From a marketing automation point of view, there are several things you'll want to keep in mind if you're going to accept Japanese input:
Will these Japanese leads touch English-based or overall global efforts? The single biggest consideration here: knowing whether or not you want to use these leads as part of other initiatives, such as segments, mass mailings, etc. will determine most of your course of action. If your Japanese business unit is operating somewhat standalone, you may want to consider setting up Workspaces in your Marketo instance.
The Nashi Effect: I have seen at least two entirely different companies have JP list uploads using なし (which basically means "absence of a value") when fields aren't known. You will either need to specifically instruct your team to leave unknown values blank or have some campaigns doing proactive data management on this front.
Titles: It's quite common for job titles that actually mean the same thing to be written in several different ways given Japanese business culture. For example, "Manager" can be written as MGR, マネージャー or 部長 in addition to Latin characters. As a result, this may break targeted job titles and roles.
Addresses: Japanese addresses often don't map 1:1 with Western fields for noting addresses and can be a pain. Wikipedia will give you a pretty good overview. It's something worth keeping in mind if you need to do regional segmentation.
Equivalencies for Predefined Value Fields: On a field that has a defined list of values (say, lead source or lead status), you want to ensure that whatever status is used is one that is either input directly in English or use your CRM to manage localized translation values for these tracking fields. Otherwise, your reporting on a global level will break.
Align Romanization: If romanization needs to be used by your Japanese business unit, they need to agree to one system for the sake of data hygiene.