequence of erythrocyte carbonic anhydrase from tiger shark

Comp. Biochem. Physiol. Vol.95B, No. 1, pp. 205-213, 1990 Printed in Great Britain

0305-0491/90$3.00+ 0.00 © 1989PergamonPress plc

PARTIAL AMINO ACID SEQUENCE OF ERYTHROCYTE CARBONIC ANHYDRASE FROM TIGER SHARK NILS BERGENHEM and UNO CARLSSON* IFM/Department of Chemistry, Link6ping University, S-581 83 Link6ping, Sweden (Tel: 013 281714) (Received 6 June 1989) Abstract--1. A partial primary structure (197 residues)of carbonic anhydrase from tiger shark (Galeocerdo

cuvieri) erythrocytes has been determined. 2. The amino acid sequence of the enzyme is identical to those of human cytoplasmic carbonic anhydrases (CA I-III) by as much as 52450%. 3. It is shown that tiger shark CA most closely resembles the CA II isoenzyme of amniotes. 4. Isoelectric focusing and inhibition studies on carbonic anhydrase from dogfish (Squalus acanthias) blood and muscle indicate the presence of the same isoenzyme in shark blood and muscle.

INTRODUCTION The interconversion between CO 2 and HCO3 is catalyzed by the zinc containing enzyme carbonic anhydrase (carbonate hydro-lyase, EC 4.2.1.1) known to be present in almost all groups Of organisms (Lindskog et al., 1971). Several isoenzymes of carbonic anhydrase (CA) have been found. In mammals at least six different CA isoenzymes are known (CA I-VI) (Hewett-Emmett et al., 1984; Fernley, 1988). According to available primary structure information the three cytoplasmic isoenzymes (CA I, II and III) seem to be equally distant from each other from an evolutionary viewpoint, thus leading to the assumption that they are the result of two gene duplications occurring close together in time (HewettEmmett et al., 1984). CA V and CA VI on the other hand appear to have a lower degree of identity with the cytoplasmic isoenzymes (Hewett-Emmett et al., 1987; Fernley et al., 1988). No amino acid sequence has yet been reported for the membrane bound CA IV. Evolutionary networks have been constructed for CA I, II and III with the maximum parsimony method (Hewett-Emmett et al., 1984). A "root" to make an evolutionary tree of the network is, however, lacking since the so far sequenced CAs have been obtained from amniotes (birds, reptiles and mammals), which seem to possess all three cytoplasmic carbonic anhydrases (Tashian et al., 1983). To root the evolutionary tree, an organism must be found in which the cytoplasmic CA gene duplications have not occurred, i.e. one with only one cytoplasmic CA isoenzyme. Classification of the type of such an *Author to whom correspondence should be addressed. Abbreviations used--DABITC, 4-N,N-dimethylaminoazobenzene-4'-isothiocyanate; DABTH, 4-N,N-dimethylaminoazobenzene-4'-thiohydantoin; CA, carbonic anhydrase; HPLC, high-performance liquid chromatography; PITC, phenylisothiocyanate; SDS, sodium dodecyl sulfate; TFA, trifluoro acetic acid; TPCK, L-l(tosylamido)-2-phenylethylchloromethyl ketone.

isoenzyme by homology comparison will make it possible to decide which isoenzyme is the archetypal cytoplasmic one. An attempt to solve this problem has been made with a 20-residue peptide sequence of spinach CA, but it could neither be classified as a CA I, II nor III isoenzyme (Hewe.tt-Emmett et al., 1984). We have chosen to work with CA from a shark (tiger shark, Galeocerdo cuvieri), because the elasmobranchs are reasonably related to the amniotes and consequently the primary structures from these groups would probably be homologous. Moreover there is a possibility that the gene duplications leading to the various cytoplasmic CA isoenzymes took place after the divergence of the elasmobranchs. The tiger shark CA has previously been reported to have a higher mol. wt than normal (38,000-41,000 instead of 29,000) (Maynard and Coleman, 1971). The structural reason for this difference is due to intermolecularly disulfide linked glutathione and cysteine, and is not the result of a longer polypeptide chain (Bergenhem et al., 1986).

MATERIALS AND METHODS Sequencing chemicals Heptane, ethylacetate, TFA (Fluka, for sequence analysis). PITC (Fluka, for sequence analysis) was transferred to small glass tubes and stored under nitrogen at -20°C. Pyridine (Fluka, puriss, p.a.) was redistilled twice, n-Butylacetate (Merck, p.a.) was redistilled. DABITC (Fluka) was recrystallized from boiling acetone. Water was triply distilled. Enzyme purification The purification of the enzyme has been described earlier (Bergenhem et al., 1986). The prepared enzyme was homogeneous as judged from SDS-polyacrylamide gel electrophoresis (Bergenhem et al., 1986). Alkylation All CA used in this study was alkylated. A total of 420nmol was reacted with ~4C-iodoacetate as has been previously described (Bergenhem et al., 1986).

205

206

N/LS BERGENHEMand UNo Q~ARLSSON

Radioactivity measurements The radioactivity of all peptides and DABTH-derivatives were measured using a lntertechnique SL 4000 liquid squintillation counter.

Carboa3'peptidase digestion Digestion with carboxypeptidases was performed mainly as described by Ambler (1967). Before use the carboxypeptidase A stock suspension (2 ~tl) (Boehringer Mannheim) was washed with triply distilled water (200/zl) and centrifuged. The water was discarded and the pellet was dissolved in 50/1l of 10% (w/v) LiCI in triply distilled water. The amount of carboxypeptidase was determined spectrophotometrically at 25~C with hippuryl-t--phenylalanine and hippuryl-k-arginine as substrate for carboxypeptidase A and B, respectively. Enzyme activity units (U) is defined as: U =izmol product/rain at pH 7.5 and 7.65 for carboxypeptidase A and B, respectively. The tiger shark enzyme (4 nmol) was digested at room temperature with 0.12 or 0.03 U of carboxypeptidase A and/or 0.12 or 0.03 U carboxypeptidase B (Boehringer Mannheim) in 150raM sodium phosphate buffer, pH 7.5 containing 0.23% (w/v) SDS and 40 # M ~-aminobutyric acid as internal standard. Aliquots of 15 gl were withdrawn at different times from 2 min to 24 hr. The aliquots were diluted to 30 #1 with triply distilled water and submerged in a boiling water bath for 2 min to terminate the digestion. The aliquots were subsequently frozen and thawed immediately prior to analysis.

Tryptt~' digestion Digestion with trypsin was done essentially as described by Smyth (1967). TPCK-treated trypsin (Boehringer Mannheim) was dissolved in 1 mM HCI and stored at 4~C for 1 hr before use. A total of 12/~g trypsin was added in two portions to 70 nmol (2 rag) tiger shark CA in 250 #I of 0.2 M ammonium bicarbonate buffer, pH 9.0. The reaction was allowed to proceed for 24 hr at room temperature before it was halted by submerging the reaction vessel in a boiling water bath for 2 rain. The digest was lyophylized and stored at - 20'C.

Endoproteinase Glu-C digestion Endoproteinase GIu-C digestion was performed mainly as described by Houmard and Drapeau (1972). The tiger shark CA (65nmol, 1.9mg) was dissolved in 300/~1 of 100mM ammonium bicarbonate buffer, pH7.8 and 47/ag endoproteinase Glu-C (Staphylococcus aureus V8 protease, Boehringer Mannheim) was added. After 8 hr at 37°C the reaction vessel was heated in a boiling water bath (2 min) and lyophylized.

Endoproteinase Arg-C digestion Cleavage at the arginine residues was done essentially as described by Schenkein et al. (1977). A total of 45/1g endoproteinase Arg-C (from mouse submaxillary gland, Boehringer Mannheim) was added in two portions to 74nmol (2.ling) CA in 0.2M ammonium bicarbonate buffer, pH 8.0. The reaction was kept at 37'C for 24hr before it was stopped in a boiling water bath and lyophylized.

Preparative peptide mapping on thin layer plates Separation of peptides on cellulose thin layer plates was performed mainly as described by Wittmann-Liebold and Lehman (1980). The lyophylized peptides from the tryptic and GIu-C digests were extracted with 70 + 40 #1 of electrophoresis buffer (Pyridine, acetic acid, acetone, water, 50/I00/375/1975, v/v, pH 4.4). After centrifugation the clear solution was spotted, 10/~1 on each, onto wet (electrophoresis buffer) cellulose thin-layer plates (Polygram Cel-300 Macherey-Nagel). Electrophoresis was performed in a LKB Multiphore lI with a large cooling plate for 2.5 hr at 400 V. After drying the cellulose plate was chromatographed in

pyridine, n-butanol, acetic acid, water (40/60/12/48). For detection the dried plate was dipped in 5% (v/v) pyridine in acetone and developed by dipping in 0.004% (w/v) fluoram (Fluka) in acetone. Peptides visible at 366 nm were scraped off. Peptides from a total of 10 TLC plates were pooled and eluted with 2 x 1 ml of 50% acetic acid. The peptides were lyophylized and stored at --20~C.

Gel .filtration Peptides from the tryptic and Glu-C digestion that were not soluble in the electrophoresis buffer were dissolved in 300gl of 50% acetic acid and chromatographed on Sephadex superfine G-25 and G-50, respectively. The eluent was 50% acetic acid in both cases. The column dimensions were 0.7 x 147 cm and the flow rate was 0.8 ml/hr. Fractions of 0.8 ml were collected. The peptides were detected at 254 nm using a LKB 4700 Uvicord I and by measuring the radioactivity of the collected fractions.

Reversed phase chromatography The Arg-C digest was dissolved in 200,ul of 50% acetic acid and samples of 50 ~I per chromatography were injected on a ProRPC HR 5/10 column (C~/Cs, Pharmacia). The HPLC system consisted of a Perkin-Elmer 3B HPLC pump and a Perkin-Elmer LC-75 variable wavelength u.v.-detector. The mobile phase used was (A) 0.1% (v/v) TFA in distilled water and (B) 0.1% TFA, 90% CH 3CN, 9.9% (v/v) distilled water. Elution was accomplished with a linear gradient 0-100% B (60 rain) at a flow rate of 0.3 ml/min. Peptides were detected at 215 nm and fractions were collected by hand as the peaks eluted.

Amino avid analyses The sample was flushed with nitrogen and the hydrolysis was performed in 6 M HCI for 24 hr at 110~'C in evacuated, sealed tubes. The hydrolysate was dried in vacuo and dissolved in 50% (v/v) ethanol in water. Prior to analysis samples of 20 ~1 were precolumn derivatized with 50#1 of o-phthaldialdehyde reagent (Lindroth and Mopper, 1979). The derivatized sample (20 pl) was injected on a Nucleosil 5~tm, Cts , reversed phase column (250 x 4.6 ram) using a Perkin-Elmer 3B HPLC-pump and a Perkin-Elmer 1000 fluorescence spectrophotometer for detection. Flow rate was l ml/min. Mobile phase: (A) 50 naM sodium phosphate buffer, pH 7.[. (B) Methanol Linear solvent gradient: 30-75% B (35min).

Amino acid sequencing Primary structure determination was performed by lhe manual DABITC/P1TC method as described by Chang et al. (1978) except that the conversion to the DABTHderivatives was performed with 50% (v/v) TFA in water. The DABTH-derivatives were identified by two-dimensional TLC on 3 × 3 cm polyamide sheets (F 1700, Schleicher & Scbiill). Leucine/isoleucine was distinguished both by the purple by-product spot of isoleucine (Von Babr-Lindstr6m et al., 1982), and by reversed phase chromatography according to Wittmann-Liebold et al. (1986). Two peptides (TrG30 and Erl:19) were sequenced on automatic sequencers (Applied Biosystem 470A and 477A, respectively).

Isoelectric jocusing of carbonic anhydrase ,from dogfish (Squalus acanthias) The sharks were sedated by approximately two teaspoons of MS 222 (Sandos) in 101 of water. The belly was opened and blood was sucked from sinus venosus with a heparine treated syringe. After centrifugation of the blood for 5 min at 3000g the plasma was removed and the erythrocytes washed with 3 times the volume of ice-chilled 0.4 M potassium oxalate, 0.4M NaCI, and centrifuged (3000g, 5 rain). The erythrocytes were hemolysed with 5 times the volume of cold distilled water and frozen. Preparative

AA sequence of shark RBC carbonic anhydrase isoelectric focusing was carried out on 300/~1 of hemolysate in an LKB 8100 focusing column with LKB ampholine (pH 3.5-10) (Vesterberg, 1971). To prepare the muscle homogenate the blood vessels were removed from the muscle tissue to avoid contamination of blood CA. The tissue (100g) was cut in pieces and homogenized with a Waring blender in 200 ml of 20 mM sodium phosphate buffer, pH 7.5. The homogenate was centrifuged at 30,000g for 20min and the pellet was discarded: 2.5 ml of homogenate was used for isoelectric focusing as described above. Inhibition of dogfish carbonic anhydrase by p-aminomethylbenzenesulfonamide The inhibition of carbonic anhydrase from blood and muscle homogenate assayed by the colorimetric CO 2hydration activity method described by Rickli et al. (1964) was investigated by varying the concentrations ofp-aminomethylbenzenesulfonamide in the assay solutions. RESULTS

The strategy employed for determining primary structure of erythrocyte carbonic anhydrase from tiger shark has been cleavage of the protein into rather small fragments which are best suited for the manual sequencing procedure used. By cleaving the protein with trypsin and endoproteinase Glu-C it has been possible to generate some peptide overlaps yielding longer sequenced fragments (Fig. 1). The fragments have then been aligned to known CA sequences by homology. The homology of the peptides with known CA sequences has been found by the use of a small computer program that we have written for a home computer (Sinclair Spectrum). The program compares the peptide sequences with the amino acid sequences of human CA I and II and 1

10

207

bovine CA III. While sliding the peptide along the protein the least number of nucleotide differences between the peptide and protein sequence is calculated for each position of the peptide. This approach has made it possible to unambiguously find the position of nearly all sequenced peptides with totally 197 residues (Fig. 2). The peptides that have not been aligned can be seen in Table 1. CA I numbering has been used throughout this work. Tryptic and endoproteinase Glu-C digest Most of the peptides were soluble in the electrophoresis buffer and were therefore separated by cellulose thin-layer peptide mapping (Fig. 3A and B). The separated peptides are termed TrA-U and V8A-Q, respectively. The separation was good, but the sequencing yield was rather low compared to the amount of material determined by amino acid analyses of the peptides. This suggests that in developing the TLC-plates with fluorarn a high degree of blocking of amino terminals occurred. The peptides that were not soluble in the electrophoresis buffer were separated on Sephadex G-25 and G-50 (Figs 4 and 5). The very low u.v. absorption of the eluate in the chromatography of the endoproteinase Glu-C digest (Fig. 5) is due to the small amount of applied sample and the long detection wavelength (254nm), but measurements at shorter wavelengths are not possible in 50% acetic acid due to the absorption of the eluent. Pure peptides were recovered by combining the u.v. transmission curve and the radioactivity measurements and by screening the chromatograms by sequencing aliquots of various fractions. The peptides were termed TrG (fraction No.) and V8G (fraction No.), respectively (Fig. 1).

20

30

E G T R Q S P I

N Q E A K F

"VSM rtR 50

--

--TrS.--

60

70

80

D S N L K P L D L ? Y D P A

YVLT

V8G26 lrF 90

G G P L C G

1~

110

- - T r B - -

-

V

8

b

-

-

~1:19. 130

VH

120

R Q F H F H W G A S D T H G S E H I L D G K T Y A A E L H L

140

V8~ T~33

150

160

E K Y S D F A E A L K A P D G L A V ? G V M L ? V C D S N P A L ,

,,V8P

V8G35 - - T ~ 3 T - -

T~31"-170

V C S L E S L K A K G D V 8 P - -

180

T~

--

1~

2~

E F K D F S P C S L L P S C L Q Y W T Y L G S L T T --VSG33--

TrA---

T~30 210

PPL

KI

220

E S V I W L V I K E P I S L G P D Q L A K F R G

~0

240

R G L E F G P N E P

V8G32"TrE ~0

2~

TrM--'~I:14-16"--

M Q N N Y R P P Q P L K G R E I R K N F E - - V 8 1 -

V8E --Tr~ "--TrL----TrN--

Fig. 1. Sequenced peptides derived by cleavage with trypsin, endoproteinase Glu-C and endeproteinase Arg-C. The tryptic peptides are termed Tr, the endoproteinase Glu-C peptides V8 and the endoproteinase Arg-C peptides Er. The extensions to the peptide code names are explained in the result section. CA I numbering is used.

208

N1LS BERGENHEM and UNO CARLSSON 1

,o

sharkHCAIIcA S H H W G

10

20

Y G K H NGP

E HWH

50

60

90 0 ,0

0 Zn

0 Zn

G~G E

S

LK

170

o o

~

D G LAV

*

o 70

* *

110

I IF

ED~TK N ~ s ~ R c

250

I ID NT QH T ~

S A CD

L H

Q

I o * 0 200

*

S c ~ DL~ P GT

T LQG S L T

230 L~--~K

A

240

L N F N G~]G

G

V

S

190

sG~Es

0 Zn 120

* GLKJTIY A A E

?

F 80

K~G

?IGIVM

180

A DS

40

Y

o

TH

o o o * 210 o 220 L~C~T~I~L K E P I S V S S E~V IEIS V~IIWILMI K E P I S L G P D •

o o

V Q Q p D G L AV L A

~Dc sV~ ED~ I L ~ T ~ K

o

100

RQFHFHW W N T

* 30

KD F P I AEK~E~VD

R G L

E[~ E['~L

G

N

260

Fig. 2. Comparison of the amino acid sequence of tiger shark CA and human CA II. Identical residues in both sequences are boxed. The three histidine residues involved in binding the zinc are labeled Zn. Residues hydrogen bonded to these histidine residues or to the zinc-bound solvent molecules are labeled (*) (Lindskog et aL, 1984; Notstrand et al., 1975). Exposed residues in the active-site cavity are also indicated (O) (Notstrand et al., 1975), Asn-62 is exposed to the active site according to the refined structure of Eriksson (1988). CA I numbering is used.

One of these peptides (TrG30) was sequenced on an Applied Biosystem 470A Protein Sequencer. Endoproteinase Glu-C has been found to be specific for glutamic acid residues under the conditions employed (Houmard and Drapeau, 1972). The amino acid residues at positions 126, 175 and 205, have been assigned as glutamic acid from the cleavage by the endoproteinase Glu-C protease. Position 126 might be a deletion as in other CA IIs and in that case the glutamic acid is residue number 125 and not 126. Sequencing of the tryptic peptide mixture that was applied on the gel filtration column showed the presence of probably four peptides. After the gel filtration only three of these could be found. The missing peptide had the tentative sequence: Gln-(Phe, Pro or Tyr)-His-Phe-His. Our assumption is that the glutamine residue in the amino terminal has reacted by cyclization to a pyroglutamate residue, which is known to occur in acidic solution (Allen, 1981). Table 1. Tiger shark CA peptides that have not been possible to align with other CSs by homology search. Peptides are termed as descrihed in the result section Peptide Trl* TrJ TrO VgC VgH

Sequence F R and L R YR CVCN LGFSLA [TR

*Trl was a mixture of two peptides and hence two sequences are given.

Endoproteinase A r g - C digest

The sequence of the tryptic peptide given above is very homologous to positions 92-96 in other carbonic anhydrases. This peptide contains two of the metal ligands His-94 and His-96 in the active site, which makes this peptide particularly interesting. Since there is an arginine residue in position 91 in the CA IIIs we presumed that the same might be the case in the shark enzyme and therefore we used endoproteinase Arg-C to obtain larger, more easily separated peptides. Reversed phase chromatography was used for separation of the peptides to avoid pyroglutamate formation. In this chromatography the peptides were exposed to 50% acetic acid for only a few minutes during the injection. Only a small number of peptides were present in this digest (Fig. 6). When fractions from the chromatography were sequenced manually only two sequences were found in low yields. The sequence Gly-Leu-Glu-Phe-GlyPro-Asn corresponding to positions 232-238 was found in fractions 14-16, and fraction 19 contained a peptide with the sequence Gln-Phe-His-Phe-His, corresponding to the peptide at positions 92--96. The latter peptide was sequenced on an Applied Biosystem 477A Protein Sequencer (Fig. 1). Since tiger shark CA contains nine arginine residues the presence of only two sequences indicate incomplete cleavages at the arginine residues by endoproteinase Arg-C, which has been noted earlier (Schenkein et al., 1977). Positions 91 and 231 have been assigned as arginine residues based on the reported cleavage specificity of endoproteinase Arg-C for arginine residues (Schenkein et al., 1977)~

AA sequence of shark RBC carbonic anhydrase

A

(~O

0

~,~

209

M ~J

~T

ELECTROPHORESIS

Fig. 3. Preparative peptide mapping of tryptic (A) and endoproteinase GIu-C peptides (B). The peptides were dissolved in electrophoresis buffer (pyridine, acetic acid, acetone, water, 50/100/375/1975) and spotted on wet cellulose thin-layer plates (Polygram Cel-300). Electrophoresis was performed for 2.5 hr at 400 V. After drying the cellulose plates were chromatographed in pyridine, n-butanol, acetic acid, water (40/60/12/48). The plates were dried and dipped in 5% (v/v) pyridine in acetone and developed by dipping in 0.004% (w/v) ftuoram in acetone. Peptides were made visible by illumination at 366 nm. The numbers indicate the relative fluorescence of the spots (5 = strongest intensity) and the letters give the code name extensions of the peptides, spot A in (A) for instance is termed TrA for tryptic peptide A.

E c 80 ,¢1" u3

lOOO

z

0

1-

_~ 9o

5 0 0 f..

g

Z

I..-

e • o • e

o_

e--

IO0

5o

i

45

55

5O

FRACTION NO.

Fig. 4. Separation of tryptic peptides on a Sephadex G-25 column (0.7 x 147 cm) eluted with 50% (v/v) acetic acid/water. Flow rate 0.8 ml/hr. Fraction volume 0.8 ml. The fraction numbers indicate the midpoint of the fractions. The radioactivity was determined on 10 #1 aliquots of each fraction. Transmission at 254 nm ( ), radioactivity ( - - O - - ) .

.¢ 8 0

300

/ ~

oz

I.-

z

N

p-

lOO

~o

~o

~o

~o

do

FRACTION NO.

Fig. 5. Separation of endoproteinase Glu-C peptides on Sephadex G-50. The column and conditions used and the radioactivity measurements were as described in Fig. 4. Transmission at 254nm ( ), radioactivity ( O - - ) .

210

NILS BERGENHEMand UNO CARLSSON

100

E u~ 2.56

j ~

. . . . .

J J 7

w 0

z

//

/

/

i/-

////

1.28 0

i

"~/// /

S

///

~

~ 0

41s1~ 20

40 R E T E N T I O N TIME (rain)

~o

do

Fig. 6. Fraclionation of the peptides obtained by endoproteinase Arg-C digestion of tiger shark CA on a Pharmacia ProRPC H R 5/10 column. Mobile phase (A) 0.1% TFA in distilled water and (B) 0.1% TFA, 90% CH3CN, 9.9% distilled water. The peptides were eluted with a linear gradient of 0-100% B over 90min at a flow rate of 0.3ml/min. Collected fractions 14-16 and 19 are indicated by bars. The radioactivity was measured on 10pl per fraction. Absorbance at 215 nm ( - - - ) , % acetonitrile ( .... ).

Carboxypeptidase digestion

Amino acid analysis

Deduction of the carboxyl-terminal from aminoterminal sequencing might lead to errors and therefore carboxypeptidase digestions of the shark enzyme were performed. As is shown in Fig. 7 the levels of free Glu, Phe, Asn, Lys, Arg and Ile increase during the course of digestion with carboxypepidase A + B. The sequence cannot be established from these results, but since there is an endoproteinase Glu-C peptide (V8E, Fig. 1) with the sequence Ile-ArgLys-Asn-Phe and since the levels of free glutamic acid rose most rapidly during carboxypeptidase digestion the car boxyl-terminal sequence can be obtained. The presence of glutamic acid at the carboxyl-terminus was further confirmed by digestion with lower amounts of carboxypeptidase A + B and by digestion with carboxypeptidase A and B independently.

The amino acid compositions of most peptides were analyzed and were found to be in agreement with the determined sequences. Residue 159 was assigned as lysine from the amino acid analysis of TrQ, since the only basic residue in TrQ was lysine.

a Lg

n., ~X-. . . . . . . . . . . .

/

"_~22"_

Dogfish carbonic anhydrase Carbonic anhydrase activity was found in both the erythrocyte hemolysate and the muscle homogenate. The enzyme from both sources had, within experimental error, the same pI and sensitivity to p-aminomethylbenzenesulfonamide (Table 2). DISCUSSION

This paper presents for the first time sufficient primary structure data of a non-amniotic carbonic anhydrase to allow comparisons with known CA sequences. The sharks have been found to have only one type of erythrocyte CA (Maynard and Coleman, 1971; Bergenhem et al., 1986), and the aim of this project has been to assign the erythrocyte carbonic anhydrase from tiger shark as a CA I, II or Ill isoenzyme. These isoenzymes are assumed to have evolved by two gene duplications occurring close together in time. Moreover, the divergence time

tU rr 0.1

/

I 2

I 4

I 6

Table 2. lsoelectric point and sulfonamide inhibition of carbonic anhydrase from dogfish blood and muscle

TIME (hi

Fig. 7. Time course of carboxypeptidase A and B digestion of tiger shark CA. The enzyme (4 nmol) was digested at room temperature with 0.12U carboxypeptidase A and 0.12 U of B in t30 it l of 150 mM sodium phosphate buffer, pH 7.5 containing 0.23% (w/v) SDS and 40pM ~-aminobutyric acid. Aliquots of 15/~1 were withdrawn at 0.25, 0.75, 1.25, 2, 4 and 8 hr. The released residues are normalized to the amount of shark CA: Glu ( --), Phe (.. 9, Asn (- - -), Lys (--.--), Arg (--.. ) and Ile ( ).

pI I50 (uM)

Blood CA

Muscle CA

5.0 1.7

5.1 1,9

Blood hemolysate and muscle homogenate were prepared as described in Materials and Methods. The pl was determined in an LKB 8100 focusing column with LKB amoholine pH 3.5 10, The ls0 value was calculated by plotting the CO2-hydration activity versus the concentration of p-aminomethylbenzenesulfonamide in the assay.

AA sequence of shark RBC carbonic anhydrase between mouse CA I and II has been estimated to be 300-320 million years (Fraser and Curtis, 1986), i.e. approximately the time of the radiation of the amniotes. This information together suggests that the sharks might possess only one cytoplasmic CA, probably resembling the archetypal cytoplasmic enzyme. In Fig. 2 the tiger shark CA sequence is aligned with the human CA II sequence. Of the sequenced active site residues all but one are the same as in other CA IIs. We have found that the three metal-ligands, His-94, 96 and 119, and all the residues that are either hydrogen bonded to the histidine ligands or the zinc-bound solvent molecules are the same as in other carbonic anhydrases. It must, however, be noted that residue 91 is an arginine. This arginine residue has been proposed, at least partly, to be responsible for the low esterase activity and weak sulfonamide inhibition in CA III (Chegwidden et al., 1984). Since the shark CA has similar esterase activity as mammalian CA II (Maynard and Coleman, 1971; Bergenhem et al., 1986) and is inhibited almost as strongly by sulfonamides as mammalian CA II (Maynard and Coleman, 1971), it seems as the replacement of a hydrophobic residue (lie, Val in CA II) by Arg in position 91 does not cause the lowering in esterase activity and sulfonamide binding. From X-ray studies on bovine CA III it has also been found that Phe-198, together with Ile-207, significantly diminish the space for inhibitor and substrate binding (Eriksson, 1988). In tiger shark CA these residues are Leu-198 and Val-207, as in all other sequenced carbonic anhydrases of type I and II. Previously, we have found that tiger shark CA contains eight cysteine residues, all of which are disulfide linked to glutathione or fi'ee cysteine (Bergenhem et al., 1986). The positions of five of these have been determined (Fig. 2, residue Nos 85, 151, 162, 182 and 188). Interestingly, all of the corresponding amino acid residues in the X-ray structure of human CA II are exposed to the surface (Vaara, 1974). This might explain why these cysteine residues are involved in intermolecular disulfide formation, rather than in intramolecular disulfide bridges. Since the sequenced tiger shark CA is a cytoplasmic enzyme, it was of interest to investigate if it is more closely related to any one of CA I, II or III. However, a comparison of the amino acid sequence of tiger shark CA with the sequence of human CA I, II and III (Table 3) shows that the differences are too small to allow a decisive isoenzyme type classification of shark CA. There are a number of unique and invariant amino acid residues in the three cytoplasmic CAs that can be used to classify a newly sequenced enzyme as a type I, II or III (Hewett-Emmett et al., 1984). Unfortunately, the number of unique and invariant residues for CA II is very low (two in the determined shark CA sequence) if all so far sequenced Table 3. Amino acid sequence identities between tiger shark CA and human CAI, II and IIl Tiger shark CA

HCA I

H C A II

HCA 111

53%

60%

52%

The sequence of human CA I, |I and IIl are obtained from Andersson et al. (1972), Hendersson et al. (1976) and Lloyd et al. (1986); Wade (1986), respectively.

211

CAs are included. For that reason we introduce the concept of unique and "consensus" residues. A "consensus" residue is defined as a residue that is found at a specific position in at least all but one of the sequenced CAs of the respective type. For the residue to be unique it must only be found in one of CA I, II or III. If these amino acid residues are compared with the corresponding amino acid residues in shark CA (Table 4) the highest proportion of matches, three out of five, is obtained with CA II. For CA I and III one out of nine and two out of 25 matches are obtained, respectively. His-200 is one unique and invariant residue with a functional role in CA I (Lindskog et al., 1984). In tiger shark CA there is a Thr-residue in this position, which makes it unlikely that the shark enzyme is a type I isoenzyme. In CA III Phe-198 and Ile-207 are two unique and invariant residues, which are also functionally important (Eriksson, 1988). The corresponding residues in shark CA are Leu-198 and Val-207. These facts, together with the sulfonamide sensitivity of tiger shark CA (Maynard and Coleman, 1971; Bergenhem et al., 1986; Sanyal et al., 1982), indicate that shark CA is not a type III isoenzyme. In CA II residue 64 (His) has a mechanistic role (Lindskog et al., 1984), and residues 67 (Asn) and 69 (Glu) are unique and invariant. Unfortunately, these residues have not been determined in tiger shark CA. Nevertheless, the above discussion leads to the conclusion that the single CA present in shark erythrocytes is most closely related to the CA II isoenzyme of amniotes. To investigate if sharks have a CA III isoenzyme in the muscle tissue we have made a preliminary study on blood hemolysate and muscle homogenate from dogfish. The carbonic anhydrase isolated from both sources had practically the same pI and equivalent sensitivity to sulfonamide inhibition. The high sensitivity towards sulfonamides indicates that the enzyme found in shark muscle is not a type III isoenzyme which is normally insensitive to sulfonamide inhibition (Sanyal et al., 1982). Carbonic anhydrase of slightly higher specific activity compared to the erythrocyte enzyme has been found in dogfish eyes (ciliary folds) and rectal gland. A higher K~ for anionic inhibitors was also noticed for the rectal gland enzyme compared to the enzyme from eye and blood (Maren and Friedland, 1978). Since we have noticed that the activity of the dogfish CA is markedly influenced by reduction of disulfides the difference in activity might be the result of different reduction states of the enzyme. The higher KI of the rectal gland CA might implicate a secreted CA that is genetically different from the one in blood and muscle. All studied amniote species so far possess a typical muscle CA (CA III) with characteristic properties. For sharks, however, the muscle CA seems to be identical, at least regarding pI and inhibition pattern, to its erythrocyte enzyme. The absence of a typical CA III in muscle as well as the presence of only one erythrocyte CA isoenzyme (CA II) might be due to the presence of only one cytoplasmic isoenzyme in the sharks. It must however be emphasized that it is impossible to totally exclude the presence of another cytoplasmic isoenzyme in any other tissue of the shark.

212

NILS BERGENHEMand UNO CARLSSON If the sharks contain only cytoplasmic C A of type II then C A II is the isoenzyme that most strongly resembles the archetypal cytoplasmic carbonic anhydrase. In this context it is interesting to note t h a t a n ancient cytoplasmic C A with an active site resembling t h a t of CA II has been proposed from the n u m b e r of fixed m u t a t i o n s in the active site ( H e w e t t - E m m e t t et al., 1984).

IN

cq

eq

¢-q

g

~, "~

~