diff -r 3877bf2e3235 Lib/email/header.py
--- a/Lib/email/header.py	Mon Mar 12 09:46:44 2012 -0700
+++ b/Lib/email/header.py	Tue Mar 13 17:46:46 2012 -0700
@@ -283,7 +283,12 @@
         # character set, otherwise an early error is thrown.
         output_charset = charset.output_codec or 'us-ascii'
         if output_charset != _charset.UNKNOWN8BIT:
-            s.encode(output_charset, errors)
+            try:
+                s.encode(output_charset, errors)
+            except UnicodeEncodeError:
+                if output_charset!='us-ascii':
+                    raise
+                charset = UTF8
         self._chunks.append((s, charset))
 
     def encode(self, splitchars=';, \t', maxlinelen=None, linesep='\n'):
diff -r 3877bf2e3235 Lib/test/test_email/test_email.py
--- a/Lib/test/test_email/test_email.py	Mon Mar 12 09:46:44 2012 -0700
+++ b/Lib/test/test_email/test_email.py	Tue Mar 13 17:46:46 2012 -0700
@@ -604,6 +604,19 @@
         msg['Dummy'] = 'dummy\nX-Injected-Header: test'
         self.assertRaises(errors.HeaderParseError, msg.as_string)
 
+    def test_unicode_header_defaults_to_utf8_encoding(self):
+        # Issue 14291
+        m = MIMEText('abc\n')
+        m['Subject'] = u'É test'
+        self.assertEqual(str(m),"""\
+Content-Type: text/plain; charset="us-ascii"
+MIME-Version: 1.0
+Content-Transfer-Encoding: 7bit
+Subject: =?utf-8?q?=C3=89_test?=
+
+abc
+""")
+
 # Test the email.encoders module
 class TestEncoders(unittest.TestCase):