diff --git a/modules/angular2/src/core/change_detection/parser/lexer.ts b/modules/angular2/src/core/change_detection/parser/lexer.ts
index aa9cf852b2..460a43c8b6 100644
--- a/modules/angular2/src/core/change_detection/parser/lexer.ts
+++ b/modules/angular2/src/core/change_detection/parser/lexer.ts
@@ -395,6 +395,18 @@ function isIdentifierStart(code: number): boolean {
   return ($a <= code && code <= $z) || ($A <= code && code <= $Z) || (code == $_) || (code == $$);
 }
 
+export function isIdentifier(input: string): boolean {
+  if (input.length == 0) return false;
+  var scanner = new _Scanner(input);
+  if (!isIdentifierStart(scanner.peek)) return false;
+  scanner.advance();
+  while (scanner.peek !== $EOF) {
+    if (!isIdentifierPart(scanner.peek)) return false;
+    scanner.advance();
+  }
+  return true;
+}
+
 function isIdentifierPart(code: number): boolean {
   return ($a <= code && code <= $z) || ($A <= code && code <= $Z) || ($0 <= code && code <= $9) ||
          (code == $_) || (code == $$);
diff --git a/modules/angular2/src/core/change_detection/parser/parser.ts b/modules/angular2/src/core/change_detection/parser/parser.ts
index 80fe041176..dbfe8f7f30 100644
--- a/modules/angular2/src/core/change_detection/parser/parser.ts
+++ b/modules/angular2/src/core/change_detection/parser/parser.ts
@@ -5,6 +5,7 @@ import {ListWrapper} from 'angular2/src/facade/collection';
 import {
   Lexer,
   EOF,
+  isIdentifier,
   Token,
   $PERIOD,
   $COLON,
@@ -105,15 +106,10 @@ export class Parser {
     if (isBlank(input)) return null;
     var prefixSeparatorIndex = input.indexOf(':');
     if (prefixSeparatorIndex == -1) return null;
-    var prefix = input.substring(0, prefixSeparatorIndex);
+    var prefix = input.substring(0, prefixSeparatorIndex).trim();
+    if (!isIdentifier(prefix)) return null;
     var uninterpretedExpression = input.substring(prefixSeparatorIndex + 1);
-
-    // while we do not interpret the expression, we do interpret the prefix
-    var prefixTokens = this._lexer.tokenize(prefix);
-
-    // quote prefix must be a single legal identifier
-    if (prefixTokens.length != 1 || !prefixTokens[0].isIdentifier()) return null;
-    return new Quote(prefixTokens[0].strValue, uninterpretedExpression, location);
+    return new Quote(prefix, uninterpretedExpression, location);
   }
 
   parseTemplateBindings(input: string, location: any): TemplateBinding[] {
diff --git a/modules/angular2/test/core/change_detection/parser/parser_spec.ts b/modules/angular2/test/core/change_detection/parser/parser_spec.ts
index 06af623177..b13d9c4ca4 100644
--- a/modules/angular2/test/core/change_detection/parser/parser_spec.ts
+++ b/modules/angular2/test/core/change_detection/parser/parser_spec.ts
@@ -233,6 +233,9 @@ export function main() {
 
         it('should parse quoted expressions', () => { checkBinding('a:b', 'a:b'); });
 
+        it('should not crash when prefix part is not tokenizable',
+           () => { checkBinding('"a:b"', '"a:b"'); });
+
         it('should ignore whitespace around quote prefix', () => { checkBinding(' a :b', 'a:b'); });
 
         it('should refuse prefixes that are not single identifiers', () => {