diff --git a/src/img2pdf.py b/src/img2pdf.py
index 4a1cd37..9723339 100755
--- a/src/img2pdf.py
+++ b/src/img2pdf.py
@@ -721,8 +721,16 @@ class pdfdoc(object):
         if engine != Engine.pikepdf:
             self.writer.docinfo = PdfDict(indirect=True)
 
+        timezone_regex = r'^([+-])([0-9]{2})([0-9]{2})([0-9.]+)?$'
+
         def datetime_to_pdfdate(dt):
-            return dt.strftime("%Y%m%d%H%M%S%z")
+            time_no_tz = dt.strftime("%Y%m%d%H%M%S")
+            # Format for `%z` specifier is [+-]HHMM(SS(\.ffffff)?)?, but the
+            # PDF format only accepts the [+-]HHMM part, and it must be
+            # formatted as [+-]HH'MM'.
+            tz = dt.strftime("%z")
+            tz_pdf = re.sub(timezone_regex, r"\1\2'\3'", tz)
+            return time_no_tz + tz_pdf
 
         for k in ["Title", "Author", "Creator", "Producer", "Subject"]:
             v = locals()[k.lower()]
@@ -752,7 +760,13 @@ class pdfdoc(object):
                 )
 
         def datetime_to_xmpdate(dt):
-            return dt.strftime("%Y-%m-%dT%H:%M:%S%z")
+            time_no_tz = dt.strftime("%Y-%m-%dT%H:%M:%S")
+            # Format for `%z` specifier is [+-]HHMM(SS(\.ffffff)?)?, but the
+            # XMP metadata only accepts the [+-]HHMM part, and it must be
+            # formatted as [+-]HH:MM.
+            tz = dt.strftime("%z")
+            tz_xmp = re.sub(timezone_regex, r'\1\2:\3', tz)
+            return time_no_tz + tz_xmp
 
         self.xmp = b"""<?xpacket begin='\xef\xbb\xbf' id='W5M0MpCehiHzreSzNTczkc9d'?>
 <x:xmpmeta xmlns:x='adobe:ns:meta/' x:xmptk='XMP toolkit 2.9.1-13, framework 1.6'>