Biochemistry 1989, 28, 7401-7408
7401
Site-Directed Mutagenesis of Histidine- 13 and Histidine- 1 14 of Human Angiogenin. Alanine Derivatives Inhibit Angiogenin-Induced Angiogenesis? Robert Shapiro and Bert L. Vallee* Center for Biochemical and Biophysical Sciences and Medicine. Harvard Medical School, Boston, Massachusetts 021 15 Received April 18, 1989; Revised Manuscript Received June 1 , 1989
ABSTRACT: The roles of His-13 and His-114 in the ribonucleolytic and angiogenic activities of human
angiogenin have been investigated by site-directed mutagenesis. Replacement of either residue by alanine (H13A and H114A) decreases enzymatic activity toward tRNA by at least 10000-fold and virtually abolishes angiogenic activity in the chick embryo chorioallantoic membrane assay. Both the H13A and H114A mutant proteins compete effectively with angiogenin in the latter assay; only a 5-fold molar excess of H13A over unmodified protein is required for complete inhibition. The His Ala substitutions, however, do not have any significant effect on the interaction of angiogenin with human placental ribonuclease inhibitor, an extremely potent inhibitor of angiogenin (Ki 7X M) previously shown to interact with another active-site residue, Lys-40. The effects of more conservative replacements-glutamine at position 13 and asparagine at position 114-were also examined. While the enzymatic activity of the H114N mutant was at least 3300-fold less than for the unmodified protein, the H13Q derivative had only 300-fold reduced activity toward tRNA and cytidylyl(3’+’)adenosine. Both substitutions substantially decreased angiogenic activity. The parallel effects on ribonucleolytic and biological activities observed with all four mutant proteins provide strong evidence that the latter activity of angiogenin is dependent on a functional enzymatic active site. The capacity of the H13A and H114A derivatives to compete with angiogenin in the chorioallantoic membrane assay suggests several additional features of the biological mode of action of this protein.
-
N
H u m a n angiogenin, a 14.1-kDa monomeric protein, induces neovascularization in the chick embryo chorioallantoic membrane (CAM)’ and rabbit cornea (Fett et al., 1985), cleaves RNA (Shapiro et al., 1986b), and exerts a number of effects on vascular endothelial and smooth muscle cells in vitro, which include activating phospholipases C and A2 (Bicknell & Vallee, 1988, 1989) and increasing cholesterol esterification (Moore & Riordan, 1989). Angiogenin displays extensive sequence homology to the pancreatic RNases (Strydom et al., 1985; Kurachi et al., 1985). Indeed, most of the active-site components of bovine pancreatic RNase A, including the critical catalytic residues His- 12, Lys-4 1, and His-1 19, are conserved in angiogenin. Nonetheless, the enzymatic activity of angiogenin differs markedly in both magnitude and specificity from that of RNase A (Shapiro et al., 1986b St. Clair et al., 1987, 1988; Rybak & Vallee, 1988). Numerous pieces of evidence suggest that the ribonucleolytic and angiogenic activities of angiogenin are related. Both activities are abolished by the human placental RNase inhibitor (PRI) (Shapiro & Vallee, 1987), which binds to angiogenin with a dissociation constant of 7 X M (Lee et al., 1989b). Mutation of Lys-40 to glutamine also virtually eliminates both activities (Shapiro et al., 1989). Conversely, a mutation that increases enzymatic activity 15-fold, Asp-1 16 His, also increases angiogenicity by 1-2 orders of magnitude (Harper & Vallee, 1988). Further, chemical modification of histidines in angiogenin by bromoacetate at pH 5.5 decreases both enzymatic and angiogenic activity (Shapiro et al., 1986b, 1987). The sites of carboxymethylation have been identified as His- 13 and
-
‘This work was supported by funds from Hoechst, A.G., under an agreement with Harvard University. *Address correspondence to this author at the Center for Biochemical and Biophysical Sciences and Medicine, Harvard Medical School, 250 L o n g w d Ave., Boston, MA 021 15.
His-1 14 (Shapiro et al., 1988b), whose corresponding residues, His- 12 and His- 1 19, are modified under these conditions in classical studies on RNase A (Stein & Barnard, 1959; Crestfield et al., 1963). However, unlike with RNase A, it has not been possible to obtain and characterize derivatives of angiogenin carboxymethylated at only a single histidine. For this reason, and because of the inherent limitations and interpretive ambiguities of chemical modification approaches, we have now employed oligonucleotide-directed mutagenesis to investigate the role of these two histidines in angiogenin. Replacement of either histidine by alanine essentially abolishes ribonucleolytic and biological activities. Strikingly, both the H13A2 and H114A mutants effectively inhibit the angiogenesis induced by angiogenin in the chick embryo CAM assay. The effects of more conservative replacements-i.e., glutamine at position 13 and asparagine at position 114-were also examined.
EXPERIMENTAL PROCEDURES Angiogenin) was obtained from a recombinant expression system in Escherichia coli as described (Shapiro et al., 1988a) and was quantitated by amino acid analysis. Bovine pancreatic Abbreviations: CAM, chorioallantoic membrane; RNase, ribonuclease; RNase A, bovine pancreatic ribonuclease A; PRI, human placental ribonuclease inhibitor; C18, octadecylsilane; HPLC, high-performance liquid chromatography; TFA, trifluoroacetic acid; Hepes, 4(2-hydroxyethyl)-l-piperazineethanesulfonic acid; Mes, 2-(Nmorpho1ino)ethanesulfonic acid; CpA, cytidylyl( 3’-4’)adenosine; SDSPAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; CpG, cytidylyl(3’4S’)guanosine; C>p, cytidine cyclic 2’,3’-phosphate. * We designate mutant proteins by the single-letter code for the original amino acid followed by its position in the sequence and the single-letter code for the new amino acid. Angiogenin expressed in E . coli differs from natural angiogenin only with respect to its N-terminus: Met-(-1)-Gln-1 versus 98% pure as judged by SDS-PAGE (not shown). Yields were 0.5 mg/L (H13Q) to 2 mg/L (H13A). Structural Characterization. Amino acid compositions of the four mutant angiogenins (Table I) support the proposed structures and indicate that all proteins contain a Met-(-1) residue. Tryptic peptide maps (H13A and H114A shown in Figure 1) indicate that in all cases the three disulfide bonds
Biochemistry, Vol, 28, No. 18, 1989 7403
Role of Histidines in Angiogenin 3*9 /I
10
I
1
v -
50
75
IO0
125
ELUTION T I M E ( m i n )
FIGURE 1: Chromatography of tryptic digests of 3.5 nmol of H13Aangiogenin (panel A) and 3.7 nmol of H114A-angiogenin (panel B) on a C18 H P L C column.
Table 11: Amino Acid Compositions of Tryptic Peptides from Angiogenin Mutants" H13A H13Q H114A H114N amino acid T7b T7 T11 TI 1 Asx 0.93 (1) 1.09 (1) 4.00 (4) 4.91 (5) Glx 1.86 (2) 2.79 (3) 2.99 (3) 2.87 (3) Ser 1.03 (1) 0.95 (1) 1.56 (1) GlY 1.11 (1) 1.32 (1) 1.17 (1) His 1,05 (1) 0.99 ( I ) 0.10 AW 0.89 (1) 1.09 (1) 1.03 (1) 1.06 ( 1 ) Thr 1.93 (2) 1.87 (2) 2.89 (3) 1.93 (2) Ala 1.95 (2) 0.98 (1) Pro 1.07 (1) 1.07 (1) 1.19 (1) 1.20 (1) TYr 1.80 (2) 1.81 (2) Val 2.95 (4) 2.96 (4) Met Ile 1.70 (2) 1.79 (2) Leu 0.98 (1) 1.04 (1) 2.03 (2) 2.16 (2) Phe 0.96 (1) 0.92 (1) 0.82 (1) 0.91 (1) 0.82 (1) 0.94 ( 1 ) LY5 1.03 (1) 0.91 (1) pmol analyzed 175 224 90 240 'Cystine and tryptophan contents were not determined. Numbers in parentheses represent compositions of the mutant proteins expected for peptide T7 (amino acids 6-21) or peptide T11 (disulfide-linked amino 102-1 2 1). Peptide T7 of H 13A-angiogenin coelutes acids 55-60 with peptides T8 and T9 during C18 HPLC (Figure 1). The amino acid composition of peptide T7 shown has been corrected for the presence of 0.22 molar equiv of T8 and 0.15 equiv of T9.
+
(in peptides T9, T10, and T11) formed correctly [see Strydom et al. (1985) for a description of angiogenin tryptic peptides]. For the H13A- and H13Q-angiogeniw amino acid compositions of peptide T7 (Table 11), containing residue 13, indicate the replacement of His by Ala or Glx, respectively, and no other changes. Peptide T7 of the H13A mutant elutes later than the natural peptide T7 during C18 HPLC, as expected for replacement of His by Ala. For the H114A- and H114N-angiogenins, amino acid compositions of peptide T11 (Table II), containing residue 114, show the replacement of His by Ala or Asx, respectively, and no other changes. The H 1 14A peptide elutes somewhat later than the corresponding peptide from angiogenin, again consistent with the His Ala replacement. Amino acid compositions of the remaining peptides are in good agreement with the proposed structures and account for the rest of the four proteins, except for Arg-32 and Arg-33 of H114A. Ribonucleolytic Activity of Angiogenin Derivatives. No enzymatic activity toward tRNA could be detected with the
-
H13A- and Hl14A-angiogenins, allowing an upper limit to be set at 0.01% activity compared with unmodified angiogenin. Initial preparations of the H13Q and H114N derivatives, however, appeared to be 0.1-1% active. In order to examine whether this activity might be due to trace amounts of a contaminating ribonuclease, both mutant proteins were rechromatographed on a C18 HPLC column employing a shallow gradient (see Experimental Procedures). Fractions collected from across each peak were lyophilized, reconstituted in water, quantitated by amino acid analysis, and tested for enzymatic activity. With the H13Q derivative, the three major protein-containing fractions had similar specific activities, 0.3% that of native angiogenin. Thus, this activity appears to be associated with the mutant protein. With the H114N derivative, the specific activity varied considerably through the peak. The earliest fractions had no measurable activity, under conditions where an activity 0.03% of that of angiogenin could have been detected. This value thus represents an upper limit on the true activity. Activity of the rechromatographed H13Q mutant toward CpA was also measured. A k,,/K, value of 0.045 M-' s-I was obtained, 0.4% of that determined for angiogenin. Owing to limitations of the assay, it has not been possible to determine whether this activity decrease reflects changes in kat, K,, or both. Rate of Nonenzymatic Cleavage of tRNA and CpA. In order to facilitate mechanistic evaluation of ribonucleolytic activities measured with mutant angiogenins, the nonenzymatic rates of substrate cleavage were examined. At pH 6.8, 37 OC, the observed reaction velocity for formation of perchloric acid soluble fragments from tRNA was 3.3 X lo4 A260units/min at a substrate concentration of 2 mg/mL. The initial reaction velocity with 0.24 pM angiogenin and 2 mg/mL tRNA was 1.2 X A260units/min. In calculation of the difference between the rate constant for the nonenzymatic reaction ( = u o / [ S ] ) and that for the angiogenin-catalyzed reaction (= Vm/ [E]) from these data, the following assumptions were made: (a) all phosphodiester bonds in the substrate are equally susceptible to nonenzymatic cleavage; (b) no more than 10% of these bonds are attacked by angiogenin in the initial stages of the reaction [see Rybak and Vallee (1988)l; and (c) the catalyzed reaction rate is approximately half of V,, (Lee & Vallee, 1989b). On this basis, the calculated rate enhancement is at least IO5. The rate enhancement by angiogenin was also examined in a less complex system, employing the dinucleotide CpA as substrate. After 11 days at 37 OC, pH 5.8, in the absence of angiogenin,
